Copyright Board Canada

Commission du droit d’auteur Canada

 

Date	2014-05-16
Citation	File: Public Performance of Sound Recordings
Regime	Collective Administration of Performing Rights and of Communication Rights Copyright Act, subsection 68(3)
Members	The Honourable William J. Vancise Mr. Claude Majeau Mr. J. Nelson Landry
Proposed Tariffs Considered	Re:Sound No. Tariff 8 – Non-interactive and semi-interactive webcasts, 2009-2012

statement of royalties to be collected for the performance in public or the communication to the public by telecommunication, in canada, of published sound recordings embodying musical works and performers’ performances of such works

Reasons for decision

I. INTRODUCTION 1

II. RIGHTS AND ACTIVITIES TARGETED: DEFINITIONS 2

III. THE PARTIES AND THEIR PROPOSALS 5

A. Re:Sound 5

B. CAB 6

C. Pandora 6

D. CBC 6

E. NCRA 7

F. Services 7

IV. EVIDENCE 8

A. Re:Sound 8

B. CAB and Pandora 9

C. CBC 11

D. NCRA 11

E. Services 12

V. ANALYSIS 12

A. Issues Relevant to All Matters Under Examination 12

i. Simulcast 12

ii. Webcasts as Core Music Consumption 13

iii. Music as a Single Ecosystem 15

iv. Webcasting as a Global Industry 16

v. User-Based vs Use-Based Tariff 16

vi. A “Greater Of” Formula 17

vii. SRPC and Related Technical Issues 18

viii. Archived programming 18

B. Rates for Webcasters Other than CBC and Community Systems 18

i. Percentage vs Per Play 18

ii. Selecting an Appropriate Proxy 20

C. American and Foreign Rates 20

D. Agreements between Record Labels and Webcasters 22

E. Agreements Between Webcasters and Re:Sound 22

F. Existing Tariffs as Proxies 27

i. Selecting the Appropriate Tariff as Proxy 28

ii. Partial Plays 31

iii. An Interactivity Premium 32

iv. Repertoire Adjustment 33

v. Minima 34

G. Rates for CBC 34

H. Rates for Webcasting Operations of Community and Non-Commercial Radio Systems 36

I. Certified Rates and Royalties Generated 36

VI. TARIFF WORDING, STRUCTURE AND TRANSITIONAL PROVISIONS 38

A. Definitions 39

B. Non-interactive and semi-interactive webcasts 39

C. Non-commercial webcaster 40

D. Webcasts 40

E. Partial plays 40

F. Channel 41

G. Transitional provisions 41

H. Interest 42

I. INTRODUCTION

[1] What we term “radio” today has expanded well beyond the delivery of audio content through Hertzian waves, to include offerings on other platforms such as over-the-air radio station simulcasts over the Internet, cable and satellite pay audio services, free standing, subscription-based mobile satellite services as well as Internet-only stations and other forms of webcasts or streams. These services are required to pay royalties for the right to transmit published sound recordings of musical works. This decision deals with three forms of audio streams, defined in paragraphs 12 to 14 below: simulcasts, non-interactive webcasts and semi-interactive webcasts.

[2] On March 28, 2008, pursuant to subsection 67.1(2) of the Copyright Act [1] (the “Act”), Re:Sound Music Licensing Company (Re:Sound) filed a statement of proposed royalties for the communication to the public by telecommunication, by way of simulcasting and non-interactive webcasting, [2] of published sound recordings embodying musical works and performers’ performances of such works for the years 2009 to 2012 (Tariff 8.A). On March 31, 2010, Re:Sound filed a statement of proposed royalties for communications by way of semi-interactive webcasting for the years 2011 and 2012 (Tariff 8.B). The statements were published in the Canada Gazette on May 31, 2008 and July 24, 2010 respectively, along with a notice concerning the right of prospective users to object to them.

[3] Objectors to one or both proposed tariffs were the Alliance des radios communautaires du Canada, the Association des radiodiffuseurs communautaires du Québec and the National Campus and Community Radio Association (collectively, NCRA); the Canadian Association of Broadcasters (CAB); the Canadian Broadcasting Corporation (CBC); Quebecor Media Inc., Rogers Communications Inc. and Shaw Communications Inc. (collectively, the Services); Pandora Media Inc. (Pandora); Digital Media Association and its Members (DiMA); and Planetnerve8 Inc. (Planetnerve). On March 29, 2012, DiMA and Planetnerve, not having provided answers to interrogatories Re:Sound had addressed to them, were deemed to have withdrawn from the proceedings.

[4] On January 26, 2011, at the request of Re:Sound and with the agreement of the Objectors, the examination of both proposed tariffs was consolidated.

[5] The matter was heard over ten days in September and October of 2012. The record was perfected on November 6, 2012 when closing arguments were heard.

II. RIGHTS AND ACTIVITIES TARGETED: DEFINITIONS

[6] The Board has already certified certain tariffs for Internet streaming. The CSI Online Music Services Tariff, 2008-2010 [3] deals with the reproduction of musical works in permanent downloads, limited downloads, and on-demand streams. SOCAN Tariff 22.A (Online Music Services), 2007-2010 [4] sets the royalties payable to communicate musical works for on-demand streams. The SOCAN-Re:Sound CBC Radio Tariff, 2006-2011 [5] targets the simulcasting of CBC’s over-the-air radio signals in addition to the original broadcast. The Internet activities of a variety of users are addressed in SOCAN Tariffs 22.B (Commercial Radio), 22.C (Non-Commercial Radio), 22.D (Commercial Television, Non-Broadcast Television, Pay Audio Services, Satellite Radio), 22.E (Canadian Broadcasting Corporation, Ontario Educational Communications Authority, Société de télédiffusion du Québec), 22.F (Audio Websites) and 22.G (Game Sites) (collectively “SOCAN Tariffs 22.B to G”). [6] Royalties for some other streaming activities are governed by private licensing deals.

[7] Streaming music over the Internet can involve as many as six rights or sets of rights. [7] These proceedings only concern the equitable remuneration to which performers and makers are entitled when a published sound recording of a musical work is communicated to the public by telecommunication. [8] These two rights [9] always trigger a single payment for any type of sound recording; 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. [10] Re:Sound administers these rights for the vast majority of eligible performers and makers.

[8] The following exclusive rights are not at play in these proceedings: the right to communicate a musical work to the public by telecommunication; the right to reproduce a musical work; the right to reproduce a sound recording; the right to reproduce any reproduction of an authorized fixation of a performer’s performance for a purpose other than that for which the authorization was given; the rights granted, on November 7, 2012, to Canadian performers and makers over the communication resulting from making available a sound recording to the public. [11]

[9] At the outset, we wish to explain the meaning we ascribe to certain terms and expressions.

[10] Generally speaking, music is available on the Internet as downloads or streams. A sound recording is downloaded when a file that contains the recording is sent to and stored on a recipient’s device. A recording is streamed when only enough data is transmitted or cached to allow the recipient to listen to the recording at the time of the transmission; in principle, [12] the recipient is unable to store the recording onto a medium or device for later use.

[11] The proposed tariffs target three types of streams: simulcasts, semi-interactive webcasts and non-interactive webcasts.

[12] A simulcast is the substantially simultaneous streaming of a signal via the Internet or another digital network. The content is identical in the two transmissions.

[13] A semi-interactive webcast enables only the (near) real-time communication of the streamed content, but allows recipients to exert some measure of influence over the content or timing of the webcast. Just how much influence can be exerted will vary. Under Re:Sound’s proposed definition, recipients would be allowed to provide to the webcaster information, such as the musical genre or artists they prefer, which in turn would be used to alter the content of a stream. Recipients would also be allowed to skip through the stream by advancing to the start of the next recording. They would not be allowed to pause, skip back, repeat, rewind or fast forward the stream, or to access a specific recording from a place and at a time of their choice. Pandora, Galaxie Mobile, Songza, Hou5e, Last.fm and Accuradio, among others, offer semi-interactive services.

[14] Generally speaking, a non-interactive webcast refers to a stream, other than a simulcast, over which the recipient exercises no control over the content of the signal or the time of the transmission. This includes the purely passive Internet radio services offered by Songza and Hou5e, as well as the numerous Internet radio “stations” offered by individuals on a platform such as Live365. Re:Sound’s proposed definition goes beyond this, to include “a webcast other than a simulcast, a semi-interactive webcast or an interactive communication.” In other words, this definition applies if no other definition does. This raises a number of issues that are addressed below. [13]

[15] An interactive communication is any form of streaming where a specific sound recording can be listened to at a time and place of the recipient’s choosing. This includes services such as Spotify and Rdio.

[16] A play [14] means the single performance of a single sound recording to a single person.

[17] Section 2 of SOCAN Tariffs 22.B to G defines a page impression as a request to load a single page from a site and an audio page impression as a page impression that allows a person to hear a sound. [15] Users can use the ratio of audio page impressions to all page impressions to lower their royalties if the ratio is lower than the default ratios that are set in the tariff.

[18] An archived program is a program which can be accessed through streaming after having initially been transmitted over the air, via satellite or via cable. The archived program will contain only what was in the original broadcast, though it may not contain everything from that broadcast: some parts may have been removed to account for copyright issues. It is offered as a single file: the recipient cannot select a specific sound recording contained therein. A podcast is a form of archived programming.

[19] While recipients are prevented in principle from storing a copy of a streamed recording for later use, technology exists that allows a person to copy and store streamed content. Stream ripping refers to processes that are used to achieve this.

[20] In turn, techniques, including cross-fading, exist to help prevent stream ripping. Cross-fading consists in shortening or eliminating the interval between songs or overlapping the end of the last song with the beginning of the next one.

[21] Two American concepts were mentioned repeatedly during the proceedings. The first is the concept of aggregate tuning hours (ATH). [16] One listener tuning in to a webcast for one hour amounts to one ATH; five listeners tuning in for two hours amount to ten ATH. ATH have been used to measure listenership when tracking the number of plays is either difficult (e.g., for some smaller webcasters) or impossible (tracking plays retroactively).

[22] The second is the Sound Recording Performance Complement (SRPC). [17] American webcasters who comply with the SRPC pay royalties at a significantly lower rate. To comply with the SRPC, a webcaster cannot, over a period of three hours, transmit more than three sound recordings from the same album if more than two such recordings are transmitted consecutively, or transmit more than four sound recordings from the same artist or compilation if more than three such recordings are transmitted consecutively.

III. THE PARTIES AND THEIR PROPOSALS

A. Re:Sound

[23] Re:Sound is the collective society that administers the remuneration rights for the communication to the public by telecommunication of published sound recordings of musical works.

[24] Originally, Re:Sound proposed that Tariff 8.A, for simulcasts and non-interactive webcasts, be 12 per cent of gross revenues, with an annual minimum of $500 per channel capped at $50,000 per site; not-for-profit services would pay $60 per month. Tariff 8.B, for semi-interactive webcasts, would be the greater of 45 per cent of revenues or 0.75¢ per play for cellular telephones and the greater of 30 per cent of revenues and 0.50¢ per play for other devices; not-for-profit sites would pay $60 per month. Tariff 8.B users would be subject to an annual minimum of $720. A service that pays according to Tariff 8.B would not pay royalties pursuant to Tariff 8.A.

