Misplaced Pages

Talk:Musepack

Article snapshot taken from Wikipedia with creative commons attribution-sharealike license. Give it a read and then ask your questions in the chat. We can research this topic together.

This is an old revision of this page, as edited by Rhobite (talk | contribs) at 15:38, 24 June 2005 (→Listening tests: self-selected). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

Revision as of 15:38, 24 June 2005 by Rhobite (talk | contribs) (→Listening tests: self-selected)(diff) ← Previous revision | Latest revision (diff) | Newer revision → (diff)

Listening tests

Let's talk about a word: "conclude". It's a big word. Statistical tests don't give us proof of anything, but they can suggest that conclusions are likely. In this case, we don't have enough data to make much of a valid conclusion. You can't sample 20 self-selected people in various listening environments and prove that one codec is better than another. These listening tests do suggest that MPC is one of the better codecs at 128 kbit/s. A 128 kbit/s test doesn't allow us to conclude anything about higher bitrates, or more recent versions of the codecs. While the ff123 results should be included, we need to note that 2001-2002 was a long time ago. Vorbis, AAC, and even MP3 have improved in quality since then. RJamorim's tests are more recent, but they still involve self-selected participants and small sample sizes. Due to the lack of random sampling we can't make any conclusions about what these data say about the general population of listeners. Rhobite 17:06, Mar 14, 2005 (UTC)

it wasn't 20 self-selected people, it was 20(+) people that volunteered. they did not just choose the people that rated MPC favorably.Plonk420 09:36, 24 Jun 2005 (UTC)
That is what "self-selected" means. They selected themselves to participate in the listening tests. It doesn't mean that they are biased, it just means that they do not represent a random sampling of a population. Rhobite 15:38, Jun 24, 2005 (UTC)

You started using the word "propaganda." I edited this page and you changed it again. You can stop trolling now. The following remark and the next one remain, as well as the page's structure: Linking to Slashdot, which is a mere news site and has nothing to do with audio formats, is propaganda. Those are objective tests done by means of ABX and ABC/HR (which you can do a search for and learn about) and the Musepack.net site is the *official* Musepack site. The link you presented is one amongst several other important tests. Omitting the others twists the facts. I will include all the links on this page instead, even though I tried to avoid this bloat by linking to the site. If you're annoyed by the fact that Musepack suffers from less problem cases than Vorbis, that's tough. I can assure you that Monty knows well about the limitations of Vorbis (and all its current forks) with transients. You can ask him on IRC or by email yourself. Vorbis 1 was released early in order to have a final version out as soon as possible. Vorbis 2 will include rewritten transient handling code. comment by User:80.74.124.55.

Actually, you were the first one to use the word "propaganda". You said that linking to Slashdot was propaganda, and claimed that since the listening tests used "ABC/HR" they are irrefutable. While I agree that blind tests are much more useful than informal ones, as I've pointed out here and in the article, they were unscientific, small sample-size tests. The tests do not support conclusions such as "MPC is the best available lossy codec". The best we can do is report the conditions and actual results of the test (sample sizes, etc), which you decided to remove from the article. I DID include all the links, even the obsolete ff123 test. Please don't accuse me of trolling. I have no bias here. The audiophiles will be shocked to learn that I've never used MPC, and personally my ears are content with LAME -V5. I'm trying to report facts objectively, and certain users appear to want to whitewash certain codecs, or remove information from this article. Rhobite 23:32, Mar 14, 2005 (UTC)
Are you even aware that those who designed the first versions of world-class codecs (such as MP3, Ogg Vorbis and aacPlus V2) never did "ABR/HR" tests before releasing them? You know why? Because they simply used their ears while designing, listened carefully, combined that with graphic analysis on I/O of several samples, tuning it till they were reasonably satisfied, and repeated that process a couple of times. 195.64.95.116 18:12, 23 Mar 2005 (UTC)
Um, they might not have used "ABC/HR" exclusively, but psychoacoustic models of codecs are tuned with double-blind testing. Graphic analysis may be useful in diagnosing problems, but it's useless in judging the effectiveness of a codec.

Patents?

When discussing the patent thing, please provide evidence. It's not helpful when all these enthusiasts drive by the article and simply remove things they don't like. Rhobite 04:41, Apr 4, 2005 (UTC)

