Copyright Board Canada

Commission du droit d’auteur Canada

 

Date	2010-07-09
Citation	Files: Public Performance of Musical Works; Public Performance of Sound Recordings; Reproduction of Musical Works; Reproduction of Sound Recordings; Reproduction of Performers’ Performances
Regime	Collective Administration of Performing Rights and of Communication Rights Collective Administration in Relation to Rights Under Sections 3, 15, 18 and 21 Copyright Act, subsections 68(3) and 70.15(1)
Members	Mr. Justice William J. Vancise Mr. Stephen J. Callary Mrs. Francine Bertrand-Venne
Proposed Tariffs Considered	SOCAN (2008-2010), RE:SOUND (2008-2011), CSI (2008-2012), AVLA/SOPROQ (2008-2011), ARTISTI (2009-2011)

Statement of Royalties to be collected by socan, Re:Sound, csi, avla/soproq and ArtistI in respect of Commercial Radio Stations

Reasons for decision

TABLE OF CONTENTS

I. INTRODUCTION 1

II. RIGHTS AT PLAY AND DRAMATIS PERSONAE 2

III. PROPOSED RATES 4

IV. POSITIONS OF THE PARTIES AND THEIR EVIDENCE 5

A. SOCAN 5

B. Re:Sound 5

C. CSI 6

D. AVLA/SOPROQ 7

E. ArtistI 8

F. ACTRA/AFM 9

G. CAB 9

V. TECHNICAL ASPECTS RELATING TO THE OPERATIONS OF BROADCASTERS 11

VI. LEGAL ANALYSIS 15

A. What is the nature of the performer’s reproduction right pursuant to subparagraph 15(1)(b)(ii)? 15

i. The meaning of subparagraph 15(1)(b)(ii) 15

ii. Does the reproduction right of performers materialize? 18

a. The Rights of Performers from the “Rest of Canada” 18

b. The Rights of Québec Performers 19

iii. The terms of the recording contracts vs. those set out in the collective agreements 21

iv. This exclusivity agreement is subject to the UDA/ADISQ Phonogram Agreement.” 23

B. In using “works” as they do, are radio stations making protected reproductions? 27

C. Are radio stations liable for these reproductions under the respective tariffs of the reproduction collectives? 32

D. Can the Board set a tariff for reproductions made before a tariff takes effect? 34

E. Can the Board, in a decision that certifies a tariff, clarify the wording of a regulation it made, or must it amend the regulation? 35

F. In setting the rate base for the Re:Sound tariff, is the Board restricted to the regulatory definition of advertising revenues pursuant to subsection 68.1? 37

G. Does section 90 of the Act preclude the Board from setting a single value or certifying an overall or single rate for broadcasting music? 38

VII. ECONOMIC ANALYSIS 38

A. The Wall/Globerman Methodology 38

B. The Bakos Methodology 40

C. Nordicity’s Methodology 41

D. Determination of Rates 42

i. SOCAN and Re:Sound Rates 44

ii. CSI 44

iii. ArtistI 47

iv. AVLA/SOPROQ 50

E. Very Low Use Rate Category 52

F. Indexation of Tiers 53

G. Rate Base 53

H. Internet-Related Activities 57

I. Individual Stations or Networks 57

J. Final Certified Tariffs and Royalties Generated 58

VIII. TARIFF WORDING 59

A. Years Covered 59

B. Rate Base 59

C. Qualifying For Low Use 59

D. Definition of Performer’s Performance 61

E. Calculating Royalties Using a “Reference Month” 61

F. Authorizing Uses by Others 62

G. Tailoring the Special Provisions of Paragraph 68.1(1)(a) 62

H. Music Use Information 62

I. Adjustments 64

J. Audits 65

K. Notices 65

L. Transitional Provisions 65

I. INTRODUCTION

[1] Copyright is not a single right, but a cluster of rights that attach to a number of “subject matters”. What to the layman is a single “thing” often is not for the copyright specialist. A music track, for example, is a sound recording in which are embedded a musical work and one or more performer’s performances.

[2] A Canadian radio station that broadcasts recorded music off a server reproduces and communicates musical works, performers’ performances and sound recordings. Four copyrights and two remuneration rights must be accounted for. This is the first time that the Board has been asked to set tariffs for all those rights at the same time.

[3] This decision deals with the proposed statement of royalties of five different collective societies targeting the use of their respective repertoire by commercial radio stations. The first two collectives filed their proposed statements of royalties pursuant to subsection 67.1(1) of the Copyright Act (the “Act”). [1] On March 30, 2007 and March 31, 2008, the Society of Composers, Authors and Music Publishers of Canada (SOCAN) filed proposed statements of royalties for the communication to the public by telecommunication of musical and dramatico-musical works in 2008 and in 2009. On March 30, 2007 and March 28, 2008, the Neighbouring Rights Collective of Canada (NRCC) [now Re:Sound Music Licensing Company (Re:Sound)] filed proposed statements of royalties for the communication to the public by telecommunication of published sound recordings embodying musical works and performers’ performances of such works in 2008 and in 2009 to 2011.

[4] The other three collectives filed proposed statements of royalties pursuant to subsection 70.13(1) of the Act. On March 30, 2007, CMRRA/SODRAC Inc. (CSI) filed, on behalf of the Society for Reproduction Rights of Authors, Composers and Publishers in Canada and of SODRAC 2003 Inc. (jointly SODRAC), as well as on behalf of the Canadian Musical Reproduction Rights Agency (CMRRA), a proposed statement of royalties for the reproduction of musical works for 2008 to 2012. On the same day, AVLA Audio-Video Licensing Agency (AVLA) and the Société de gestion collective des droits des producteurs de phonogrammes et de vidéogrammes du Québec (SOPROQ) (jointly AVLA/SOPROQ) filed a proposed statement of royalties for the reproduction of sound recordings in 2008 to 2011. Finally, on March 31, 2008, ArtistI filed a proposed statement of royalties for the reproduction of performers’ performances in its repertoire in 2009 to 2011.

[5] All the proposed statements of royalties were published in the Canada Gazette with a notice informing potential users or their representatives of their right to object to the statements. Only the Canadian Association of Broadcasters (CAB) filed an objection on behalf of its members.

[6] On the eve of the hearings, ACTRA Performing Rights Society (ACTRA PRS) and the American Federation of Musicians of the United States and Canada (AFM Canada) (together ACTRA/AFM) were granted intervenor status with the right to file written comments.

[7] At the request of the CAB, the Board consolidated the examination of all these proposals. The order in which parties filed their evidence was somewhat unusual. The CAB, AVLA/SOPROQ and ArtistI filed first, the CAB because it was challenging the status quo and the collectives because it was their first proposed tariff. SOCAN, Re:Sound and CSI, who on the whole favoured the status quo, filed second. The CAB, AVLA/SOPROQ and ArtistI replied. The hearings were held over a sixteen-day period between December 2, 2008 and January 29, 2009.

II. RIGHTS AT PLAY AND DRAMATIS PERSONAE

[8] These proceedings involve six rights or sets of rights. They are reviewed in the order in which the Board was asked to set a tariff.

[9] The first set of rights is the exclusive right of the owner of the copyright in a musical work to communicate it to the public by telecommunication and to authorize such a communication. [2] SOCAN administers these rights in Canada for virtually all copyright owners. SOCAN is subject to sections 67 to 68.2 of the Act (the “SOCAN regime”) and if it does not file a proposed tariff, it cannot, in practice, collect royalties. [3]

[10] The second and third rights are the remuneration rights that performers and makers each enjoy when a published sound recording of a musical work is communicated to the public by telecommunication. [4] These two [5] rights are treated together because they always trigger a single payment; in the case of sound recordings of musical works, that payment is always made to a collective society authorized by the Board to collect it. [6] Re:Sound administers these rights for the vast majority of eligible performers and makers. It too is subject to the SOCAN regime. The remuneration rights are subject to a number of conditions. [7]

[11] The fourth set of rights is the exclusive right of the owner of the copyright in a musical work to reproduce it and to authorize such a reproduction. [8] Together, SODRAC and CMRRA administer most, but not all, of this repertoire in Canada. SODRAC represents the vast majority of rightsholders in Québec and most works written in French by Canadians as well as many of its foreign counterparts. CMRRA represents a large number of Canadian and foreign English-language music publishers. Both collectives are subject to sections 70.1 to 70.6 of the Act (the “general regime”) and as such, they can negotiate licensing agreements directly with users or ask the Board to certify tariffs. Where both collectives opt to use a tariff, CSI acts for them.

[12] The fifth set of rights is the exclusive right of the owner of the copyright in a sound recording to reproduce it and to authorize such a reproduction. [9] SOPROQ represents mostly Francophone record producers from Québec. AVLA acts for the major record companies and for many independent labels, artists and producers. Together, they represent the vast majority of the repertoire. These collectives are subject to the general regime.

[13] The sixth set of rights is the exclusive right of the owner of the copyright in a performer’s performance to reproduce any reproduction of an authorized fixation of the performance for a purpose other than that for which the authorization was given and to authorize such a reproduction. [10] Three collective societies administer these rights: ArtistI, ACTRA PRS and AFM Canada. Only ArtistI, which acts predominantly but not exclusively for French-speaking performers from Québec, has filed a tariff. ACTRA PRS represents professional cinema, television, radio and recording artists who work in English. Its mandate includes the collection and distribution of fees, royalties, residual fees and all other forms of compensation or remuneration to which members and permit holders of the Alliance of Canadian Cinema Television and Radio Artists (ACTRA) may be entitled. AFM Canada, acting for musicians in the United States, Canada and other countries, collects and distributes government mandated or other compulsory royalties or remuneration that are subject to collective administration. These collectives are subject to the general regime.

[14] All collectives except AVLA/SOPROQ obtain exclusive rights in the repertoire they administer for the purpose of the commercial over-the-air radio market.

III. PROPOSED RATES

[15] Most commercial radio stations pay SOCAN 3.2 per cent on the first $1.25 million of annual advertising revenues and 4.4 per cent on the rest. Low-use commercial radio stations [11] pay only 1.5 per cent. SOCAN asks that the tiered rate be abandoned. Its proposed tariff sought rates of 2.6 per cent for low-use stations and 6 per cent for the rest. Its statement of case lowered the request to 4.7 per cent for the higher rate, on account of an increase in music use; the low music use rate would remain at 1.5 per cent. During oral argument, counsel suggested a higher rate of between 4.7 and 5.2.

[16] Low-use commercial radio stations pay royalties to Re:Sound at a rate of 0.75 per cent. Other commercial radio stations pay 1.44 per cent on the first $1.25 million of annual advertising revenues and 2.1 per cent on the rest. In its proposed tariffs, Re:Sound asked that the low-use rate be increased to 0.86 per cent on the first $625,000 of annual revenues, 1.72 per cent on the next $625,000 and 2.58 on the rest. For other radio stations, the rates would be 2, 4 and 6 per cent, respectively. Re:Sound now requests that the low-use rate remain at 0.75 per cent and that other radio stations pay 1.54 per cent on the first $1.25 million of annual revenues and 2.24 on the rest.

[17] Low-use commercial radio stations [12] pay royalties to CSI at a rate of 0.12 per cent on the first $625,000 of annual gross income, 0.23 per cent on the next $625,000 and 0.35 per cent on the rest. For other radio stations, the rates are 0.27, 0.53 and 0.8 per cent, respectively. Subject to a repertoire adjustment, CSI requests that low-use stations pay at a rate of 0.42 per cent of all their gross income; other stations would pay 0.96 per cent of the first $1.25 million and 1.32 per cent on the rest.

[18] AVLA/SOPROQ seeks royalties from low-use radio stations at rates of 0.48 per cent on the first $625,000 of annual gross income, 0.96 per cent on the next $625,000 and 1.40 per cent on the rest. For other stations, the rates would be 1.33, 2.67 and 4 per cent, respectively.

[19] Initially, ArtistI proposed a royalty of 0.06 per cent of gross income from low-use radio stations; others would pay 0.13 per cent on their first $1.25 million of revenue and 0.18 per cent on the rest. In the end, ArtistI claims 2.43 per cent of gross income subject to a repertoire adjustment.

[20] The CAB proposed the amount of royalties, subject to repertoire adjustments, be set 2.9 per cent for all the rights a station requires to play sound recordings of musical works off its server. As an alternative, the CAB proposed rates of 2.9 per cent for SOCAN, 1.4 for Re:Sound and, subject to repertoire adjustments, 0.91, 0.45 and 0.45 respectively for CSI, AVLA/SOPROQ and ArtistI.

IV. POSITIONS OF THE PARTIES AND THEIR EVIDENCE

A. SOCAN

[21] SOCAN argued that there was no basis for lowering the current rates. It contended that, taking into account the increase in music use and the methodology used by the CAB, the main rate should be increased from 4.4 per cent to between 4.7 and 5.2 per cent. SOCAN did not propose a specific methodology, but rather submitted a report prepared by Dr. Stanley Liebowitz, Professor of Economics at the University of Texas at Dallas, reviewing the methodology proposed by the CAB.

[22] SOCAN submitted a report from Erin Research Inc. that found that music use had increased by 6 per cent since the 2005 decision, from 76.1 per cent to 80.6 per cent. SOCAN suggested that based on this factor alone, the 4.4 rate should be increased to 4.7 per cent.

[23] With respect to the rate base, SOCAN contended that the Board should revert to the situation as it stood before 2005 and use gross income, including the full amount received pursuant to turn-key advertising contracts. In support of this position, SOCAN presented a report prepared by Mr. Rob Young, from PHD Canada, reviewing how Canadian commercial radio stations sell airtime to their advertiser clients. Mr. Young argued that the charges incurred by clients with turn-key contracts only reflect the true value of the airtime sold, and not the cost of producing the advertisement. This, according to SOCAN, suggests that all revenues associated with turn-key contracts should be included in the rate base.

[24] Dr. Walid Hejazi, Professor at the Rotman School of Business of the University of Toronto, examined the economic appropriateness of deducting from the rate base the fair market value of production services provided under turn-key contracts. Given the current business model in the radio industry, where stations do not charge separate fees for airtime and production services, Professor Hejazi concluded that it is not possible for radio stations to identify and allocate revenues to production services.

B. Re:Sound

[25] Re:Sound argued that in order to account for an increase in music use, its rates should be increased by 6.7 per cent, from 1.44 to 1.54 per cent on the first $1.25 million of gross income, and from 2.1 to 2.24 per cent on the rest. Re:Sound did not present any specific valuation methodology, but rather submitted a report prepared by Dr. John McHale, then Professor of Economics at Queen’s University, analyzing the model presented by the CAB.

[26] Re:Sound opposed the CAB’s contention that the regulatory definition of “advertising revenues” entitles radio stations to deduct from the rate base the fair market value of the cost of production services provided to advertisers pursuant to turn-key contracts. Re:Sound argued that allowing this would result in the exclusion of revenues related to air-time-based sales of advertisements and give a windfall to commercial radio stations. In support of this position, Re:Sound relied on the report of Mr. Alan T. Mak, accountant at Rosen Associates Limited, who reviewed information supplied by radio stations on their production services. He concluded that the stations charge their fees based on air-time, and do not charge additional or separate fees for production services. Stations do not record revenues related to production services. Furthermore, it is common practice to combine the cost of production services with other operating costs.

[27] Ms. Doris Tay, Distribution Manager for Re:Sound, addressed the CAB proposal to introduce a new rate category for radio stations broadcasting music for less than 5 per cent of their broadcast time. She stated that the addition of a very low music use category would make it significantly more difficult and costly for Re:Sound to administer the tariff, especially if stations were to migrate between the very low and low categories as their music use varied over time.

C. CSI

[28] CSI advanced three reasons to ask that its rate be increased to 0.96 per cent on revenues not exceeding $1.25 million and 1.32 on the rest. First, the SOCAN rate has increased. Second, music use has increased. Third, CSI now holds a larger share of the repertoire.

[29] Mr. Paul Audley, President of Paul Audley and Associates Ltd., and Dr. Marcel Boyer, Bell Canada Professor of Economics at the Université de Montréal, examined the value of the use of CSI’s repertoire by commercial radio stations and commented on the approach proposed by the CAB. The authors also concluded, based in part on the Erin Research study already mentioned, that music use, as a percentage of broadcast time, had increased since 2005 from 76.1 to 81.2 per cent. CSI proposed that this increase be reflected in the rate.

[30] Mr. Benoît Gauthier, President of Circum Network Inc., examined the importance of music for radio stations as a means to attract listeners. His study showed that without music, 66 per cent of the current listeners would switch from radio to prerecorded CDs. This rate reached 74 per cent when considering only listeners to commercial radio stations predominantly broadcasting music.

[31] CSI also presented a study prepared by Paul Audley & Associates Ltd. to examine the use of the CSI repertoire by commercial radio stations. The study concluded that CSI represents 89.52 per cent of the music broadcast by commercial radio stations.

[32] CSI and AVLA/SOPROQ jointly commissioned a study prepared by Dr. Michael J. Murphy, Professor at Ryerson University School of Radio and Television Arts, to explain radio broadcasting technology. The content of his report is addressed as part of the evidence of AVLA/SOPROQ.

[33] Finally, CSI relied on the studies prepared for SOCAN by Mr. Rob Young and Professor Walid Hejazi to argue that its rate base should continue to be “gross income” and not “advertising revenues”.

D. AVLA/SOPROQ

[34] Ms. Stéphanie Duquette, Assistant Executive Director of SOPROQ, and Mr. Richard Pfohl, Vice-President and General Counsel of AVLA described the membership of each collective and the scope of the authorization secured from rightsholders.

[35] AVLA/SOPROQ argued that the reproduction rights administered by AVLA/SOPROQ, CSI and ArtistI are separate and distinct. Furthermore, neither the Act [13] nor earlier decisions of the Board [14] prescribe a ratio or approach to be used in valuing the right to reproduce a sound recording. Accordingly, AVLA/SOPROQ submitted that the Board should assess the right to reproduce a sound recording on its own merits using proper valuation methodologies. To assist the Board in doing so, AVLA/SOPROQ provided two reports in support of its tariff rate proposal.

[36] The first, prepared by Messrs. Peter Lyman and Dustin Chodorowicz, both from Nordicity Group Ltd., analyzed the economic value of reproducing sound recordings for commercial radio stations. The second, prepared by Dr. Yannis Bakos, Associate Professor of Management at the Stern School of Business, New York University, estimated the economic surplus generated from the use of sound recordings by commercial radio stations.

[37] AVLA/SOPROQ argued that the source used by radio stations to make copies (whether through CD ripping [15] or downloads) has no impact on the value of that copy. It also argued that the relative value of the reproduction right has increased as radio stations automated their operations. In support of these positions, AVLA/SOPROQ relied on Professor Murphy, who provided an overview of the technology currently used to create and deliver commercial radio programming in Canada. He explained that the introduction of digital automation technology has had a positive impact on commercial radio’s overall operational efficiency. He also said that the majority of stations now directly download from the Internet in digital form using secure Web server technology.

[38] Professor Murphy explained that the use of an automated digital system involves adding songs to the main music library as well as making copies for audio logging, system backups and CRTC requirements. Other reproductions may also occur for manipulation of songs (compression, metadata, etc.), evaluation, deployment of a remote voice tracking server to facilitate the recording of voice-tracking shifts and delivery of an Internet stream.

E. ArtistI

[39] Ms. Annie Morin, ArtistI’s Director, explained that ArtistI is a collective society that administers the reproduction rights of its performer members. To support its right to file a tariff, ArtistI contends that performers in Québec retain the reproduction right in their performances embedded in sound recordings as a result of collective agreements reached pursuant to An Act respecting the Professional status and conditions of engagement of performing, recording and film artists (the “SAA”). [16] The combined effect of this statute and of the agreements prevents performers from assigning to record producers anything beyond the right to commercially exploit the sound recording. ArtistI argued that reproductions made by radio stations are beyond the scope of that exploitation.

[40] Ms. Morin also presented the results of a repertoire study demonstrating that the use of ArtistI’s repertoire as a percentage of air play of musical works went from 3.6 per cent in September 2008 to 5.6 per cent in November 2008. Based on this information, ArtistI asked to receive 6.9 per cent of what it would be entitled to if all prerecorded performances played on radio were in its repertoire. According to ArtistI, this percentage would be obtained if all of its members had assigned their rights to it.

[41] To determine that entitlement, ArtistI proposed to use the CSI rate as a benchmark, adjusting it for the ratio of revenues obtained by performers relative to composers in the CD market.

[42] Messrs. Louis Landreville, a lawyer specializing in the entertainment industry, and Marc Ménard, Professor at l’École des médias of the Université du Québec à Montréal, presented evidence in order to establish that ratio.

F. ACTRA/AFM

[43] ACTRA/AFM made the following comments in writing. Outside Québec, record producers secure sweeping rights in the performers’ performances. Rights in those performances are subsumed in AVLA’s repertoire. A few performers retain rights in their performances, for whom no tariff has been filed. To determine precisely which performers fall in which group would require analyzing all recording agreements. A practical alternative would be for ACTRA/AFM to secure the rights in all performances not already subsumed in AVLA’s repertoire and to assign them to AVLA, subject to an undertaking that AVLA would then remit to ACTRA/AFM royalties attributable to performances that are “owned” by these collectives.

G. CAB

[44] The CAB’s case is based on the premise that music is a single input to radio. Its value reflects what can be earned from using it and should not depend on the number of rights involved. The CAB relied on a report prepared by Dr. Gerry Wall and Mr. Bernard Lefebvre, both from Wall Communications Inc. The report used the economic approach developed for the CAB by Professor Steven Globerman and adopted by the Board in Commercial Radio (2008). [17] This approach consists of an estimation of the changes in the broadcaster’s music reservation price, defined as the maximum price that a broadcaster would be willing to pay to obtain the music as an input. This maximum price reflects the average productivity of music, the advertising rates per unit of music audience and the number of hours of broadcasted music.

[45] The authors of the report modified the Globerman model. For example, they suggested that the model should include a seller’s reservation price, or a minimum price under which music rights owners would refuse to sell. They estimated that since rightsholders gain significant benefits from radio airplay, they should be willing to compensate radio stations to play their music. This implies a negative seller’s reservation price.

[46] Taking this negative price into account, as well as other modifications, Dr. Wall and Mr. Lefebvre concluded that the SOCAN rate should be 10 per cent less than the 3.2 per cent certified in 1987, or 2.9 per cent. Based on the argument that the price of music should be the same without regard to the number of rights, the CAB argued that this rate should be the sum total of what commercial radio stations pay for their music. The CAB took no position on the allocation of this amount among the collectives.

[47] In the alternative, the CAB used the evidence of Dr. Wall and Mr. Lefebvre to propose that the SOCAN rate be set at 2.9 per cent. The CAB then proposed that the reproduction right be compensated according to the ratio of 1 to 3.2 that effectively exists between the CSI and SOCAN rates, and that sound recordings trigger the same amount of royalties as musical works. This led to proposed tariffs of 2.9 per cent for SOCAN, 2.9 per cent for Re:Sound, 0.91 per cent for CSI, 0.45 per cent for AVLA/SOPROQ and 0.45 per cent for ArtistI, subject to repertoire adjustments.

[48] Finally, Dr. Wall and Mr. Lefebvre argued that since radio stations receive sound recordings almost exclusively in digital format directly from digital music delivery services (MDS), as opposed to copying tracks from CDs on a hard drive, a 50 per cent reduction should apply to the CSI, AVLA/SOPROQ and ArtistI rates. This position is founded on the CAB’s argument that downloading files from a MDS does not involve making “material” copies, resulting in a decrease in the number of protected copies made.

[49] The CAB commissioned the Solutions Research Group (SRG) to conduct studies on the use of music on Canadian commercial radio stations and on how music is received and copied at music-based radio stations. The study found that 80.6 per cent of program time, excluding advertising, station promos and IDs, is devoted to sound recordings. Regarding the reception of music, all the music-based radio stations surveyed admitted using a digital music playback system and 90 per cent of stations that added music to their digital playback system in the past week reported some form of electronic delivery as the original source.

[50] Mr. Dean Sinclair, a broadcast consultant specializing in private radio, provided detailed information on the station-level, operational activities of Canada’s private broadcasters as they relate to music use. He explained that a MDS, the most common of which is DMDS, is a replacement medium for record companies and distributors to get new music into the hands of radio stations without the need for delivering physical copies. DMDS is described in detail in the next section.

[51] Mr. Bruce Wilkinson, an engineering consultant, described from a technical perspective how music files are used in a radio station environment. He explained that all commercial radio stations in Canada currently use digital playout systems, and that music is integrated into the digital delivery system either through CD ripping or downloads from a MDS.

[52] Dr. James Dertouzos, Professor of Economics at the RAND Graduate School of Policy Studies, presented a study looking at the impact of radio airplay on the music industry. While he recognized that different studies have come up with conflicting results, he believed that radio play stimulates record sales. This conclusion is consistent with anecdotal evidence, including the fact that record companies pay large sums to promote their releases. In addition, consumer surveys indicate that the exposure to radio is a primary method to learn about music.

[53] To support this position, the CAB presented additional evidence from Jupiter Research that found that radio remains the most powerful means of music discovery. In addition, a 1996 study from Angus Reid found that radio emerged more often than any other factor as having influenced CD buyers’ most recent music purchase.

[54] Mr. Patrick Grierson, President of Canadian Broadcast Sales, presented evidence showing that advertisers attach a higher value to talk content than to music. Mr. Stephen Armstrong, President of Stephen M. Armstrong Consulting Inc., examined the programming expenses and profitability of radio stations and concluded that radio stations in general, and smaller stations in particular, are in a much more difficult financial situation than they were at the beginning of the decade. Only larger, music-oriented stations increased their profitability between 2002 and 2007. According to Mr. Armstrong, over the period 2008 to 2012, the financial situation of radio stations is expected to get significantly worse, no matter the size or type of content broadcasted. Mr. Pierre-Louis Smith, Vice-President, Policy and Chief Regulatory Officer of the CAB, explained that radio stations are increasingly using talk content as a differentiating factor, and that talk is the most valuable factor for attracting audiences.

[55] Finally, Mr. Stéphane Gilker, Partner at Fasken Martineau DuMoulin, provided an overview of the prevailing contractual arrangements, in the Province of Québec, between songwriters and publishers and between performers and record labels. His examination led him to conclude that ArtistI should receive no more than 7.3 per cent of what CSI is paid, before any repertoire adjustment.

V. TECHNICAL ASPECTS RELATING TO THE OPERATIONS OF BROADCASTERS

[56] Canadian private broadcasters now commonly use a digital music playback system [18] to deliver their music programming. Initially, the cost associated with making the changeover to the new digital technology precluded smaller and less profitable broadcasters from modernizing their systems. Since 2003, the arrival of improved and more affordable products has made it possible for smaller broadcasters to digitize their operation systems. Broadcasters have thus changed their physical music libraries to digital ones, which are collections of digital music files stored on hard drives in conjunction with metadata (such as title, genre, name of the album and performer) contained in a computer database. Programming can now be delivered in a fully automated fashion, or in a partially automated method (commonly known as live-assist). It can also originate locally, or be provided from satellite or other communication network feeds.

[57] Digital technology has transformed not only the way broadcasters deliver their programming, but also how they obtain musical material. The last time the Board examined the reproduction activities of radio stations, CD ripping was more prevalent than digital downloading and broadcasters still had the option of not reproducing music for broadcast. [19] Label representatives still routinely delivered newly released CDs to radio stations. If a broadcaster decided to air the material it received, it would either play it on air or “rip” the CD onto its server to incorporate it into their playback system.

[58] Today, visits from labels are rare or serve other purposes. Although some field staff still delivers music on CD in the largest markets, label representatives mostly come in with a file and play it off their system. [20] In mid-size markets, in-market servicing is infrequent and when done is usually related to promotional projects such as concerts, artist promotion and product giveaway. [21] Nearly all stations, regardless of market size, now rely on some form of electronic delivery as a source for new music; MDS constitute the dominant electronic source. [22]

[59] In the Anglophone market, broadcasters rely predominantly on DMDS, a MDS owned by Yangaroo Inc. In the Francophone market, the broadcasters use the services of 45tours.ca (45tours). According to the evidence, 85 per cent of surveyed stations use DMDS. The services are sufficiently similar that describing one of them will be sufficient for our purposes.

[60] DMDS allows broadcasters to download digitally new music releases through its Web site. Generally, the music or program director of a station opens an account with DMDS and selects a variety of music formats in order to create the package of music targeted by the station. Once membership is approved, DMDS provides the broadcaster the software (decryption module or DMDS agent) required to use the products it will thereafter download to its server from the DMDS Web site.

[61] Record labels, their agents and some independent artists decide what material is available for download. Whenever they upload new material on the DMDS Web site that corresponds to a subscriber’s music format, the subscriber is notified by email. The subscriber can then log onto the DMDS server via the Internet by using an encrypted password to either audition or download the released material.

[62] The audition feature makes it possible for a user to listen to music without making a copy by previewing the recording from DMDS through an audio stream. To enable the audition feature, the user clicks on the “Listen” button, which prompts the image of a player to appear on the screen as the song is played. The DMDS patent describes the technique enabling the audition feature as follows:

The music director may choose to listen to sample tracks (step 200-6), which are then streamed unencrypted to the recipient computer system, for example in MP3 format (step 200-7). These sample tracks are of insufficient length or quality to be used for radio play and thus, pose little risk insofar as the security of the single is concerned. [23]

[63] Dr. Murphy testified that “[i]n the streaming with DMDS there’s just a temporary Internet file that exists on your computer for as long as you need to listen to it and then it’s flushed away, it doesn’t remain permanently on.” [24]

[64] The DMDS stream can be assimilated to a stream or to an on-demand stream as described in Tariff 22.A [25] and in CSI – Online Music Services (2007). [26] This feature reduces the number of reproductions made for the purpose of evaluating sound recordings on a preliminary basis at music meetings since instead of downloading the song to an iPod or other medium, those in attendance can simply listen to the material directly from the DMDS Web site. The extent of the reduction however, is not quantifiable at this time.

[65] When a user clicks “download”, the process by which broadcasters obtain a copy of the selected musical file is launched. In the first phase of the “download”, a temporary copy of the encrypted file residing on the DMDS server is made in the temporary memory of the radio station’s computer. Anyone who tried to grab that temporary encrypted copy off the broadcaster’s computer would not easily be able to do anything with it.

[66] There is contradictory evidence as to where this temporary copy is created. Dr. Murphy stated, in his report [27] and during his testimony, [28] that the temporary copy of the DMDS file is created on the station’s server, not on the DMDS server and explained how it was copied by referring to the DMDS patent. By contrast, Mr. Wilkinson testified that he “believe[d] it would take place on the DMDS server”; [29] his written report did not address the issue. Dr. Murphy’s evidence accords with the Board’s long standing understanding of how the Internet operates. That evidence also is founded on, among other things, an examination of the DMDS patent and on more than a belief. It is more convincing and therefore, we find that the temporary copy is created on the broadcasters’ server.

[67] During the second stage of the process, the DMDS decryption module decrypts the file after confirming with the DMDS server that the station is an authorized user. Once confirmation is obtained, there is no need for the station to remain connected with the DMDS server. The file is decompressed into a broadcast quality format and a watermark is inserted to allow DMDS to monitor the broadcaster’s subsequent use of the file. Hence, from the temporary encrypted file, the decryption module generates a separate, usable, decrypted copy. Apart from the watermark addition, the decrypted copy is identical to the content of the original encrypted file, which remains on the DMDS server.

[68] The resulting copy is “placed” in what is commonly called the “download directory”. The requirements of each radio playback system dictate how the station manipulates the decrypted digital music files. These requirements differ from one system to the other, as do the manipulations required before a decrypted file can be uploaded from the download directory to the playback system.

[69] When all systems are fully compatible with DMDS and well-integrated, the decrypted file can be copied directly to the automation system hard drive. However, most popular playback systems require some form of file manipulation or editing of the metadata; the underlying digital audio content is not generally subject to manipulations. When manipulations are necessary, a copy of the file will be made on an operator local workstation. Generally, once the “editing” process is completed, a copy is made and the original remains on the local hard drive of the PC that has performed the editing. Depending on the technical requirements of the various playback systems and the desired level of redundancy, multiple copies may be stored on the server and other workstations and there may exist multiple servers. The music selection practices of broadcasters will also play a factor. The file may be copied to various internal computers or other devices to allow convenient individual or group listening and to facilitate the evaluation and selection process.

[70] It is considered poor engineering to design a system that is completely dependent on a single server and continuous network availability because it exposes the on-air playback systems to instant failure in the event of any interruption in network connectivity or in the operation of the server or its storage subsystems. Accordingly, most broadcasters rely on workstations that have a local copy of the audio files that are not server or network dependent. It is thus reasonable to assume that even if it is theoretically possible to download the DMDS copy directly onto the playback systems, broadcasters will choose not to do so and will first save a copy on a local workstation before adding it to the playback system.

[71] Finally, although the audio file becomes part of a broadcaster’s complete digital system only when it is uploaded to the playback system, the file is nevertheless capable of manual playback from the time it is copied into the download directory. It does not need to be acknowledged by a playlist. From that moment, broadcasters are not technically restricted in their use of the file, subject to the terms of their service agreement.

VI. LEGAL ANALYSIS

[72] After the parties concluded the presentation of the evidence, we issued a list of legal questions we wanted addressed by the parties in their written arguments. Not all of these questions need to be addressed to determine the terms and conditions of the tariff to be certified. The following questions are relevant:

1. What is the nature of the performer’s reproduction right pursuant to subparagraph 15(1)(b)(ii)?
2. In using “works” as they do, are radio stations making protected reproductions?
3. Are radio stations liable for these reproductions under the respective tariffs of the reproduction collectives?
4. Can the Board set a tariff for reproductions made before a tariff takes effect?
5. Can the Board, in a decision that certifies a tariff, clarify the wording of a regulation it made, or must it amend the regulation?
6. In setting the rate base for the Re:Sound tariff, is the Board restricted to the regulatory definition of advertising revenues pursuant to section 68.1?
7. Does section 90 of the Act preclude the Board from setting a single value or certifying an overall or single rate for broadcasting music?

We now proceed with the analysis of each.

A. What is the nature of the performer’s reproduction right pursuant to subparagraph 15(1)(b)(ii)?

i. The meaning of subparagraph 15(1)(b)(ii)

[73] ArtistI’s tariff is based on subparagraph 15(1)(b)(ii) of the Act. It reads as follows:

15(1) Subject to subsection (2), a performer has a copyright in the performer’s performance, consisting of the sole right to do the following in relation to the performer’s performance or any substantial part thereof:

[…]

(b) if it is fixed,

[…]

(ii) where the performer authorized a fixation, to reproduce any reproduction of that fixation, if the reproduction being reproduced was made for a purpose other than that for which the performer’s authorization was given, and […] [our emphasis]

[74] Thus, to determine whether we can certify the proposed ArtistI tariff, we must, as set out in the passage we emphasized, decide two things. First, to what uses did performers consent when they authorized the fixation of their performances onto sound recordings of musical works? Second, what is the impact of authorizing a fixation for certain purposes but not for others? The answer to the first question depends on the facts, more specifically on the terms of the contracts between performers and makers. The answer to the second question depends on an interpretation of the Act.

[75] Most parties, in interpreting the scope and nature of the performers’ reproduction right, relied on the Rome Convention [30] and corresponding documents. In our opinion, these documents are not the appropriate interpretive tools by reason that Canada has gone beyond its obligation under the Rome Convention by granting exclusive and positive rights to performers. This was achieved by coupling the right itself with the right to authorize. When one authorizes, one allows someone to act, one does not simply prevent the act.

[76] ArtistI claims that the performers’ reproduction right materializes when a reproduction of a reproduction of a fixation is made for a purpose other than that for which the fixation was authorized. An authorization “for all purposes” cannot be enforced. It is contrary to the existing contractual scheme. An authorization for a purpose other than that for which the fixation was made cannot be assumed. A performance is linked with the personality right, which has certain consequences. Namely, consent is exhaustive. The performer’s right cannot be extinct at the time of the contract since it is the subsequent infringement that gives rise to the right.

[77] AVLA/SOPROQ and the CAB contend that the existence of the 15(1)(b)(ii) right depends on the fulfilment of specific conditions. For the right to be triggered, the performer must have explicitly restricted the purposes for which the performer authorizes the reproduction of the fixation of the performance. Unless restrictions are put on the purposes of the fixation, the right never comes into existence.

[78] The CAB concedes that in theory, the Act permits performers to limit the purposes for which reproductions of the fixation may be made; however, it claims that in practice, all-purpose authority is granted and accordingly, the maker assumes all rights.

[79] ACTRA/AFM contends on the other hand that, notwithstanding the wording of the relevant provision, the right does not arise only upon the commission of an unauthorized act and can be licensed through a tariff. It follows, in their submission, that as soon as authorization is given in a tariff, the right is extinguished.

[80] This is not how the Act reads. The rights granted to performers pursuant to paragraph 15(1)(b) arise only if and when the performance “is fixed”. The performance itself is fleeting; it can only be fixed as it occurs. The authorization to fix it however, almost always is granted before the performance occurs (hence, probably, the use of the past tense “authorized”). Once this is understood, it becomes easier to understand why paragraph 15(1)(b) grants the rights that it grants. Subparagraph 15(1)(b)(ii) comes into play only if a performance is fixed with the performer’s authorization. Absent such an authorization, the performer’s control over the fixation flows from subparagraph (i), not (ii). Furthermore, performers who consent to their performances being fixed generally also authorize the person making the fixation to use it for some purpose or other. As a result, the performer’s reproduction right cannot materialize to the extent that the performer granted the maker the right to engage in those uses before the fixation even existed. For all other purposes, however, subparagraph 15(1)(b)(ii) grants to performers the exclusive right to control the reproduction of the performance embodied in the fixation. Naturally, subparagraph 15(1)(b)(i), which applies to a performance that is illegitimately fixed, does not circumscribe the reproduction right of the performer, since the performer consented to nothing in the first place. The performer retains the exclusive right to control all subsequent uses of the unauthorized fixation of the performance.

[81] Consequently, it would be wrong to contend that makers of a sound recording acquire a performer’s right of reproduction when they receive authorization to use the fixation of the performance. [31] One cannot acquire something that never existed. However, producers are entitled to the economic value of the performance embodied in their sound recording. The transfer of value occurs when performers authorize the fixation pursuant to subparagraph 15(1)(a)(iii) of the Act.

[82] In our opinion, the reproduction right of performers is a full right. Makers can use the fixation of a performance only in ways that the contract allows either expressly or by implication. Furthermore, the uses so authorized are limited to what is required to fulfil the objects of the recording contract. Our reasons to so conclude are as follows.

[83] First, where the legislator intended the performer to abandon control over reproductions of a fixation when authorized, it has so provided. That is the case in subsection 17(1) of the Act, which reads as follows:

Where the performer authorizes the embodiment of the performer’s performance in a cinematographic work, the performer may no longer exercise, in relation to the performance where embodied in that cinematographic work, the copyright referred to in subsection 15(1).

[84] Second, as not all fixations are sound recordings, we should be careful about making generalizations about the performance-fixation interface based solely on an analysis of the performance-recording interface.

[85] Third, performers exploited their performances commercially before the Act recognized the rights. It is unlikely that the recognition of those rights was intended to prejudice them.

[86] Finally, generally speaking, producers are more able to dictate conditions than artists. Recording contracts are, to a large extent, adhesion contracts, especially where background artists are concerned. Both common and civil law generally resolve any ambiguity in such contracts in favour of the party who did not dictate their terms. This also favours resolving doubts in favour of artists.

[87] It remains to be seen how this applies in this case. It is only by looking at individual recording contracts that we could determine with any certainty whether or not performers control the reproductions made by radio stations. However, no individual contracts were filed and we are limited to relying on the boiler-plate recording contracts filed as evidence. Insofar as the rights of ArtistI performers are concerned, we must also examine the collective agreements between the Union des Artistes (UDA) and the Association québécoise de l’industrie du disque, du spectacle et de la vidéo (ADISQ) and between the Guilde des Musiciens du Québec (Guilde) and ADISQ.

ii. Does the reproduction right of performers materialize?

a. The Rights of Performers from the “Rest of Canada”

[88] It is necessary to look at the rights of performers from “the rest of Canada” even if they are not targeted by a proposed tariff. If the performer’s subparagragh 15(1)(b)(ii) right does not materialize, any economic value that the performance may have as an input to the sound recording in which it is incorporated must be “owned” by the maker and ALVA/SOPROQ is entitled to collect royalties; if it is the performer, we cannot set a tariff since none was proposed.

[89] Although the wording of the sample agreements provided by AVLA in response to interrogatories varied considerably, they allow us to conclude that performers outside Québec almost invariably grant to makers of sound recordings the exclusive right to use the fixation of their performances for “all purposes”. Under such agreements, a performer’s reproduction right never materializes. Consequently, as we just explained, the maker is entitled to monetize the economic value in the performance as part of the maker’s exclusive right over the sound recording.

b. The Rights of Québec Performers

[90] The situation is different for Québec performers. While the sample recording agreements suggest that Québec artists authorize producers to use their performances’ fixation in as broad a fashion as elsewhere in Canada, the agreements may be unenforceable if they provide for less favourable terms than those stipulated in collective agreements reached pursuant to the SAA. The SAA provides that if a contract between an artist (here, the performer) and a producer (here, the maker) does not contain terms and conditions that meet or exceed the minimum standard set in the relevant collective agreement, the collective agreement will prevail.

[91] The CAB and AVLA/SOPROQ contend that if the SAA deals with copyright, it is ultra vires with the result that an ArtistI member who signs a standard recording agreement retains nothing and cannot rely on the collective agreements governed by a provincial legislation to reclaim the rights they have assigned to the label. They further contend that various decisions emanating from the SAA regime recognize the constitutional limits that constrain the legislative reach of the provincial regime, which thereby preclude collective agreements from dealing with the terms of exploitation of performers’ copyright. [32]

[92] ArtistI relies on other decisions [33] as well as doctrinal authorities [34] in support of its position that it is possible, indeed essential, to deal with the terms of exploitation of performers’ performances created in the process of providing a service to a producer in the collective agreements.

[93] In Guilde et AHGM (1990), several organizations representing authors asked that the Commission de reconnaissance des associations d’artistes et des associations de producteurs to circumscribe the scope of the bargaining sector for which the Guilde was seeking certification. In defining the bargaining sector as « tous les artistes qui pratiquent l’art de la musique instrumentale […], excluant tout le champ des droits d’auteur », [our emphasis] the Commission gave effect to the various agreements reached between the Guilde and the intervenors. In those agreements the Guilde recognized that its authority as a bargaining agent did not extend to the rights of performers in their capacity as authors (les droits d’auteurs) of musical works.

[94] In our opinion the Guilde did not agree to exclude matters of copyright altogether. In Guilde et AHGM (1990), the only issues relating to copyright concerned musical and dramatico-musical work. Nothing else was in issue. This is not surprising given that performers had no copyright in their performances in 1990. The object of the definition of the bargaining sector and the subsequent collective agreement provisions was to make it explicit that insofar as performers who are also authors, their rights as authors are not covered by the collective agreements. Most importantly, since the certification occurred in 1991, its terms should be interpreted in accordance with section 58.1 of the 1997 Copyright Amendment Act: [35] the limitation does not clearly target performers’ future copyrights and therefore, must be interpreted to target only the rights of authors and not all copyright subject matters that existed or were created since.

[95] It would appear that the inconsistency in the jurisprudence stems from diverging grammatical interpretation of the phrase « les droits d’auteur » found in Guilde et AHGM (1990), which notably the Commission replaced by « les droits d’auteurs » in Guilde et AHGM (1991), when it finally granted certification to the Guilde. Based on the existing legislative context at that time, in conjunction with section 58.1 of the 1997 Copyright Amendment Act and section 16 of the Act, the phrase should be interpreted as the rights of authors, not copyright matters.

[96] Insofar as the constitutional argument is concerned, in our opinion, the SAA is not ultra vires the provincial government. The “pith and substance” doctrine recognizes that it is in practice, almost impossible for a legislature to exercise its jurisdiction over a specific matter effectively without incidentally affecting matters within the jurisdiction of another level of government. [36] The primary purpose of the SAA is to adapt labour law to the particularities of the cultural industry by establishing specific rules for collective bargaining between artists and producers. The pith and substance of the legislation is a matter (labour law) that falls within provincial jurisdiction. The SAA regulates, incidentally, the way in which copyright subject matters created in the context of the contractual relationship between the performer and the label can be transacted. It would indeed be paradoxical if a statute governing collective agreements between artists and producers could not contain stipulations relating to copyright.

[97] This interpretation of the SAA is compatible with the Act. Section 16 provides that “Nothing in section 15 prevents the performer from entering into a contract governing the use of the performer’s performance for the purpose of broadcasting, fixation or retransmission.” There are no restrictions. Performers can enter into contracts individually or collectively. [37] The general rules of contract apply to transactions involving copyright. It is therefore within the power of the province to provide a legislative framework to balance the bargaining power of performers and producers in their contract negotiations. It would be ultra vires only if the legislation imposed rules that are contrary to the Act.

[98] This is not the case and it is therefore necessary to assess the existence or scope of the performer’s reproduction right having regard to the collective agreements.

iii. The terms of the recording contracts vs. those set out in the collective agreements

[99] We have already found that individually, Québec performers generally grant to producers the right to use the fixation of their performances “for all purposes”. As a result, the performers’ 15(1)(b)(ii) right does not come into existence and ArtistI does not own the rights it claims unless the standard agreements contravene the UDA/ADISQ and Guilde/ADISQ collective agreements.

[100] In support of its claim, ArtistI notes that the reproductions at issue in these proceedings are made by the broadcasters, who are not party to the agreement. Therefore, even if producers were entitled to reproduce the fixation of a performance for the purposes in question in these proceedings, broadcasters are not, directly or indirectly. Moreover, the collective agreements provide that nothing in the agreement should be interpreted as a renunciation by the performers of any right to collect royalties pursuant to other legislative or contractual schemes.

[101] The CAB and AVLA/SOPROQ challenge the validity of ArtistI’s claim on the ground that the collective agreements cannot preclude producers from obtaining the right to reproduce the fixation of performances embodied in a sound recording in a recording contract. The UDA/ADISQ collective agreement authorizes the use of fixations for the purposes of commercial exploitation, promotion and marketing. According to the CAB and AVLA/SOPROQ, these terms are sufficiently broad to cover a broadcaster’s reproductions. The Guilde/ADISQ collective agreement provides that performers grant producers the exclusive, perpetual and irrevocable right to commercially exploit all fixations, by whatever means and in any medium. Therefore, in their submission, ArtistI does not own the rights it claims.

[102] Here again, for the purpose of our analysis, we assume that unless the agreement provides to the contrary implicitly or explicitly, an artist only grants the producer what is necessary to pursue the object of the recording agreement, that is, the commercial exploitation of the sound recording.

[103] The UDA/ADISQ collective agreement became effective on December 1, 1997. It is still in effect by reason of the automatic renewal provided for in the agreement until a new agreement is signed. Since it was concluded after April 25, 1996, section 58.1 of the 1997 Copyright Amendment Act does not apply to it. It is therefore possible that the agreement grants an interest in a copyright without specific stipulation in the agreement.

[104] In our opinion, the UDA/ADISQ agreement allows an artist to consent to considerably less than what is usually granted in the standard recording contracts. The agreement treats lead singers differently than background vocalists. For example, only lead singers sign exclusivity agreements. These differences do not have, however, any effect on our interpretation.

[105] We have examined a number of provisions in interpreting the collective agreement, which can be grouped into three categories: (1) the definitions, (2) the provisions that grant exploitation rights to producers, and (3) the provisions that recognize the rights of authors and performers pursuant to other legislative or contractual schemes. These provisions are reproduced in order to facilitate an understanding of our decision.

Definitions:

Clause 1-1.23 – Commercial exploitation: [TRANSLATION] “Manufacture, promotion, marketing, distribution and retail sale of any phonogram produced from a master tape.”

Clause 1-1.35 – Phonogram: [TRANSLATION] “Any medium enabling the reproduction of the acoustic performance of a work fixed in a master tape, by any means, that is intended for retail sale.”

Clause 1.01 of Schedule A – Multimedia phonogram: [TRANSLATION] “In the recording industry, any medium combining digital data in the form of text, sound and/or images (still or motion) that is intended for industrial reproduction for retail sale. For the purposes of this agreement, it may consist of the following: a) a recording primarily of sound that replaces the traditional phonogram and to which value has been added (for example, text and/or images) in order to make the featured artist more familiar to his or her audience; [...] ”

Exploitation rights – Producers:

Clause 3-1.02 – [TRANSLATION] “The producer may not commercially exploit phonograms [...], unless that phonogram has been produced in accordance with the provisions of this agreement.”

Clause 6-7.01 – [TRANSLATION] “The producer is responsible for the promotion of the featured artist’s phonogram. He shall ensure that the record label with which he deals respects the conditions provided for herein and in the exclusivity agreement.”

Clause 8-3.08 – [TRANSLATION] “In accordance with common industry practice, phonograms shipped as free goods by the producer’s distributor or licencee to promote sales do not incur the payment of royalties. However, only those free goods that have in fact been provided freely and shipped are exempt from the payment of royalties, and they may represent a maximum of 10 per cent of the phonograms sold in the two (2)-year period beginning on the date the phonograms enter the market.”

Clause 6.1 of Schedule A – This clause is preceded by the heading [TRANSLATION] “Information Highway” and reads as follows: [TRANSLATION] “For the purposes of promoting a phonogram, the use of a master tape or video on the Information Highway does not incur any additional payment.”

Authors’ and Performers’ Rights:

Clause 8-1.03 – [TRANSLATION] “Any exclusivity agreement between a producer and an artist is governed by this collective agreement and shall include the following clause:

iv. This exclusivity agreement is subject to the UDA/ADISQ Phonogram Agreement.”

Clause 8-3.06 – [TRANSLATION] “Each time an artist receives royalties from a collective society, pursuant to copyright (or neighbouring rights) legislation or resulting reciprocity agreements, for any form of commercial exploitation of the master tape other than the sale of a phonogram or a video, the royalties received by the producer for said other forms of exploitation are not considered gross royalties subject to recuperation.

Each party maintains his respective rights over said other forms of exploitation mentioned herein.”

Clause 8-4.01 – This clause is preceded by the heading [TRANSLATION] “Equitable Remuneration” and reads as follows: [TRANSLATION] “Nothing herein shall be interpreted as a waiver, in favour of the producer, of an artist's right or power to collect any amount owed to him or her personally by virtue of any legislation, Canadian or foreign, or resulting from a current or future agreement between users and collective societies.”

Clause 7.1 of Schedule A – [TRANSLATION] “Nothing in this letter of agreement shall be interpreted as a waiver of an artist’s right or power to collect any amount owed to him or her personally by virtue of any legislation, Canadian or foreign, or resulting from a current or future agreement between users and collective societies. Nothing shall prevent the artist from receiving through a collective society of which he would be a member the royalties owed to him by virtue of legislation or this collective agreement.”

[106] Based on these provisions, we find that ArtistI members who are subject to the collective agreement between UDA and ADISQ “own” some of the rights at play in these proceedings. Before 2000, to the extent that performers consented to producers using the fixation of their performances for purposes of commercial exploitation, promotion and marketing of the sound recording in which it is embodied, the authorization extended only to physical media. One reason for so finding is that the parties felt the need to supplement the agreement, with Schedule A, which targets Internet uses. Had “dematerialized” formats been covered in the original agreement, there would have been no need to sign the Schedule.

[107] Since the coming into force of Schedule A, performers can authorize producers to use the fixation of their performances on the Internet for promotional purposes. In light of the evidence respecting the changes in the business models of the labels and of the broadcasters, this includes the right to upload a copy of a phonogram to a MDS and to provide it with the authority to allow broadcasters to download a copy onto their servers. However, Schedule A does not extend to authorizing the labels to authorize broadcasters to make any subsequent copies using the MDS copy. When Schedule A was signed, it is unlikely that the parties had in mind the possibility that radio stations would play music off their servers as a matter of course. As a result, AstistI “owns” the right to authorize those subsequent copies.

[108] We note that since there is no 15(1)(b)(ii) right insofar as uses to which performers consent in the collective agreement by virtue of it never having materialized, clauses 8-3.06, 8-4.01 and of Schedule A, on which ArtistI bases its claim, do not apply.

[109] Given what we know about the life expectancy of most sound recordings played over the radio, it is likely that broadcasters mostly use only the part of ArtistI’s repertoire that is subject to Schedule A, that is, any fixation that was made since 2000.

[110] The Guilde/ADISQ Agreement came into force on the date it was signed, April 25, 1996, and is still in effect by virtue of the automatic renewal provisions in the agreement until a new agreement is signed. As we already concluded with the UDA agreement, since the Guilde/ADISQ agreement was signed on April 25, 1996, section 58.1 of the 1997 Copyright Amendment Act does not apply to it.

[111] This agreement, like the UDA/ADISQ agreement, applies to lead and background musicians although they are treated slightly differently. However, unlike the UDA/ADISQ agreement, it is less equivocal that their being treated differently has no consequence on our interpretation.

[112] Clause 12.02 states that [TRANSLATION] “However, the provisions of clause 12.01 do not prevent the featured artist from negotiating, in his or her exclusivity agreement, restrictions on the exploitation rights granted to the PRODUCER under clause 12.01.”

[113] Clause 12.09 states that [TRANSLATION] “Nothing in this agreement shall be interpreted as a waiver or assignment in favour of the PRODUCER of the musician’s right or power to collect or keep any amount to him or her by virtue of any legislation, convention or agreement, whether Canadian or foreign, or, whether current or future, in relation with the exploitation of the performance delivered under this agreement, as long as such exploitation is subject to neither the musician’s authorization nor the PRODUCER’s, such as, in particular, equitable remuneration or private copying.”

[114] The rights and obligations relating to the exploitation of the sound recording (phonogram) and the fixation of the performance embodied therein are dealt with in clauses 11.07, 11.08, and in chapter V, section 12 in addition to the Lettre d’entente No. 2 and 4.

[115] Clause 12.01 is preceded by the heading [TRANSLATION] “Remuneration through Royalties and Exclusivity Agreements” and provides that [TRANSLATION] “The signing of a contract for services entails assignment in favour of the PRODUCER of the right to fix any performance delivered by the musician pursuant to such contract for services, and, subject to payment of the fee provided for therein, conveys to the PRODUCER an exclusive, perpetual, irrevocable authorization to exploit any fixation so produced by any means and in any medium, known or yet unknown, without restriction as to time or territory.” [Our emphasis] This exclusive authorization to exploit the fixation of the performance amounts to an “all purposes” authorization, which would prevent the materialization of the musicians’ 15(1)(b)(ii) right.

[116] Clause 12.03, on the other hand, states the following: [TRANSLATION] “This chapter applies solely and exclusively to the lead musician. In this respect, any exclusivity agreement shall be initialed prior to the first rehearsal or recording session of the master tape in question [...].” Artist I submits that if, as clause 12.03 suggests, section 12 applies solely to lead musicians, clause 12.01 does not apply to accompanists. However, a reading of section 12 and the agreement as a whole does not support this interpretation.

[117] Whenever a provision in section 12 deals specifically with matters concerning exclusivity agreements, such as royalties (“redevances”) and exploitation rights of the lead musician and the obligations of the producers, the term “musicien vedette” is used. Indeed, only lead musicians sign exclusivity agreements with producers. As a result, they are the only ones who benefit from the “advantages” that flow from such agreements; conversely, producers do not face the same commitment and obligations vis-à-vis lead musicians and background musicians insofar as the commercial exploitation of the master. In our opinion, clause 12.03 serves to explicitly convey this reality rather than limit the application of clause 12.01 to lead musicians.

[118] Conversely, when a provision addresses an exploitation issue that applies notwithstanding the existence of an exclusivity agreement and therefore, applies to all musicians, the term “musicien” is employed, for example in clauses 12.01 and 12.09. Clause 12.01 speaks of “musicien”, not “musicien vedette” and of “contrat de service”, which all musicians must sign before the first recording session, as opposed to “entente d’exclusivité”, which only a “musicien vedette” signs.

[119] Clause 12.09 supports the proposition that some provisions of section 12 apply to background musicians. That provision is the only one that preserves the performers’ right to collect monies payable for the exploitation of the performance [TRANSLATION] “as long as the exploitation can be made without the authorization of either the musician or the producer [...]”. From a Canadian perspective, this is a clear reference to the remuneration right provided in section 19 of the Act. Section 19 grants a remuneration right to both front line and background musicians. If section 12 applied only to front line musicians, then it would be possible to conclude a contrario that producers can ask background musicians to give away their remuneration rights.

[120] Given the nature of the right to remuneration, it is not inconsistent to find that musicians would grant the producer, pursuant to clause 12.01, the right to use the fixation of their performances embodied in a phonogram for “all purposes”, thereby precluding the materialization of their 15(1)(b)(ii) right.

[121] The agreements peripheral to the main collective agreement serve to confirm our interpretation. Thus, the Lettre d’entente No. 2 relative aux nouvelles utilisations provides that the question of whether or not new uses of the master not foreseen in the collective agreement warrant additional remuneration and if so, the amount, will be determined by the joint committee or, failing this, an arbitrator. The reference to additional remuneration implies that the producer need not additional permission to engage in new forms of exploitation.

[122] In our opinion, the Lettre d’entente No. 2 is a counter balance to clause 12.01 pursuant to which the producer is vested with the exclusive right to use the fixation of the performance for the purpose of exploiting the phonogram in which it is incorporated. Finally, the Lettre d’entente No. 4 relative à la définition de “phonogramme” indicates that the uses at play in these proceedings are well within the scope of the exploitation activities covered by the agreement.

[123] ArtistI noted that the reproductions at issue in these proceedings are made by the broadcasters, who are not party to the UDA or Guilde agreements. This is not relevant. When the agreement between a performer and a producer is such that the producer controls a certain use of the fixed performances, the producer, not the performer, decides who can make that certain use and who cannot.

[124] We find that the Guilde/ADISQ agreement does not prevent Guilde members from dealing with their performances in a manner that precludes their 15(1)(b)(ii) right from materializing. Standard agreements allow record producers to use sound recordings on which performances are fixed in any manner they see fit. Consequently, we find that Guilde members cannot have assigned to ArtistI a right they themselves did not have. Our interpretation reconciles the ambiguities in a way that is consistent with both the legislative framework and the business realities of the recording industry.

B. In using “works” as they do, are radio stations making protected reproductions?

[125] The CAB contends that some of the broadcasters’ reproduction activities are not protected under the Act and should not attract royalties. The issue is not whether digital copies are in a “material form” but rather, whether all forms of digital reproduction contain the constitutive elements of a “reproduction” protected by the Act. The question is not whether the reproduction right is engaged but rather, what volume of reproductions broadcasters are liable for.

[126] There is no controversy or disagreement about the “subsequent copies”, which include the backups and other working copies made after the music file has been incorporated into the digital playback system. We do not need to deal with podcasts or other similar activities for two reasons: first, no one seeks to target those activities and second, there is no evidence on record in these proceedings that the broadcasters currently use music in their podcasting. Therefore, they do not require a licence for this programming delivery medium. Our analysis is thus limited to the MDS copy, simulcasting and auditioning.

[127] For an activity to be a “reproduction” protected by the Act, the act in question must involve: 1) the act of copying the protected work; [38] 2) the copying must be “substantial”; and 3) the resulting copy must be in a material form.

[128] Insofar as the first element, Théberge v. Galerie d’Art du Petit Champlain inc. [39] states that multiplication of the copies is a necessary consequence of the concept of reproduction. [40] It must involve “the physical making of something which did not exist before”. [41]

[129] With respect to the second element, the term “substantial” is not defined in the Act. Richard J.’s summary of what is involved in deciding what is substantial and what is not has stood the test of time:

“[w]hat constitutes a ‘substantial part’ is a question of fact and, in this respect, the courts have given more emphasis on the quality of what was taken from the original work rather than the quantity. [citation omitted] Some of the matters that have been considered by Courts in the past include:

(a) the quality and quantity of the material taken;

(b) the extent to which the defendant’s use adversely affects the plaintiff’s activities and diminishes the value of the plaintiff’s copyright;

(c) whether the material taken is the proper subject-matter of a copyright;

(d) whether the defendant intentionally appropriated the plaintiff’s work to save time and effort; and

(e) whether the material taken is used in the same or a similar fashion as the plaintiff’s.” [42]

[130] As for the third component, the Act provides no definition of “material form”. The jurisprudence dealing with this concept speaks to the question of copyright subsistence rather than copyright infringement and is therefore of limited assistance. For example, Canadian Admiral Corp. v. Rediffusion, Inc., [43] stands for the proposition that “for a work to be protected by the Act, a work must be expressed in some material form capable of identification and having a more or less permanent endurance.” [44]

[131] “Material form” insofar as copyright infringement is conceivably broader in scope since the case law recognizes digital temporary incidental copies as acts of reproduction. [45] Moreover, in Eros-Équipe, [46] the Federal Court defined material form in the context of copyright infringement. The Court stated that the ordinary meaning of material form meant something that was palpable, tangible and perceptible. [47] We were also referred to A & M Records Inc. v. Napster Inc. [48] to establish that downloading a musical file infringes the exclusive right of reproduction.

[132] There are two MDS copies. The first is the encrypted digital file that is “downloaded” from the MDS Web site and saved on the temporary memory of the broadcaster’s server. The second is the decrypted file that the decryption module located on the broadcaster’s server creates, using the encrypted file located on that same server, after the user provides the appropriate password to the MDS server. The CAB submits that broadcasters are not liable for the MDS copies on two grounds. First, they argue that the MDS, not the broadcasters, make the reproductions. The copies are made on the MDS server. Thereafter, the broadcasters merely save or store into their playback system the file that the MDS delivers to them in a non-material form. The CAB compares the MDS to an online music service.

[133] In the context of Internet transmissions, while words such as “copy” and “reproduction” probably have their ordinary meaning, others such as “transfer”, “deliver”, “download”, “send” and “file” are more metaphors than actual facts. The Internet itself is not used to transfer files but to transmit bits of data organized in packets. A file that is copied from a source computer to a destination computer never leaves the source computer as such. Neither is it made on the source computer and shipped to the destination computer. A file is usually stored not altogether but in separate sectors located throughout the relevant memory device.

[134] The contention that music files downloaded off a MDS Web site are created on the MDS server, “delivered” to a radio station and then “stored” on the station’s computer is factually flawed. The broadcaster creates both the encrypted and decrypted copy on the broadcaster’s server. It initiates the download; the decrypting software resides on its server. No file is “downloaded” from a MDS server onto a broadcaster’s server unless the broadcaster signs up for an account and thereafter, logs into its account and clicks on the download icon. The first constitutive element of a reproduction that is protected by the Act is satisfied.

[135] The CAB is correct to compare MDS to online music services, to the extent that MDS, like music services, enable subscribers to create music files on their computer using instructions sent to that subscriber by the service. This does not mean, however, that broadcasters are not engaged in making protected reproductions. For the purposes of its online music services tariff, CSI chose to target the services not only for the copies they make, but also for those their clients make. As a result, CSI does not need to deal with the end user. Here, as is its prerogative, CSI targeted the end-user, the broadcasters.

[136] The CAB contends in the alternative that the MDS copy is not protected by the Act until it is accessed as a digital file embodying music to be incorporated in a broadcaster’s music playback system. The MDS copy is a digital copy which emanates from another digital file. [49] Since the originating file is not protected by the Act by virtue of being in a non-material form, the downloaded copy cannot trigger the application of the Act. The MDS copy achieves the required level of materiality once it is incorporated in the broadcaster’s playback system and only thereafter do the digital reproduction activities trigger the application of the Act. The CAB claims that in a sense, the MDS copy that is added to the playback system is an “original” file.

[137] The CAB contrasts this delivery method with CD ripping. In CD ripping, the resulting digital copy emanates from a CD, which is in a material form and therefore, protected by the Act. Accordingly, the ripped copy that broadcasters add to their playback system immediately engages the reproduction right. Since broadcasters rely predominantly on MDS downloads to obtain new music, the rate or the tariff base should account for a decreased level of reproduction activity.

[138] This argument is untenable. The file sitting on the broadcaster’s server is not an “original”. The only original that exists is the master; all subsequent copies are reproductions and reproductions of reproductions. The encrypted digital file residing on the MDS server, the temporary encrypted file created on the broadcaster’s server and the decrypted file also residing on the broadcaster’s server before being incorporated into the playback system all are copies of the sound recording and underlying works and performances. Therefore, what is at issue is the existence of a reproduction of the sound recording, not the existence of a sound recording protected by copyright.

[139] The fact that a reproduction results from an “internal, fully digital, technical process” is not relevant. [50]

[140] The downloaded copy does not need to be incorporated into the playback system to trigger the application of the Act. The evidence establishes that the audio file is available for playback immediately after it is downloaded and it does not need to be ingested in the digital playback system. [51] Setting aside whether the temporary encrypted file may have met the “materiality exigency” since it was unusable at that stage, [52] that is not the case once the file is decrypted. After decryption, the broadcaster is able to use the MDS copy to make as many additional copies as it wishes.

[141] The MDS copy is just as much a reproduction as a permanent download. The online music service user who, after downloading a song, moves it to the “recycling bin” without ever playing it nevertheless reproduced the musical file. The same logic applies to the MDS copy that is downloaded onto the broadcasters’ server, whether or not it is ingested in the playback system.

[142] The MDS copy does not need to be incorporated into the playback system to be in a material form. Therefore, we conclude that broadcasters in downloading audio files from MDS services engage in reproductions that are protected by the Act.

[143] The song audition function is another service that MDS offer broadcasters. It allows them to listen to a song before deciding whether or not they want to download a certain song. According to the experts who testified, no downloading occurs when a station “auditions” or “previews” a song; that function is enabled by streaming. As noted in the past, [53] some forms of streaming nonetheless involve the creation of a temporary (or permanent) Internet file on the temporary folder of the end user’s computer. Once the file has been played, the temporary copy is “flushed away” or appears to be erased but remains indefinitely in the temporary Internet file folder.

[144] In this instance, this does not appear to be the case. Just as with the buffer in Satellite Radio Services, [54] it fails to reproduce a substantial part of the copyright subject matter. Consequently, we conclude that the right of reproduction is not engaged in auditioning songs from the MDS server.

[145] Finally, little attention was paid to the question of whether simulcasting, targeted by CSI’s proposed tariff, engages reproductions that are protected by the Act. No evidence was presented to address the issue; the comments made relating to simulcasting were intended merely to describe one of the relatively new media by which broadcasters deliver their programming.

[146] In Tariff 22.A, the Board defined simulcasting as “a portmanteau of ‘simultaneous broadcast’; it generally involves broadcasting a program or event at the same time on more than one medium. For the purpose of this decision, simulcast refers to streaming a radio or television signal simultaneously to its transmission over the air, on cable or by satellite.” [55]

[147] In a 2006 decision in respect of CSI for Commercial Radio, [56] the Board certified a tariff for simulcasting at the request of the parties. Consequently, the Board did not assess whether this technique met the requirements of the Act. Given that there is no objection to the inclusion of simulcasting and the lack of evidence adduced, we assume that radio broadcasters need and want the authorization to reproduce works embodied in sound recordings for simulcasting purposes. However, it would be helpful in the future if technical evidence was presented so that the Board could undertake such analysis.

C. Are radio stations liable for these reproductions under the respective tariffs of the reproduction collectives?

[148] To reproduce sound recordings of musical works as they do, radio stations must clear copyright with sound recordings makers, authors and, to some extent, performers. None of those copyright owners directly licence radio stations to make reproductions. However, the CAB submits a MDS already authorizes broadcasters to make these reproductions. The collectives contend that even if a MDS purports to authorize the broadcasters to download a copy of the tracks available on its Web site, it does so without any authority.

[149] The issues to be decided are whether and to what extent a MDS has the required authority to grant licences.

[150] To answer yes to the first question, there must be an agency relationship between the MDS and a rightsholder. This can arise in one of four ways: express mandate, implied mandate, ratification or apparent authority. In the present case, only express mandate and mandate by ratification can come into play.

[151] Insofar as authors are concerned, neither CSI nor its members have any relationship with any MDS. Labels have a relationship with the MDS for distribution purposes. However, the terms of the mechanical licences pursuant to which labels are authorized to reproduce musical works are too narrow for anyone to claim that copyright owners in those works have authorized the labels to reproduce them so as to allow any MDS to operate. A tariff has been in place for the use of works in CSI’s repertoire by commercial radio stations since 2001. As the broadcasters are well aware, CSI exclusively owns the right to licence the broadcasters’ reproductions of musical works. Consequently, any licence a label might grant to a MDS does not obviate the need for a licence from CSI to reproduce works embodied in sound recordings.

[152] Similarly, performers who are ArtistI members have no relationship with any MDS. Furthermore, their 15(1)(b)(ii) right, to the extent that it has materialized, is owned exclusively by the collective. Consequently, performers, like authors, cannot licence the use of their performances other than through ArtistI. Here again, the MDS licence cannot convey the right to use a performance; broadcasters need ArtistI’s authorization.

[153] On the other hand, owners of sound recordings have a direct relationship with the MDS. AVLA/SOPROQ argues that by filing the tariff, the labels made it clear that they do not authorize stations to reproduce their repertoire. They further submit that the Board does not have the power to interpret private licensing arrangements once the collective society has established its right to licence the uses in question and accordingly, MDS licences are inconsequential to the Board’s decision to certify a tariff.

[154] AVLA and SOPROQ cannot insist on users dealing with them unless the terms of their arrangements with rightsholders are such that they administer the rights to the exclusion of all others. AVLA and SOPROQ act as non-exclusive agents for their members; the labels remain free to licence their sound recordings directly. To the extent that a station is authorized, directly by a label or indirectly by a label’s agent, to make a copy of a sound recording, it does not need to deal with the label’s collective. To the extent broadcasters’ reproduction activities do not require the collectives’ permission, they should not pay royalties to the collectives. Accordingly, we must examine the contractual relationships that exist between the labels and each MDS and between each MDS and the broadcasters.

[155] As previously mentioned, labels predominantly use two MDS systems to distribute their sound recordings to broadcasters: DMDS in the Anglophone market, and 45tours in the Francophone market.

