Re: all you MP3 people out there... a fix for lars!

[Lars Magne Ingebrigtsen <larsi@gnus.org>]

Harald Meland <Harald.Meland@usit.uio.no> writes:

> I'd guess Lars is using L.A.M.E., found on
> <URL:http://student.uq.edu.au/~s302585/>.

I'm actually using Mike Cheng's version of 8hz, but that's the same
thing, only different.

> (Just guesswork, of course, but he _is_ getting thanks in the NEWS
> section :-)

I had to do *something* to relax before the Gnus talk a couple of
weeks ago, so I hacked away at LAME.  :-)

> L.A.M.E. also claims to be much faster than BladeEnc, whereas BladeEnc
> seems to be claiming the opposite :)

Here are the tests I did; 8hz and LAME are virtually identical
speed-wise. 

orig.wav (/ 56812604.0 (* 44100.0 4)) => 322.07
blade-128.mp3 (/ (+ (* 29 60) 18) 322.07) => 5.46
8hz-128.mp3 (/ (+ (* 18 60) 54) 322.07) => 3.52
encode-128.mp3 (/ (+ (* 25 60) 95) 322.07) => 4.95
blade-256.mp3 (/ (+ (* 24 60) 16) 322.07) => 4.52
8hz-256.mp3 (/ (+ (* 18 60) 6) 322.07) => 3.37
encode-256.mp3 (/ (+ (* 20 60) 31) 322.07) => 3.822
m3enc-128.mp3 (/ (+ (* 9 60) 39) 30) => 19

This is on a K6.  These number say that 8hz compresses at a 3.52:1
time rate, while blade does it on a 5.46:1 rate, which means that
blade is, er, 70% slower than 8hz.

This is with Cheng's version of 8hz, mind you -- the original 8hz
version from the 8hz people is slower.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen