Hi all, advice please I'm using lamedrop to produce cbr 192 mp3s from wav files No matter what bitrate I use, lamedrop produces joint stereo mp3s, even at 320 Is this OK, I understood that above 128K stereo should be used, not joint. Thanks
JointStereo is beneficial at any bitrate. LAME incorporetes a very high quality JS algorithm and it should be used ay any bitrate.
Some would argue it depends on what type of music you listen to. From an ogg vorbis fan in his most enlightening article "I Just Finished Compressing 700+ CDs To Ogg Vorbis Format" @ http://papaya.altamente.com/ogg.html : "....no this was different, the high end was definately there, but something else, too - "stereo separation". Now this is something new. MP3 makes some of its best gains through the use of cleverly comparing left and right channels and optimizing where they are very similar. Good in theory, but what you end up with is a lost stereo separation. It's cool for rock/pop, but classical absolutely needs stereo separation. In fact, encode some classical music (any classical music) in mp3, and then in Ogg. [[without using channel-coupling or 'joint-stereo']] You'll never go back." There are some fascinating tidbits in that article. Compact disc storage is so incredibly dirt cheap these days, why not preserve the stereo image intact? If you're just archiving your cds, it costs virtually nothing extra. -- K.A. --
> Some would argue it depends on what type of music you listen to. They would be wrong. Valuable info about joint stereo in Lame: http://www.hydrogenaudio.org/forums/showthread.php?s=&threadid=1081
Hello, kdo, and how are you today? Thank you for the link. I read most of (Dibrom's) long tirade, and I have not the slightest doubt that as the (inventor? original writer?) of the alt-preset specific tunings, he knows what he's talking about. I don't use LAME. If I *were* going to use mp3s in general (I use .ogg instead), I *would* use LAME without hesitation, and nothing else. His defence of his joint-stereo presets seems to border on the fanatical, and I do agree with almost everything he said except for one point: (Item # 5): "Joint stereo IS needed even at bitrates of 320 kbps to achieve the best quality in some critical cases." (what nonsense) It seems to me that in the best-possible-case senario, all that joint-stereo (or 'channel-coupling') could hope to achieve is what independently-processed channels already achieves by default: TWO INDEPENDENT INDIVIDUALY-PROCESSED CHANNELS !! How in Christ can you improve upon that? To suggest that joint stereo "IS needed" to "achieve best quality" is nonsense. Now, having said that, I can tell you that I listen to Rock/Pop music, not classical. With OggVorbis, whether using channel-coupling or not, the resulting music files sound equally fine to me. If you'll notice in my above post, I said "[[some]] people might disagree; 'some people' is not necessarily me. (And it's not necessarily *not* me, either; I just haven't experimented much with classical music), but what the author says in the link I provided above, makes a lot of sense to me. It was in *his* experience, that LAME produced inferior results regarding stereo-image, and I'll take him at his word. Some other stuff for you to think about: With .ogg format, channel-coupling takes place by default for any quality level from '0.00' up through '4.99'. From quality level '5.00' and up, channel coupling is turned off and becomes lossless. (not lossy). There is a *huge* filesize increase between level 4.99 and 5.00 as a result. Normally, an increase of .01 in quality-level is so tiny, that any increase in filesize would be virtually nonexistant, but because 4.99 ---> 5.00 is at the crital point, you can see just how *much* information is being discarded at level 4.99. Is all this lost information audible? I dunno. Also, as I mentioned above, current-day costs for recordable compact discs is so cheap, that I (or you) can easily afford to forego lossy channel-coupling (or 'joint-stereo' or whatever else you want to call it) *however* good or nice or pefect or inaudible it may be. (With .ogg I have firmly settled on using level 7.50 - which for my ears, is so close to the original wave, I doubt my own abilities to discern a difference. Merry Christmas. - KlingonAgent -
Hey! > I don't use LAME Neither do I, generally. I use MPC. :-O > How in Christ can you improve upon that? Improvement is when there is no need for full stereo. If there is a bunch of samples with little stereo separation (mono, in extreme case), why force full stereo on those samples, better use joint stereo and save the bits for quantisation. The "smartness" of Lame alt-presets is tuned enough to detect such situations. So eventually, lame jstereo is more similar to ogg lossless coupling. Not quite the same, but reasonably close. But, of course, ogg is better than mp3. Nobody argues. (Just a good link here: http://www.xiph.org/ogg/vorbis/doc/stereo.html) W.r.t. classical vs. rock. Encoder (any - mp3,ogg) operates at the level where one cannot tell whether the sound belongs to classical, or rock, or whatever. If encoder has some stereo-separation problem, it may show up in any music genre. If algorithm works ok, then it works ok on any genre. This has been discussed so many times before that it is rather banal now. Btw, that guy from your link didn't mention any blind listening tests in his essay. I wouldn't trust his conclusions if they are not backed-up by ABX tests or similar. > I said "[[some]] people might disagree; 'some people' is not necessarily me I had no intention to argue with you or anybody personal. I just didn't like the article: a lot of hand-waving, a few questionable statements and little proof. Cheers /konstantin
Agreed. I was more interested in his colourful wording and caught up in his obvious excitement at having finally found "a codec to grow old with". He compares, as you noticed, his love of music with the love of a woman, which I thought was precious. <gg> I don't think he was trying to prove anything in a technical sense. >> I don't use LAME. > Neither do I. Generally I use MPC. (Shhhhh...., that's 'ok'. I won't tell anyone if you don't; it'll be our little secret). Seriously though, I'm looking at joint-stereo from my own point of view - basically archiving a lot of music cds to cdr, where space-concerns aren't very critical. I can see where files that are going to be made available for streaming or downloading (or p2p file-sharing)on the net - where dial-up modems are still the norm - need to be kept as small as possible. Cheers. -- Mike --
Yes, Joint Stereo is the only way to break the '320kbps barrier' of MP3 and it is recommendable because there are cases where 320SS is not enough. And alt-preset's utilize the newer JS routines by Naoki Shibata which are adjustable and even more sensitive. I would be more than happy to see OGG used by everyone and supported by every program. The problem is that OGG doesn't yet deliver. The optimization process has been soooo long and we are yet to see a final. There are various reports about problem samples where OGG still fails. At the same time Musepack is highly tuned, produces extremely high quality with *stability*. Let's hope that OGG get's there asap.
You're right about that, Lasse, vorbis is s-u-r-e taking it's time getting to that final 1.0 release, which was promised over a year ago. The more I read about it, the more I learn, but right now, it sounds just *super*, is simplicity in itself to use, and I simply adore the free-source-no-patent philosophy behind it, assuming the "has-no-patents" assertion remains intact. It will be very interesting to see where this thing leads..... -- K.A. --