[156] In our opinion, the wording of agreements between labels, DMDS and broadcasters make it clear that DMDS is an agent of the labels. Labels grant DMDS a no-charge, non-exclusive license throughout Canada authorizing it to allow third parties to download and/or stream their music solely for the purpose of providing the Service, defined “as delivering music [...] via uploading and downloading to authorized users.” DMDS authorizes broadcasters to audition music and to download tracks for the purpose of adding them to their playlist.

[157] These agreements are public. Everyone knows their terms. A label confirms its participation in the licensing chain every time it uploads a music file on the DMDS server. Consequently, the labels, through DMDS, already licence the creation of the encrypted file onto the broadcaster’s server.

[158] That being said, the terms of the DMDS-broadcasters agreement limit the authorization to the initial encrypted and decrypted copies; subsequent copies are not covered by the licence. The agreement even stipulates that any other use will prejudice DMDS and its clients (the labels) and that it may pursue legal remedies in the event that the user breaches the agreement. In our opinion, the DMDS licence does not relieve broadcasters from having to comply with AVLA/SOPROQ’s tariff insofar as the subsequent copies are concerned.

[159] In our opinion, 45tours is also an agent of the labels. According to the evidence, the terms of agreement between labels and 45tours are implied. The labels merely create an account allowing them to upload their material onto the 45tours Web site and pay the amounts invoiced by it. The agency arises either implicitly or by ratification. It is implicit if we accept that as soon as it uploads music onto the 45tours server, a label implicitly confers upon 45tours the authority to act as its agent for some purposes. It is by ratification if, instead, we link the creation of the agency relationship to a label’s payment of the invoice of 45tours. The source and timing of the agency does not affect our analysis.

[160] In the absence of an express mandate, the scope of the resulting authorization must be assessed in light of the following facts. 45tours “signs” electronic agreements with broadcasters, but no copy of the agreements was placed in evidence and the agreements are not readily available. Therefore, the labels may or may not know its precise terms (though one would think that they would inquire before agreeing). The undisputed testimony of Mr. Sinclair [57] is that other MDS systems attach similar conditions to their licences as DMDS does.

[161] In the absence of clear evidence and given the foregoing factors, we favour an interpretation that is similar to the scope of the authorization given under the DMDS licence. Therefore, 45tours’ authority only extends to the MDS copy and excludes all subsequent copies. The broadcasters cannot rely on the 45tours licence to fully escape liability under the AVLA/SOPROQ tariff.

D. Can the Board set a tariff for reproductions made before a tariff takes effect?

[162] The AVLA/SOPROQ tariff will take effect on January 1, 2008, the ArtistI tariff one year later. On those dates, the broadcasters’ hard drives will already be loaded with most of the digital musical files in their current music library. Those copies will have been made before the tariffs take effect and without authorization from the rightsholders.

[163] ArtistI asks that its tariff apply to reproductions made before the tariff takes effect.

[164] AVLA/SOPROQ contends that the broadcasters need a licence to perform their regular backups of existing copies, and that the musical files are often modified for a variety of purposes. They argue we should not discount the tariff on account of pre-tariff copies.

[165] The CAB contends that the Board cannot certify tariffs for uses that occurred at a time for which no tariff proposal was filed. Furthermore, it contends the Board cannot monetize prior uses under the guise of requiring royalties for post-tariff uses.

[166] We agree partially with the CAB. The Board cannot certify tariffs for years for which no tariff proposal was filed. ArtistI’s reliance on sections 70.17 and 70.18 of the Act is misguided. Those provisions apply only once a tariff has been certified. The collectives’ remedy for copies made before a tariff takes effect is either to negotiate a licence or to sue broadcasters for copyright violation. If the CAB could prove that once a file is added to the musical library, it is never copied again, this would have to be taken into account in setting the tariff. Such is not the case, however. Broadcasters require a licence to make back ups, in the process of which the entire “library” is regularly reproduced. Furthermore, it may be within the discretion of the Board, in setting the terms and conditions of the tariff, to make paying for the use of preexisting copies a condition of making new copies; however, given the broadcasters’ copying practices, there is no need to use this approach in this instance. Radio stations will pay the full rate of the new tariffs from the first day they apply.

E. Can the Board, in a decision that certifies a tariff, clarify the wording of a regulation it made, or must it amend the regulation?

[167] Re:Sound’s royalties have always been calculated as a function of a station’s “advertising revenues” as this term is defined in the Regulations Defining "Advertising Revenues", [58] adopted by the Board pursuant to subsection 68.1(3) of the Act. These regulations provide that:

2.(1) For the purposes of subsection 68.1(1) of the Copyright Act, “advertising revenues” means the total compensation in money, goods or services, net of taxes and of commissions paid to advertising agencies, received by a system to advertise goods, services, activities or events, for broadcasting public interest messages or for any sponsorship.

(2) For the purpose of calculating advertising revenues, goods and services shall be valued at fair market value.

[168] The Regulations are needed because subparagraph 68.1(1)(a)(i) of the Act entitles broadcasters to pay Re:Sound only $100 in royalties on their first $1.25 million of annual “advertising revenues”. The Act does not impose any restriction or direction in determining what the rate base is for a SOCAN tariff. Prior to 2003, SOCAN’s royalties were set as a function of a station’s “gross income”. When the Board consolidated the SOCAN and Re:Sound tariffs, it chose to harmonize the rate base and apply the regulatory definition to both SOCAN and Re:Sound based on the assumption that both definitions represented the same rate base. [59]

[169] In 2006, Standard Radio Inc. (Standard) applied to the Board for a ruling that the Regulations permitted radio broadcasters to deduct the fair market value of production services provided under advertisement turn-key contracts from the rate base. The Board dismissed the application on the ground that it did not have the jurisdiction to interpret its own regulation outside the exercise of its core mandate. [60]

[170] Some broadcasters then applied for a declaration from the Federal Court to the same effect. After a trial, Zinn J. concluded that “[t]he Regulations Defining “Advertising Revenues”, permits a radio broadcaster to exclude the fair market value of the production services that it provides to advertisers from the revenues it generates from the broadcast of the ads to which those production services relate and upon which royalties are to be paid under the NRCC 1998-2002 Radio Tariff and the SOCAN-NRCC Commercial Radio Tariff 2003-2007.” [61]

[171] The Federal Court of Appeal allowed the appeal from the trial division after the record of these proceedings was perfected, and granted a declaration in the following terms:

The Regulations Defining "Advertising Revenues" [...] permits radio broadcasters to exclude from “advertising revenues” upon which they must pay royalties [...] any revenues that they derive from the production of advertisements. However, the mere fact that radio broadcasters incur costs in the production of advertisements under turn-key contracts or that their service is of value to advertisers does not prove that broadcasters have production revenues which must be excluded from “advertising revenues”. [62]

[172] The Federal Court of Appeal added that the characterization of revenue is largely a factual issue that must be determined at trial. [63] For reasons that will become clear, it was unnecessary to reopen this matter as a result of the decision on appeal.

[173] Re:Sound requested that we clarify when a station can deduct the fair market value of production service. In its submission, it should be limited to the amount that exceeds the fair market value of the airtime purchased by the advertisers. It contends that in so clarifying the interpretation we would be acting within our implied powers necessary to meet and be consistent with our statutory role. SOCAN essentially takes a similar position to the extent that the interpretation of the Regulations does not change its meaning.

[174] The CAB objects to the Board clarifying the Regulations primarily on the ground that issue estoppel precludes the Board from providing an interpretation different from that of the Federal Court. The CAB argues that the Board may explain or clarify the intended significance of the language used in the Regulations as long as in the course of doing so, it does not change or misstate its meaning and effect. Finally, it notes the Board has previously found that the resolution of disputes between parties concerning the interpretation of regulations is not such an exercise nor is it ancillary to such an exercise. [64]

[175] We agree for the most part with the CAB. The parties confuse interpretation and clarification. The power to clarify a regulation, even when the Board is properly seized of the issue in the course of exercising its mandate, is not one the legislator has conferred upon the Board for the reasons set out in the Standard Application. [65] The only tool available to the Board to clarify the Regulations is the same available to it to change the Regulations: a rule-making process.

[176] That being said, for the reasons which follow, Re:Sound is correct in pointing out that nothing in the Act requires us to use “advertising revenues” as defined in the Regulations as the rate base for Re:Sound.

F. In setting the rate base for the Re:Sound tariff, is the Board restricted to the regulatory definition of advertising revenues pursuant to subsection 68.1?

[177] Re:Sound proposed that we change its rate base to “gross income” while allowing broadcasters to pay 100$ in royalties on their first $1.25 million of advertising revenues.

[178] Subparagraph 68.1(1)(a)(i) of the Act provides:

68.1(1) Notwithstanding the tariffs approved by the Board under subsection 68(3) for the performance in public or the communication to the public by telecommunication of performer’s performances of musical works, or of sound recordings embodying such performer’s performances,

(a) wireless transmission systems, except community systems and public transmission systems, shall pay royalties as follows:

(i) in respect of each year, $100 on the first 1.25 million dollars of annual advertising revenues, and […]

[179] This subsection, allowing broadcasters to pay $100 on the first $1.25 million of annual advertising revenues, applies “notwithstanding” the tariffs as approved by the Board. Put another way, that subsection supersedes the tariff. It does not fetter the Board’s discretion to choose any rate base as long as it is reasonable. It simply allows broadcasters to carve out from the rate base an amount, as defined in the Regulations, and to treat this amount as if it did not exist for the purposes of the tariff.

G. Does section 90 of the Act preclude the Board from setting a single value or certifying an overall or single rate for broadcasting music?

[180] The CAB contends that as broadcasters consume music as a single input and not in a fragmented fashion, the Board should fix a single value and fix a single tariff for the collectives. The collectives oppose such an approach by reason that they each own individual and distinct rights, which by law must be valued separately.

[181] The collectives are correct but we have the discretion to set a single rate for the benefit of the user, as long as we apportion the royalty so established among the collectives so that the Board exhausts its jurisdiction.

VII. ECONOMIC ANALYSIS

[182] We have examined the methodologies proposed by the parties to set the rates and have decided not to use any of them. We will however analyze some of their underlying assumptions to explain why we have rejected them. We do so because we believe that economic modeling should remain central to the Board’s tariff determination.

A. The Wall/Globerman Methodology

[183] The Globerman model, which the Board used in Commercial Radio (2008), analyses changes in average productivity of music, advertising rates per unit of music audience and number of hours of broadcasted music to derive a value for music. Using a revised version of the Globerman model, Dr. Wall finds that the value of music has decreased by 10 per cent since 1987. Since SOCAN’s tariff was 3.2 per cent in 1987, the rate should be cut to 2.9 per cent.

[184] Dr. Wall then adds that music is a single input, whose price should be the same without regard to the number of rights required to use it. Relying on this assumption, Dr. Wall concludes that radio stations should pay 2.9 per cent for all the music rights they need, before any repertoire adjustment.

[185] Dr. Wall’s conclusion is not tenable for three reasons. First, as was explained in more detail in NRCC Tariff 3, [66] the Board sets rates for the use of rights attached to copyright subject matters, not for the use of copyright subject matters and even less for the use of clusters of such subject matters. Second, in Commercial Radio (2008), the Board proceeded in a manner that simply does not accord with Dr. Wall’s thesis. It used that methodology to value a single right, the right to communicate a musical work; only then did it derive an additional value for the right to communicate the performance and the recording. Third, the Globerman model can only serve to estimate the change in the value of identical inputs and therefore, in the value of the rights for which a price was being set at the start of the period, not of rights that either were not being priced (the reproduction right existed in 1987 but was not priced) and even less of rights, such as the performers’ rights, that did not even exist. Viewed from another angle, we somewhat doubt that the CAB would have endorsed Dr. Wall’s approach had SOCAN used it 20 years ago to seek an increase in royalties to account for the value of rights in music that SOCAN did not own, that radio stations did not use and that may not have even existed.

[186] In the alternative, Dr. Wall proposes to use the Globerman model to establish the SOCAN rate at 2.9 per cent. Dr. Wall’s use of the modified Globerman model raises a number of issues. One is the introduction of the notion of a seller’s reservation price, to reflect the value of the benefits which, according to Dr. Wall, radio airplay generates for rightsholders. The proportion of net benefits to be allocated to rights owners (or “pass-through”) is another. The number of years on which the benefits should be calculated (1987-2007 or 2005-2007) is a third. There is no need, however, for us to address these issues, for the reason that using the Globerman model, whether as originally conceived or as revised by Dr. Wall, would not be appropriate in this instance.

[187] In Commercial Radio (2008), the Board used the Globerman model to determine whether there was a historical undervaluation of music and to fix the amount of the efficiencies achieved through the use of music by radio stations between 1987 and 2005. The Board did so after concluding that the Globerman model provided a useful tool to examine long-term changes in the value of music. [67]

[188] The CAB maintains that the Globerman model remains a good approach and that its use provides stability and continuity in the Board’s analytical approach. Professor Liebowitz contends on the contrary that the Globerman model should not be used in this instance by reason that the variation in profits is what dictates changes in the tariff rate. In his opinion, using the Globerman model in 2008 made sense because the Board had already concluded that music was undervalued; as a result, profitability of the radio stations was also expected to increase. Using the Globerman model in other circumstances is problematic by reason that any movement in profits would lead to a change in the tariff rate, whether or not these changes are related to the value of music.

[189] We agree with Professor Liebowitz. The use of the Globerman model on a regular basis to set the rates would lead to regular rate changes, up or down, which is not desirable. Stability in the rates is generally beneficial to all. Until circumstances are fundamentally altered, tariff rates should remain the same.

B. The Bakos Methodology

[190] Professor Bakos, called by AVLA/SOPROQ, estimated the economic surplus generated from the use of sound recordings by commercial radio stations to determine the value that should accrue to all owners of copyright in music before proposing a tariff only for the reproduction of sound recordings. Professor Bakos found that the total value of the tariffs for the use of music by commercial radio stations should be between 21.1 per cent and 23.2 per cent of radio stations’ total revenues.

[191] Professor Bakos relied partly on a model proposed by Messrs. Audley, Boyer and Stohn in Commercial Radio (2005). The Board declined to use the model because it was based on unproven assumptions and because it was volatile. Professor Bakos refined the model in an attempt to address the Board’s concerns.

[192] The revised model assumes that radio stations allocate music and talk content so as to maximise their profit; at equilibrium, marginal net revenues of music and talk will be equal. Professor Bakos, using this framework, estimated that net revenues generated by music and talk represent respectively 74.3 and 7.2 per cent of total advertising revenues, the rest being accounted for by variable costs. Professor Bakos then allocated fixed costs to music in proportion to its net revenues. By subtracting these fixed costs from total variable costs due to music, he concluded that music royalties should represent 23.2 per cent of total revenues.

[193] Professor Bakos proposed a second approach, which he preferred, where royalties would be the result of a bargaining process. In such a scenario, the difference between net revenues generated by both music and talk and fixed costs (25.4 per cent) is shared between radio stations and music rights owners. In the digital music market, record labels appropriate a large proportion of the economic surplus associated with the sale of music. Professor Bakos assumes that music copyright owners would achieve the same result in the commercial radio market. The application of this calculation results in a proposed rate of 21.1 per cent for the bundle of rights.

[194] Professor Bakos also relied on the digital music industry to calculate the relative value of each right. Professor Bakos determined that 15.3 per cent of revenues in the digital music industry are paid to rightsholders of musical works. This lead him to conclude that musical works should get 15.3 per cent of commercial radio revenues while sound recordings should get the rest.

[195] Professor Bakos did not propose how to allocate the royalties attributable to sound recordings between the communication and reproduction rights. He simply concluded that the 4 per cent rate requested by AVLA/SOPROQ would be fully justified unless the new tariff for the communication of sound recordings was more than 13.5 per cent.

[196] For the reasons the Board expressed in Commercial Radio (2005), we will not use the methodology proposed by Professor Bakos. The adjustments he proposed do not address the main concern raised in 2005. Given the comments that Messrs. Audley and Boyer made on the Board’s 2005 decision, we think it may be helpful to explain again why the model is based on “unproven assumptions”.

[197] The Bakos model allocates revenues between music and talk based on advertising-adjusted music use. Music as a proportion of program time is used for each day part. These numbers are then adjusted to account for the amount of paid advertising and the amount of listener hours in each day part. The underlying assumption, as it was in Commercial Radio (2005), [68] is that a correlation exists between music use (when weighted by listening patterns) and the role of music in generating revenues. It is that assumption that is unproven, and even counterintuitive.

[198] Listening patterns indicate the number of persons who are tuned in to a radio station at any point in time. They do not provide any indication of why any person tunes in to a given station in the first place. Listeners must first be attracted to a particular station, and then be retained, before they generate advertising revenues. We cannot assume that if music, even weighted by listeners, accounts for 74 per cent of total advertising revenues, it necessarily follows that the same percentage of listeners are attracted to the station and retained because of music.

[199] A popular morning talk show with a star host is a useful example of this. Even if 60 per cent of the programming time is music, it is likely that many listeners are attracted to the station because of the host, and at least to some extent, listen to the music only “incidentally”, while waiting for the host to come back on-air. In this particular case, the weighted share of music in program time would overestimate the contribution of music to advertising revenue.

C. Nordicity’s Methodology

[200] AVLA/SOPROQ called Messrs. Lyman and Chodorowicz to estimate the cost savings commercial radio stations realize as a result of copying music, and to use these savings to derive a rate for the reproduction of sound recordings. Their assumption is that while all stations adopted and benefited from radio automation technology, only music stations benefited from the music reproduction component of radio automation. Therefore, they use the difference in cost savings between music and talk stations to estimate the value of music reproductions for music stations.

[201] The authors define the preautomation period as 1993 to 1996 and use 2000 to 2004 as the postautomation period based on the technical reports presented in this hearing. Their analysis is limited to savings in program and technical costs between the two periods.

[202] According to this model, music reproductions produced a cost savings of 5.2 per cent of total revenues for music stations. The authors contend that in any negotiation with the radio stations over the savings generated by the reproduction of sound recordings, rightsholders would be able to obtain 85 per cent of the savings because of their stronger bargaining position. According to their calculations, this would result in a rate of 4.5 per cent, increased to 4.7 per cent when expressed as a proportion of gross income.

[203] The model’s final adjustment is to allocate the reproduction royalty between the musical work and the sound recording. They agree with Messrs. Audley and Boyer that the full value of the reproduction right of musical works is about 1.3 per cent, which should be taken out of the 4.7 per cent. However, since a tariff for the reproduction of musical works was already certified for the postautomation period, the calculation already includes an effective CSI rate, which the authors estimate to be 0.61 per cent of total revenues. Therefore, the 4.7 per cent rate needs to be reduced only by 0.7 per cent, the difference between the full value of the reproduction right of musical works and the effective rate. In the end, Messrs. Lyman and Chodorowicz propose that the reproduction of sound recordings generate royalties of 4.0 per cent of gross income.

[204] The CAB outlined a number of problems with the model. Differences in the financial performance of music and talk stations can result from many factors which the model does not take into account, including audience levels, station size, broadcasting frequency band and reliance on on-air personalities. Cost differences between music and talk content could also be explained by increases in the cost of talk content. Finally, there is no solid basis for the choice of the pre- and postautomation periods, and yet this choice greatly affects the results. We agree with the CAB and, as a result, we conclude the model cannot be used to set rates.

D. Determination of Rates

[205] In 2005, the Board increased the SOCAN and Re:Sound rates significantly, from 3.2 per cent to an effective rate of 4.2 per cent in the case of SOCAN, and from 1.44 to 2.1 per cent in the case of Re:Sound. To do so, the Board relied on three reasons. First, there was a historical undervaluation of music that justified a correction of 10 per cent. Second, the use of music by radio stations had increased by 10.6 per cent since 1987. Finally, authors should receive 7.5 per cent more to account for the fact that radio stations realized efficiency gains from the use of music.

[206] The CAB successfully challenged the Board’s decision. The Federal Court of Appeal found that the reasons given were insufficient to justify the adjustments for the undervaluation of music and for the efficiency gains from the use of music. The Board was directed to explain how it arrived at these adjustments.

[207] The Board reheard the matter in 2008. Using, with some changes, the Globerman model proposed by the CAB, the Board found that the rates set in 2005 were appropriate and certified them again. More specifically, the Board noted that “What is clear is that the rate established in 2005 is consistent with the general trend generated by the Globerman model, and is as appropriate now as it was then.” [69]

[208] Until 2005, the SOCAN rate had remained basically unchanged for more than 25 years. The Re:Sound rate was first set in 1999, as a function of the SOCAN rate. In 2005 and 2008, using different methodologies and considering a number of factors that could impact on the value of music, the Board increased the effective SOCAN rate to 4.2 per cent and derived the Re:Sound rate using the same formula as in 1999. A new balance was created. Parties are now asking us to disrupt this recently created balance.

[209] There are benefits to the parties to having some stability in the rates. There are administrative costs associated with changes in tariff rates, structure and terms and conditions. The Board should try to minimize these costs as much as possible. It is also desirable for the Board to use a careful and reasoned approach to ensure that there is a degree of permanency in the changes affecting the rates before such changes form the basis for a change in the tariff rates, whether it be a rate increase or decrease. In this instance, the parties have failed to convince us that any of the methodologies they proposed ought to be used to recalculate the royalties payable to SOCAN or to Re:Sound.

[210] That leaves the issue of music use. The evidence submitted in this hearing is that music use by commercial radio stations has increased since 2005 by 6 to 6.5 per cent. This is consistent with the Board’s finding of 2005 that the increase in music use of “10.6 per cent Re:Sound derives from the Erin music study represents a minimum for any increase on this account. An adjustment of 13 per cent or more might better reflect the actual increase in music use.” [70]

[211] Having said that, the evidence does not establish whether this is a sufficiently permanent trend to warrant a permanent adjustment in the rate. Consequently, and again in the interest of stability, we will not adjust the SOCAN and Re:Sound rates.

i. SOCAN and Re:Sound Rates

[212] SOCAN and Re:Sound proposed the status quo, subject to an adjustment to account for an increase in music use. For the reasons we have just stated, we opt for the status quo. Consequently, for SOCAN, the effective rate for large stations will remain at 4.2 per cent, and the rate for low music use stations at 1.5 per cent.

[213] Currently, radio stations pay SOCAN 3.2 per cent on revenues not exceeding $1.25 million. SOCAN wants to abandon this lower rate, which the CAB would maintain. In support of its position, the CAB submitted evidence that continues to show that low revenue stations have negative profitability. It also requested that the $1.25 million threshold be indexed to inflation.

[214] Lower rates for lower-revenue stations remain justified for the reasons given in 2005. [71] As the Board did in 2005, and to maintain the effective SOCAN rate at approximately 4.2 per cent, we certify for SOCAN a rate of 3.2 per cent on the first $1.25 million of annual revenues and of 4.4 per cent on the rest.

[215] The rates for Re:Sound remain at 1.44 per cent on the first $1.25 million of annual revenues and at 2.1 per cent on the rest. The rate of 0.75 per cent applies to low music use stations. These rates are based on the finding that 50 per cent of all sound recordings played by radio stations are part of Re:Sound’s repertoire, a proportion agreed upon by Re:Sound and the CAB.

[216] The Re:Sound rates remain subject to subparagraph 68.1(1)(a)(i) of the Act which sets the amount broadcasters pay on the first $1.25 million of advertising revenues at $100 per year.

ii. CSI

[217] CSI argues that it should continue to receive 31.25 per cent (1 ÷ 3.2) of SOCAN’s rate, adjusted to reflect a 6.7 per cent increase in the use of music and to account for the fact that CSI’s repertoire represents 90 per cent of commercial radio air play. This results in a rate of 1.32 per cent.

[218] The CAB uses essentially the same approach, except that it uses its proposed rate of 2.9 as a starting point, makes no adjustment for increased music use and applies a repertoire adjustment of 80 per cent. The CAB also reduces the rate by a further 50 per cent to account for its argument that when a radio station uses a MDS to acquire digital files of music, no “material” reproductions are involved and as a result, the reproduction right is used at best for purely ancillary purposes. The final result is a rate of 0.37 per cent.

[219] We use the same methodology as the parties and use the SOCAN rate as the starting point to establish the CSI rate.

[220] When the existing CSI rate was set at 1 per cent before repertoire adjustment, the SOCAN rate was 3.2 per cent. The 1 to 3.2 ratio was never used as such; it resulted from the approach taken in 2003, by which the Board arrived at a rate of 1 per cent. The SOCAN rate has now increased to 4.4 per cent. Therefore, we should reevaluate whether to use the same ratio as before.

[221] The SOCAN rate was increased based on three factors: higher music use, higher historical value of music and higher efficiency of the use of music in generating advertising revenues. Both the communication and the reproduction rights are used by radio stations to generate higher advertising revenues arising from the higher value of music. It would be inconsistent for us to change the relative value of both rights when such a fundamental factor as the value of music affects both rights.

[222] In CSI – Commercial Radio (2003), [72] the Board found that new broadcasting technology meant lower costs for radio stations and that music reproduction on a hard drive allowed stations to optimize their broadcasting efficiencies. The evidence presented on both sides in this instance confirms that reproduction technologies allow stations to increase their efficiency and profitability.

[223] Therefore, all three factors justifying an increase in the SOCAN rate apply to CSI. CSI’s repertoire remains as undervalued as SOCAN’s repertoire was. A music use increase permanent enough to trigger a SOCAN rate increase should do the same for the CSI rate, as should efficiencies such as those described in 2005. We therefore apply the 1 to 3.2 ratio to the SOCAN rate of 4.4 per cent and establish a rate of 1.375 per cent for CSI.

[224] For the reasons given in considering the same adjustment for SOCAN, we decline to further adjust the CSI rate to reflect any increase in music use since 2005.

[225] That leaves the repertoire adjustment. CSI’s analysis demonstrates that its repertoire currently represents 89.52 per cent of the plays of musical works. SRG, having reviewed this analysis for the CAB, identified a number of concerns with respect to three issues: first, the possible over representation of musical plays for the station for which the Broadcast Data Systems (BDS) data was used; second, the possible duplication of some radio stations; and third, completeness of the BDS data. As a result, the CAB argued that the Board should make the repertoire adjustment using data from earlier decisions. CSI replied there is no over representation because of the weighting process, and there is no reason to believe that the reliance on the BDS data is problematic, since this data is widely recognized and used, including in some of the documents filed by the CAB.

[226] We accept CSI’s analysis. In 2003, the Board expected the level of representation of CSI’s repertoire to increase as additional rights were assigned to it. The evidence that radio broadcasters use CSI’s repertoire for 90 per cent of air play is consistent with this expectation. The application of this repertoire adjustment to the 1.375 rate results in a final CSI rate of 1.238 per cent for revenues above $1.25 million.

[227] The last certified CSI tariff applied different rates to revenues not exceeding $625,000, above $625,000 but not exceeding $1.25 million, and above $1.25 million. CSI proposes only one lower rate for revenues not exceeding $1.25 million, as is the case for SOCAN. The CAB objects to this on the basis that this would result in an implicit rate increase, but does not propose any alternative or specific methodology or rate.

[228] CSI’s proposal is for both an implicit and explicit increase. The removal of a lower revenue rate category constitutes an implicit rate increase, and CSI’s proposal also increases the rate to apply to both lower revenues categories. When the Board increased both SOCAN and Re:Sound rates, the rates were actually increased only on revenues above $1.25 million. Both SOCAN and Re:Sound rates remained unchanged for revenues not exceeding $1.25 million. We will therefore apply the same principle to CSI.

[229] The last certified CSI rates were 0.27 and 0.53 per cent for revenues not exceeding $625,000 and those above $625,000 but not exceeding $1.25 million, respectively. Removing the repertoire adjustment of 80 per cent done at the time leads to unadjusted rates of 0.338 and 0.663 per cent. We leave these unadjusted rates unchanged. The new, adjusted rates are obtained by applying a repertoire adjustment of 90 per cent. The new rates we certify for the lower revenues are therefore 0.304 and 0.597 per cent.

[230] Setting a single, lower rate for CSI on revenues not exceeding $1.25 million would help simplify tariff administration. However, there is no simple way of achieving this without implicitly increasing or decreasing the rate on lower revenues. Setting a single rate at 0.45, the average between the rates of 0.3 and 0.6 per cent, increases royalties for all very low revenue stations. Setting a single rate at 0.3 would decrease royalties for all stations with revenues above $625,000 but not exceeding $1.25 million. We thus prefer for now to remain with the current structure for CSI.

[231] We see no reason to consider a lower rate for radio stations that use a MDS. As stated earlier, downloading audio files from a MDS involves reproductions that are protected by copyright. A reproduction made using a MDS server is not different at law from a reproduction ripped from a physical CD, and has the same value for the broadcasters.

[232] For low music use stations, CSI proposes to apply the 1 to 3.2 ratio to the SOCAN rate of 1.4 per cent that existed for low music use stations prior to the 2005 decision. This results in a rate of 0.44. CSI then applies the adjustment the Board made to the SOCAN rate in 2005 that resulted in a 1.5 per cent low-use rate, to obtain 0.47 per cent. A final repertoire adjustment of 0.9 results in a final rate to 0.42 per cent. The CAB proposes to use the current discount of about 44 per cent to obtain the low-use rate. The CAB’s methodology imposes an implicit increase that was not applied to SOCAN.

[233] In 2005, the Board did not increase the low-use rate to account for higher music use or for additional music efficiencies. The SOCAN low-use rate was only increased by 10 per cent, to account for the historical undervaluation of music. This 10 per cent increase will be applied in this case in the following manner:

1. The last certified low-use rates for CSI are 0.12, 0.23 and 0.35 per cent, respectively, for revenues not exceeding $625,000, above $625,000 but not exceeding $1.25 million, and above $1.25 million. These rates, uncorrected for the repertoire, are 0.150, 0.288 and 0.438, respectively;

2. The 10 per cent increase applies to the higher of these rates, resulting in 0.482 per cent;

3. The other two rates are left unchanged, in line with the 2005 decision of leaving rate unchanged for low-tier revenues.

[234] The repertoire adjustment of 0.9 results in final rates of 0.135, 0.259 and 0.434 per cent, which we certify for low-use stations.

iii. ArtistI

[235] ArtistI proposed as a starting point the unadjusted rate of 1.47 per cent proposed by CSI. It then increased that rate to account for the difference in authors’ and performers’ revenues in the CD market. According to ArtistI, authors receive $1.14 per CD reproduced, while performers receive $1.88 or more. The resulting ratio is 1.65 (1.88 ÷ 1.14) which, when applied to the starting rate of 1.47 per cent, leads ArtistI to propose a rate of 2.43 per cent (1.65 × 1.47).

[236] The CAB challenges ArtistI’s approach on two fronts. First, since the Board has ruled that performing rights in sound recordings are worth the same as in musical works, the same should be true of the reproduction right. Second, ArtistI’s data overstates payments to performers arising from retail sales.

[237] In this instance, we have ruled, consistent with earlier decisions of the Board, that all things being equal, SOCAN’s and Re:Sound’s repertoires are worth the same. The CAB proposes that we use that same principle in setting the combined rate for AVLA/SOPROQ and for ArtistI. The CAB also proposes that we attribute equal values for the reproduction rights of makers and performers. Since the CAB proposes a rate (unadjusted for repertoire) of 0.91 per cent for CSI, ArtistI should receive 0.45 per cent.

[238] ArtistI argues that, similarly to what the Board has done in private copying, the value of the performer’s reproduction right should be higher than that of authors, reflecting the relative remuneration that can be observed in the CD market. We do not agree. The CD market is clearly a relevant proxy for private copying. The private copying regime in fact, exists precisely because it is a substitute for the CD market, which it could tend to replace. In addition, the basis for valuing the private copying tariffs is the value to the consumer, for which the physical CD market is a good proxy. Such is not the case in this instance. We are setting the relative value of the reproduction right of both makers and performers in the broadcasting industry. The basis for the valuation of these rates is the value that the use of the rights generates for radio stations, which is unrelated to the amounts performers receive in the CD market. [73]

[239] We agree with the CAB’s basic methodology, and believe that, as is the case for the communication right, the value to radio stations of the reproduction right of musical works is equal to that of the sound recording, which encompasses both the maker’s and the performer’s reproduction right.

[240] Having established a repertoire-unadjusted rate of 1.375 per cent for CSI, we use that rate to determine the value of the other elements contained in a sound recording: the performers’ performances and the recording itself. Therefore, the starting point for each is 0.688 per cent.

[241] ArtistI contends that it would administer 6.9 per cent of the repertoire if all its members had assigned to it their 15(1)(b)(ii) right. Moreover, the evidence presented by ArtistI showed that as the number of assignments rises, so does the use of its repertoire by commercial radio stations, which increased from 3.6 per cent to 5.6 per cent from September to the end of November 2008.

[242] ArtistI’s tariff commences in 2009 and for this reason, it is appropriate to use the most recent data. However, while it may be that ArtistI has the potential to attain a repertoire of 6.9 per cent of performances reproduced by radio stations, we have no evidence indicating that ArtistI will represent such use of their repertoire during the duration of this tariff. We therefore apply the 5.6 per cent adjustment to the 0.688 rate to obtain 0.039 per cent.

[243] ArtistI’s figures on repertoire use suppose that all its members own the 15(1)(b)(ii) right in all of their performances. We concluded earlier on that this is correct in respect of performances governed by the UDA collective agreement, but not of performances governed by the Guilde collective agreement. ArtistI filed evidence in this regard, at the request of the Board. Even though ArtistI itself recognizes the weaknesses of this evidence, it nevertheless allows us to estimate that approximately 40 per cent of ArtistI’s members are also subject to the Guilde collective agreement. This number is arrived at by calculating the proportion of all ArtistI’s members that are chefs d’orchestre, musiciens accompagnateurs or musiciens solistes. Removing this proportion leads to a rate of 0.023 per cent (0.039 × 0.6). Ideally, we should have reopened the matter to ask all the parties to comment on the approach we are using to segregate performances that are in ArtistI’s repertoire from those that are not. We decided against this for two reasons. First, the correction has no impact on the total amount of royalties paid by broadcasters. What is not due to ArtistI goes to AVLA/SOPROQ. Second, the amounts involved are simply too small to necessitate any length of debate on whether a better approximation can be made at this time: we leave the matter to the next hearings.

[244] We have used the CSI rate as the proxy to determine ArtistI’s main rate. We intend to do the same for setting the rates for revenues not exceeding $1.25 million and for low music use stations. The repertoire-unadjusted CSI rates for revenues not exceeding $625,000 and above $625,000 but not exceeding $1.25 million are 0.338 and 0.663 per cent, respectively. Starting from those rates, we bring an adjustment to account only for the performers’ share (divided by 2), for the repertoire (multiplied by 0.056) and for the proportion of performers that are members of the Guilde (multiplied by 0.6). The final rates are 0.006 and 0.011 per cent, respectively.

[245] CSI’s repertoire-unadjusted rates for low music use stations are 0.15, 0.288 and 0.482 per cent for revenues not exceeding $625,000, exceeding $625,000 but not over $1.25 million and above $1.25 million, respectively. Applying the adjustments described in the preceding paragraph leads to final rates of 0.003, 0.005 and 0.008 per cent, respectively.