[25] Re:Sound’s statement of case proposed changing both the rates and the tariff structure. Simulcast-only services would pay between 0.23¢ to 0.40¢ per play. Services offering any other combination of simulcasts, non-interactive webcasts and semi-interactive webcasts would pay the greater of 25 to 30 per cent of revenues and 0.23¢ to 0.40¢ per play. All services would be subject to an annual minimum of $500 per channel capped at $50,000 per site. Not-for-profit rates would remain the same. The royalty payable for simulcasts and non-interactive webcasts would be capped at the amount originally requested in proposed Tariff 8.A.

[26] In its reply statement of case, Re:Sound proposed three significant changes. First, the simulcast rate would be lowered to 0.089¢ per play for 2009 to 2011 and 0.096¢ for 2012. Second, services offering any other combination of simulcasts, non-interactive and semi-interactive webcasts would pay the greater of 21.75 per cent of revenues and 0.089¢ per play (0.096¢ for 2012) if they complied with the SRPC, and the greater of 25 per cent of revenues and 0.23¢ per play if they did not. Third, simulcasters and non-interactive webcasters (but not semi-interactive webcasters) would have the option to pay according to the number of ATH. Re:Sound proposed an ATH of 3.0 plays for low music use channels, and 12.09 for music-based channels.

[27] During closing arguments, Re:Sound proposed that the rates for simulcasts be 25 per cent less than for other webcasts. Re:Sound also proposed to introduce a microcaster rate of $250 per year.

B. CAB

[28] CAB represents commercial radio stations. Most stations simulcast their over-the-air signal. Some also provide non-interactive streaming of archived and original programming.

[29] Originally, CAB proposed to use as proxy SOCAN Tariff 22.B (Commercial Radio) for simulcasts and SOCAN Tariff 22.F (Audio Websites) for webcasts. During closing arguments, CAB proposed that simulcasts be subject to the same rate of 2.1 per cent as over-the-air radio set in paragraph 6(b) of Commercial Radio (2010). With respect to webcasts, CAB still proposed using Tariff 22.F as proxy, with a 27 per cent repertoire adjustment, yielding a rate of 1.5 per cent; it proposed to abandon the page impression adjustment Tariff 22.F provides [18] but to limit the rate base to “relevant revenues”. [19] The same minima of between $28 and $100 per year, per site as found in Tariff 22.F would apply.

C. Pandora

[30] Pandora offers two semi-interactive services (though not in Canada). [20] One is free and advertising-supported, the other ad-free and subscription-based. Roughly 98 per cent of Pandora’s listeners, and 87 per cent of its revenues, are currently associated with its free service. As of May 9, 2012, Pandora had surpassed 150 million users in the United States across multiple platforms.

[31] Pandora proposed to set the simulcast rate at 2.1 per cent (with no minimum fee) and the non-interactive and semi-interactive webcast rate at between 1.5 per cent and 2.65 per cent (with a minimum fee of $100 per site, per year), depending on the choice of repertoire adjustment.

D. CBC

[32] CBC is Canada’s public broadcaster. Its simulcasts its over-the-air radio signals, operates Internet radio services and offers a variety of other webcasts and podcasts.

[33] CBC proposed that its simulcasts be addressed in its over-the-air radio tariff, without any additional payment.

[34] For the post-broadcast, non-interactive webcasting of its radio programs, CBC proposed paying 10 per cent of what it pays pursuant to its over-the-air radio tariff, adjusted by the ratio of audio page impressions which include published sound recordings to all page impressions; using a ratio of 0.15 as default, this would yield royalties of approximately $20,000 per year.

[35] For its semi-interactive services, CBC offered to pay on the same basis as SOCAN Tariff 22.F: 5.3 per cent of the relevant Internet-related revenue, subject to the same repertoire adjustment as commercial radio. CBC proposed adding a further 20 per cent discount to account for CBC’s unique programming practices. The minimum would be $100 per channel.

E. NCRA

[36] NCRA represents community and campus over-the-air radio stations. These stations provide community access, opportunities for volunteers to participate in the programming and governance of the station, locally relevant news and information, and exposure to new and local artists. In these proceedings, NCRA’s interest is limited to simulcasts and non-interactive streaming of archived programs. [21]

[37] Paragraph 68.1(1)(b) of the Act sets at $100 the royalties community services pay to Re:Sound. These services probably include all stations represented by NCRA. [22] NCRA proposed that non-commercial stations pay nothing more for simulcasts, and proposed a $100 annual royalty for all non-interactive webcasting activities. In no instance should that fee exceed what a station already pays for its over-the-air broadcasts, [23] since the total number of listeners online always is a small fraction of the number of listeners to the station’s over-the-air broadcasts.

F. Services

[38] The Services offer to their cable television subscribers a variety of audiovisual and pay audio services. To complement this, several now provide certain audio services online, including the Galaxie music service operated by Stingray.

[39] The Services proposed that the rate for both tariffs be between 1.5 per cent and 2.5 per cent of revenues. They did not comment on minimum fees except to oppose CAB’s proposition that the revenue base never be less than 10 per cent of the total revenues of the website.

IV. EVIDENCE

A. Re:Sound

[40] Ian MacKay, President of Re:Sound, explained how the collective developed a webcasting rate card and used it in its negotiations. He outlined the terms of agreements concluded between Re:Sound and webcasters, and explained why Re:Sound considers them to be proper proxies for the proposed tariffs. He commented on the potential effect of the Objectors’ proposals on the Canadian recording industry, including the cannibalization of consumers’ music purchases that streaming may cause.

[41] Adrian Capobianco, President, Quizative Inc., described the various forms, growth and benefits of online and mobile advertising, the targeting that such advertising facilitates, and the prospects for continued growth in Canada over the next few years.

[42] Lucy Otterwell, Vice President Digital Strategy and Business Development, Warner Music Group, described the global nature of the webcasting industry. She discussed Warner’s licensing and negotiation practices and commented on Warner’s agreements with webcasters.

[43] Dr. Michael Murphy, professor of Engineering at Ryerson University’s School of Radio and Television Arts, described various types and business models of Internet webcast services. He discussed technologies used to limit their unauthorized use (geo-blocking, anti-stream-ripping software, cross-fading). He commented on a number of advances since 2006 he considered relevant, including the move away from portals to platforms, the development of mobile devices and apps and the integration of Internet radio technologies into factory-installed car entertainment systems. He suggested that page impressions no longer reliably measure the amount of music being listened to over the Internet; technology designed to record the number of plays now provides a better measure for this.

[44] Mark Eisenberg is an American attorney and the Co-Founder and Chief Operating Officer of LatticeWorks Media. He described the American statutory licensing framework and the evolution of American rates for webcasting and compared them to the proposed tariffs. He testified to the ubiquity of Internet radio. He offered the view that competition has become more substantial and differentiation less discernible. He concluded that music streaming now acts as a substitute for the purchase of CDs or downloads.

[45] Joshua McIvor, of Sony Music and formerly of BMG Music, was of the opinion that Mr. Eisenberg’s description of the American digital marketplace is applicable to Canada. He offered the view that if we certify rates that do not reflect an equilibrium price ratio between the semi-interactive and on-demand streaming markets, Canadians will be unable to have access to many of the world’s best digital music services.

[46] Dr. George Ford, President of Applied Economic Studies, Chief Economist of the Phoenix Center for Advanced Legal & Economic Policy Studies and Adjunct Professor at Samford University, provided the economic underpinning for Re:Sound’s proposal. He suggested that royalties should serve to compensate the parties for the risks and costs they bear. Since those risks and costs are much higher for sound recordings makers than for composers, compensation for each will differ in a market setting.

[47] At first, Dr. Ford proposed to use as proxies agreements involving record labels as copyright owners on the one hand, and webcasters, on-demand service providers and two large Canadian media companies as copyright users on the other. Since Re:Sound no longer relies on these agreements, there is no need to set out what Dr. Ford thought of them or the results he arrived at using them as proxy.

[48] Later on, Re:Sound filed five agreements it signed with webcasters in Canada. Having reviewed only the two available at the time he wrote his second report, Dr. Ford found them to be excellent proxies not commonly found in rate regulation proceedings and providing a more suitable proxy than tariffs established for other right holders. Relying on the agreements, he recommended that the royalties for SRPC-compliant non-interactive and semi-interactive webcast services be set at the greater of 25 per cent of revenues and 0.089¢ per-play (0.096¢ in 2012). Dr. Ford recommended the same per-play rates for simulcasts.

[49] Using as illustrations the revenue allocation practices of a few fully-integrated media companies, Alan Mak, Principal, Rosen & Associates, offered the view that financial reporting standards for allocating revenues are so flexible that a tariff based solely on a percentage of revenue could be gamed by over-allocating revenue to the non-musical parts of a website.

B. CAB and Pandora

[50] Since CAB and Pandora were represented by the same counsel and presented the same expert witnesses, their evidence and argument are examined together. As appropriate, arguments and evidence are attributed to either or both.

[51] CAB and Pandora jointly called two witnesses: Bruce Joseph, partner in the law firm of Wiley Rein LLP and Dr. David Reitman, Vice-President at Charles River Associates, an economics and business consulting firm.

[52] Mr. Joseph testified largely in rebuttal of Mr. Eisenberg’s evidence. He provided his own overview of the American regime. He sought to demonstrate that American agreements are not an appropriate proxy for the proposed tariffs. The differences between the regimes are too great. Furthermore, American rate setting is affected by a series of unique political and economic dynamics.

[53] Dr. Reitman criticized Dr. Ford’s analysis on a number of grounds. He dismissed the agreements as unreliable. He considered it inappropriate that the proposed tariffs may require users to pay amounts that exceed their total revenues. He proposed using SOCAN Tariffs 22.B and 22.F as proxies, in part because these tariffs simply are derivatives of SOCAN Tariff 1.A, adapted to the webcasting market. He proposed a repertoire adjustment of 50 per cent for simulcasts and of between 27.8 per cent to 50 per cent for webcasts. He did not propose to set different rates for semi-interactive and non-interactive webcasts: to the extent one had more value than the other, prices for these services would reflect these differences, and the same percentage rate would generate different amounts per subscriber, thus reflecting any added value.

[54] Testifying for CAB, four witnesses offered their views on the operational, technical, and economic realities associated with the use of music on radio station websites.

[55] Rob Farina, Executive Vice-President, Content, Astral Radio, explained the experimental nature and purpose of the webcasting licences Astral secured from record labels. He described an experimental webcasting service available on the websites of Astral radio stations and spoke to the costs and difficulties in developing content accessible on numerous mobile platforms.

[56] Earl Veale, Director of Interactive Technology, Corus Radio, discussed some technological and financial aspects of its webcasting business. He noted that simulcasting had little impact on listenership or advertising revenues. He suggested that technological changes relating to streaming have not had a material impact on the ability of Canadian conventional broadcasters to monetize interactive content. He views simulcasting as a form of brand extension and considers that it cannot compete with semi-interactive webcasting businesses because of the imbalanced regulatory environment.

