Wikipedia talk:Sound/Archive2

Length of Excerpts of Copyrighted Sound

I posit that for a three to four minute song, twenty to forty seconds is ideal, more than a minute too long by far. A comparable ratio should be maintained, so an approximately ten to twenty second clip would be good for a 2 to three minute song and a minute to a minute ten or twenty would be good for a song at least six minutes in length. This is my opinion purely with regards to the copyright legality of the clips, not in terms of their suitability for Wikipedia more generally. Tuf-Kat 06:09, Feb 9, 2004 (UTC)

I wouldn't be comfortable uploading sample of more than 20sec for any kind of song. One of the Jazz songs is over 13 minutes, but it sounds pretty repetitive to me, so if we were to include a 2 - 3 min sample of that, it would most likely be infringing. It would be nice to have a more expertly opinion on this. I will let Alex know about this talk page and see if he can give us any hints. Dori | Talk 01:38, Feb 10, 2004 (UTC)

I don't think there are any quick and fast rules about fair use of sound clips. Use of single samples of sounds have found to be infringing when the sound is a result of a distinctive performance whereas having a radio playing in the background of documentary film footage might be allowable if there is a considerable excerpt but the foreground was some kind of significant activity that could not be separated from the background. I think Dori's limit of 20 sec. is more likely to be considered reasonable as the idea is the sound clip is being used for informational purposes, 20 sec. gives one an idea of the sound, recognizing the melody, etc., but it is a limited copy that could not really be copied again for any other purposes that would fall outside of fair use. The longer the clip (irrespective of the length of the original) would give a downstream copyist opportunity to do more potentially infringing things ith the clip. Of course this is also related to the resolution of the sample, i.e. lower resolution sampling is less likely to be reused. — Alex756 [http://en.wikipedia.org/wiki/User_talk:Alex756 talk] 08:31, 10 Feb 2004 (UTC)

Tuf-kat, I use Audacity as well (available for several platforms including Windows) to edit the sound clips and save them as Wave files. I then used CDex [1] to batch-encode the files into the Ogg Vorbis format. Dori | Talk 21:34, Feb 17, 2004 (UTC)

I just want to note that I really think 20 sec is not a good timeframe to aim for. Wal-Mart.com has thirty second clips from every single song on this album, including two songs for which that is almost half of the entire recording. If this can be fair use for them, I don't see why we don't have at least as much latitude (that is, of course, assuming Walmart doesn't have special permission from every artist in their database, which seems unlikely). Tuf-Kat 01:16, Feb 19, 2004 (UTC)

It may not be fair use though. They might have an agreement. Also, they've got the lawyers to test the waters, we don't, so I think we should be cautious. I agree that 20 sec is not enough sometimes, but this is supposed to be just a sample. If people want to hear more, they have to buy the album or the song, it's that simple. Dori | Talk 01:50, Feb 19, 2004 (UTC)

According to fair use, audio samples are never fair use, regardless of length:

One of the few cases where this factor is irrelevant is in sampling a piece of a copyrighted sound recording. If no permission is obtained to use a sample, then no matter how small the sample, an infringement has been committed.

This, of course, is contrary to intuition (you'd think samples would be promotive as long as the artist is given credit), but if it's law then it has to be followed. Tuf-Kat noted that the article doesn't include any citations to back up the assertion, so perhaps someone should look into it. - Fredrik 08:31, 10 Mar 2004 (UTC)

What plays Ogg Vorbis?

I was dismayed to find that Wikipedia wants sound files in "ogg" form. It's all very nice to have an open-source tool to convert into ogg, but what browsers play ogg natively? For those that don't, are there plugins that can be installed? This strikes me as important information for the use of sound in Wikipedia. After, ideally, there should be orders of magnitude more "readers" than "contributors", so it should be as easy as possible to listen to such sound resources. Is my use of Opera (and my MSIEphobia) blinding me to some common knowledge here? ☺ -- Jeff Q 14:09, 30 May 2004 (UTC)

I have been led to believe that ogg vorbis can be read by most any music program. The only one I can actually vouch for is iTunes, however. Tuf-Kat 19:40, May 30, 2004 (UTC)

We don't seem to have the same purposes in mind. I was under the impression that sound clips might be played through one's browser. (The concern that brought me here was the desire to find a recording of someone correctly pronouncing "Abu Ghraib" and make it available on appropriate pages.) You appear to feel it's sufficient if Wikipedia sound clips can be downloaded into a music program.