It's the same case as Vorbis, nobody can really prove it's patentless unless some court lawyers comes and do a deep check. So until this happens, as vorbis, musepack is believed to be patentless.
Well, er, no: AOL's patent lawyers pored over Vorbis' claims before allowing Nullsoft's Winamp to be bundled (still to this very day) with a Vorbis codec, since any existing patent claim would have made AOL a juicy target for a lawsuit & damages for unlawfully distributing patented code w/o a license. That's why an OGG-Vorbis codec is fearlessly bundled with most audio software, and why it is actually used in the real world in "MP3/OGG players", such as most of Samsung's , ISM's , iRiver's, and more -- because AOL's mighty patent lawyers' seal of approval showed it safe enough for anyone to use w/o paying Fraunhofer. And that's why you still don't have MPC bundled with WinAmp, nor used in the real world. MPC needs to try and push an official codec into WinAmp in order to have the "AOL effect" work for them too. Until that, MPC won't be considered patentless by investors, while Vorbis is. -- #6
Well, to be considered patentless is still different from *being* patentless. Anyway, the current article (2005-04-15) does explain the patent issue. It's now up to the reader to make his own opinion
"Well, er, no: AOL's patent lawyers pored over Vorbis' claims"
Do you have any proof on that claim? I only heard about AOL looking for Vorbis patents in unverified news sources, like the Xiph mailing lists and slashdot (IIRC) rjamorim
Quickly, from authoritative sources (Xiph, and Vorbis founder Monty):
-- #6 20:33, 28 Apr 2005 (UTC)
Nope, these aren't good sources. They are actually very biased sources. I have a hard time trying to trust Monty when he claims Vorbis is patent-free. Did you expect him to ever admit it is patented?
Authoritative, for me, would be AOL or Xiph releasing their legal documents stating yes, vorbis is free of patents. Everything else is hearsay and FUD in itself.
rjamorim

Hard Truths

If MPC is an improvement on MP3 remains to be seen. Listen to these: http://www.xs4all.nl/~jult/ogg/ Especially noteworthy are the first 14 seconds of that sample.. It's very clear that it's not much better, in fact, MPC today is still not really improving on LAME MP3 in my testing. It does really bad things with the audio, actually. So please, don't believe the hype, and don't just put lies on pages in wikipedia..

Is that your site? That's cute, you think ABX testing is worthless. Rhobite 19:52, Mar 11, 2005 (UTC)
My hearing is better than yours, obviously. Throwing in a defense-shield by using words like "abc" "hr" or "abx" means you are incapable of using your ears when dealing with sound.
Yes, your hearing probably is better than mine. Is that some kind of insult? I'm not an irrational tube amp-worshipping audiophile, I just want to keep your obvious bias out of encyclopedia articles. Rhobite 02:42, Mar 12, 2005 (UTC)
My obvious bias? Are you THAT deaf, that you can't hear the crap MPC put in that given example? Really, if you still dare write that it sounds "better than " with that sample, you are insane. Nobody needs a test for that. You simply LISTEN and you'll hear that the MPC file contains loads of strange EXTRA sounds that aren't in the original! Geez, you people write all sorts of long-winding nonsense about the validity of "blind listening tests" while NEVER having heard the sample once. In case you forgot: It is ABOUT AUDIO quality, in order to make judgements on that, you'll USE YOUR EARS. By the way, why would I be biased? I have no financial gain in promoting ANY codec, in case you wondered. 195.64.95.116 18:04, 23 Mar 2005 (UTC)
Please see Misplaced Pages:No personal attacks. Thank you. Rhobite 00:20, Mar 24, 2005 (UTC)
I deleted this because it is totally irrelevant, and frankly, I don't see where the personal attacks here aren't appropriate. Discussing your precious "personal attacks" does not apply to some page about MPC. Really, nobody cares anyway, MPC is a stupid in-between format which will never grow big. JuliusThyssen 12:50, 8 Jun 2005 (UTC)
Civility is required of Misplaced Pages contributors. In no case is a personal attack appropriate. If you're not able to conduct yourself here without resorting to petty name-calling, you will be asked to leave at some point. Rhobite 20:01, Jun 8, 2005 (UTC)
"...Are you THAT deaf, that you can't hear the crap MPC put in that given example? Really, if you still dare write that it sounds "better than " with that sample, you are insane..." I'd just like to point out that as someone with freakish hearing (I can hear things like subsonic elephant rumbles, frequency shifts in cathode ray emissions inside television tubes, and a number of other things that are typically regarded as being outside the range of human hearing), I can clearly hear the superior fidelity of an MPC file when compared to an MP3 file of identical bitrate. There are less compression-based artefacts and generally a more robust bass and crisper treble with MPC files in the 128kbps-192kbps bitrate window.
Sorry, but that's a lie. Learn to use your ears They don't lie, if they're any good. 195.64.95.116 16:03, 19 May 2005 (UTC)
When encoding at bitrates above 192kbps there isn't even a comparison -- MusePack is way out in front of every other audio codec with probably Ogg Vorbis trailing somewhere in second place. But don't take it from me, a lone individual, do a web search for any number of conclusive listening and spectrum analysis tests detailing the superiority of MusePack over MPEG-1/2 Layer III. And, to be fair, MPC shouldn't be regarded as some sort of "holy grail" of audio compression since MP3 and Ogg Vorbis are still viable. However, MPC *is* a successful refinement of what MP3 started in MPEG audio compression standards. And I think that's all people are trying to get across.
  • Hello, Julius. Having fun here at WikiPedia after being banned from most audio-related forums? God speed. rjamorim
    • I have not been banned, I've been censored, quite a difference. As far as I know only from one audio-related forum, you know the one, the hardrocker kid's life-project hydrogenaudio, created by that teenager who claims his ears are miraculously alien, even though he frequented many a heavy metal concert, pre-deafening his allegedly superior hearing, that way assuring us his judgements must be those of a true expert. And that was way AFTER I contributed to wikipedia, smartass. In case you forgot, I'm from 1966 195.64.95.116 16:00, 19 May 2005 (UTC)