[57] Tom Irwin, Director of Astral Radio Digital, explained how page impressions are tracked and why he considers such impressions provide the best measure for calculating revenues associated with a website’s music content.

[58] Dave Lehman, Manager of Online Media, Golden West, explained the dynamics of small webcasting markets in Canada. He views the incremental costs of simulcasting as significant, and considers that the purpose of simulcasting is to retain audiences, not grow them. He described the online success Golden West achieved by focussing on hyper-local, non-musical content. He shared Mr. Irwin’s views as to the usefulness of page impressions. He considered that technological evolutions have not changed the inherent nature or attractiveness of simulcasting.

[59] Two witnesses testified for Pandora: Joseph Kennedy, then President and CEO, Pandora Media, and Mark Ramsey, President, Mark Ramsey Media.

[60] Mr. Kennedy discussed the scope and nature of the services offered by Pandora, its business model and competitive environment. He explained the impact of American royalties and the potential impact of Re:Sound’s proposed rates on Pandora’s willingness and ability to bring its services to Canada. He considers his primary competitor to be the over-the-air commercial radio.

[61] Mr. Ramsey testified that semi-interactive webcasting services operate in a similar landscape and use music in a manner similar to over-the-air radio. He also argued that far from cannibalizing sales of sound recordings, such services indirectly promote the purchase of digital music tracks. In support of this view, he pointed out that as the number of American registered users of Pandora increased from 13 million in 2009 to 150 million in March 2012, total CD and digital music-track sales rose 4 per cent in 2011.

C. CBC

[62] Chantal Carbonneau, then Director, Legal Affairs, Intellectual Property, described agreements that CBC has entered into with performers’ unions and rights holders’ associations for the use of content over the Internet and explained how they could be used to set CBC’s webcasting royalties.

[63] Chris Boyce, Executive Director, Radio and Audio, English Services, described CBC’s mandate, the role that its radio operations play within the Canadian broadcasting system and its radio simulcasting and non-interactive webcasting services.

[64] Dominique Gagné, Head of Web Content, explained the simulcast of Première Chaîne and Espace Musique. She discussed the Webtrends data on simulcast, including the impossibility of determining the exact number of listeners at any given time.

[65] Denise Desgagnés, Director of Administration, French Radio, discussed the financial performance and the design of the websites for espace.mu and bandeapart.fm.

[66] Michael Mooney, Senior Director, Corporate Finance and Administration, filed a statement reviewing CBC’s financial situation and offering his views as to the impact of Re:Sound’s proposed rates on CBC’s ability to fulfil its statutory mandate.

D. NCRA

[67] Roxanne Casey, Station Manager, Canoe FM, explained the station’s mandate, operations, expenses and revenue sources, staff, and programming. She discussed the station’s website and online music services, including their goals, costs of development and operation, revenue, capacity, and reporting functions.

[68] Shelley Robinson, Executive Director, NCRA, summarized common characteristics and practices across the not-for-profit campus and community radio sector, particularly with respect to finances, operational practices and online music services.

E. Services

[69] The Services filed no evidence and proffered no witnesses.

V. ANALYSIS

[70] As extensive and informative as the evidence may have been, the analytical framework and the approach we use to derive the tariff make it unnecessary to review all the evidence in detail here. In particular, there will be no need to comment on issues such as the input cost pricing method; the difference in costs between conventional radio and webcasting; or the impact of recent technology advances on the value of the communication of sound recordings.

A. Issues Relevant to All Matters Under Examination

[71] In this part, we address issues that are relevant to all uses targeted in the tariff.

i. Simulcast

[72] In our view, the Internet simulcast of a radio station’s over-the-air signal is a textbook example of an ancillary use that should be dealt with at the same time as the main use. The value per listener for simulcasting and for over-the-air broadcasting is the same. That value is best achieved by attaching the ancillary use to the main one. Accordingly, the royalties should be proportional to the relative audience. [24]

[73] Setting a simulcast rate according to this principle requires certain information. One is the extent to which audience data already accounts for simulcast listenership. If it does, then this is probably already reflected in a station’s advertising rate card and a simulcast probably should not attract additional royalties. If it does not, we need to know the audience of the over-the-air broadcast and of the simulcast. Since none if this information is part of the record, and since we wish to firmly tie together royalties for simulcasting and for over-the-air broadcasting, we will defer setting all simulcast rates until the next time the Board considers the relevant tariffs, except for simulcasting by non-commercial webcasters and simulcasting of content not subject to another tariff.

[74] We agree with the Objectors that simulcasting remains essentially local, even if we were to accept that the music industry is global. Local stations deal with local issues, of interest only to locals and expatriates. A few radio stations may be very popular outside their region; this does not mean that all stations that simulcast their signal will suddenly reach a new audience.

[75] Re:Sound proposed to limit the definition of simulcast to the streaming of the over-the-air [25] signal of a radio station. The effect would be that the simultaneous streaming of the signals of a satellite service would be treated as a non-interactive webcast. We see no reason to treat these two transmissions differently. On the other hand, any Internet offering that is not identical to the original signal will be treated as a non-interactive webcast.

[76] Dealing with simulcasts as we do makes it unnecessary to address Re:Sound’s proposition to set a 25 per cent differential between simulcasts on the one hand, and non-interactive and semi-interactive webcasts on the other.

[77] Re:Sound rightly pointed out that it receives no royalties from commercial radio stations on account of advertising that is subject to the preferential rate set out in subparagraph 68.1(1)(a)(i) of the Act when a station’s over-the-air signal is simulcast. Re:Sound is also probably correct in stating that the preferential rate concerns only the over-the-air transmission of the station’s signal. Again, this matter can be addressed during the next examination of the commercial radio tariffs.

ii. Webcasts as Core Music Consumption

[78] According to Re:Sound, webcasts are a form of core music consumption that cannibalizes sales of CDs, downloads and on-demand streams. Music listening is not multi-tasked: rare are those who listen to more than one song at a time. Furthermore, portability and the ability to customize streams facilitate cannibalization: once Internet streaming is untethered from a desktop computer, consumers listen to streamed music on the go. This has two consequences, one direct and one indirect. First, money spent on a subscription service cannot be spent on downloads or CDs: there is a direct monetary substitution effect between streams and other forms of core music consumption. Second, even in the case of free streaming services, there is a form of cannibalization: consumers stream music instead of buying it. In other words, according to Re:Sound, advertising-supported streams supplant sales of music.

[79] Re:Sound argues that sales revenue lost through cannibalization must be made up through streaming royalties. Making up lost royalties cannot be achieved by setting Re:Sound royalties based on SOCAN tariffs; it requires using the ratios that exist in the markets being cannibalized. In those core music consumption markets, performers and makers are paid multiples of what composers are paid. This reflects their higher investment and greater risk. Using the one-to-one ratio fails to account for these investments and risks, which will result in a reduced supply of sound recordings and therefore higher costs of recordings for consumers.

[80] In the past, the Board has declined to use the ratio of royalties paid to composers and sound recording makers for the sale of CDs and downloads as a proxy to allocate the value of the communication right among these rights holders, in part because these were different markets involving two different rights. In Re:Sound’s opinion, that rationale cannot hold in this instance, since the two markets are converging, as demonstrated by the fact that streams cannibalize sales.

[81] The Objectors asked that we reaffirm the one-to-one relationship between SOCAN and Re:Sound tariffs. In their opinion, Re:Sound has not offered any justification to depart from this well-established practice. They reject the claim that non-interactive and semi-interactive webcasting replace sales of sound recordings; on-demand streaming, which is not at issue in these proceedings, is the only form of webcasting that might do so. Even if the claim were true, the Objectors disagree that, as a result, royalties paid to Re:Sound should reflect lost sales; lost sales caused by webcasting should be mitigated through the licensing of the labels’ exclusive reproduction right.

[82] We agree that portability has value. The ability to stream music over mobile devices increases the attractiveness of music streaming. This is also true of most forms of music listening. Radio sets and CD players were not always portable. Music royalty rates did not increase when portable devices emerged or when they became commonplace. Consequently, portability cannot of itself justify higher royalties.

[83] We are unconvinced that non-interactive and semi-interactive streaming cannibalizes sales of CDs or downloads. Though the Objectors’ evidence and arguments in this respect are not without contradictions, [26] we agree with them, for the reasons set out in paragraph 157 below, that non-interactive webcasting is similar to over-the-air radio. We find that neither over-the-air radio nor non-interactive webcasting is likely to cannibalize music sales; if anything, they are likely to stimulate them.

[84] The same seems true of semi-interactive webcasts. Pandora’s American free and paying subscribers are about twice as likely to purchase CDs or downloads as are non-subscribers. Furthermore, while purchases by Pandora’s subscribers are declining, the decline is not as steep as for non-subscribers. [27] Furthermore, the fact that music listening is not multi-tasked does not mean that listeners will select one source of music at the expense of another. At times, they will tune in to a semi-interactive service to listen to new music that matches their musical tastes without having to pre-select each track. At other times, they will prefer to hear a particular CD or download. These products target not so much different types of consumers as they do consumers according to their different moods.

[85] As requested by the Objectors, and for reasons that need not be repeated here, we find that the ratio between SOCAN and Re:Sound rates should be maintained at one to one.

[86] We find that there is no evidence allowing us to conclude that the forms of webcasting addressed in this decision cannibalize sales of CDs and downloads. Consequently, there is no need to account for the relative risks and costs of record makers and composers in setting the rates we certify today. Re:Sound’s concerns in this respect cannot be rejected out of hand. They do however require evidence, not simple assertions, before we can act on them. In addition, even assuming that cannibalization through legitimate streaming did exist, it probably would be difficult to segregate its impact from that of non-legitimate forms of Internet transmission of music, including illegal downloads. Until then, these claims are, in the words of the Services, interesting but undemonstrated.

iii. Music as a Single Ecosystem

[87] Re:Sound argues that there is now a single music ecosystem. Semi-interactive and on-demand services compete. Unless both types of services pay comparable rates, one type will gain an unfair competitive advantage over the other. Canadian rates for on-demand streaming are already set. If the rates for semi-interactive and non-interactive webcasts are set too low, on-demand services will be at a competitive disadvantage and may not launch in Canada. In turn, if expected returns are too low, there will be no incentive for the Canadian recording industry to create new sound recordings. We disagree, for three reasons.

[88] First, just as we are unconvinced that webcasts will cannibalize sales of CDs, we are unconvinced that non-interactive or semi-interactive webcasts may become a substitute for fully interactive services, regardless of the tariff we set. At times, consumers may opt for the lean-back experience a semi-interactive service affords. At other times, they will prefer being able to hear a particular sound recording at a particular time.

[89] Second, if a single ecosystem does exist, this would tend to drive the cost of music to a single price point for all participants within the system, unless differences are justified through cogent, market-based evidence. In turn, such evidence would require the Board to make an assessment of the ecosystem as a whole, something which is not possible in this instance.

