Does anyone know how audible can get a file to be 300MB when it is 24 hours long? I have been looking into different codecs and I just do not understand how they can compress such a long audio file into a small size. I have read about FLAC vs Opus vs AAC and even Audible's proprietary AA codec, but nowhere can I find some theoretical or even practical explanation. I thought you could not compress files that are already compresses (such as MP3, FLAC, etc) so how do they do it??
Asked
Active
Viewed 301 times
1 Answers
3
My guess is that it's a combination of audio filtering and codec algorithm. Spoken-word audio books don't need to be as hi-fi as a symphony recording to sound acceptable to us. Spoken-word audio files can:
- be mono instead of stereo, which presumably reduces the bytes needed.
- filter out unnecessary frequencies that are either inaudible or not provided in the recording. If one person is speaking, the frequency spectrum will be narrow, which also saves space.
- use variable bitrate encoding to keep portions of files small for speech-only but also provide more bandwidth for portions of files with sound effects or music.
- Opus is pretty amazing and has speech mode which is specifically engineered to compress speech.
Using the 15.18s Opus music files below as an example, 5,692 files are needed to comprise 24 hours (86,400s) of audio.
https://opus-bitrates.anthum.com/
- 32 kbit/s stereo:
5,692 * 62,239 bytes =337.85 MiB (24hrs) - 16 kbit/s mono:
5,692 * 31,959 bytes =173.48 MiB (24hrs)
anthumchris
- 8,245
- 2
- 28
- 53