Compression and stereo image

I note from reading earlier postings that compressing to MP3 can sometimes cause damage to the stereo image.

This leads to a number of questions:

  1. Is the MP3 stereo image damage less at higher bit rates?

  2. Does Apples iTunes AAC compression produce the same stereo image damage - and if so is that ameliorated by using higher bit rates?

  3. Does iTunes Apple Lossless compression produce the same stereo image damage?

  4. Does AIFF compression produce the same stereo image damage?

  5. And do I assume correctly that exporting to WAV will produce no damage to the stereo image?

Basically I am trying to decide what format to choose to re-rip my classical music collection (currently a mix of AAC-192 and MP3-192) now that I have recently acquired a 160gb iPod that gives me much more storage space (though not enough still to do everything in WAV format - at bitrate 192 the iPod is already over 1/3 full).

Any suggestions and insights gratefully received,

The killer for “vocal removal” is the “joint stereo” option in MP3. Instead of producing two distinct channels (one for left and one for right), it uses a “centre/side difference” method. Do you remember koz’s explanation about FM stereo? - it’s similar to that, but because the data is also compressed it does not reconstruct quite the same stereo field when it is decoded. Selecting the “stereo” option should improve the situation, though I would not say that sounds that are centre stage will necessarily remain centre stage after compressing/un-compressing. (I’ve not thoroughly checked out the “stereo” option, so I’m just guessing that it may be better).

There are some issues with MP3 encoding in the current version of Audacity (incorrect compression settings sometimes occur), so if you go for MP3 format you may be better to use an external compressor. I use “LameDrop” which is very good.

Lossless formats (compressed or not) should not cause any problem - (they should be… lossless)

Does the iPod support FLAC or Ogg?
FLAC is lossless, Ogg is lossy, but can do very high quality (possibly better than MP3 at high bit rates).

Update on “Joint vs. stereo” for MP3 export.

I’ve done a few experiments and the “stereo” setting seems to offer no advantages. The file size is larger, but the stereo image is no better - if anything it is worse !

Using high quality settings (“Insane” preset, with joint stereo) the stereo imaging is pretty good and “vocal removal” works almost as well as on the original WAV file.


it doesn’t look like iTunes/iPod supports FLAC or Ogg - the formats available in my import dropdown are: AAC, AIFF, Apple Lossless, MP3 and WAV.

Thanks for your thoughts on this - I may need to do some experiments with lossless formats and then “do the math” to see whether that classical collection would fit that way still leaving space for all the other stuff …

BTW I’m not worried about vocal removal - just about getting the best possible stereo image from a portable player or PC.


I didn’t really think you were, but it’s a common issue: “why doesn’t vocal removal work on my poorly recorded MP3 that I downloaded”.
Also, vocal removal seemed like a reasonable test for how accurately the stereo imaging is maintained after lossy encoding.

If lossless is too big (expect no better than 50% of the original file size), then top quality MP3 is a good option, 320 kbps CBR. For high bit rates like this, CBR is better than VBR. (VBR is better at lower bit rates).

If Audacity MP3 export is working for you, just select the “Insane” preset, or specify “Constant / 320 / Joint”.


MP3 export works fine for me - but I now prefer to use AAC (exporting as WAV then importing into iTunes as AAC) as bitrate for bitrate AAC is reckoned to be better than MP3. So I may go for AAC at 256 0r 320.


That depends on who you ask :smiley:

At high bit rates there is very little to choose between AAC, MP3 and Ogg.
This survey from December 2005 put them all on par, (with a less than significant lead from Ogg Vorbis)

I sometimes wonder why Apple developed their own format, when they have been unable to improve on Ogg Vorbis, which is free and open source. It would have been far less expensive for them to adopt Ogg Vorbis, instead of developing their own proprietary format, and far more convenient for hardware manufacturers, software developers and the public.


thanks for that link - an interesting set of results.

I suspect the AAC proprietary choice by Apple may be to do with DRM and an attempt to facilitate the locking down of stuff …