Subject: Re: [linux-audio-dev] LADSPA hard_rt_capable
From: Steve Harris (S.W.Harris_AT_ecs.soton.ac.uk)
Date: Tue Dec 12 2000 - 13:55:17 EET
On Mon, Dec 11, 2000 at 09:40:57PM +0200, Kai Vehmanen wrote:
> 17.63user 0.05system 0:17.69elapsed 99%
Erk!
Um, well looking at the code in the light of 32 sample blocks I can
definatly improve on that:
On a PIII500, 1024 sample blocksize:
[swh_AT_inanna]$ time ecasound -b:1024 -t:5 -i testdata/f242.wav -o null -eli:1191,2,4,1,0 -kl:3,0.5,100,5.0
user 0m0.530s
sys 0m0.040s
real 0m0.755s
32 sample blocksize:
[swh_AT_inanna]$ time ecasound -b:32 -t:5 -i testdata/f242.wav -o null -eli:1191,2,4,1,0 -kl:3,0.5,100,5.0
user 0m0.610s
sys 0m0.030s
real 0m0.825s
And that wasn't cheating by optimising the case where the modulation
speed is always chaged from block to block either. I did have to cut a
few corners with the modulation, but it doesn't seem to be audible.
Re the subject line, I think a lot more of my plugins are going to loose
thier HARD_RT_CAPABLE flag :(
Oh well, that'll keep me busy for a while. Thanks for the test case Kai.
- Steve
This archive was generated by hypermail 2b28 : Tue Dec 12 2000 - 14:40:21 EET