How did the Wikipedia community decide for what purposes sound files would be used? Was there a discussion on what should be uploaded and what format? Can such a discussion be opened or reopened? Inquiring minds (well, one at least) want to know. ☺ -- Jeff Q 22:30, 30 May 2004 (UTC)

Oh, I don't know. I recommend asking at the Village Pump. I think the reason for insisting on ogg vorbis is because it is freely distributable, and Wikipedia content can not be encumbered by patents or it risks violating the GNU FDL license. I'm not sure if it was ever discussed or not, but someone at the VP will probably know. Tuf-Kat 05:49, May 31, 2004 (UTC)

Using sound on Wikipedia

In response to my efforts to determine exactly how to pronounce "Abu Ghraib", someone suggested uploading a sound clip to Wikipedia, much as one can upload images for articles. I found, however, that there seems to have been little discussion on the uses of sound on Wikipedia. The Special:Upload page, which discusses images at length and provides useful links, only makes a passing reference to sound files, provides no link, and gives an example that seems to require the use of "OGG" files, which are unquestionably much less well-known than either WAV or MP3 files and are not supported natively by browsers.

After some digging, I found the article Wikipedia:Sound and Wikipedia, which calls itself "a place to discuss issues surrounding using sound in Wikipedia". It seems, though, to have the exclusive POV of a community of users who merely want to download music into portable players (possibly just iPods!). I thought a much more logical use of sound in Wikipedia was to enhance articles (for pronunciations and obvious sound-related issues, not the ghastly practice of forcibly "entertaining" Website visitors!). When I asked why OGG and why not something more browsable, Tuf-Kat suggested I bring it up here in the Village pump.

So I put it to you, the Wikipedia community:

• Should it be possible to add sound files to provide useful content for Wikipedia articles? Should they just be stuff to download for player use? Are there other uses besides these two that should be considered?

• What format or formats should be accepted/required for sound files? OGG's primary advantage in being unencumbered by licensing issues, apparently. WAV, MP3, and other formats are much more likely to be useful for augmenting browsable pages, but may have restrictions. (Actually, it's not clear to me how format licensing issues would be a problem, since the ability to use a file format depends on the browser or other application, which must be licensed by the reader, not the repository. But I'm hardly an authority on licensing issues. ☺)

• Shouldn't there be a more specific place to have this discussion? The logical one, Wikipedia:Sound and Wikipedia, is not currently as general as it claims. Should it be generalized? Should another article — say, [[Special:Upload sound files]] or something like it — be created and linked with Special:Upload?

(By the way, please don't post your responses to these questions inside this single multi-paragraph posting. This is very confusing. Please create your own contiguous posting and excerpt mine as needed.) I invite your opinions and suggestions. -- Jeff Q 12:02, 31 May 2004 (UTC)

See Wikipedia:sound help for a little more discussion.

The specific case for Vorbis sound hasn't been made in detail anywhere; it's something that happened about 100 Wikipedia years ago :) Firstly, there's nothing in itself illegal about using MP3, WMA, or other encumbered codecs to post otherwise-legal material on Wikipedia. WAV (as well as FLAC, a lossless compressed audio format) are unencumbered. The trouble with using the encumbered formats is that they limit accessibility to the format (as a Linux user, it's very difficult for me to play WMA and AAC audio, for instance, and most Linux distributions refuse to ship MP3 encoders because of patent issues). Ogg, however, requires, at most, one free download for support on any platform. Uncompressed wav files are, indeed, more accessible than Ogg, as native support for them is available, without any extra download, on just about every platform in the world. However, their size makes them horribly inconvenient for modem users. Finally, it's more convenient for everyone to minimize the number of different file formats for content used on Wikipedia. Therefore, Ogg Vorbis remains the only sensible choice as far as I'm concerned.

As far as the uses are concerned, if a sound file adds useful information to an article, and doesn't pose any copyright problems (because it's available under the GFDL, is in the public domain, or is fair use), then it can be added. A good example is on V1 flying bomb, where there is a recording of the V1's noise.

