|
My biggest complaint about audiobooks is that they are most often recorded using an mp3 or WMA (or some other) encoder intended for music. Most audio books contain almost exclusively speech. There is a large difference, from the theory to the implementation, of music and speech codecs.
Anyway, the results are that audio files tend to be either very large or of poor quality. Audio book recorders should be using a modern speech codec for the bulk of the recording, consuming something like and average of 12.65 (AMR-WB) kbps of bandwidth. Speech codecs can perform very well given this much bandwidth, while mp3 is miserable to listen to.
It doesn't help to have somethng like AMR-WB or AMR-WB+ available if the original recording is in mp3 format. The conversion process results in unacceptable artifacts in the audio. One needs access to the original WAV or similar oncompressed audio files and go straight to the final compressed audio to keep quality as high as possible at minimum bit rates (and therefore file sizes).
|