I recently bought a Rio 500 on eBay. While very happy with the device itself, I am convinced that, with some research (and your help), I can achieve the same or greater fidelity audio at smaller file sizes than I am currently have. For the purpose of this thread, my only concern is the best format/setting for "books on tape"/spoken word.
The device can read MP3, MPEG2.5, Audible AA (G.723), and ADPCM formats.
Right now, I covert a CD to mp3 format, using LAME alt-preset 80. This is considerably better sounding than any other LAME generated file (particularly the LAME ACM) [Thanks, Dibrom!]. And it generates files that are 114% the size of MP3pro 64Kb and 133% the size of AAC Psytel -Internet. While suprisingly competitive in file size, the quality of MP3pro and AAC is MUCH better sounding--particularly if the original audio had any hint of echo. In general, and particularly with echos, alt-preset 80 sounds significantly more "metallic", with overtones of a speech synthesizer. MP3pro and AAC, by contrast, are difficult to distinguish from the source material even on good quality speakers.
Quality-wise, I can withstand something as low as Audible 4 though I prefer better. In a perfect world, I would use AAC for high fidelity encodings (regular commute) or G.723 for extended duration encodings (long plane rides).
I know that knowledgable people on this board have said that LAME is intended for higher bit rate encodings and suggested MP3pro for these lower bit rate scenarios (given its smaller file size, I don't know why AAC wasn't suggested). However, until I figure out how to reprogram the Rio 500 codec or Sonic Blue releases firmware including either MP3pro or AAC (hah!), I don't have that option.
Therefore:
* I have spent a lot of time on this board but perhaps I missed a better LAME setting/MP3 encoder for spoken word?
* If not, can LAME be tweaked [hint, hint Dibrom]? I don't understand why LAME couldn't have a setting that beats any competitive MP3 encoder for spoken word.
* Alternatively, is there a means of encoding in MPEG 2.5 (which I believe will support lower bit rates) or G.723 that is as easy to do as MP3? I have found very little on the Internet regarding either of the latter formats, particularly with respect to generally available encoders as opposed to telecom implementations. I also searched specifically for an Audible format encoder, with no luck.
* Finally, if anyone knows of links to reprogramming the Rio 500 codec (which I understand to be a general purpose processor), I am interested, particularly with respect to AAC.