I think CC could have a web page with collected wisdom about audio formats. While there are a lot of pages about formats, ours would be from the point of view of on-line collaboration. No doubt there are other topics we could treat the same. The Sultan's Tent seems like the perfect place for this type of conversation. Here are a set of point I believe and think would be helpful. I encourage everyone to add their wisdom and correct my mistakes or misjudgements. Later, I'll be happy to edit the consensus opionions into a web page, suitable for posting on CC. I'd like to assume a basic background of obviousness for DAW users, but would summarize the main points of each format to establish a baseline. WAV (.wav) files 1) used in DAWs, therefore the common denominator. 2) Mono and Stereo. Stereo takes twice the space as mono for the same quality. 3) Different sample frequencies a) CDs are 44.1 kHz, so most work will be this speed when done. b) (notes on other speeds, like for video work, etc.) c) While theorists claim that anything over 44.1 kHz is wasted, this depends on ideal pre-A/D and post-D/A filters. 96 kHz allows sub-ideal filters to produce best results, and may reduce artifacts introduced by processing. On the other hand, 96 kHz more than doubles the demand on processing resources, for subtle differences that are debated by the experts. Not commonly used except on the more demanding professional work. Large file size inhibits collaboration. 4) Different bit depths. 8 bits is insufficient for DAW work. 16 bits is the target for CDs, and is supported by all usable gear. 24 bits has a lot of advantages over 16 bits: a) Increased headroom during recording, making level setting easier. b) Increased tolerance to processing. If a lot of processing of audio tracks is required, the quantization errors introduced in 16-bit recordings mount up. While noise introduced by these errors are as low as 96 dB below the max signal to start with, several passes though EQ and FX can multiply them and degrade audio quality. With 24 bits, these quantization errors still multiply but are so low to start with that much more processing can be done before the resulting noise affects the quality. Also, it only increases .wav file size by 50%, and in many DAWs doesn't affect the CPU load very much. On the other hand, many collaborators don't have 24-bit capability, and some don't have recording gear that delivers better than 96 dB S/N ratio anyway. In this case, the advantages are debatable (but not dismissable!) Most collaborators, especially amateurs, use 16 bits for most work, because of the reasonably high sound quality, universal compatibility, and reduced upload/download times. Pros generally use 24 bits. MP3 Two popular encoders, Blade and Lame, both freely available. Blade produces very poor results at low bit rates (128k or below?), but supports higher rates (224, 256, 320 kHz). Lame produces much better results at lower rates, but maxes out at 192 kHz. I don't know how Blade and Lame compare at 192, or how the higher Blade rates sound -- I avoid Blade in general. Any opionions? Lame supports Variable Bit Rate, which dramatically reduces space for files that contain lots of silence -- and this happens a lot when collaborating. MP3 also supports "joint stereo". For a lot of music (especially before much processing has been done), the left and right channels are very similar. Joint Stereo takes advantage of this and, in theory, produces higher quality stereo sound at the same bit rate. Any opinions from the golden ears crowd? MP3 is the linga franca on the Internet, e.g., on MP3.com. For this reason, it may be the best for posting to the general public. However, MP3 has one problem for collaborators: it adds about 25 msec of silence to the beginning of the file (after reconverting to WAV). WMA (Windows Media Audio) Freely available. Microsoft product, may not be supported on non-Windows systems. The majority of folks feel WMA sounds better at the same bit rate. I feel strongly that this is the case at low bit rates. For example, I find 64K WMA adequate for "walkman" type use, or even home stereo when I'm not paying attention. There are noticeable artifacts but not enough to ruin my experience. (I'm not an audiophile.) At 64K, I find MP3 annoying. I haven't done careful comparisons at 160 kBps and higher, and I'm not sure my ears are well enough trained or undamaged enough to be critical, but I tend to use WMA out of faith, because it sounds so much better at lower bit rates. Does not cause sync problems. APE (Monkey's Audio) The APE compression is a totally lossless process, meaning that a WAV file reduced to APE and expanded again will duplicate sample-by-sample and bit-by-bit the original WAV file. It compresses 16-bit mixdown files to 30-40% of their original size, and single-instrument recordings even more (depending on the complexity of the sound of the instrument). There's really no good reason not to use APE files to exchange WAV files. For both MP3 and WMA, a mono file has twice the quality of a stereo file at the same bitrate. The files are the same size. Therefore, below, I'll use "64S/32M" to mean 64K stereo or 32K mono, where the quality is the same. Appropriate uses: For the following, I'll assume roughly the same quality for MP3 and WMA files when the bit rate is over 128K. When I talk about posting a part, I mean posting a recording of a single instrument. 64S/32M WMA Good for posting parts or portions of songs where the intent is to convey the idea, and sound quality is not important. OK for a stereo quick mix, so folks can say whether they like where the song is going and very general comments about the mix (e.g., guitar is too loud). Not good enough for any kind of critique of imaging, although a terrible image will be obvious, and a good mix will usually image nicely -- but not necessarily the way it was intended. This is especially true if phase/delay queues are used in the mix. If the mix is muddy, it won't always be easy to tell why. 160S/80M WMA, or 192S/96M MP3 Good enough for posting a part on a song where there won't be a lot of processing and the sound quality demands for the project are not high.