[90] Third, we agree with the Objectors that the possible effects that any rate imbalance between interactive and other forms of webcasting may have on the market is a matter better pursued by the labels, in the free market, using the exclusive reproduction right.

[91] Again, Re:Sound’s concerns, while not unreasonable, must be supported by evidence before we can act on them.

iv. Webcasting as a Global Industry

[92] Re:Sound argues that webcasting is a global industry. Because of this, Canadian rates must be consistent with Canada’s major trading partners, especially the United States.

[93] Foreign rates for webcasting sound recordings vary by a factor of up to 10 to 1. The American rate is the highest, at USD 0.0021 per play in 2012; in New Zealand, the rate is NZD 0.0002 or USD 0.00016. [28] Such variations among independently-functioning national markets tend to show that this is not a global industry, [29] at least not yet, and that fears about the possible consequences of variations between Canadian and foreign rates are unfounded.

v. User-Based vs Use-Based Tariff

[94] Re:Sound argued that a user-based tariff is no longer appropriate given the rapidly developing and changing ways in which simulcasters and webcasters use sound recordings. In its statement of case, CAB argued that the Board should continue to set user-based tariffs, at least for radio stations. Their Internet activities are ancillary to over-the-air broadcasting. Their website is an extension of their conventional operation and is operated primarily to extend their brand and to retain listeners, not to derive independent revenues. Streaming has not resulted in notable revenue growth. CAB did not pursue the matter further. Other objectors did not raise it.

[95] We have already dealt with simulcasts. As for the rest, we find that the time has come to move from user-based to use-based tariffs for most forms of webcasting under examination. The developments we anticipated that would lead to such a shift have largely occurred. [30] Internet transmissions now are an integral part of the overall business strategy of many music users; for some, it is their only source of revenue. Better, cheaper monitoring tools have been developed. Business models have diversified; to allow certain users (radio stations for example) to shelter behind their main business model could lead to the creation of unintended competitive advantages over new entrants.

vi. A “Greater Of” Formula

[96] Re:Sound proposes that the tariff be the greater of a percentage-of-revenue rate and a per-play rate. The Objectors oppose this for several reasons. Previously, the Board refused to use the greatest of costs or revenues as a rate base for a tariff. In the United States, per-play has become the real payment mechanism in a similar context, since they most often exceed what would be payable as a percentage of revenue.

[97] In this instance, a “greater-of” tariff would not be fair and equitable, because it would provide an undue advantage to Re:Sound on two counts.

[98] To be fair and equitable, a tariff should neither overcompensate nor under compensate rights owners. If set correctly, neither a per-play rate nor a percentage-of-revenue rate will tend to do so, to the extent that each captures a (different) measure of usage. On the other hand, a tariff set at the greater of those two rates is hedged in favour of the collective. It may prevent undercompensation if a service has low revenues; it does not prevent overcompensation in the case of a high-revenue service that uses few sound recordings.

[99] A greater-of formulation also burdens users with an unfair share of risks. Re:Sound benefits if there are high revenues and a large number of plays, if there are high revenues and a small number of plays, and if there are low revenues and a large number of plays. Only if there are low revenues and a small number of plays does the user benefit. By contrast, either a per-play or a percentage-of-revenue tariff, with or without a minimum fee, allocates risk between Re:Sound and the users more evenly.

[100] We disagree with the Objectors when they state that a “greater of” formula necessarily gives rights holders an excessive share of related revenue. This is largely a function of how high the rates are, not of the tariff formula.

[101] It is important not to confuse a greater-of formulation with a minimum fee or a recoupable advance. Under these scenarios, the user pays the greater of a fixed amount and the amount a formula yields. Here, Re:Sound is proposing that the user pay the largest amount of royalties yielded by two different formulas. The purpose of a minimum fee is to ensure that users cannot use protected sound recordings for free. The purpose of Re:Sound’s proposed formula is to maximize royalties as a function of a user’s business model and success.

vii. SRPC and Related Technical Issues

[102] Re:Sound asked that certain limitations on programming be imposed and other steps be taken to protect the unauthorized reproduction of sound recordings included in webcasts. It also proposed that services who do not comply with the SRPC pay at a higher rate. The Objectors took issue with this. They challenged the relevance or feasibility of other technical measures such as cross-fading, maximum gaps between plays and other forms of deterrence of stream ripping. They also argued that the Board cannot, pursuant to the equitable remuneration regime, impose restrictions or obligations other than to pay the relevant remuneration. Re:Sound is not a licensor, and has no power or authority to impose or to request the imposition of these SRPC limits or any other limits.

[103] We disagree with the Objectors about the extent of the Board’s powers. At a minimum, we could encourage webcasters to apply technical measures to prevent stream ripping by offering them a discount. However, we will not do so in this instance, for the following reasons. First, sufficient doubts were raised about the feasibility of some measures to require further evidence before they can be imposed or encouraged. Second, there is no evidence that stream ripping is in fact a problem in Canada. Third, some reproductions of streamed content almost certainly involve “users’ rights” (e.g., reproducing a stream as part of fair dealing for an allowable purpose), the exercise of which should be allowed for free. Fourth, the SRPC and other technical measures seek to protect rights which are not targeted by this tariff and which are not administered by Re:Sound; as we said in paragraph 90 above, this a matter better pursued by the labels, in the free market, using the exclusive reproduction right unless an intervention on our part becomes absolutely necessary to restore balance in the market.

viii. Archived programming

[104] The playing of archived programs is a type of non-interactive webcast and should, in principle, be subject to the per-play rate we set below. However, by a letter to the Board on May 20, 2008, Re:Sound advised that it no longer intended to seek a tariff for podcasts, which were defined in the original proposed tariff to target the activities we define here as archived programs. As such, we are not certifying a tariff for the communication of sound recordings in that fashion.

B. Rates for Webcasters Other than CBC and Community Systems

i. Percentage vs Per Play

[105] If asked to choose between a percentage-of-revenue rate and a per-play rate, Re:Sound prefers a per-play tariff, to ensure that rates are tied to music use over the ability to generate revenues. According to Re:Sound, a revenue-based tariff is a poor solution for many reasons. It forces rights holders to subsidize a user’s decision to give music away for free without properly monetizing it. Also, no accurate means exist to apportion a website’s revenue between visitors who access sound recordings and those who do not; this provides an opportunity to allocate revenue so as to minimize royalties. A per-play rate avoids both problems.

[106] The Objectors favour a percentage-of-revenue rate. This approach is consistent with analogous tariffs. It also reflects a webcaster’s ability to generate revenue with the sound recordings and allows it to select different business models, prices and uses: the increased value inherent in an increased consumption of bandwidth is reflected in higher revenue which in turn will lead to more being paid to Re:Sound. A per-play rate does not self-adjust. Also, no one earns revenue on a per-play basis. Finally, there is no evidence that webcasters or any other users are allocating revenues so as to game the tariff.

[107] Each approach charges the user for a measure of usage (revenues or plays). Each has its advantages.

[108] A per-play rate correlates with usage more directly than a percentage rate: it monetizes music given away for free. On the other hand, a percentage rate scales with the retail value of the service; a per-play rate does not.

[109] A percentage rate requires selecting an appropriate rate base. This may prove difficult in this instance, especially for free, advertising-supported services offering a wide variety of music and other content. Setting percentage rates for new markets without creating imbalance with existing ones can be challenging. A per-play rate avoids these difficulties; the price per unit is the same irrespective of business model, sources of income or success. On the other hand, a percentage rate self-adjusts to price increases, while a per-play does not; this is truly only important when inflation is high.

[110] Some percentage rates, including this one, can be gamed. A per-play rate cannot. This problem increases with the size and diversification of the user. It is a greater concern with a nascent industry than with a mature one.

[111] With a percentage-of-revenue rate, the implicit per-play rate necessarily varies according to the relative success of the service. Of two services that transmit an equal number of plays, the one with the most revenue will pay more per play. This difference may not be fair if the two services compete, or if they should be expected to pay the same price per unit of music input.

[112] A key disadvantage of a percentage rate relates to the fact that Internet music is often not offered as a separate product. If rates are to be tied to music use, such use must somehow be measured. We agree with Re:Sound that, with the development of web pages with dynamic elements and rich Internet applications, page impressions no longer measure reliably the amount of music being listened to over the Internet. A per-play rate only counts music use; there is no need to differentiate what contains music and what does not, or to determine the share of value that music delivers when a consumer accesses music and other content at the same time. All that remains then is to account for music that is not embedded in an eligible sound recording. In this instance, this is addressed later, through a repertoire adjustment.

[113] Per-play rates are used in the United States and elsewhere [31] for webcasting; this implies that the technology for measuring plays exists, is affordable and is available. While it may be necessary to have transitional provisions to address the difficulty of counting plays retroactively, this does not present a problem on a prospective basis.

[114] The key advantage of per-play rates is that they are strictly correlated with usage. They are a type of transactional price. Transactional prices are used in other Internet tariffs the Board has certified, including in subsections 5(3) and (4) of the CSI Online Music Services Tariff, 2008-2010. There are thus precedents for the use of transactional pricing in a related market.

[115] In this instance, we opt for a per-play tariff because it is better correlated with usage, because it monetizes music given for free, because usage is more readily and reliably measurable, because it is technologically possible and because it is a transactional price.

[116] Since we do not set a percentage rate, there is no need to discuss the rate base to which that percentage should apply.

ii. Selecting an Appropriate Proxy

[117] We are certifying tariffs of first impression, which generally requires either an economic model usable to set the tariff or an appropriate proxy. No one offered an economic model; accordingly, we turn to the selection of a proxy.

[118] Re:Sound offered as possible proxies agreements between major labels and webcasters operating in Canada and agreements between itself and these webcasters. CAB and Pandora proposed using SOCAN Tariff 22.F as proxy. None of these proxies are perfect. By construction, a proxy comes from another market or another tariff, which may be more or less relevant to the instant tariff. In the end, for the reasons set out below, we will use SOCAN Tariff 1.A.

C. American and Foreign Rates

[119] No one is proposing outright that we use American rates as proxies. Nevertheless, since we use them at paragraphs 149 to 154 below to assess the reasonableness of Re:Sound’s proposal, we wish to note the following differences, which tend to explain at least in part the significant divergences in Canadian and American outcomes.

[120] The American right to broadcast a sound recording over the Internet is an exclusive one, subject to some statutory licences. The Canadian right to communicate a sound recording is a remuneration right. Though this may not impact the value of each right directly, it affects the ways in which royalties are determined. In the United States, most users licence the right to broadcast sound recordings over the Internet without going through the Copyright Royalty Tribunal. In Canada, royalties for communicating sound recordings can only be collected pursuant to a tariff set by the Board.

