Argued and Submitted April 15, 1999 — Pasadena, California.
Decided June 15, 1999
Hadrian R. Katz (argued), Lawrence J. Hutt, Arnold Porter, Washington, D.C., for the plaintiffs-appellants.
Andrew P. Bridges (argued), Wilson Sonsini Goodrich Rosati, Palo Alto, California, for the defendant-appellee.
John B. Wyss, Wiley, Rein Fielding, Washington, D.C., for the amicus Consumer Electronics Manufacturers Association.
Appeal from the United States District Court for the Central District of California, Audrey B. Collins, District Judge, Presiding, D.C. No. CV-98-08247-ABC.
In this case involving the intersection of computer technology, the Internet, and music listening, we must decide whether the Rio portable music player is a digital audio recording device subject to the restrictions of the Audio Home Recording Act of 1992.
This appeal arises from the efforts of the Recording Industry Association of America and the Alliance of Artists and Recording Companies (collectively, "RIAA") to enjoin the manufacture and distribution by Diamond Multimedia Systems ("Diamond") of the Rio portable music player. The Rio is a small device (roughly the size of an audio cassette) with headphones that allows a user to download MP3 audio files from a computer and to listen to them elsewhere. The dispute over the Rio's design and function is difficult to comprehend without an understanding of the revolutionary new method of music distribution made possible by digital recording and the Internet; thus, we will explain in some detail the brave new world of Internet music distribution.
The introduction of digital audio recording to the consumer electronics market in the 1980's is at the root of this litigation. Before then, a person wishing to copy an original music recording — e.g., wishing to make a cassette tape of a record or compact disc — was limited to analog, rather than digital, recording technology. With analog recording, each successive generation of copies suffers from an increasingly pronounced degradation in sound quality. For example, when an analog cassette copy of a record or compact disc is itself copied by analog technology, the resulting "second-generation" copy of the original will most likely suffer from the hiss and lack of clarity characteristic of older recordings. With digital recording, by contrast, there is almost no degradation in sound quality, no matter how many generations of copies are made. Digital copying thus allows thousands of perfect or near perfect copies (and copies of copies) to be made from a single original recording. Music "pirates" use digital recording technology to make and to distribute near perfect copies of commercially prepared recordings for which they have not licensed the copyrights.
Until recently, the Internet was of little use for the distribution of music because the average music computer file was simply too big: the digital information on a single compact disc of music required hundreds of computer floppy discs to store, and downloading even a single song from the Internet took hours. However, various compression algorithms (which make an audio file "smaller" by limiting the audio bandwidth) now allow digital audio files to be transferred more quickly and stored more efficiently. MPEG-1 Audio Layer 3 (commonly known as "MP3") is the most popular digital audio compression algorithm in use on the Internet, and the compression it provides makes an audio file "smaller" by a factor of twelve to one without significantly reducing sound quality. MP3's popularity is due in large part to the fact that it is a standard, non-proprietary compression algorithm freely available for use by anyone, unlike various proprietary (and copyright-secure) competitor algorithms. Coupled with the use of cable modems, compression algorithms like MP3 may soon allow an hour of music to be downloaded from the Internet to a personal computer in just a few minutes.
These technological advances have occurred, at least in part, to the traditional music industry's disadvantage. By most accounts, the predominant use of MP3 is the trafficking in illicit audio recordings, presumably because MP3 files do not contain codes identifying whether the compressed audio material is copyright protected. Various pirate websites offer free downloads of copyrighted material, and a single pirate site on the Internet may contain thousands of pirated audio computer files.
RIAA represents the roughly half-dozen major record companies (and the artists on their labels) that control approximately ninety percent of the distribution of recorded music in the United States. RIAA asserts that Internet distribution of serial digital copies of pirated copyrighted material will discourage the purchase of legitimate recordings, and predicts that losses to digital Internet piracy will soon surpass the $300 million that is allegedly lost annually to other more traditional forms of piracy. RIAA fights a well-nigh constant battle against Internet piracy, monitoring the Internet daily, and routinely shutting down pirate websites by sending cease-and-desist letters and bringing lawsuits. There are conflicting views on RIAA's success — RIAA asserts that it can barely keep up with the pirate traffic, while others assert that few, if any, pirate sites remain in operation in the United States and illicit files are difficult to find and download from anywhere online.