WRT the music clips uploaded in the past, they are all, to my knowledge, short excerpts from songs rather than complete ones, and I think are regarded as fair use. --Robert Merkel 04:31, 1 Jun 2004 (UTC)

As an American monolinguist, I often have trouble pronouncing the names of subjects from another language (e.g. Mohammed Mossadegh, Abu Ghraib, José María Aznar, etc). It would be helpful for me (and, I suspect, many other English-only speakers) to have a small ogg or wav or something of the pronounciation of foreign names. Currently, this would be relatively difficult. You would have to save a sound file of you voice, download an ogg converter (since most people don't have one), convert the file to ogg, upload it, and link it as a pronunciation. Should this be encouraged? Is there any to make it easier for someone to do? Quadell (talk) 15:35, Jun 2, 2004 (UTC)

Robert Merkel's rationalization of the OGG format neatly demonstrates the nature of the problem here. His entire argument assumes without ever explicitly acknowledging it that the main purpose for providing sound clips on Wikipedia is to download music into players. I challenge that assumption. I suggest that the main purpose for sound on Wikipedia should be to augment articles, and that music download is necessarily a niche topic within the Wikipedia community. For that reason, I feel that the Wikipedia:Sound and Wikipedia article is inappropriately narrow and doesn't even admit to it. Quadell's comment indicates that I am not alone in wanting some reasonable sound support from Wikipedia, at least for pronunciation guides, and I'm sure that's only the tiniest possibility of usefully sound-enhanced articles.

I also suggest that OGG is a far from obvious choice, even for player download. Everyone has heard of WAV and MP3, but OGG, while perhaps popular among the Linux crowd, is not terribly well-known. Yes, Wikipedia:sound help describes how to play OGG files, but not a word about how to play them in browsers, the very tool required to use Wikipedia. Is there any browser that supports OGG natively, or even with a plugin? Even with external applications, OGG surely rates much lower on the popularity list than WAV, MP3, WMA, AIF[F] (Mac/Unix), and AU (Unix).

Finally, the argument against multiple formats is rather meager. Wikipedia has no problems supporting more than one image format. The real issue is what does one wish to do with Wiki sound, and what formats are needed to accomplish this. For my own admittedly niche application of pronunciation, WAV is the obvious choice, since all browsers support it, and its size is inconsequential for 1-2 second clips. Size only becomes an issue for much longer clips, where the excessive size of WAV justifies the trouble one must go through, both to encode and to play the files. Offhand, the only situation I can think of that should warrant trouble is music files.

I suggest we give the music downloaders a special area to do their thing and reclaim the topic of "Sound and Wikipedia" for true general use, with WAV files being acceptable for uploads and attachment to articles, at least to begin with. -- Jeff Q 07:03, 3 Jun 2004 (UTC)

Popularity is not an issue here, it is simply wrong to store something in WMA here because it is a proprietery format and directly interferes with out goal of creating a _free_ encyclopedia. Ogg is a free format that is not limited by patents and should therefor be used. --195.8.26.136 17:51, 3 Jun 2004 (UTC)

Jeff, you seem to be reading things into my comments that I never actually said. If you want to use sound for word pronounciations, a three-headed booby's mating call, or any other encyclopedic purpose, I'm sure Wikipedians will approve - I certainly do. Whether that ends up making the "majority" of sound use on the Wikipedia or not is an open question. If you want to use WAV files for short clips (and I mean probably a maximum file size of 50KB or so), I don't see a problem with that either. What I do have a strong objection to is the use of patent-encumbered media formats such as WMA or MP3 on Wikipedia for any purpose where there is a suitable alternative available, or the use of large WAV files making it unrealistic for modem users to download content. --Robert Merkel 15:18, 5 Jun 2004 (UTC)

Heated agreement, indeed! (Reminds me of my childhood, when my father and I would have lively, hours-long arguments during which we would find ourselves unexpectedly trading positions.) Wikipedia should allow whatever legal use is desired by its community, and the "majority" will no doubt shift over time and changing use, no matter what the subject. My real concern is that, in reading this article, one is easily discouraged from anything but OGG file use, which effectively rules out browser-based sound. The real solution is for me to stop complaining and to attempt to revise the article to make it more WAV friendly, including guidelines on what not to upload (like large clips) — after I figure out how to upload WAV files despite the lack of guidelines. (I suspect it won't be hard, but I haven't tried it yet.) I'll make one more gripe below in response to Raul654's posting and then shut up until I can report progress. -- Jeff Q 16:52, 5 Jun 2004 (UTC)

The one thing I don't understand is why Ogg poses such an accessibility problem for browsers. If your browser is set up correctly, you click on the sound file, it starts the helper app (XMMS in my case) and the sound plays. Are you telling me that Internet Explorer won't do the same? --Robert Merkel 00:47, 6 Jun 2004 (UTC)

Keywords being 'if your browser is set up correctly' - by default, IE/Windows do not support ogg (that's 90%+ of the internet population out there). Some version of winamp support it, while others do not and require a plug in. →Raul654 00:49, Jun 6, 2004 (UTC)

Actually, finding and installing a plugin is not the biggest problem, though it's certainly a deal-killer for many non-techies. (Wikipedia is for the general Internet population, right?) The real problem is having to put up with an external application with a control panel that compares favorably with a 767 cockpit simply to hear a 1- or 2-second sound bite. Considering that all browsers play WAV files without forcing their users to descend into geekdom (of which I am a card-carrying member, by the way — and my browser is Opera, not MSIEry), until browsers become sufficiently user-friendly for multimedia purposes (about another 5-10 years, I'd guess), WAV is the only practical way to go for this specific use. — Jeff Q 01:42, 6 Jun 2004 (UTC)

Somebody did that sorta thing with Poutine. Quite humorously (not that humour is necessarily helpful in this situation).

I think it's a great idea to add a pronunciation bit to exotic names. Unless you are into news, chances is the accepted pronunciation of many foreign names will always be a mystery to you. And dictionaries are no help with modern proper names. So, having sounds for such words would increase the informativeness of the article. --Menchi 07:19, 3 Jun 2004 (UTC)

Without being egotistical, I think I'm probably one of the most experience users when it comes to sounds on wikipedia (last time I checked, I have 4 or 5 of the top 10 largest files uploaded). To answer your question about format - mp3 would be the ideal choice for its support, but it is patent encumbereed. The other formats (real, quicktime, microsoft) are proprietary and encumbered. That leaves ogg. Ogg is supported in the newest version of winamp, and you can get free programs that convert it to other formats (I *STRONGLY* recommend Audacity). →Raul654 18:00, Jun 3, 2004 (UTC)

I would like to re-iterate the call for using WAV for short clips, and Ogg for large files. Pronunciation and the sound made by an animal or object are the most common uses of sound in an encyclopedia that I can think of, and they require only short clips. Wikipedia aims to be browser-accessible and, as I understand, WAV is the most widely browser-supported of the formats that do not pose any licensing problems for encoding and decoding sound. Ogg is supported by some players, but not natively by browsers. If Ogg becomes widely supported, the WAV files can easily be automatically converted to Ogg. Pgan002 04:29, 4 Jun 2004 (UTC)

Perhaps Raul654 is the most experienced sound user on Wikipedia, but having "4 or 5 of the [redundantly] top 10 largest files upload" is not evidence of it. (I can beat that in a few minutes by uploading four audiobook chapters, each of whose content is 2 or more hours. If I do it using WAV, it'll be even more pointlessly impressive.) Even if Raul654 is the most experienced, that doesn't mean his experience is a good measure of general use of Sound and Wikipedia. It merely means that he's completely focused on incredibly large files — music files, I'd guess. This argument not only ignores the uses I've described, it actually dismisses them because the files are so small. (I guess among the existing crowd, size does matter. Sorry — I couldn't resist. ☺)

As I mentioned a few paragraphs up, though, continued argument on this subject is much less useful than a concrete attempt to make Wikipedia friendly to in-browser sound. Since I'm the loudmouth on this subject, it's incumbent on me to do something about it. I'll see if I can't upload some useful stuff, link it into appropriate pages, and then update both this article and Special:upload to let others know how to do it. -- Jeff Q 17:12, 5 Jun 2004 (UTC)

Jeff, I hate to break it to you, but there's a 2 meg upload limit per file. However, only one or two files even come close to that. →Raul654 17:59, Jun 5, 2004 (UTC)