[246] The CAB contends that the ArtistI tariff should be payable only by French language stations and that a nominal ArtistI rate should be set for English language stations to reflect the possibility of occasional repertoire use. We disagree with the CAB for the following reasons. First, the repertoire adjustment is based on a survey that included both French and English language stations. The adjustment thus already takes into account the fact that the use of ArtistI’s repertoire by English language stations is low. Second, stations not using ArtistI’s repertoire do not have to pay royalties to ArtistI. Third, ArtistI’s rate is already minimal and a nominal rate to reflect the possibility of occasional repertoire use would be insignificant.

iv. AVLA/SOPROQ

[247] AVLA/SOPROQ offered two approaches to justify its proposed tariff of 4 per cent: we rejected both.

[248] AVLA/SOPROQ represents makers of sound recordings. For the reasons we just set out, the starting point for makers should be the same as for performers, 0.688 per cent.

[249] A correction must be made to account for the fact that services such as DMDS, acting as the makers’ agent, authorize subscribers to make the initial reproduction of the sound recordings they download from the service. AVLA/SOPROQ should not be paid for those copies.

[250] The technical experts called by AVLA/SOPROQ and the CAB provided detailed descriptions of how the various reproductions take place at various stages in the broadcasters’ operations. The survey prepared by SRG for the CAB also provides an evaluation of the quantitative importance of each type of reproductions. According to SRG, an average of 4.9 songs per week are copied to the system by stations that add songs to their digital playback system. Of these additions, 2.8 songs come from DMDS, 0.8 from another electronic delivery system, and the rest from CDs or other sources. [74] A weekly average of 5 songs are transferred elsewhere, 3.4 copies are made for some specific purposes and 3.7 songs are edited for some specific purposes. [75] The total number of copies that radio stations make at this stage is on average, 17 per week.

[251] In addition, the report notes that almost all the stations surveyed back up some of their computer files, most on a daily basis. Finally, about half of the stations back up all the files each time while the other half back up only the files that have changed since the last back up. [76] If we assume, for stations that only back up the files that have changed, there is one relevant back up copy for every reproduction of musical files done during the week, we have a total of 34 reproductions in a week. This number is clearly a minimum because it will be much higher for stations that maintain more than one back up file, and for stations that back up all of the files each time.

[252] The total number of initial reproductions obtained through electronic delivery services is 3.6. The maximum discount that can be applied with respect to these copies is 10.6 per cent (3.6 ÷ 34). We have no evidence to permit us to measure precisely the relative importance of the digitally delivered, initial reproduction to the total number of music files that are being copied in a typical week, by reason that we do not know the total number of files copied each week. More importantly, the copies actually added to the playback system and used to broadcast music are more useful, and therefore more valuable than the initial, appraisal copies, some of which will never be used.

[253] In Satellite Radio Services, the Board made a similar adjustment to account for the fact that one service made assessment copies but no playback copies in Canada. The Board set the relative importance of the assessment copies at 5 per cent. In the absence of any contrary evidence, we use the same discount in this case for the MDS copies. In this instance, only 73 per cent (3.6 ÷ 4.9) of assessment copies are authorized by makers. This yields an adjustment of 3.7 per cent and an adjusted rate of 0.663 per cent (0.688 × 0.963).

[254] Here again, the CAB asked that this rate be reduced by half to account for the “non material” nature of MDS copies. For the reasons already given, we decline to do so.

[255] That leaves only the repertoire adjustment. AVLA/SOPROQ filed evidence that they represent 93.72 per cent of the sound recordings played by commercial radio stations which is uncontested by the CAB. This adjustment to the rate of 0.663 per cent, results in a final rate of 0.621 per cent.

[256] Makers who have secured the performers’ permission to exploit the sound recording in this market “own” as makers (and not as successors in title of performers) the economic value associated with the performance. The performance has an economic value whether a performer’s right attaches to it or whether that value is subsumed in the recording. That value must be accounted for in setting a rate for the repertoire of AVLA/SOPROQ. We have already concluded that, all things being equal, performers and makers should be paid the same. Given the nature of the evidence available to us, the simplest way to achieve this is to use the rate of 0.688 per cent as a starting point.

[257] First, we must remove what ArtistI receives on account of performances that are on AVLA/SOPROQ recordings, 0.023 percentage points. The result is 0.665 per cent.

[258] Second, we must remove performances on AVLA/SOPROQ recordings that are “owned” by other performers. According to the record of these proceedings, most producers outside Québec secure rights over the performance that are broad enough to include the reproductions targeted in these proceedings. The same is certainly true of American recordings, but may not be so of recordings made in some other countries. Absent any other evidence, we will assume that only 5 per cent of radio air play involves performances for which the reproduction right is still owned by the performer. The result is 0.632 per cent. That number must be adjusted for MDS copies and for repertoire. The result is 0.571 per cent. The final rate for AVLA/SOPROQ is therefore 1.192 per cent (0.621 + 0.571).

[259] We set the AVLA/SOPROQ rates for low income and low music use as we did for ArtistI. Starting from CSI’s low revenues rates of 0.338 and 0.663 per cent for revenues not exceeding $625,000 and revenues above $625,000 but not exceeding $1.25 million, respectively, we account for the maker’s share (divided by 2), for the MDS copies (multiplied by 0.964) and for the repertoire (multiplied by 0.9372). The final rates are 0.153 and 0.299, respectively.

[260] To account for the value of performances “owned” by the maker, we again follow similar steps. The starting points are the same CSI’s low revenues rates of 0.338 and 0.663 per cent. Taking only into account the performer’s share (divided by 2), removing ArtistI’s share (subtracting 0.023), the performances owned by other performers (multiplied by 0.95), the MDS copies (multiplied by 0.964) and adjusting for the repertoire (multiplied by 0.9372) leads to rates of 0.125 and 0.265, respectively. The final rates for lower revenues are thus 0.278 (0.153 + 0.125) and 0.564 (0.299 + 0.265) per cent for revenues not exceeding $625,000 and revenues above $625,000 but not exceeding $1.25 million, respectively.

[261] The rates for low music use stations are set in the same way. The starting rates are CSI low music use rates, unadjusted for repertoire. The adjustments described in paragraph 259 lead to final “sound recording” rates of 0.068, 0.13 and 0.218 respectively for revenues not exceeding $625,000, above $625,000 but not exceeding $1.25 million, and above $1.25 million. The adjustments described in paragraph 260 lead to “performances” rates of 0.045, 0.104 and 0.187, respectively. The final low music use rates are 0.113 (0.068 + 0.045), 0.234 (0.13 + 0.104) and 0.405 (0.218 + 0.187), respectively.

[262] Our task would have been considerably simpler had all collectives representing performers and sound recording makers formed some sort of alliance for the purposes of these proceedings, as CMRRA and SODRAC (and others) have done in the past. That would have made it unnecessary for us to speculate on the content of individual recording deals based on the terms of boiler-plate contracts and on certain assumptions about the extent to which these agreements truly represent what occurs in the market. Had AVLA/SOPROQ, ArtistI, ACTRA/PRS and AFM presented a united front, these issues would have become largely irrelevant. We can understand the need for these collectives to get some legal and interpretive issues clarified before sitting at the same table. We can only hope that this decision may provide these clarifications.

E. Very Low Use Rate Category

[263] Some talk stations have a music use level that is lower than 5 per cent. The CAB submits that these stations pay more than their use of music justifies. It proposes the creation of a new, very low music category, at half the rate of low-use stations. The CAB contends stations can easily identify and measure production and other compensable music to determine if the 5 per cent threshold applies.

[264] Re:Sound argues that the administration of the tariff would be complicated with the addition of a new music use threshold. It submits that for a significant proportion of low-use stations, music use fluctuates above and below 5 per cent on a daily basis. SOCAN and CSI agree with Re:Sound.

[265] In 2005, the Board eliminated the talk category because of the difficulties associated with the administration of a “no music use” category. Although the administration cost of a “5 per cent” category might be lower, we agree with Re:Sound that they are not insignificant.

[266] There are just as many reasons (if not more) to create a very high music use category as there are to create one for very low music use. This however would lead to an unduly complicated tariff. In certifying the current tariff, we have made decisions that tended to be consistent with the objective of harmonization and simplification of the tariffs as they relate to radio stations. Creating a very low-use category would go contrary to this objective and therefore, we will not do so.

F. Indexation of Tiers

[267] The CAB proposes that the revenue-tier levels be indexed to inflation starting in 2008. We will not do that. The Act sets at $100 the amount stations pay Re:Sound on the first $1.25 million of advertising revenues. The Re:Sound rate we certify adopts the same tier level of $1.25 million to define the lower revenue tier to which the reduced rate would apply in the absence of this subsection in the Act. Indexing the limit of the category would de facto introduce a third rate category, between $1.25 million and whatever indexed level we would set except for those few stations that may be able to segregate their advertising production revenues pursuant to Astral (Appeal). This is contrary to our objective of simplification. In addition, because of the current low rate of inflation, any benefits this would provide to the broadcasters are very small. In this case, our position is that simplification in the administration of the tariffs should be our primary objective. This issue however should be reexamined from time to time as economic conditions evolve.

G. Rate Base

[268] The CAB proposes that advertising revenues, as defined in the Regulations, be used as the rate base for all tariffs, thus allowing radio stations to exclude the fair market value of production revenues from the rate base.

[269] SOCAN requests the Board reintroduce the definition of gross income that was in use until Commercial Radio (2005). It also wants us to clarify that revenues pursuant to turn-key contracts are included in the rate base. SOCAN argues that, in any event, the revenues radio stations receive from turn-key contracts are revenues from the sale of air-time, not production services. SOCAN provided evidence that the change in the definition of “advertising revenues” could result in a significant decrease in the total royalties paid by broadcasters, close to 10 per cent.

[270] Re:Sound also requests that a common rate base be established for all collectives, based on gross income. This, Re:Sound claims, would standardize and simplify the tariff. Alternatively, if the Board were to use advertising revenues as the rate base, Re:Sound requests that broadcasters be allowed to deduct the fair value of production services, but only in excess of the fair value of the air-time purchased through the turn-key contract.

[271] CSI requests that its rate base be left unchanged, at gross income. CSI however adds that it would agree to the addition of wording in the definition of gross income that would allow broadcasters to exclude amounts received for the production of commercial advertisements.

[272] In its statement of case, AVLA/SOPROQ appears to agree with the CAB’s proposal to use advertising revenues as the rate base for all collectives. Finally, ArtistI is asking for gross income to be used as the rate base for its tariff.

[273] The rate base for Re:Sound’s commercial radio tariffs has always been advertising revenues as defined in the Regulations. Prior to 2003, SOCAN’s commercial radio tariffs were based on “gross income”.

[274] In 2005, the Board certified tariffs for both SOCAN and Re:Sound with respect to the years 2003-2007 using advertising revenues as the rate base. SOCAN had maintained that “gross income” could be broader than “advertising revenues”. However, in its 2005 decision, the Board stated that its understanding was that both definitions represented the same rate base and that if proved to be incorrect, SOCAN would have the opportunity to demonstrate this at a later date. [77]

[275] Following the 2005 decision, some stations took the position that the regulatory definition of advertising revenues excludes the fair market value of production services provided to advertisers, and contended they had overpaid royalties to Re:Sound since 1998 and to SOCAN since 2003. On October 24, 2008, the Federal Court concluded “that the value of production services in such a situation is not included in “advertising revenues”. [78] On January 18, 2010, the Federal Court of Appeal ruled that while production income is not part of advertising revenues, the value of production services is not the same as production income. [79] As a result, some of the arguments the parties advanced concerning the choice of rate base were rendered moot. Still, we think it useful to review them.

[276] While Re:Sound and SOCAN acknowledged the trial judge’s ruling, they believed that there is no evidence to support a fair market value of production services other than zero. The exclusion of any amount from the rate base on account of the value of production services would be arbitrary and inappropriate.

[277] Having reviewed the radio stations’ responses to certain interrogatories, Mr. Alan Mak concluded that stations do not perceive production to be a distinct revenue-generating function. Moreover, most radio stations indicated that air-time rates are not discounted for customers who do not require production services. This suggested to Mr. Mak that the fair market value of the production service is nil.

[278] Professor Hejazi examined the appropriateness of deducting the fair market value of production services from advertising revenues. In his opinion there is no economic justification to apply such an adjustment because there is no recorded accounting value for the production services. He added that since radio stations do not bill differently whether advertisers have air-ready commercials or not, the bargaining price or the fair market value of the production service is zero.

[279] We disagree with Professor Hejazi that the current fair economic value of production services is zero. While costs may be incurred without direct associated revenues (such as when a hotel offers free airport shuttle service), it does not mean that there is no economic value resulting from the service offered. Many businesses offer specific services that are not explicitly remunerated. The costs of these services are being nevertheless recouped in some other way, for instance when these “free” services happen to stimulate the sale of other services.

[280] Revenues associated with production services are not advertising revenues. They do not derive their value from the use of music per se. It is therefore difficult to argue that these revenues should remain in the rate base.

[281] Professor Hejazi reviewed five methodologies proposed by Standard in 2006, as part of a submission to the Board, to estimate the fair market value of production services. These are based on:

1. The fee charged to the advertiser when an ad produced by Standard is used on other radio stations.

2. The fee actually charged by third party production houses in the same market to produce a similar ad.

3. The cost of producing the spot, and the applicable wage scales of Standard radio employees.

4. The actual costs incurred by Standard in 2005 for producing spots for advertisers.

5. The 15 per cent agency discount or commission.

[282] Professor Hejazi concluded that each of these methodologies is flawed, and none represent valid measures to derive the fair market value. The first two methodologies cannot be used because neither commercials produced to be aired on other stations, or by third party production houses, is a perfect substitute for in-house production. The third and fourth methodologies are problematic because costs cannot be measured with sufficient accuracy, and even if they were, costs do not necessarily translate into fair market value. The 15 per cent commission referred to in the fifth methodology covers other factors than production services, such as the fact that agencies are registered and can back-stop the station’s receivables from the advertiser in the case of non-payment.

[283] We agree with Professor Hejazi’s conclusion that none of these proposed methodologies can be used to determine the fair market value of production services. We note however that the 15 per cent agency discount could represent the maximum estimation of the fair market value since production services are one of the benefits that radio stations receive from working with advertising agencies.

[284] As an additional potential avenue to measure the fair market value of production services, both Professor Hejazi and Mr. Mak examined the CAB’s responses to interrogatories. Professor Hejazi concluded that it is overwhelmingly the case that advertisers with air-ready commercials are charged the same price as those that do not. Responses to interrogatories also show that production costs of these services are not clearly tracked.

[285] Mr. Mak comes to similar conclusions with respect to the general trends observed in the responses to interrogatories. The Fairchild Radio Group was the only exception that separately records and reports revenues and costs of its production services and provides information about the price billed to an advertiser for the production services. Even there, the production fee is not charged to all customers; in any event, it constitutes between 2 and 10 per cent of total revenues.

[286] The Astral decisions interpret the Regulations and nothing more; they do not constrain us to exclude the costs, value or revenues associated with production services from the rate base. As a matter of principle, we believe that they should be. However, there is no evidence in these proceedings that would allow us to proceed with the adjustment, precisely because of the way production services are bundled with advertising contracts. Thus, in practice, they cannot be excluded from the rate base. We do not expect to be able to unbundle the value of these services until and unless the industry changes its current business practices.

[287] When the Board changed the rate base in its 2005 Decision from gross income to advertising revenues, it was of the opinion that both definitions represented the same rate base. In our opinion this is essential in reaching our decision regarding the appropriate rate base. SOCAN’s rate base was modified to achieve uniformity among the different radio tariffs. It was not the Board’s objective then to reduce royalties to be collected by the collectives. If we were now to achieve uniformity by applying advertising revenues as the base rate for CSI, AVLA/SOPROQ and ArtistI, royalties would be unintentionally decreased for all of these collectives. In our opinion, it is more appropriate to therefore use gross income as the rate base for all tariffs, including the Re:Sound tariff.

[288] Thus, we will use gross income as the tariff rate base for all the collectives including Re:Sound. However, the Re:Sound gross income rate base will factor in the statutory discount on the first $1.25 million of advertising revenues. Accordingly, radio stations will pay to Re:Sound $100 on the first $1.25 million of advertising revenues, as well as the set percentage of their gross income from which advertising revenues will be subtracted until the statutory threshold is surpassed. Thereafter, the radio stations will include advertising revenues as part of their gross income.

H. Internet-Related Activities

[289] Our goal is to achieve tariffs that are as uniform and easy to administer as possible. Hence, we would prefer to certify a tariff that targets the same activities, with a single rate base, for all collectives. However, this is not possible in this instance, since the ambit of the collectives’ proposals filed is not identical.

[290] The proposed tariffs of SOCAN and Re:Sound target over-the-air broadcasting: we cannot go beyond that.

[291] The proposed tariffs of AVLA/SOPROQ and ArtistI target all of a station’s broadcasting activities without further specification. This includes most or all of a radio station’s Internet transmissions. Yet, these collectives do not ask that these transmissions be targeted in the tariff; presumably, they want to licence those rights through negotiations, as they are entitled under the general regime. Given the total absence of evidence allowing us to set a fair rate in this respect, we prefer to leave the matter to the market, at least for the time being.

[292] This leaves CSI’s request that the tariff continue to apply to simulcasting, which we grant. The CAB is asking that to the extent that the CSI tariff applies to Internet activities, it provide for the same deductions as in SOCAN Tariff 22.B. There is no need to do so. Tariff 22.B is not limited to simulcasting; it targets all of a station’s Internet activities. The discount accounts for the presence of pages without any audio content on a radio station’s Web site. Simulcast only concerns the station’s broadcast (i.e. 100% audio) signal.

I. Individual Stations or Networks

[293] AVLA/SOPROQ proposed that revenues be calculated on a network or business basis instead of on a radio station basis. The CAB saw this proposal from AVLA/SOPROQ as an attempt to do an end-run around income-based tiering.