Whether or not piracy causes such financial harm is a subject of dispute. Critics of the industry's piracy loss figures have noted that a willingness to download illicit files for free does not necessarily correlate to lost sales, for the simple reason that persons willing to accept an item for free often will not purchase the same item, even if no longer freely available. See Lewis Kurlantzick Jacqueline E. Pennino, The Audio Home Recording Act of 1992 and the Formation of Copyright Policy, 45 J. Copyright Soc'y U.S.A. 497, 506 (1998). Critics further note that the price of commercially available recordings already reflects the existence of copying and the benefits and harms such copying causes; thus, they contend, the current price of recordings offsets, at least in part, the losses incurred by the industry from home taping and piracy. See id. at 509-10.
In contrast to piracy, the Internet also supports a burgeoning traffic in legitimate audio computer files. Independent and wholly Internet record labels routinely sell and provide free samples of their artists' work online, while many unsigned artists distribute their own material from their own websites. Some free samples are provided for marketing purposes or for simple exposure, while others are teasers intended to entice listeners to purchase either mail order recordings or recordings available for direct download (along with album cover art, lyrics, and artist biographies). Diamond cites a 1998 "Music Industry and the Internet" report by Jupiter Communications which predicts that online sales for pre-recorded music will exceed $1.4 billion by 2002 in the United States alone.
Prior to the invention of devices like the Rio, MP3 users had little option other than to listen to their downloaded digital audio files through headphones or speakers at their computers, playing them from their hard drives. The Rio renders these files portable. More precisely, once an audio file has been downloaded onto a computer hard drive from the Internet or some other source (such as a compact disc player or digital audio tape machine), separate computer software provided with the Rio (called "Rio Manager") allows the user further to download the file to the Rio itself via a parallel port cable that plugs the Rio into the computer. The Rio device is incapable of effecting such a transfer, and is incapable of receiving audio files from anything other than a personal computer equipped with Rio Manager.
Generally, the Rio can store approximately one hour of music, or sixteen hours of spoken material (e.g., downloaded newscasts or books on tape). With the addition of flash memory cards, the Rio can store an additional half-hour or hour of music. The Rio's sole output is an analog audio signal sent to the user via headphones. The Rio cannot make duplicates of any digital audio file it stores, nor can it transfer or upload such a file to a computer, to another device, or to the Internet. However, a flash memory card to which a digital audio file has been downloaded can be removed from one Rio and played back in another.
RIAA brought suit to enjoin the manufacture and distribution of the Rio, alleging that the Rio does not meet the requirements for digital audio recording devices under the Audio Home Recording Act of 1992, 17 U.S.C. § 1001 et seq. (the "Act"), because it does not employ a Serial Copyright Management System ("SCMS") that sends, receives, and acts upon information about the generation and copyright status of the files that it plays. See id. § 1002(a)(2). RIAA also sought payment of the royalties owed by Diamond as the manufacturer and distributor of a digital audio recording device. See id. § 1003.
At the time the preliminary injunction was sought and denied, the Rio did not incorporate SCMS; Diamond asserts that it has now incorporated such a system into the Rio Manager software, though not into the Rio itself.
The district court denied RIAA's motion for a preliminary injunction, holding that RIAA's likelihood of success on the merits was mixed and the balance of hardships did not tip in RIAA's favor. See generally Recording Indus. Ass'n of America, Inc. v. Diamond Multimedia Sys., Inc., 29 F. Supp.2d 624 (C.D. Cal. 1998) ("RIAA I"). RIAA brought this appeal.
The initial question presented is whether the Rio falls within the ambit of the Act. The Act does not broadly prohibit digital serial copying of copyright protected audio recordings. Instead, the Act places restrictions only upon a specific type of recording device. Most relevant here, the Act provides that "[n]o person shall import, manufacture, or distribute any digital audio recording device . . . that does not conform to the Serial Copy Management System ["SCMS"] [or] a system that has the same functional characteristics." 17 U.S.C. § 1002(a)(1), (2) (emphasis added). The Act further provides that "[n]o person shall import into and distribute, or manufacture and distribute, any digital audio recording device . . . unless such person records the notice specified by this section and subsequently deposits the statements of account and applicable royalty payments." Id. § 1003(a) (emphasis added). Thus, to fall within the SCMS and royalty requirements in question, the Rio must be a "digital audio recording device," which the Act defines through a set of nested definitions.