[121] American rates are subject to a detailed set of statutory instructions. They are intended to reflect competitive market outcomes and to account for the promotional value of air play. In Canada, the only overreaching requirement is that the rates be fair. A market price is but one possible rationale for setting a tariff; the Board is not bound to set tariffs only on that basis. [32] Also, the Board has repeatedly refused to take into account the promotional value of air play. [33]

[122] The American regime was expressly “directed at preventing the diminution in record sales through outright piracy of music or new digital media that offered listeners the ability to select music in such a way that they would forego purchasing records.” [34] The Canadian regime was not designed with this in mind.

[123] The American regime mandates that webcasters deploy a variety of mechanisms, including the SRPC, to deter such copying. The Canadian regime mandates no such measures.

[124] We agree with Mr. Joseph that American rate setting has been affected by a series of unique political and economic dynamics, steeped in controversy that has led to interventions on the part of Congress. These dynamics and controversy have not arisen in Canada, at least not to the same extent: the Canadian Parliament has not sought to dictate the rates the Board has set for online music uses.

[125] A significant discrepancy exists between the rates paid in the United States for broadcasting a musical work and a sound recording. The Board has concluded in the past, and we are concluding again, that such a difference should not exist in Canada.

[126] In addition to American rates, the parties supplied us with a variety of other foreign rates for webcasting sound recordings. As we noted in paragraph 93, these rates vary greatly from country to country, with the American rates being the highest. There is no reason to believe that American rates are any better a proxy for our purposes than other foreign rates. Furthermore, since we know nothing of the market environment within which these other rates exist, they cannot serve as proxies.

D. Agreements between Record Labels and Webcasters

[127] Initially, Re:Sound proposed using Canadian licensing agreements between webcasters and major record labels to set the tariff. The Objectors challenged this suggestion on several counts, including the following.

[128] First, the agreements licence other rights in addition to the right to communicate a sound recording: everything else being equal, it should cost more to licence two rights than to licence one. As a result, the agreements are prima facie suspect as proxies for setting only the communication right unless one discounts the value of the additional rights or demonstrates that they are of little or no value.

[129] Second, the agreements only licence the recordings in each label’s catalogue; whether the rates were adjusted accordingly is not always clear. Using the agreements as proxies may require adjusting the prices they set both to reflect that Re:Sound’s repertoire includes the catalogue of all major labels and to account for the fact that not all recordings in the labels’ catalogues trigger equitable remuneration in Canada.

[130] Third, the parties clearly considered some of the agreements to be experimental. A quality proxy should be drawn from a mature business.

[131] Fourth, the agreements are closely aligned with their American counterparts. [35] Differences between the Canadian and American regimes are important enough to make U.S.-based agreements unsuitable as proxy to set Canadian rates.

[132] We would have found, for the above reasons, that the agreements between labels and Canadian webcasters were not a suitable basis to establish rates in these proceedings. However, since Re:Sound no longer relies on these agreements, there is no need to further elaborate on this point.

E. Agreements Between Webcasters and Re:Sound

[133] In reply, Re:Sound proposed using as proxy a new set of agreements it reached with certain webcasters at rates much lower than those found in the labels’ agreements. These webcasters, wishing to offer their services in Canada before the Board certified a tariff, apparently found some measure of comfort in these agreements. All contain essentially the same percentage of revenue and per-play rates, in a “greater of” structure. All set a minimum royalty, though the amount and structure can vary significantly. Limitations on licensed uses and technical restrictions also vary; so do expiration dates, though if all options are exercised all would end at the same time. Re:Sound argued that the rates we set must reflect market realities of the webcasting industry, and that these agreements reflect the exact marketplace and value of rights targeted by the tariffs.

[134] The agreements state that they were reached through arms-length negotiations and that they are based on independent legal advice. Yet they may not reflect a competitive value. Such will be the case, for example, when the agreements are all signed by a single entity (here, Re:Sound) once that entity has decided the price-point (or “sweet spot”) at which the entity will be willing to deal with the intention of establishing a high (or low) proxy for later use as precedent.

[135] CAB and Pandora argued that the negotiations leading to the agreements were of the “take-it-or-leave-it” type, and that the prices they set correspond to Re:Sound’s “sweet spot”. We disagree.

[136] The fact that all agreements set the same rates is unconvincing. In a highly competitive market, all users will tend to purchase identical goods at the same price. Furthermore, collectives are generally expected to deal with similar users on a similar basis: in some countries, they are required by law to do so. At most, rate uniformity may suggest that on the pricing element, Re:Sound was reluctant to deviate from its set rates. The fact that the agreements were all concluded within a short period of time is also a rational explanation for such uniformity.

[137] Other terms (duration, limitations, technical restrictions) vary. Minimum fees are not all the same: the effective rates will not be uniform if, in the end, only the minimum is payable. There is no evidence that anyone refused the terms found in the agreements or that Re:Sound only sought certain deals or initiated negotiations only with parties of its choosing.

[138] One webcaster had earlier discussions with Re:Sound but failed to reach an agreement. The terms found in the Re:Sound agreements were not offered to that webcaster, as they were developed at a later date. However, Mr. Mackay testified to Re:Sound’s willingness to enter into agreements on market-based rates with any webcaster.

[139] The first arbitration panel to set the American webcasting royalties found that agreements filed by the Recording Industry Association of America (RIAA) were in substantial conformity with the Association’s “sweet spot”. [36] RIAA deployed extraordinary efforts to pursue a predetermined strategy intended to develop rate precedents. Most licensees were smaller entities which were unlikely to endure; they paid no royalties, paid nothing beyond the prescribed minimum or quickly went out of business. Other licensees, including one that was already engaged in litigation with the RIAA, were motivated to pursue a deal at almost any cost. None of these factors is present here.

[140] We therefore find that the “sweet spot” analogy with the situation that unfolded in the United States is lacking and unconvincing.

[141] A second issue raised by the Objectors is that the Re:Sound agreements grant nothing. In Canada, the communication of a sound recording only triggers a right to equitable remuneration, which cannot be the subject of a licence. The agreements may be unenforceable; at best they are a form of insurance that some webcasters bought because of the apprehension caused by the high rates Re:Sound proposed initially. As such, the Objectors argue, they are unreliable and do not reflect the proper value the Board is asked to assess.

[142] The agreements should not be discarded for that reason alone. They involve sophisticated parties. They were driven by legitimate commercial reasons: a webcaster may end up paying less than the tariff for at least part of the term of the agreement. Nothing indicates that the potential unenforceability had an effect on the way the parties negotiated. Furthermore, the agreements may well constitute an enforceable promise not to collect more than the agreement if the rate we certify is higher.

[143] On the other hand, none of the webcasters who signed a Re:Sound agreement testified as to their understanding of what it provided them. A webcaster may have been aware of the potential legal validity issues but still gone ahead and agreed to the terms as a form of insurance against future royalty claims. In such a case, the consideration paid may not accurately reflect the value of the remuneration right: from the webcasters’ point of view, value could reside in more than the use of Re:Sound’s repertoire. For example, a service may be willing to accept a rate, even if the certified rate ends up being lower, because this provides certainty; in turn, this certainty may enable the service to secure investments that will allow it to launch earlier than other services. Also, it is certainly conceivable that the higher rates Re:Sound initially proposed served as an incentive to accept the rates found in the agreements: two were agreed before Re:Sound lowered its rate proposals, and two shortly thereafter.

[144] A third issue concerns the newness of the Canadian webcasting market and of the agreements. The Canadian market for webcasting is nascent and underdeveloped. Relying on the agreements as proxies carries a greater risk than within a mature market, warranting a more cautious approach in assessing their suitability as proxies. Furthermore, since the agreements were very recent at the time of the hearings, no data were available regarding their workings, such as whether the per-play rate or the percentage-of-revenue rate was triggered or whether any options were exercised by the parties.

[145] Taken as a whole, and even though the above analysis of the Re:Sound agreements fails to address some of our concerns, we are hesitant to discard them for setting the tariffs under examination. We must therefore decide on another basis whether or not to use them as proxies.

[146] The Re:Sound agreements can serve to set the tariffs only if we conclude that their rates are reasonable. While there are no direct comparables available for those rates, reasonableness can be assessed indirectly. This can be done as follows.

[147] Once the size of the relevant repertoire has been accounted for, the rates set in the Re:Sound agreements are closely aligned with what American webcasters pay to broadcast sound recordings. The per-play rate the agreements set for 2012 (0.096¢), which Re:Sound proposes we certify, is 48 per cent of the approved American rate (0.20¢) for that same year. In Canada, only Rome sound recordings [37] are entitled to equitable remuneration. [38] Furthermore, Re:Sound is entitled to collect royalties only in respect of such recordings that are in its repertoire. [39] As a result, the Board has found that recordings in the repertoire of Re:Sound are involved in between a quarter and a half of the communications effected by various broadcasters. [40] By contrast, virtually all sound recordings attract webcasting royalties in the United States. [41] Therefore, for Re:Sound to collect the same amount per eligible recording by applying a per-play rate on all recordings, whether or not eligible, the Re:Sound rate should be, all other things being equal, half the American rate or less. Conversely, if the eligible American repertoire were the same as in Canada, the total amount or royalties collected in the United States would be 50 per cent or less of what it currently is. This shows that Re:Sound’s proposed rates are the same as, or higher than, the American rates, per play of eligible recording. Consequently, the Re:Sound proposed rates are reasonable in Canada only if the existing American rates are reasonable for Canada.

[148] The reasonableness of the American rates in the Canadian market cannot be assessed directly. After all, the rates to be paid to webcast sound recordings in Canada are precisely those whose reasonableness is at issue.

[149] We must therefore turn to another market. One possibility is to compare how much commercial radio stations pay SOCAN with how much they would pay if the SOCAN rate was the same as the American per-play rate for webcasting sound recordings. That comparison is both appropriate and possible.

[150] The comparison is appropriate for the following reasons.

[151] First, the Board has decided in the past, and we take as a given in this instance, that the rate the Board set for Canadian commercial radio in SOCAN Tariff 1.A is fair. Second, the Board has decided in the past, as we have done in this instance, that the use of music in a radio signal should attract the same amount of royalties, whether over-the-air or in a simulcast. Third, the Board has decided in the past, and we have already decided in this instance, that the right to communicate a sound recording of a musical work is worth the same as the right to communicate that musical work. Fourth, the American rate for webcasting an Internet-only signal and for simulcasting an over-the-air radio signal over the Internet [42] is the same. Fifth, Canadian commercial radio essentially plays only musical works in the SOCAN repertoire and sound recordings that are protected under American copyright legislation.

[152] Applying the above findings, the American webcasting rates are reasonable for Canada if, when applied to commercial radio, they yield approximately the same amount as Tariff 1.A. Put another way, either as a matter of fact or as a result of our findings, (a) a sound recording and a musical work attract the same amount of royalties, (b) a radio station’s simulcast and over-the-air broadcast attract the same amount of royalties, (c) the American rates for webcasting and simulcasting are the same, (d) the American repertoire of protected sound recordings and the Canadian repertoire of protected musical works occupy the same share of a Canadian commercial radio station’s air play, and (e) SOCAN Tariff 1.A is fair. Given these findings, what an American webcaster pays per play to broadcast a sound recording is reasonable for Canada if it is the same as the price per “listen” (i.e., the number of time a song is listened to by an individual) a Canadian commercial radio station pays to broadcast a musical work.

