To: AlienTech who wrote (78 ) 8/6/1999 3:21:00 PM From: William T. Katz Read Replies (2) | Respond to of 165
AlienTech, I know MP3 = MPEG audio but that's not what ADBL is using. I just read over their IPO prospectus and figured out a little more. ADBL has licensed a proprietary and unnamed codec from someone, and then added their own security features to the codec. There are all sorts of ways of compressing data, depending on what kind of data you are dealing with. There are non-MP3 ways of doing compression on spoken audio, maybe they can take advantage of the fact that human voices are in a relatively limited bandwidth and that dead time (no sound between words) is an inherent part of the media. There are probably a lot of other things particular about speech, but all I know is that it's a different signal and hence can use a more optimized codec. (http://195.27.241.3/pcmag/supp/1998/voice_data/20.html) MP3 was not designed for spoken audio, in my understanding, but is geared toward general audio, esp. music. Under MP3, you get roughly 60 min music for 32 MB of RAM. You can double that at least when you encode for spoken audio since you decrease the sampling. But that is still a far cry from getting 16 hours using the Audible format. ADBL currently has some MP3 versions of their content because Diamond (and Creative) hasn't come out with an AudibleReady format of Rio yet. This is primarily a stop-gap before AudibleReady formats are played by Rio. You might take note, also, that Audible.com can change formats if new standards for spoken audio come on the scene. They allude to such a possibility in the prospectus. But for now, their proprietary format (licensed spoken audio codec + proprietary security layer) is much better for handling spoken audio than generic MP3 codecs. Hope that clears the issue up more. -Bill