The Act defines a "digital audio recording device" as:
any machine or device of a type commonly distributed to individuals for use by individuals, whether or not included with or as part of some other machine or device, the digital recording function of which is designed or marketed for the primary purpose of, and that is capable of, making a digital audio copied recording for private use. . . .
Id. § 1001(3) (emphasis added).
A "digital audio copied recording" is defined as:
a reproduction in a digital recording format of a digital musical recording, whether that reproduction is made directly from another digital musical recording or indirectly from a transmission.
Id. § 1001(1) (emphasis added).
A "digital musical recording" is defined as:
a material object
(i) in which are fixed, in a digital recording format, only sounds, and material, statements, or instructions incidental to those fixed sounds, if any, and
(ii) from which the sounds and material can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.
Id. § 1001(5)(A) (emphasis added).
In sum, to be a digital audio recording device, the Rio must be able to reproduce, either "directly" or "from a transmission," a "digital music recording."
We first consider whether the Rio is able directly to reproduce a digital music recording — which is a specific type of material object in which only sounds are fixed (or material and instructions incidental to those sounds). See id.
The typical computer hard drive from which a Rio directly records is, of course, a material object. However, hard drives ordinarily contain much more than "only sounds, and material, statements, or instructions incidental to those fixed sounds." Id. Indeed, almost all hard drives contain numerous programs (e.g., for word processing, scheduling appointments, etc.) and databases that are not incidental to any sound files that may be stored on the hard drive. Thus, the Rio appears not to make copies from digital music recordings, and thus would not be a digital audio recording device under the Act's basic definition unless it makes copies from transmissions.
Moreover, the Act expressly provides that the term "digital musical recording" does not include:
a material object
(i) in which the fixed sounds consist entirely of spoken word recordings, or
(ii) in which one or more computer programs are fixed, except that a digital recording may contain statements or instructions constituting the fixed sounds and incidental material, and statements or instructions to be used directly or indirectly in order to bring about the perception, reproduction, or communication of the fixed sounds and incidental material.
Id. § 1001(5)(B) (emphasis added). As noted previously, a hard drive is a material object in which one or more programs are fixed; thus, a hard drive is excluded from the definition of digital music recordings. This provides confirmation that the Rio does not record "directly" from "digital music recordings," and therefore could not be a digital audio recording device unless it makes copies "from transmissions."
The district court rejected the exclusion of computer hard drives from the definition of digital music recordings under the statute's plain language (after noting its "superficial appeal") because it concluded that such exclusion "is ultimately unsupported by the legislative history, and contrary to the spirit and purpose of the [Act]." RIAA I, 29 F. Supp.2d at 629. We need not resort to the legislative history because the statutory language is clear. See City of Auburn v. United States, 154 F.3d 1025, 1030 (9th Cir. 1998) ("[W]here statutory command is straightforward, `there is no reason to resort to legislative history.'" (quoting United States v. Gonzales, 520 U.S. 1, 6 (1997))). Nevertheless, we will address the legislative history here, because it is consistent with the statute's plain meaning and because the parties have briefed it so extensively.
We can, of course, affirm on any grounds supported by the record, see Gemtel Corp. v. Community Redevelopment Agency of City of Los Angeles, 23 F.3d 1542, 1546 (9th Cir. 1994), thus, we can affirm even if the lower court relied on incorrect grounds or faulty reasoning, see Aronson v. Resolution Trust Corp., 38 F.3d 1110, 1114 (9th Cir. 1994).
There is no precedent (other than the district court's order) to guide the panel's interpretation of the Act. The Act has only been discussed once in a published opinion by another federal court, and there, only to explain why it had no effect on the Copyright Act provisions at issue in that case. See ABKCO Music, Inc. v. Stellar Records, Inc., 96 F.3d 60, 65-66 (2d Cir. 1996) (rejecting the contention that the Act changed or affected the definition of "phonorecord" in the Copyright Act).
The Senate Report states that "if the material object contains computer programs or data bases that are not incidental to the fixed sounds, then the material object would not qualify" under the basic definition of a digital musical recording. S. Rep. 102-294 (1992), reprinted at 1992 WL 133198, at *118-19. The Senate Report further states that the definition "is intended to cover those objects commonly understood to embody sound recordings and their underlying works." Id. at *97. A footnote makes explicit that this definition only extends to the material objects in which songs are normally fixed: "[t]hat is recorded compact discs, digital audio tapes, audio cassettes, long-playing albums, digital compact cassettes, and mini-discs." Id. at n. 36. There are simply no grounds in either the plain language of the definition or in the legislative history for interpreting the term "digital musical recording" to include songs fixed on computer hard drives.
