[pulseaudio-discuss] [PATCH 8/8] Improve performance of peaks resampler

Maarten Bosmans mkbosmans at gmail.com
Mon Nov 28 01:45:02 PST 2011


2011/11/27 Colin Guthrie <gmane at colin.guthr.ie>:
> 'Twas brillig, and Maarten Bosmans at 02/11/11 20:54 did gyre and gimble:
>> This is mainly achieved by special-casing the common 1ch float case,
>> which is used by applications such as pavucontrol.
>
> OK, I pushed this and the previous patch too.
>
> I didn't check the maths so I'll trust your analysis here (and the peaks
> resampler isn't really super important either, hence my slightly
> lackadaisical stance :D)

Thanks Colin.

Unfortunately you pushed an earlier version of the patch. This isn't
entirely your fault, as these two patches have been floating on the
list in various versions at 16-11, 2-11 and 14-10 (note to readers
from the USA: try reversing those numbers ;-) and you picked the
middle one.

I rebased the latest patch on master and the diff that remains has
been sent to the list. It's fairly insignificant by itself, that's why
I rolled it up in the earlier patches to start with.

Maarten


More information about the pulseaudio-discuss mailing list