Hey, no worries — I have no intention of trying something so absurd. Your position is secure. ☺ Seriously, I've already run into a problem. I have only two immediate reasons for in-browser sound uploads: pronunciation of "Abu Ghraib" (which I don't have) and a sample of the sound of Brood X cicadas, which I won't have until the unseasonably cool weather in Northern Virginia recedes. Anyone else have any GFDL-eligible WAVs for useful augmentation of Wikipedia articles that they'd volunteer to let me create & test a WAV upload process? -- Jeff Q 18:38, 5 Jun 2004 (UTC)

.WAV and .MP3 are good to use - don't use Ogg

I think Wiki's attempting to convert the world to Ogg is doomed to silliness, and that at least the two common sound formats, .WAV and .MP3, are good to use.

My reasoning is as follows:

• Ogg is not supported 'out of the box' by Microsoft Windows Media Player, Apple Quicktime, nor RealPlayer. So OGG enabled users are rare, and very few will be able to use .OGG files!
• If we start worrying about potential legal issues around licence claims over WAV or MP3 formats, then you also need to ban the use of image formats such as GIF, JPG and just about everything else. There are probably a few pending on PNG - if not I could go start one tomorrow
• I think the Unisys claim on GIF is the best example. Some years ago Unisys found that they owned the GIF license (someone literally found it in a drawer), and started enforcing it. However, they allow free distribution and viewing of GIF files. They only claim royalties from software that actually makes GIF files. So if you write something like Adobe Photoshop, you need to pay Unisys royalties. When an end user buys Photoshop, part of the money goes to Unisys for a GIF license. If you downlaod IE, Unisys allows you to view GIF files for free. Similarly, if anyone ever was to gain ownership of .WAV or .MP3, it is most likely that it will be Microsoft, Apple and RealMedia that will burden the initial cost. End users will have any required .WAV or .MP3 licences as part of the software they run.

--Zarni02 10:55, 19 Jun 2004 (UTC)

I disagree. We should encourage movement towards formats that are as free as possible. For the moment, I suggest downloading audio samples in both MP3 and OGG. No-on will support OGG without OGG content, and OGG content is unusable without OGG players; we can kick-start the process by making sure that OGG content is available on Wikipedia alongside proprietary formats.

By the way, your understanding of the Unisys patent issue is out of date, unless you live in Canada: see GIF for more information. -- Anon.

The best solution to the problem, would be to provide several versions of each sound. This should be done server-side, so that the copyright holder / uploader only needs to provide one version of the sound (suggested: WAV or OGG).

What you say about Unisys and GIF, Zarni02, is not a good example. Remember that Mediawiki is free software, and it seems unlikely today that it would ever be possible for the Foundation to pay license fees for each copy of Mediawiki. Also, free sound editing programs are often unable to encode MP3 (because MP3 is already patent encumbered, by Frauenhofer). WAV is not as likely to be patent encumbered, but when you take note that Microsoft tries to get license fees from FAT16 and FAT32....

Also, you cannot know what license requirements a hypothetical IP owner would make. They could very well choose to go after individual users, like SCO has threatened to do with large companies.

-- David Remahl 11:23, 19 Jun 2004 (UTC)

The argument about MP3 encoding is also specious, as MediaWiki neither encodes nor decodes MP3. All it does it store a bunch of bits. The original file creator must have the rights to encode it, and file players must have the rights to decode it if they want to listen to the content. The process of uploading and downloading such files doesn't even require a license, so contributors and readers without MP3 licenses (however provided) aren't even in violation of MP3 rights just by posting and fetching the files. Of course, if the material that is posted or fetched (like a copyrighted recording) has its own content license issues, that's a separate matter, and OGG format offers zero protection from that encumberance. -- Jeff Q 16:05, 19 Jun 2004 (UTC)

What use is uploading and downloading MP3 files if you can neither create nor use them? - Fredrik (talk) 16:16, 19 Jun 2004 (UTC)

I was responding to David Remahl's implication that "the Foundation" would have to pay license fees for MediaWiki if it used an encumbered format like MP3. This is a specious argument, because MediaWiki does not in any way use these formats; it simply provides bits. People encoding and decoding these encumbered bits are the ones obligated to pay licensing fees (if any), just as they are obligated to pay for the computer hardware they use to access the bits. (Or are people advocating the "liberation" of computer equipment, too, to make information free? ☺)