The Senate Report discusses the original term "audiogram," which was replaced by the term "digital music recording," but the two definitions are nearly identical, with the only difference being the deletion from the "audiogram" definition of examples of material objects in which things other than sounds are fixed. Compare S. Rep. 102-294, at *4-5 ("An `audiogram' is a material object (i) in which are fixed, by any method now known or later developed, only sounds (and not, for example, a motion picture or other audiovisual work even though it may be accompanied by sounds), and material, statements or instructions incidental to those fixed sounds, if any, and (ii) from which the sounds and material can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.") (emphasis added), with 17 U.S.C. § 1001(5)(A) ("A `digital music recording' is a material object — (i) in which are fixed, in a digital recording format, only sounds, and material, statements, or instructions incidental to those fixed sounds, if any, and (ii) from which the sounds and material can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device."). Thus, comments in the legislative history regarding the "audiogram" definition are relevant to our interpretation of the "digital music recording" definition.
RIAA contends that the legislative history reveals that the Rio does not fall within the specific exemption from the digital musical recording definition of "a material object in which one or more computer programs are fixed." 17 U.S.C. § 1001(5)(B)(ii). The House Report describes the exemption as "revisions reflecting exemptions for talking books and computer programs." H.R. Rep. 102-873(I) (1992), reprinted at 1992 WL 232935, at *35 (emphasis added); see also id. at *44 ("In addition to containing an express exclusion of computer programs in the definition of `digital musical recording'. . . .") (emphasis added). We first note that limiting the exemption to computer programs is contrary to the plain meaning of the exemption. As Diamond points out, a computer program is not a material object, but rather, a literary work, see, e.g., Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1240, 1249 (3d Cir. 1983) ("[A] computer program . . . is a `literary work.'"), that can be fixed in a variety of material objects, see 17 U.S.C. § 101 ("`Literary works' are works . . . expressed in words, numbers, or other verbal or numerical symbols or indicia, regardless of the nature of the material objects, such as books . . . tapes, disks, or cards, in which they are embodied.") (emphasis added). Thus, the plain language of the exemption at issue does not exclude the copying of programs from coverage by the Act, but instead, excludes copying from various types of material objects. Those objects include hard drives, which indirectly achieve the desired result of excluding copying of programs. But by its plain language, the exemption is not limited to the copying of programs, and instead extends to any copying from a computer hard drive.
Moreover, RIAA's assertion that computer hard drives do not fall within the exemption is irrelevant because, regardless of that portion of the legislative history which addresses the exemption from the definition of digital music recording, see id. § 1001(5)(B)(ii), the Rio does not reproduce files from something that falls within the plain language of the basic definition of a digital music recording, see id. § 1001(5)(A).
The district court concluded that the exemption of hard drives from the definition of digital music recording, and the exemption of computers generally from the Act's ambit, "would effectively eviscerate the [Act]" because "[a]ny recording device could evade [ ] regulation simply by passing the music through a computer and ensuring that the MP3 file resided momentarily on the hard drive." RIAA I, 29 F. Supp.2d at 630. While this may be true, the Act seems to have been expressly designed to create this loophole.
Under the plain meaning of the Act's definition of digital audio recording devices, computers (and their hard drives) are not digital audio recording devices because their "primary purpose" is not to make digital audio copied recordings. See 17 U.S.C. § 1001(3). Unlike digital audio tape machines, for example, whose primary purpose is to make digital audio copied recordings, the primary purpose of a computer is to run various programs and to record the data necessary to run those programs and perform various tasks. The legislative history is consistent with this interpretation of the Act's provisions, stating that "the typical personal computer would not fall within the definition of `digital audio recording device,'" S. Rep. 102-294, at *122, because a personal computer's "recording function is designed and marketed primarily for the recording of data and computer programs," id. at *121. Another portion of the Senate Report states that "[i]f the `primary purpose' of the recording function is to make objects other than digital audio copied recordings, then the machine or device is not a `digital audio recording device,' even if the machine or device is technically capable of making such recordings." Id. (emphasis added). The legislative history thus expressly recognizes that computers (and other devices) have recording functions capable of recording digital musical recordings, and thus implicate the home taping and piracy concerns to which the Act is responsive. Nonetheless, the legislative history is consistent with the Act's plain language — computers are not digital audio recording devices.