[153] The comparison is also possible because the parties provided us with all the required data. [43] In 2011, Canadian commercial radio paid approximately $56 million to SOCAN. We can estimate the total number of commercial radio “listens” at over 258 billion. The American sound recording webcasting rate for 2011 was 0.19¢ per play; if commercial radio paid that price for their use of musical works, the royalties would be in the order of $490 million which is slightly less than 9 times more than what commercial radio actually paid to SOCAN during that year.

[154] Applying the American rate would result in SOCAN royalties that are almost nine times greater than what they really are. Consequently, the American rate is prima facie not suited to the Canadian market. Since the Re:Sound proposed rates are closely aligned to the American rates, the Re:Sound proposed rates also are inherently suspect and improper for use as proxies in this case.

F. Existing Tariffs as Proxies

[155] Re:Sound argues that existing tariffs do not offer any appropriate proxy. SOCAN Tariff 22 does not reflect current business models, which have matured and evolved considerably. Its rates do not reflect the current market value of communicating a sound recording; they do not accord with rates paid elsewhere by many of the same services to be licensed under the proposed tariffs. The CSI Online Music Services Tariff also is not an adequate proxy, largely for the same reasons. SOCAN Tariff 1.A is not an appropriate proxy because webcasting is not the same as over-the-air radio; their cost structures are simply too different. Re:Sound’s commercial radio tariff is even less reliable, because of the distortion created by the fact that subparagraph 68(1)(a)(i) of the Act sets at $100 the amount of royalty a commercial radio station pays on the first $1.25 million of advertising revenues.

[156] Conceptually, we agree with the Objectors that SOCAN Tariff 22.F is an appropriate proxy. It is simple to understand. It allows us to maintain the usual SOCAN to Re:Sound tariff ratio. We agree with Re:Sound that technology has evolved since Tariff 22.F was certified, but we find that no hard evidence was provided as to what effect such changes (or other differentiating factors invoked by Re:Sound) have had on the value of the communication of sound recordings.

[157] Most importantly, the Tariff 22.F proxy is fairly robust because it is itself based on SOCAN’s long-standing commercial radio tariff. We agree with Mr. Ramsey that non-interactive and semi-interactive webcasters offer services, and operate in a landscape, similar to over-the-air radio. They are supported by advertising, play a wide selection of music, are consumed when listeners seek a relatively passive music experience, are largely used in mobile environments and provide value based on the music mix. Mr. Kennedy put it in another way that is just as convincing, when he stated that over-the-air radio and semi-interactive services share many attributes: each communicates sound recordings to the public; each presents a curated music listening experience; neither permits consumers to choose the recordings they listen to or the time that recordings are transmitted; and each generates revenue predominantly from advertising. We also note Mr. Eisenberg’s reliance on Pandora’s declaration, in its 10-K filing of March 19, 2012, that it competes with over-the-air radio in concluding that lean-back and customized radio Internet offerings compete with one another. [44]

[158] We agree that the costs structures of conventional radio and Internet webcasting are significantly different. We also agree that Canadian commercial radio operates under regulatory constraints over content that do not apply to Canadian webcasters and do not exist in the United States. We remain unconvinced that the first factor is relevant in setting a rate in this instance, especially as the rate we certify is per play. We find no evidence that the second factor might unduly lower the price commercial radio pays to SOCAN for its music, to Re:Sound’s prejudice. On the contrary, such constraints would increase demand for certain types of music, which in turn may tend to drive up the price of music, including Re:Sound’s royalties.

[159] We address other issues Re:Sound raised with respect to SOCAN Tariffs 1.A and 22.F (user-based tariff, substitution effects, relative interactivity, tariff gaming, page impressions as measure of music use, imputed page impression ratio, foreign rates, royalties for free streams) elsewhere in these reasons.

i. Selecting the Appropriate Tariff as Proxy

[160] The application of the principles on which the Board usually relies to select a proxy would normally lead us to use SOCAN Tariff 22.F to set the tariffs under examination, except for simulcasts. In this instance, however, we find it preferable, and more practical, to use SOCAN Tariff 1.A, for the following reasons.

[161] First, we have decided to set a per-play tariff. In order to transform Tariff 22.F, a percentage-based tariff, into a per-play tariff, we would need to know or estimate the number of tracks played pursuant to this tariff; we do not have this information, or the data required to perform the estimation. By contrast, Re:Sound (at our request) and CAB (of its own motion) provided both the data required to derive a per-play tariff from Tariff 1.A and a calculation thereof. [45] In both instances, the approach was identical to the one we use below.

[162] Second, we have already noted in paragraphs 130 and 144 above the problems that may arise when using a proxy from a nascent and underdeveloped market. We also found in paragraph 157 above that if Tariff 22.F is a conceptually valid proxy, it is principally because it is based on Tariff 1.A. Finally, there were fewer than 60 licensees under Tariff 22.F, as compared with more than 600 under Tariff 1.A. Under the circumstances, using Tariff 1.A makes the most sense.

[163] Third, we have found that semi-interactive and non-interactive webcasts compete with conventional radio, irrespective of cost structure. It follows that a per-play rate for radio is a solid starting point for setting the rate for webcasts.

[164] Fourth, deriving a per-play rate for webcasts from SOCAN Tariff 1.A reaffirms, as the Objectors asked, the one-to-one relationship between SOCAN and Re:Sound tariffs.

[165] Tariff 1.A involves four rates: three explicit and one implicit.

[166] The lowest rate, 3.2 per cent, is intended to benefit only commercial radio stations whose yearly revenues do not exceed $1.25 million. This rate was certified based on evidence of the inability of smaller stations to pay an increase in the rate from 3.2 to 4.2 per cent. [46] In this instance, we have neither evidence that smaller commercial webcasters would face similar difficulties nor information that would allow us to decide how to tier such a rate.

[167] The highest rate, 4.4 per cent, is what larger stations pay on their yearly revenue over $1.25 million. This rate was certified to claw back from larger stations the benefit of applying the 3.2 per cent rate to all stations instead of only smaller ones. [47] Such claw back is unnecessary if a single rate applies to all commercial webcasters.

[168] A third rate is the percentage (approximately 3.9 per cent) SOCAN actually receives from the revenues of all the Canadian commercial radio industry. This rate is a derived rate. A derived rate is as unstable as the data used to derive it. In this instance, the total amount of royalties commercial radio pays to SOCAN is revised continuously, making the result of any calculation using that rate unstable.

[169] The fourth rate is what the Board has described before as the effective rate of 4.2 per cent. [48] This is the rate the Board arrived at before taking into account the specific situation of smaller commercial radio stations, based on evidence that was available then and absent now. [49] This is the rate we use here.

[170] The calculation of the per-play rate using SOCAN Tariff 1.A as starting point is as follows.

Deriving a webcasting per-play rate from SOCAN Tariff 1.A (2011 data)

Deriving a webcasting per-play rate from SOCAN Tariff 1.A (2011 data) For data sources see Exhibit Re:Sound-78 at p. 15 Calcul du taux par écoute visant les webdiffusions selon le tarif 1.A de la SOCAN (données de 2011) Pour les sources de données, voir la pièce Ré:Sonne-78, p. 15 (en anglais)
Weekly radio hours tuned Nombre d’heures d’écoute de la radio par semaine	17.7
Yearly radio hours tuned (1 x 52) Nombre d’heures d’écoute de la radio par année (1 x 52)	920.4
Number of Canadians aged 12+ Nombre de Canadiens de 12 ans et plus	29,996,689
Canadian commercial radio share of listening Part des heures d’écoute des radios commerciales canadiennes	0.774
Number of 4-minute sound recordings playable during an hour Nombre possible d’enregistrements sonores de 4 minutes dans une heure	15
Percentage of programming time using sound recordings Pourcentage du temps de programmation consacré aux enregistrements sonores	0.806
Number of sound recordings “plays” on commercial radio (2 × 3 × 4 × 5 × 6) Nombre d’« écoutes » d’enregistrements sonores à la radio commerciale (2 × 3 ×4×5×6)	258,355,190,971
Commercial radio gross revenues ($) Revenus bruts des radios commerciales ($)	1,613,000,000
Royalties @ 4.2 per cent ($) Redevances : 4,2 % ($)	67,746,000
SOCAN royalties, per “plays”, in 2011 (9 / 7) ($) Redevances à verser à la SOCAN par « écoute » en 2011 (9 / 7) ($)	0.00026222

[171] CAB suggested that if we derive the per-play rate from SOCAN Tariff 1.A, we should adjust the amount to account for the fact that a minute of music generates less revenue than a minute of spoken word. [50] We disagree, for two reasons. First, CAB’s preferred proxy is SOCAN Tariff 22.F. That tariff was derived from SOCAN Tariff 1.A without performing CAB’s proposed adjustment. Therefore, it would appear that CAB itself concedes that no adjustment is warranted. Second, all parties agree that webcasters have a superior ability to target an audience so as to increase that audience’s value to advertisers. We find that this ability will tend to shrink or eliminate any difference in the relative ability of music and spoken word to attract advertising dollars.

[172] Re:Sound proposed increasing the rate by 8 per cent (from 0.089¢ to 0.096¢) in 2012. We will not, since no reason was offered for doing this.

[173] The per-play rate we set, before repertoire adjustment, is one-fifth of the $0.0013 rate set in 2012 for SOCAN in respect of free on-demand streams. [51] Even if we were to find that on-demand streams should attract a significant premium over non-interactive ones, such a ratio is prima facie suspect and will require further examination at some time. This being said, we remain satisfied that the rate we arrive at today should not be set aside solely based on this comparison. The SOCAN rate for on-demand streaming is a percentage of a service’s subscription revenue; consequently, some provision, however imperfect, had to be made for services that collect no such revenue. The rate for free streams is capped at the minimum payable by subscription-based services. As a result, any prejudice caused by setting a too high per-stream rate is largely contained. Indeed, under such a scenario, it was preferable to set too high a rate rather than one too low; such is not the case in this instance.

ii. Partial Plays

[174] Semi-interactive services allow consumers to skip tracks. Since we set a per-play tariff, we must decide whether partially-played tracks should attract royalties. At this time, we agree with Re:Sound that partial plays should attract full royalties, for the following reasons.

[175] First, even though the Board expressly alluded to this issue, no one suggested that partial plays should attract lower royalties than full plays.

[176] Second, we agree with Re:Sound that while the technology exists to track a stream every 10 seconds, the duration of plays is not generally tracked; that the administrative burden of such a measure would outweigh any benefits; and that since the amount of skipping is minimal, the benefits to users would be insignificant.

[177] Third, requiring the payment of royalties for partial tracks gives value to the interactivity that semi-interactive services provide over non-interactive ones: see paragraphs 179 to 181.

[178] It is probable that some partial plays involve an insubstantial, non-compensable taking of a sound recording. A per-play tariff does not prevent webcasters from pursuing the matter before the ordinary courts of law.

iii. An Interactivity Premium

[179] To set the tariff, we use as starting point SOCAN Tariff 1.A. Commercial radio is a non-interactive medium. This leaves open the possibility that semi-interactive services should pay a premium to account for their greater ability to customize and personalize webcasts.

