MP3 Encoder Torture Test

In preparing a track for artist review I ran into a piece that really fell apart in the MP3 conversion from AA.  I thought it might be a good test track for different MP3 encoders.

I exported the same WAV file out of AA, Samplitude, and then used LAME.  I selected the highest quality option in all cases, 128 kbps CBR.  I have links here to the original WAV file and the three resulting MP3 files.  IMHO, Lame really shines on this one.

FYI

MP3_Torture_Test.wav

MP3_Torture_Test_AA_CBR_128.mp3

MP3_Torture_Test_Sam_CBR_128.mp3

MP3_Torture_Test_Lame_CBR_128.mp3

Just curious, what do you mean by "I selected the highest quality option in all cases, 128 kbps CBR."?

All MP3 files are 128kbps (I know that's not "HIGH" quality, but it's usually adequate for my purposes).  Each of the MP3 encoders have speed vs quality options - "Do you want it FAST, or do you want it GOOD?"

I selected the best QUALITY, slowest ENCODING performance in all cases.
In Lame that meant "lame -q 0 -b 128"

[all]

All MP3 files are 128kbps (I know that's not "HIGH" quality, but it's usually adequate for my purposes.  Each of the MP3 encoders have speed vs quality options - "Do you want it FAST, or do you want it GOOD?"

I selected the best QUALITY, slowest ENCODING performance in all cases.
In Lame that meant "lame -q 0 -b 128"

There are many different versions of the lame encoder.  The one that has been tested to be of highest quality thus far is version 3.97b3.

I'm going out on a limb here, but I would expect any recent release of LAME to be superior in all cases to anything encoded with AA's fraunhofer codec.

There are many different versions of the lame encoder.  The one that has been tested to be of highest quality thus far is version 3.97b3.

I used 3.96.1.  I've been using LAME as I've been working on moving my CD collection to a computer-based setup, but I had not run into any previous examples of one MP3 encoder being so dramtically different from another.

http://wiki.hydrogenaudio.org/index.php?title=Lame_Compiles

Are you aware that VBR, at least with LAME, is generally considered to provide the best quality for the least storage space. LAME presets are recommended over any switches for most uses.

Some players don't handle VBR well or at all, so I hesitate to jump in that direction.  At this point I'm ripping everything first to WMA lossless and will recode to a compressed format later, so I can change my mind - do the whole thing as a batch to a different format and I won't be out anything.

Regarding the file I posted for this thread, I just found it to be a dramatic demonstration of the differences between encoders - that's all.

FWIW, I've always used switches over presets too Mr Zemlin and never regretted it. CBR as well. Usually its -q 0 -m s -b 320 and I add CRCs too (yeah I'm a dork).