Indeed, Diamond asserted at oral argument (and supports the assertion with the affidavit of a direct participant in the negotiations and compromises that resulted in the final language of the Act) that the exclusion of computers from the Act's scope was part of a carefully negotiated compromise between the various industries with interests at stake, and without which, the computer industry would have vigorously opposed passage of the Act.
In turn, because computers are not digital audio recording devices, they are not required to comply with the SCMS requirement and thus need not send, receive, or act upon information regarding copyright and generation status. See 17 U.S.C. § 1002(a)(2). And, as the district court found, MP3 files generally do not even carry the codes providing information regarding copyright and generation status. See RIAA I, 29 F. Supp.2d. at 632. Thus, the Act seems designed to allow files to be "laundered" by passage through a computer, because even a device with SCMS would be able to download MP3 files lacking SCMS codes from a computer hard drive, for the simple reason that there would be no codes to prevent the copying.
Again, the legislative history is consistent with the Act's plain meaning. As the Technical Reference Document that describes the SCMS system explains, "[d]igital audio signals . . . that have no information concerning copyright and/or generation status shall be recorded by the [digital audio recording] device so that the digital copy is copyright asserted and original generation status." Technical Reference Document for the Audio Home Recording Act of 1992, II-A, P 10, reprinted in H.R. Rep. 102-780(I), 32, 43 (1992) (emphasis added). Thus, the incorporation of SCMS into the Rio would allow the Rio to copy MP3 files lacking SCMS codes so long as it marked the copied files as "original generation status." And such a marking would allow another SCMS device to make unlimited further copies of such "original generation status" files, see, e.g., H.R. Rep. 102-873(I), at *47 ("Under SCMS . . . consumers will be able to make an unlimited number of copies from a digital musical recording."), despite the fact that the Rio does not permit such further copies to be made because it simply cannot download or transmit the files that it stores to any other device. Thus, the Rio without SCMS inherently allows less copying than SCMS permits.
In fact, the Rio's operation is entirely consistent with the Act's main purpose — the facilitation of personal use. As the Senate Report explains, "[t]he purpose of [the Act] is to ensure the right of consumers to make analog or digital audio recordings of copyrighted music for their private, noncommercial use." S. Rep. 102-294, at *86 (emphasis added). The Act does so through its home taping exemption, see 17 U.S.C. § 1008, which "protects all noncommercial copying by consumers of digital and analog musical recordings," H.R. Rep. 102-873(I), at *59. The Rio merely makes copies in order to render portable, or "space-shift," those files that already reside on a user's hard drive. Cf. Sony Corp. of America v. Universal City Studios, 464 U.S. 417, 455 (1984) (holding that "time-shifting" of copyrighted television shows with VCR's constitutes fair use under the Copyright Act, and thus is not an infringement). Such copying is paradigmatic noncommercial personal use entirely consistent with the purposes of the Act.
Even though it cannot directly reproduce a digital music recording, the Rio would nevertheless be a digital audio recording device if it could reproduce a digital music recording "from a transmission." 17 U.S.C. § 1001(1).
The term "transmission" is not defined in Act, although the use of the term in the Act implies that a transmission is a communication to the public. See id. § 1002(e) (placing restrictions upon "[a]ny person who transmits or otherwise communicates to the public any sound recording in digital format") (emphasis added). In the context of copyright law (from which the term appears to have been taken), "[t]o `transmit' a performance or display is to communicate it by any device or process whereby images or sounds are received beyond the place from which they are sent." 17 U.S.C. § 101. The legislative history confirms that the copyright definition of "transmission" is sufficient for our purposes here. The Act originally (and circularly) provided that "[a] `transmission' is any audio or audiovisual transmission, now known or later developed, whether by a broadcast station, cable system, multipoint distribution service, subscription service, direct broadcast satellite, or other form of analog or digital communication." S. Rep. 102-294, at *10. The Senate Report provides a radio broadcast as an example of a transmission. See id., at *119 (referring to "a transmission (e.g., a radio broadcast of a commercially released audio cassette)."). The parties do not really dispute the definition of transmission, but rather, whether indirect reproduction of a transmission of a digital music recording is covered by the Act.