[180] Skipping sound recordings one does not wish to hear has value; it helps webcasters to better tailor what is streamed to the listener’s tastes. Skipping is minimal and tends to be rare after the first month of having created a personalized radio channel; yet two considerations suggest that selection continues to have value notwithstanding this decrease. First, there is value to the option of skipping. Second, the absence of skipping, which implies that the listener likes the songs being played, is also an indicator of value. In a tariff that uses a percentage rate, it can be assumed that the market takes care of this; in a per-play tariff, any additional value must be reflected in the rate.

[181] For four reasons, we decline to set an interactivity premium. First, no one proposed a higher rate for semi-interactive webcasts than for non-interactive webcasts. Second, we have no reliable evidence to assess what the premium should be. Third, as must be clear by now, we are inclined to think that playing sound recordings should attract a single price in this particular market. Fourth, given Re:Sound’s proposed definition of non-interactive webcast, introducing a premium would engender perverse results. If a non-interactive webcast is “a webcast other than a simulcast, a semi-interactive webcast or an interactive communication”, then this includes some streaming that is too interactive to be semi-interactive yet not enough to meet the proposed definition of “interactive communication”. Espace.mu is a case in point. Listeners cannot access a specific recording at a time of their choice, but can pause and skip forward and back within the recording being streamed. A single rate for non-interactive and semi-interactive avoids the need for finer distinctions. We invite the parties to file evidence next time to help us better understand the various forms of interactivity and their relative values.

iv. Repertoire Adjustment

[182] Currently, Re:Sound is entitled to collect royalties only in respect of Rome sound recordings that are in its repertoire. [52] Since SOCAN’s repertoire is almost universal and since we use a SOCAN tariff as proxy, we must perform a repertoire adjustment.

[183] No evidence was adduced as to the use of the repertoire of Re:Sound in webcasts other than simulcasts. Re:Sound proposed using the average of the commercial radio, satellite radio, pay audio and CBC radio adjustments, or approximately 50 per cent. The Objectors proposed using only the 27.8 per cent adjustment as for satellite radio. Re:Sound replied that the Objectors offered very little to support using the satellite radio adjustment over, say, the pay audio adjustment of 45 per cent.

[184] We use the midpoint between 27.8 and 50 per cent, namely, 38.9 per cent. The Canadian Radio-television and Telecommunication Commission (CRTC) regulatory requirements tend to increase the use of Re:Sound’s repertoire. These requirements are minimal for satellite radio and significant for commercial radio. Using the satellite radio repertoire adjustment, 27.8 per cent, accounts for the fact that webcasters are not required to comply with regulatory requirements. Using the commercial radio adjustment, 50 per cent, accounts for the facts that webcasters compete with commercial radio and that some of them will attract listenership by offering alternative music that may largely originate from outside the United States.

[185] It would not be necessary to apply a repertoire adjustment to the per-play rate we set under two scenarios. The first supposes that users are able to determine in advance whether a sound recording is in Re:Sound’s repertoire. The second supposes that webcasters submit play reports to Re:Sound who would then invoice webcasters in a manner analogous to the sales reports submitted pursuant to the CSI Online Music Services Tariff, 2008-2010. The parties agree that it is not currently (and may never be) possible to segregate eligible recordings.

[186] As of December 31, 2012, the date at which the tariff we certify ceased to have effect, the provisions of the Copyright Modernization Act [53] that will vastly expand the repertoire of sound recordings that is eligible to receive equitable remuneration were not in force. Consequently, it is unnecessary to assess the impact these provisions will have on the proportion of air time involving Re:Sound’s repertoire.

v. Minima

[187] Re:Sound proposed an annual minimum fee for webcasting of $500 per channel capped at $50,000 per site. The Objectors individually proposed annual minimum fees of $100, without any cap. CAB also proposed minimum fees lower than $100 for stations that play less than 80 per cent of Re:Sound’s repertoire.

[188] We agree with Re:Sound that minimum fees are generally required to cover the cost of issuing licences. We agree with the Objectors that Re:Sound’s proposed minimum fees are inherently excessive. They exceed any minimum fee ever certified by the Board. They also reflect the US rate structure, which is very different from the Canadian one. Finally, we disagree with CAB that minimum fees should reflect repertoire usage. An important purpose of a minimum fee is to cover the costs of issuing the licence; the cost of the licence does not vary with the user’s usage of repertoire.

[189] We set the minimum royalty per webcaster at $100 per year. This is the same as the minimum royalty payable under SOCAN Tariff 22.F (Audio Websites). Since a minimum fee should not vary with usage of repertoire, there is no reason to set a different minimum fee for SOCAN and Re:Sound.

[190] Given the minimum we set, we see no need to provide a separate rate for microcasters. This issue may need to be revisited in the future.

G. Rates for CBC

[191] CBC proposed to model the tariff for its webcasts, other than previously aired programs, on SOCAN Tariff 22.F. Re:Sound proposed that CBC pay the same price as other webcasters.

[192] We agree with Re:Sound. In CBC Radio (2011), the Board de-linked CBC radio and commercial radio royalties, largely because CBC radio does not sell audiences to advertisers; as a result, it cannot achieve certain efficiencies linked to niche marketing. [54] Put another way, CBC delivered a different radio than its competitors. By contrast, while CBC’s webcasts may be different than those of other webcasters, there is no reason to believe that the variety of genres it offers is any different than what other webcasters provide as a group. The fact that CBC’s webcasts are advertising-supported only serves to reinforce the link between them and commercial webcasts. CBC will pay the same webcasting rate as commercial webcasters, before repertoire adjustment.

[193] The repertoire adjustment we use for commercial webcasts would be too great for CBC. By its mandate, CBC is expected to play much more Canadian content, most of which is in Re:Sound’s repertoire. We would have tended towards a figure of 70 per cent or so. However, we wish to account for the fact that some of the sound recordings CBC offers over the Internet (e.g., Radio 3, Bande à Part) may be eligible but not in Re:Sound’s repertoire. [55] Consequently, we will use for CBC the 50 per cent adjustment Re:Sound proposed for all webcasts other than simulcasts.

[194] CBC also asked that account be taken of live musical, non-musical, and audio-visual content, for which Re:Sound is not entitled to collect royalties by reason that such content does not involve the use of a published sound recording. Since there was no obvious way to do so, CBC proposed a further discount of 20 per cent. Re:Sound relied on CBC Radio (2011) to argue that “having a unique mandate does not necessarily entitle CBC to be treated differently than commercial radio.” [56]

[195] Re:Sound’s argument misses the point. CBC is asking a further discount to account not for its special mandate, but for its unique content, some of which does not involve any use of a published sound recording. Accounting for such content, while clearly an issue in a percentage-of-revenue tariff, may not be one in a per-play tariff if users are able to differentiate content that does not trigger equitable remuneration. We do not know whether, and to what extent, CBC is able to differentiate plays of sound recordings from other content; CBC even asserted that it is currently unable to calculate a per-play rate. While we find at paragraph 113 above that affordable technology is available to achieve this, we do not know whether such technology can discriminate between sound recordings and other audio content. Consequently, we should account in some measure for CBC’s unique content that does not warrant equitable remuneration. Absent any evidence on the issue, we find that the reasons on which we rely to lower the repertoire adjustment from 70 to 50 per cent already account sufficiently for this factor.

[196] CBC is not different from the commercial webcasters in this case; as such we set the minimum royalty at $100 for CBC, the same as for commercial webcasters.

[197] The rate for CBC’s simulcasts will be set with its over-the-air radio tariff. For previously aired programs, we do not set a rate for reasons explained at paragraph 104. For all other webcasts, CBC will pay a per-play rate of $0.000131 ($0.000262 x 0.5), with a minimum fee of $100 per year.

H. Rates for Webcasting Operations of Community and Non-Commercial Radio Systems

[198] Apparently, NCRA assumed (incorrectly) that Re:Sound expected to be paid two royalties of $60 per month if a non-commercial station offering non-interactive or semi-interactive webcasts in addition to simulcasts. Once that assumption is corrected, the main arguments of NCRA can be outlined as follows. First, $60 a month is $720 per year, which can be a large sum for a small station. Second, pursuant to paragraph 68.1(1)(b) of the Act, community stations only pay Re:Sound $100 for their broadcasting activities; it would be unreasonable that they pay more for services that reach far fewer listeners. Third, the reporting requirements are much more onerous here than under SOCAN Tariff 1.B (Non-Commercial Radio). Small stations do not have the ability to track which portions of websites were accessed or which musical selections were played.

[199] We agree with NCRA that $720 per year is too high. By comparison, Re:Sound is asking for a $500 minimum annual payment per channel for commercial webcasts and simulcasts. We do not consider this to be fair.

[200] With respect to simulcasts, NCRA asks that no additional fee be charged beyond the $100 paid pursuant to paragraph 68.1(1)(b) of the Act. It argues that this would parallel the situation with SOCAN tariffs where their stations do not have to pay an additional simulcast fee if they include the costs associated with the simulcast in their calculations of Tariff 1.B royalties. Re:Sound correctly states that paragraph 68.1(1)(b) only applies to the over-the-air operations of a system. However, while the Board has expressed serious misgivings about the fairness of a royalty of only $100 on the first $1.25M of income for a commercial radio station, we do not have the same reservations with respect to community systems.

[201] The other webcasting operations of community systems and other not-for-profit webcasters are minimal. To impose on them a per-play or any percentage rate would be unnecessarily costly or complicated. For the time being, they should only be required to pay an annual flat fee for all webcasts, including simulcasts, targeted in the tariff: we set such fee at $25. This fee is appropriate for what are essentially incidental uses by these stations. It also avoids any unnecessary reporting burden.

[202] Some Internet offerings that are meant at first to be non-commercial and to target a very limited public sometimes go viral. Given that the tariff we certify will apply entirely retroactively, we see no need to account for this at this time. The issue will need to be addressed in the next proceedings.

I. Certified Rates and Royalties Generated

[203] The rates we certify are as follows.

CERTIFIED RATES AND ROYALTIES GENERATED

Item	Rates Taux
CBC / SRC
Webcast /	13.1¢ per thousand plays
Webdiffusion	13,1 ¢ par millier d’écoutes
Minimum fee /	$100 per year
Redevance minimale	100 $ par année
Commercial webcasters / Webdiffuseurs commerciaux
Webcast /	10.2¢ per thousand plays
Webdiffusion	10,2 ¢ par millier d’écoutes
Minimum fee /	$100 per year
Redevance minimale	100 $ par année
Community and non-commercial webcasters / Webdiffuseurs communautaires et non commerciaux
Webcast /	$25 per year
Webdiffusion	25 $ par année

[204] The rate of 10.2¢ per thousand plays we certify corresponds to a rate of 26.2¢, before repertoire adjustment. By contrast, Re:Sound was asking for an unadjusted rate of 96¢ if SRPC compliant, or $2.30 if not. Pandora proposed a rate of 5.3 per cent of revenues. This can be converted to a per-play rate using Pandora’s U.S. revenues and plays for 2012 in the following way. Pandora’s revenues were $274.3 million, with 8.2 billion listener hours. With 15 plays per hour, this translates into 123 billion plays. Pandora’s equivalent proposed rate per thousand plays is thus 11.8¢ ($274.3M × 0.053 ÷ 123 million).

