QUOTE
Mac wrote:
To me, engineering and mathematical approaches lead to things such as judging encode quality by using a frequency biased mathematical comparison of waveforms.
Yeah, but this is not the same thing. A look at how each player processes the audio data will make it clear why foobar2000 has at least a
theoretical advantage over Winamp in terms of sound quality (and notice that Dibrom didn't say it's actually audible).
Winamp does this (I'm not sure if it's 100% correct, but it should get close

):
Decode audio, with bitdepth depending on the input plugin (most only support 16 bits)
==> processing performed by input plugin (like ReplayGain or dithering in the MPC or Vorbis plugins)
==> DSP processing, with bitdepth depending on the DSP plugin (most DSP plugins only support 16 bits, so if your input plugin can decode to 24 bits, you can't take advantage of that feature)
==> output.
As you can see, all audio processing is done in 16 bits (unless you use a combination of plugins that's compatible with other bitdepths), and the whole thing is just... messy, because too many tasks are left to the input plugins which should be carried out at a later stage. For example, it should not be necessary to include a hard limiter in the input plugin (which, for instance, is the case with the Vorbis plugin) - this type of processing should be done after all other processing (except for dithering) is completed. The purpose of this limiter is avoid clipping, but it's no use if it is applied
before DSP processing (since there's no guarantee that the DSP plugin won't clip the audio data). The same thing goes for dithering - it should be performed after any other kind of processing.
Foobar handles the data like this:
Decode audio to 64 bits
==> apply ReplayGain
==> DSP processing
==> dither down to 16 bits (or whatever you've selected)
==> output.
All audio processing is done in 64 bits, so it will be more precise than Winamp's 16 bit processing, and dither will be performed as the last step before output (and limiting before that if you've configured it correctly). This is the main theoretical advantage in sound quality everybody is speaking of. Again, noone firmly said this is audible (leaving die-hard clueless zealots out of it for a moment

).
Apart from the theoretical improvement in sound quality (which isn't my argument for using foobar2000), it's just more convenient to have one central control for ReplayGain options, limiting, tag display etc., instead of having to configure it individually for each input format.
I'm not a foobar2000 zealot, I'm not bashing Winamp, I'm trying to be reasonable and point out where its flaws are compared to foobar2000 (since you asked), that's all.