RIAA asserts that indirect reproduction of a transmission is sufficient for the Rio to fall within the Act's ambit as a digital audio recording device. See 17 U.S.C. § 1001(1) (digital audio recording devices are those devices that are capable of making "a reproduction in a digital recording format of a digital musical recording, whether that reproduction is made directly from another digital musical recording or indirectly from a transmission") (emphasis added). Diamond asserts that the adverb "indirectly" modifies the recording of the underlying "digital music recording," rather than the recording "from the transmission." Diamond effectively asserts that the statute should be read as covering devices that are capable of making a reproduction of a digital musical recording, "whether that reproduction is made directly[,] from another digital musical recording[,] or indirectly[,] from a transmission."
While the Rio can only directly reproduce files from a computer hard drive via a cable linking the two devices (which is obviously not a transmission), the Rio can indirectly reproduce a transmission. For example, if a radio broadcast of a digital audio recording were recorded on a digital audio tape machine or compact disc recorder and then uploaded to a computer hard drive, the Rio could indirectly reproduce the transmission by downloading a copy from the hard drive. Thus, if indirect reproduction of a transmission falls within the statutory definition, the Rio would be a digital audio recording device.
RIAA's interpretation of the statutory language initially seems plausible, but closer analysis reveals that it is contrary to the statutory language and common sense. The focus of the statutory language seems to be on the two means of reproducing the underlying digital music recording — either directly from that recording, or indirectly, by reproducing the recording from a transmission. RIAA's interpretation of the Act's language (in which "indirectly" modifies copying "from a transmission," rather than the copying of the underlying digital music recording) would only cover the indirect recording of transmissions, and would omit restrictions on the direct recording of transmissions (e.g., recording songs from the radio) from the Act's ambit. This interpretation would significantly reduce the protection afforded by the Act to transmissions, and neither the statutory language nor structure provides any reason that the Act's protections should be so limited. Moreover, it makes little sense for the Act to restrict the indirect recording of transmissions, but to allow unrestricted direct recording of transmissions (e.g., to regulate second-hand recording of songs from the radio, but to allow unlimited direct recording of songs from the radio). Thus, the most logical reading of the Act extends protection to direct copying of digital music recordings, and to indirect copying of digital music recordings from transmissions of those recordings.
Because of the arguable ambiguity of this passage of the statute, recourse to the legislative history is necessary on this point. Cf. Moyle v. Director, Office of Workers' Compensation Programs, 147 F.3d 1116, 1120 (9th Cir. 1998) ("[I]f the statute is ambiguous, [this court] consult[s] the legislative history, to the extent that it is of value, to aid in [its] interpretation."), cert. denied, 119 S.Ct. 1454 (1999). The Senate Report states that "a digital audio recording made from a commercially released compact disc or audio cassette, or from a radio broadcast of a commercially released compact disc or audio cassette, would be a `digital audio copied recording.'" S. Rep. 102-294, at *119 (emphasis added). This statement indicates that the recording of a transmission need not be indirect to fall within the scope of the Act's restrictions, and thus refutes RIAA's proposed interpretation of the relevant language. Moreover, the statement tracks the statutory definition by providing an example of direct copying of a digital music recording from that recording, and an example of indirect copying of a digital music recording from a transmission of that recording. Thus the legislative history confirms the most logical reading of the statute, which we adopt: "indirectly" modifies the verb "is made" — in other words, modifies the making of the reproduction of the underlying digital music recording. Thus, a device falls within the Act's provisions if it can indirectly copy a digital music recording by making a copy from a transmission of that recording. Because the Rio cannot make copies from transmissions, but instead, can only make copies from a computer hard drive, it is not a digital audio recording device.
We further note that any transmission reproduced indirectly must pass through a computer, as an MP3 file, to reach the Rio. As we explained in part III.B.2, supra, computers are exempted from the requirement of reading and transmitting SCMS codes, and MP3 files do not incorporate such codes. Thus, requiring the Rio to implement SCMS because it can indirectly reproduce a transmission of a digital music recording would be, as the district court concluded, "an exercise in futility." RIAA I, 29 F. Supp.2d at 632. SCMS would not alter the Rio's ability to reproduce such transmissions, just as it would not alter the Rio's ability to reproduce digital music recordings uploaded to a computer hard drive.
For the foregoing reasons, the Rio is not a digital audio recording device subject to the restrictions of the Audio Home Recording Act of 1992. The district court properly denied the motion for a preliminary injunction against the Rio's manufacture and distribution. Having so determined, we need not consider whether the balance of hardships or the possibility of irreparable harm supports injunctive relief.