[205] We estimate the royalties that Pandora would pay in Canada as follows. In its first year of operation as a webcaster, 2010, Pandora had 1.8 billion U.S. listener hours, which amounts to 27 billion plays. Adjusting for a smaller market size by using a ratio of 9 to 1 which reflects the relative population sizes of Canada (35 million) and the U.S. (314 million), we estimate that Pandora would have 3 billion plays in Canada in its first year of operations. To be sure, this estimate is conservative, since Pandora may have more name recognition in Canada in 2013 than it had in the United States in 2010. Based on this usage, Pandora would pay $306,000 (3 million × 10.2¢). We estimate that this would constitute between 4 and 5 per cent of Pandora’s Canadian revenues. [57]

[206] As a comparison, Pandora paid in fiscal year 2012, the last fiscal year available, the Pureplay rate (i.e. the rate that must be paid when a webcaster’s sole business is webcasting) of $1.10 per thousand plays. As such, U.S. royalties amount to $136.5 million, or about 50 per cent of its U.S. revenues.

[207] We also estimate the amount of royalties to be paid by CBC under this tariff. As part of its evidence, CBC filed three different observations on weekly hours listened for its English webcasting service, CBC Music: the average of these three observations is about 345,000. Multiplying this by the number of weeks in a year (52) and the number of plays per hour (15) leads to the total number of plays per year. Multiplying this number by the per-play rate of $0.000131 yields an annual estimate of about $35,300. CBC also filed its monthly visitors and average time per visit to its French webcasting service, Espace.mu. This amounted to 20,000 visitors for 66 minutes each, on average. Multiplying the average monthly hours listened by the number of months in a year (12), and by the number of plays per hour (15) yields the number of yearly plays. At the per-play rate of $0.000131, this works out to about $500. Accordingly, we estimate that CBC will pay about $36,000 per year under this tariff.

[208] Our estimate of the total quantum of the Re:Sound 8 tariff is about $500,000, taking into account other semi-interactive webcasters that have launched or may launch in Canada, as well as non-interactive webcasters. This estimate is arrived at knowing that the music streaming market share of Pandora in the U.S is about 70 per cent, and that its Canadian market share could be lower, given that some other firms have already been operating in Canada.

[209] We believe that royalties to be paid in respect of this tariff are fair and equitable for both the users and the copyright owners. These royalty rates will not be an impediment for webcasters to do business in Canada.

VI. TARIFF WORDING, STRUCTURE AND TRANSITIONAL PROVISIONS

[210] While Re:Sound originally filed two proposed tariffs, namely its Tariff 8.A in respect of simulcasting and webcasting for the years 2009-2012 and Tariff 8.B in respect of Semi-interactive webcasting for the years 2011-2012, we are certifying a single tariff for the period from 2009 to 2012 to deal with simulcasting, non-interactive webcasting and semi-interactive webcasting. So doing does not raise a problem since, although only the second proposed tariff referenced explicitly semi-interactive webcasting, the definition we ascribe to semi-interactive webcast in the tariff we certify was implicitly included in the first proposed tariff. Furthermore, as proposed, there could have been an overlap between the two tariffs for some semi-interactive webcasting for the years 2011 and 2012. The fact that we certify the same rate for both non-interactive and semi-interactive webcasting is a consequence of the evidence on the record, as we explained at paragraph 181 above, rather than a result of the merging of the two proposed tariffs; it does, however, simplify such merging.

[211] On January 27, 2014, we consulted the parties on the draft wording of the definitions and the reporting requirements of the tariff.

A. Definitions

[212] Some definitions in the tariff use terms such as “communication” and “performance” rather than the more general term “transmission” to describe activities targeted by the tariff. This is done deliberately for a few reasons, with the intent to remove uncertainties with respect to the scope and application of the tariff.

[213] First, the only rights involved in this proceeding are the public performance and the communication to the public by telecommunication, in Canada, of published sound recordings embodying musical works and performers’ performances of such works.

[214] Second, since the decision of the Supreme Court of Canada in Entertainment Software Association v. Society of Composers, Authors and Music Publishers of Canada, [58] downloads of musical works or other subject-matter do not engage the right of communication to the public by telecommunication but only the reproduction right. Therefore, the distinction between a communication and a transmission, when referring to copyright-protected content, is now very important. A communication is a type of transmission but not all transmissions are communications.

[215] Third, Re:Sound only administers rights of public performance and the communication to the public by telecommunication for the benefit of sound recording makers and performers.

[216] Fourth, the intent behind a definition in a tariff is to delineate clearly the scope of the tariff, and, as such, where a clearer and more accurate term or expression is available, it should be favoured.

B. Non-interactive and semi-interactive webcasts

[217] The definitions of non-interactive webcast and semi-interactive webcast explicitly exclude the simulcast of programming to which the CBC Radio Tariff (SOCAN, Re:Sound), the Commercial Radio Tariff (SOCAN, Re:Sound, CSI, AVLA/SOPROQ, ArtistI), the Pay Audio Services Tariff (SOCAN, NRCC), or the Satellite Radio Services Tariff (SOCAN, NRCC, CSI) applies. As explained at paragraphs 72-75, we intend to set the rates for simulcasting consisting of an ancillary use of a primary activity already targeted by a tariff at the same time we certify rates for such primary activity. This is in line with the approach taken by the Board in recent decisions dealing with ancillary uses on the Internet. As Re:Sound commented, and we agree, it is nevertheless important to define the activities in a way that would not create any void in the scope and application of the tariff, namely for simulcasts of programming not otherwise subject to an existing tariff. A case in point would be the simulcasting in Canada of a foreign terrestrial radio station’s programming. As a result of the tariff structure, such an activity would be subject to the non-interactive webcasting royalty rate under this tariff.

C. Non-commercial webcaster

[218] Re:Sound submitted that the Board modify the definition of non-commercial webcaster to ensure that the webcaster be legally incorporated or otherwise created as a not-for-profit entity. Re:Sound’s concern was that webcasters may otherwise claim that they fall under the non-commercial category simply because they have not generated a profit in a particular year. We do not believe such a change is necessary, as the concept of non-profit or not-for-profit organization unambiguously refers to the constitution of an organization and not to its profitability. It is a generally accepted concept, including by the Canada Revenue Agency.

D. Webcasts

[219] Two comments received from the parties with respect to the definition of the term webcast warrant clarification.

[220] Re:Sound expressed a concern that the definition adds the requirement that a copy be made and argued that while a webcast does typically involve a buffering copy, the reproduction right is outside the scope of the proposed tariff. Including the requirement of a copy in the definition would unnecessarily narrow the scope of the tariff, particularly in light of rapidly evolving technologies which could result in the ability to stream without making any copy. We disagree with that interpretation of the definition. In our view, the wording “only to the extent required” qualifying the reproduction of the file being webcast clearly allows a certain type of reproduction of a very limited nature but certainly does not make reproduction a requirement.

[221] Pandora raised the concern that the definition of webcast in the tariff appears to exclude from the tariff any service that permits caching of its streams. Again, we are of the view that the definition can accommodate such caching by the effect of the wording “to the extent required to allow listening to the files at substantially the same time as when the files are communicated.” The definition is flexible enough to adapt to some technological developments but circumscribed enough to limit its application to activities that fall under the breadth of the rights of public performance and communication to the public by telecommunication.

E. Partial plays

[222] The wording of the definitions of the terms play and file reflects our decision to include partial plays within the ambit of the tariff, for the reasons discussed at paragraphs 174 to 178 above. The definition of file further accounts for the fact that the rates we set in this tariff are adjusted for repertoire. The issue of whether to include partial plays as part of this tariff may need to be revisited in the future, in light of the evidence then available. It should also be noted that the definition of file also conveys our decision to certify a repertoire adjusted per-play rate and therefore sound recordings of musical works for which Re:Sound is not entitled to equitable remuneration will nevertheless trigger a royalty payment, regardless of whether they are in repertoire, eligible, or in the public domain.

F. Channel

[223] The parties were specifically asked how the term “channel” should be defined as it applies to non-interactive and semi-interactive webcasts, and whether there should be a distinction between a channel containing non-interactive webcasts and one containing semi-interactive webcasts. The responses received by the parties on this issue confirmed some difficulties we had already anticipated in coming up with an appropriate definition. As a result, we decide not to use the concept of channel in the context of the tariff we certify.

G. Transitional provisions

[224] The tariff contains certain transitional provisions made necessary because it takes effect on January 1, 2009, while being certified much later and because the tariff structure, based on per-play rates, does not necessarily reflect past practices of users subject to the tariff. The period under examination has already ended. For those reasons, transitional provisions are required in the tariff to account for music use for webcasting which was carried out in the past by webcasters which did not track the number of plays and also for a certain period of time in the future to allow webcasters to make the necessary changes to their operations to be able to track plays.

[225] Users cannot be expected to provide information that they did not keep or track. On the other hand, we see no reason to dispense a user from providing information that it has; that will help Re:Sound to effectively monitor the use of its repertoire and distribute the royalties it collects. Therefore, the reporting requirements for the period affected by the transitional provisions of the tariff are to be provided on an “if available” basis, with the exception of the ATH, the total number of programming hours and the total number of listeners, which are to be provided on a mandatory basis. The normal reporting requirements provision stipulates that additional information is to be provided on a mandatory basis for periods of time not subject to the transitional provisions.

[226] The application of the transitional provisions has some limitations. From the outset, they are only applicable for periods of time during which a webcaster did not have the capability of accounting for its use of music on a per-play basis. Furthermore, their application is limited in time for a few reasons. First, we found that the best rate base for this tariff is a per-play rate. Second, the evidence showed that the technology necessary to implement an accounting of a webcaster’s plays is available and affordable. Third, such a per-play accounting has been in place for several years in the United States, where several webcasters who will be subject to the tariff are already required to keep track of their number of plays. Fourth, when the transitional provisions are in force, and therefore ATH are used to calculate music use, it will be difficult to provide most information required for reporting purposes, making the administration of the tariff and redistribution difficult for Re:Sound.

H. Interest

[227] A table included in the tariff sets out the multiplying interest factors to be used on sums owed in a given month. The factors were derived using month-end Bank Rates. A second table sets out the multiplying interest factors to be used on sums owed on an annual basis. These factors were derived using the annual average of the monthly rates. We consider that a penalty over and above the interest factor should not be imposed on retroactive payments in this matter, as there was no way for users to estimate the amounts payable until the tariff was approved. Interest is not compounded. The amount owed for a reporting period is the amount of the certified tariff multiplied by the factor set out for that period.

Gilles McDougall

Secretary General