The question is not so much whether there is a licensing fee attached at some point to any particular format; it's whether the content is legally and practically usable by a significant portion of the Wiki readership. Requiring the use of OGG turns Wikipedia sound into a techie members-only club, in which the secret handshake is the ability to purchase hardware and/or install OGG-based software — none of which exists for in-line augmentation of browsers, which are used for nearly all Wiki access — that can make use of OGG. Both WAV and MP3 are in common use, and the onus is on the user (and the content contributors) to ensure that licenses are paid. WAV is even supported inline by all browsers. (MP3 is more problematic, as I've seen no players that operate seamlessly within any browsers. They all seem to require plugins that pop up their own external applications and force their absurdly complex user interfaces on you, rather than provide a simple interface worthy of a browser supplement. But that's a whole 'nother issue.)

MediaWiki and its user systems are not involved in MP3 licensing issues. That argument demonstrates a confusion between software licensing and content licensing, the latter of which MediaWiki is involved in regardless of the encoding format. For that, we have the GFDL and the "fair use" doctrine, which must be scrupulously observed to protect MediaWiki-based information networks. -- Jeff Q 05:09, 20 Jun 2004 (UTC)

Aargh - why can't everyone agree with me! Seems to me David Remahl's suggestion to provide multiple formats will:

• keep the average user happy - they can use the .wav or .mp3 files
• keep the .ogg people happy
• will not be too onerous on the contributor

I now think that each sound file should be provided in .wav and/or .mp3 format, as well as .ogg. It would be great if MediaWiki could convert files automatically, but I suggest this is tricky and is more likely to run into license issues. So I think this should be the contibutor's job.

Also, thanks for the update on the GIF issue. I had not realised the Unisys GIF patents had already expired in the US and are due to expire Worldwide in July 2004. --Zarni02 03:23, 20 Jun 2004 (UTC)

My objection to OGG, besides the technical ones discussed above, is that there is an existing crowd of Wikipedia sound users who seem bound and determined to discourage anything but OGG. This is manifested by the utter lack of information or support for uploading anything but OGG files, and the vociferous arguments (some meaningful, some specious) against anything but OGG. Actually, I suspect that Wikipedia will support any sound format, as long as people wish to upload such files.

I have wanted to test this theory by uploading one of two WAV files to augment two articles, but have had practical problems getting my hands on either. As soon as I have the means to test it, I expect to do so, after which I will post the results and modify any pages necessary (like Wikipedia:Sound and Wikipedia and Special:Upload) to make Wikipedia more friendly to other sound formats. I even volunteer to do this with someone else's sound files, but only if the sound clips add meaningful content to some Wikipedia article, like the pronunciation of "Abu Ghraib" or the sound of a periodical cicada (my two projects). I have fully-licensed conversion software for nearly all formats currently in use on the Internet — except OGG — so I can convert any legally-usable sound clip anyone wants to volunteer. -- Jeff Q 05:37, 20 Jun 2004 (UTC)

I have created a page, Noisy Miner, that has a sound .wav file and a .ogg file. The .wav file works just fine. Clicking the .wav file in my IE browser plays the file. The .ogg file does not play in IE, but I guess you could get it working by configuring a mime type or something?? Mozilla allows you to choose a player for a file type, so it was pretty easy to configure it to play the .ogg file with Winamp, although I am not sure most users could do this without reading all the Ogg stuff. --Zarni02 06:57, 20 Jun 2004 (UTC)

If people are happy with the way of doing things as on the Noisy Miner page, I propose to change the Wiki Project Page to relax the 'Ogg only' policy to be a .wav/.mp3 and .ogg policy. --Zarni02 06:57, 20 Jun 2004 (UTC)

I have asked for comments on the Wikipedia_talk:Requests_for_comment page. If there is no strong dissent I intend making these policy changes --Zarni02 04:52, 5 Jul 2004 (UTC)

Replaced by Wikipedia:Sound

As per the discussion above and since there was no dissent after letting sit on the request for comment page, have created a new page: Wikipedia:Sound that redefines the Wikipedia sound policy. The list of Ogg Vorbis samples I put on the Vorbis page. --Zarni02 09:09, 16 Jul 2004 (UTC)