[294] We agree with the CAB. AVLA/SOPROQ did not present evidence necessary to justify such an adjustment. It may be difficult to define networks and radio companies in this context. Finally, such an adjustment would have needed a review of the impact on the total royalties paid and whether it would be appropriate to accordingly adjust the rates. We therefore reject the AVLA/SOPROQ proposal and maintain the current “per station” tariffs.

J. Final Certified Tariffs and Royalties Generated

[295] The annexed table summarizes the rates we are certifying for all collectives, in percentage points.

[296] Using the financial evidence provided by the CAB, [80] we estimate that the rates we certify will generate approximately $85 million in royalties for all five collectives for the year 2009. This estimate takes into account the reduced payment of $100 per year to Re:Sound on the first $1.25 million of advertising revenues. Of the total amount, about $51 million would go to SOCAN, $13 million to Re:Sound, $11 million to CSI, $10 million to AVLA/SOPROQ and $200,000 to ArtistI.

[297] Using again the financial evidence provided by the CAB, total revenues of radio stations are slightly over $1.5 billion for 2009. With total royalties amounting to $85 million, the effective total royalty rate of radio stations is 5.7 per cent. This is the measure of the proportion of revenues that are paid in royalties to all collective societies, when the lower rates for low income and low music use stations, as well as the Re:Sound reduced payment of $100, are taken into account.

[298] Using the rates certified for 2007, we estimate that radio stations would have paid royalties in the amount of about $72 million for 2009. This decision thus increases by $13 million the amount of royalties that will be paid by radio stations. We recognize that this is an important increase, even though it is barely one-fifth of the increase the collectives asked for. [81] We have not made any change to the fundamental value of music. The increase in royalties is essentially the result of the introduction of two new rates, for the reproduction of sound recordings and of performers’ performances, which broadcasters already used but for which they had yet to pay. We have also increased the CSI rate, to account for the higher value of music already set for SOCAN in earlier decisions and for the increased use of its repertoire.

VIII. TARIFF WORDING

[299] The following comments may help the reader to better understand the wording of the tariff. As is now the rule with any new tariff, as this is in part, we consulted the parties twice on this matter before reaching a final decision.

A. Years Covered

[300] The tariff’s application periods are staggered: 2008-2010 for SOCAN, 2008-2011 for Re:Sound and AVLA/SOPROQ, 2008-2012 for CSI and 2009-2011 for ArtistI. This is inherent in the regimes governing the proposed tariffs filed by the collectives. We cannot certify a SOCAN tariff for 2011: the proposed tariff was only filed in late March, 2010 and the 60-day period during which potential users can object has not expired. We cannot certify an ArtistI tariff for 2008 since it did not file a proposed tariff for that year. The tariff we certify deals with all the years for which the Board was seized at the time of issuing the decision.

B. Rate Base

[301] As we explained earlier, gross income is the tariff’s rate base. We use the expression as defined in earlier SOCAN and CSI tariffs. The wording of the provision is somewhat cumbersome and could be simplified. That being said, we kept changes to a minimum to make it clear that we wish the provision to be interpreted as it used to be before some stations attempted to carve out from the rate base gross amounts received by a station pursuant to turn-key contracts with advertisers. In order to leave no doubt in this respect, we specifically provide in paragraph (a) of the definition that this form of income is to be included in the rate base.

[302] The only other notable change concerns paragraph (c). The earlier version of this provision specified the station’s obligation to provide upon request documents confirming that revenues carved out of the rate base pursuant to the provision did qualify. As this is in the nature of a reporting provision, we moved it with the other administrative and reporting provisions in the tariff.

[303] As explained earlier, the rate base includes simulcast income only in respect of CSI.

C. Qualifying For Low Use

[304] Currently, stations that use music in SOCAN’s repertoire less than 20 per cent of air time pay royalties at a lower rate to both SOCAN and Re:Sound. Stations that use CSI’s repertoire for less than 20 per cent of air time also are entitled to a lower rate on the royalties payable to CSI. Both the CAB and the collectives agree that low use should attract lower royalties in all cases. The collectives proposed a number of formulas, some of which would require separate calculations. Keeping track of a station’s use of each repertoire is at least burdensome, and conceivably impossible. Some measure of simplification is required.

[305] We have split low uses in two categories. The use of the SOCAN repertoire will be the benchmark for SOCAN and CSI and the use of published sound recordings for all other collectives.

[306] There is no need to explain why a station should qualify for SOCAN low-use on the basis of its actual use of the repertoire of SOCAN. Qualifying for CSI low-use on the same basis also makes sense. CSI’s repertoire represents close to 90 per cent of air play. Every protected musical work enjoys both the communication and reproduction rights. The fact that CSI owns less music than SOCAN is already accounted for in the rate for normal use, which is the starting point to set the low-use rate. The approach is also simpler as it eliminates one calculation.

[307] Currently, Re:Sound low-use is based on the use of SOCAN’s repertoire. This is fair only if a close link exists between what is in that repertoire and the protected sound recordings commercial radio plays. That link is inherently questionable, since almost all American sound recordings do not qualify for the purposes of section 19 of the Act, while virtually all music found on American sound recordings played on commercial radio are in SOCAN’s repertoire.

[308] Furthermore, whatever the link may now be between the SOCAN and Re:Sound repertoires, it can only grow weaker with time. A classical music station uses little SOCAN music but plays mostly sound recordings that are still protected if they were fixed in a Rome country. On the other hand, oldies stations play a fair amount of public domain recordings of protected music, and this will only increase with time. Consequently, the low-use tariff for Re:Sound should be a function of some measure of use of published sound recordings.

[309] Re:Sound low-use should be a function of the use of all published sound recordings, not just of recordings qualifying pursuant to section 20 of the Act, for three reasons. First, the fact that American sound recordings do not qualify for so-called neighbouring rights royalties is already accounted for in the rate for normal use, the starting point to set the low-use rate. Second, the evidence tends to establish that Re:Sound’s repertoire includes virtually all protected sound recordings heard on commercial radio stations. Third, given how difficult it can be to determine on a case by case basis what qualifies and what does not, using a simple measure (is the sound recording published?) makes the tariff simpler to administer.

[310] The repertoire of AVLA/SOPROQ represents 94 per cent of air play. This much higher number than for Re:Sound is due in part to the fact that the eligible repertoire extends beyond recordings made in a Rome country, to include recordings that are released in a Rome country within 30 days of being published in a non-Rome country. [82] Virtually all non-Rome music that commercial radio plays is released in a Rome country simultaneously. As a result, measuring low use as a function of all published sound recordings also makes sense in the case of AVLA/SOPROQ.

[311] ArtistI’s repertoire is essentially composed of performances fixed on published sound recordings of musical works. Because Canada protects any performance fixed on a sound recording released in a Rome country within 30 days of publication in a non-Rome country, [83] virtually all performances fixed on non-Rome recordings that receive any air play on commercial radio are protected in Canada. As for Re:Sound, the low-use rate is already discounted for repertoire. It would appear therefore simpler and fair to use all published sound recordings to determine the low-use threshold.

[312] Until now, radio stations could also qualify for the CSI low-use rate if they did not copy musical works on a hard drive. Information supplied by the parties at the request of the Board confirms that no station relies solely on that provision to qualify for the low-use rate. Consequently, that provision can be removed. The CAB asks that the tariff expressly provide that a station that does not copy musical works need not pay the CSI tariff. That rule is implicit in all tariffs and need not be stated explicitly here.

D. Definition of Performer’s Performance

[313] ArtistI proposed a definition of “performer’s performance” that did not reflect the definition found in the Act. To the extent possible, we rely on the statutory definition. The sole purpose of the definition we use is to exclude performances that are not previously fixed so as to reflect the limited ambit of subparagraph 15(1)(b)(ii).

E. Calculating Royalties Using a “Reference Month”

[314] Currently, a station pays on March 1 royalties that are due for that month on the basis of its January income: in that instance, January is the “reference month”, used both to calculate royalties and to determine reporting obligations. As a result, a station obtains a free licence for its first two months of operation. Some collectives asked that this come to an end.

[315] The Board inquired of the parties whether they considered the problem to be serious or marginal. CSI raised a number of instances where it claimed the provision had created difficulties. In fact, as the CAB pointed out, almost all these instances related to enforcement difficulties that have nothing to do with the use of the reference month to calculate royalties.

[316] The Board also asked whether the problem, if it exists, might be resolved by requiring that royalties be calculated using the income of the actual month of broadcast but be payable at the end of the following month so as to allow stations to calculate their royalties. This is similar to the approach used in the retransmission tariffs. SOCAN agreed. The CAB seemed to favour the current system. Others did not comment.

[317] We continue to believe that the current approach is preferable. It requires that royalties be paid in advance, something more in accord with current practices in most copyright markets. Furthermore, since broadcasters are unlikely to earn much income during their first two months of operations, the issue probably remains marginal.

F. Authorizing Uses by Others

[318] Subsection 3(2) of the tariff entitles a station to authorize certain uses that facilitate the station’s own licensed uses. This provision is borrowed from earlier CSI tariffs and allows, for example, a MDS to reproduce a CSI work if the reproduction is used to send to the station the copy it will use in turn in its broadcast operations. We have extended the provision to all collectives except AVLA/SOPROQ, since these collectives informed us that such a provision would go beyond the scope of the rights they are mandated to license under the proposed tariff. As helpful as the provision might be to facilitate the stations’ operations, we cannot licence uses that a collective does not administer.

G. Tailoring the Special Provisions of Paragraph 68.1(1)(a)

[319] Re:Sound’s proposed tariff included a provision that attempts to limit the ambit of the regulatory definition of “advertising revenues”. Quite apart from whether this would be desirable, a tariff cannot limit the ambit of a regulation.

H. Music Use Information

[320] The discussion surrounding music use information raised three main issues.

[321] The first is whether radio stations should provide sequential lists of the music they play and if so, how this would interface with other music use information requirements. Everyone agrees that sequential lists should be provided where available. They ensure more accurate payouts and generate significant savings in music use monitoring. That being said, to require all stations to provide these lists may not be practical in this market, at least not at this time. While the SRG report confirms that the number of smaller stations using programming software has increased, we do not know whether all stations currently operate with a computerized system, or whether those that do are able to generate sequential lists. Also, compatibility issues may exist that cannot be resolved in the short term.

[322] The Board asked the parties whether stations that provide playlists should be exempt from responding to any further music information requests on the part of the collectives. Generally speaking, collectives agreed that a station should be so dispensed only if the playlist contains all the information to which a collective is entitled when a “paper” request is made.

[323] We agree that a station should not be exempt from complying with a music use information request if it provides sequential lists that contain less than the minimum required. Therefore, the tariff will provide that if a list is available and does not comply with paragraph 10(1)(a), the station must provide the list but is still subject to requests for information on repertoire use.

[324] Re:Sound proposed that a station be required to make an election at the beginning of the year. We think on the contrary that if a station can provide a list, it should not have the option to withhold it.

[325] Re:Sound also suggested that sequential lists should be mandatory for stations using electronic systems and should be provided in a standard format to be agreed upon by the parties. We disagree. In this decision, we wish to leave a fair measure of flexibility to stations, for one last time. As for format, parties are free to agree on one, but we cannot compel them to come to such an agreement, as this may involve some form of illegal subdelegation. For the same reason, we cannot, as AVLA/SOPROQ suggested, allow the collectives to determine whether a playlist is “satisfactory”.

[326] The second issue is whether the requirement to provide music use information should be increased. Currently, SOCAN and Re:Sound are entitled to a total of 14 days information per year. CSI is entitled to its own 14 days, though the Board stated from the outset its expectation all three collectives cooperate with each other. The CAB asked that reporting obligations remain capped at 14 days in total and that more time be allowed to report. The collectives countered that the current requirement is insufficient for comprehensive distribution and falls well below both international standards and other existing tariffs, including satellite radio, pay audio and online music.

[327] We agree that the current requirement is insufficient, if only because of the rising expectation on the part of rightsholders that royalties “follow the dollar”. We therefore increase the number of reporting days to 28. We will limit the impact of this increase by providing that music use information is to be shared among all collectives.

[328] The third issue is whether the requirement to provide information should be limited to what is available to a station. According to the CAB, the date and time of broadcast, title of work, name of performers or performing groups and duration are always available, the name of author and composer, title of album and label are sometimes available, and the Universal Product Code (UPC) and the International Standard Recording Code (ISRC) never are. Since CSI is already entitled to be informed of the UPC and ISRC “if included in the station’s computerized logging system”, we inquired about the extent to which that information is actually supplied. CSI answered that it never gets the information.

[329] The CAB’s answers lead us to conclude that broadcasters do not currently comply with the tariffs. Radio stations are already required to provide at a minimum the date and time of the broadcast, the title of the work, the name of its author and composer and, where applicable (not where available), the title of the album, the name of the performer or performing group and the record label. Yet the CAB informs us that the name of author and composer, title of album and label are not always available. Furthermore, the answers of CSI and of the CAB taken together would tend to show that broadcasters are resisting updating their databases to suit the collectives’ purposes.

[330] The CAB argued once more that stations should not be compelled, beyond reasonable means, to provide information that requires additional significant expenditures to generate. Once more, we conclude that in the long run, all of the information mentioned above is necessary to ensure that repertoire use is properly monitored (something necessary to ensure the fairness of future tariffs) and that royalties are paid to the appropriate rightsholders. In the long run, then, what the CAB proposes is not acceptable.

[331] For one last time, we are willing to allow radio stations some measure of flexibility. The reporting requirements will remain as in the past and as a result, stations will not be required to report the UPC and ISRC if these are not included in their computer logging system. We agree that comparing radio stations to services that were digitized from the beginning (satellite radio, online music) is at best imperfect. We also know that sooner, rather than later, we intend radio stations to comply with similar information requirements as in those tariffs, especially if, as the record seems to indicate and a future panel will no doubt wish to confirm, radio stations in other countries already comply with similar requirements as a matter of course. We remain convinced that it is the broadcaster’s responsibility to request and obtain from their music suppliers the information collectives are entitled to under a tariff and to input all the information included on or with a physical CD when digitizing their library. The fact the Canadian Broadcasting Corporation, which operates 2 talk networks, already provides a complete sequential list of music use for the entire year shows it is possible to do so with the proper equipment even for talk, news and sports stations.

I. Adjustments

[332] Re:Sound, AVLA/SOPROQ and ArtistI asked that they be allowed to collect royalty underpayments indefinitely, but that stations may no longer ask a refund for overpayments after 12 months. According to these collectives, it is up to the station to discover an overpayment since the information required to establish this always is in the hands of the broadcasters. It would be unsustainable to have to pay back royalties which have already been distributed. In contrast, any limitation for underpayments could incite broadcasters to misstate their revenues, thereby enhancing the need for auditing and other inefficiencies. We agree with the last point but not with the others. No existing tariff provides for such a time limit. Normal limitation time periods should suffice to minimize disruptions in the collectives’ internal operations.

J. Audits

[333] Currently, the CSI tariff specifies that the collective is entitled to have access to the broadcast day recordings of a low-use station. AVLA/SOPROQ, and ArtistI asked that this provision also apply to them. The SOCAN-Re:Sound tariff is silent in this respect. The provision is unnecessary and potentially counter-productive. A low-use station that does not comply with the audit provisions remains a low-use station. A station that does not comply with the definition of a low-use station, which includes agreeing to make the recordings available, no longer is a low-use station.

K. Notices

[334] More recent tariffs allow information to be sent by file transfer protocol and payments to be made by electronic bank transfer. All agree with respect to information. Collectives ask that bank transfers be left to agreements between collectives and stations. Electronic bank transfers can be inconvenient to manage and transaction costs are quite high, particularly given the number of unique payments generated by the commercial radio tariff as compared with tariffs governing smaller groups of licensees. SOCAN already allows commercial radio stations to pay in this manner. ArtistI proposed setting a minimum. It is preferable to leave the use of electronic bank transfer to agreements between collectives and stations.

L. Transitional Provisions

[335] This tariff is a significant departure from the previous tariffs. To help broadcasters absorb the increase of the rates more easily, the tariff allows them to pay additional amounts owed for past periods free of interest, over a period of one year with one exception. The amounts owed to ArtistI is modest enough that radio stations ought to pay them all at once.

Gilles McDougall

Acting Secretary General

ANNEX / ANNEXE

Rates Certified by the Board, In Percentage of Gross Income
Taux homologués par la Commission, en pourcentage des revenus bruts

Rates Certified by the Board, In Percentage of Gross Income

	SOCAN	Re:Sound Ré:Sonne	CSI	AVLA/ SOPROQ	ArtistI
Low Music Use Stations Stations utilisant peu de musique
For revenues not exceeding $625,000 Pour des revenus ne dépassant pas 625 000 $	1.5	0.75	0.135	0.113	0.003
For revenues above $625,000 but not exceeding $1.25M Pour des revenus supérieurs à 625 000 $ mais ne dépassant pas 1,25 million de dollars	1.5	0.75	0.259	0.234	0.005
For revenues above $1.25M Pour des revenus supérieurs à 1,25 million de dollars	1.5	0.75	0.434	0.405	0.008
Other Stations / Autres stations
For revenues not exceeding $625,000 Pour des revenus ne dépassant pas 625 000 $	3.2	1.44	0.304	0.278	0.006
For revenues above $625,000 but not exceeding $1.25M Pour des revenus supérieurs à 625 000 $ mais ne dépassant pas 1,25 million de dollars	3.2	1.44	0.597	0.564	0.011
For revenues above $1.25M Pour des revenus supérieurs à 1,25 million de dollars	4.4	2.1	1.238	1.192	0.023

Note: The Re:Sound rates are subject to subparagraph 68.1(1)(a)(i) of the Act which sets the amount broadcasters pay on advertising revenues not exceeding $1.25 million at $100 per year.