[Asterisk-Dev] Re: lightning fast 9:1 compression ratio audio codec, anyone?

[Luke Kenneth Casson Leighton]

p.s. example 30-second wav file and its encoded-and-decoded
form at:

	http://hands.com/~lkcl/ahab-pcm

just re-listened to it - it's not _that_ bad :)

due to the way the algorithm works, it's the plosives that
are _particularly_ adversely affected - something i _could_
mitigate against with a bit more work.

p.p.s. this codec would be _unbelievably_ simple to implement
in hardware (again - mentioned for the archives / completeness
as i don't honestly believe too many hardware people to be _that_
interested in ahab-pcm - or software ones either :).


On Thu, May 12, 2005 at 03:04:15PM +0100, Luke Kenneth Casson Leighton wrote:

> in a panic last month i tried to write an audio codec.
> 
> it uses a hamming principle over-and-above what adpcm does.
> 
> it _does_ have clipping artefacts (which would be mitigated
> against when using higher sampling rates) , but it is blindingly
> fast (utilises even less CPU than adpcm), and has a compression
> ratio of about 9:1.