[pulseaudio-discuss] [PATCH 2/2] core: Add ARM NEON optimized sample conversion code

[Peter Meerwald]

Hello Arun,

> > > Possibly we're talking about different things here -- I'm referring to
> > > the float -> s16le conversion.
> > I've never seen deviation between NEON and C code for float -> s16le
> > my claim for one-off was for s16le -> float only

there is a deviation for values with .5 which are rounded differently

> You should be able to see it with the code at:
> http://cgit.collabora.com/git/user/arun/pulseaudio.git/log/?h=neon-wip

please see the 3 patches on top of your branch I just posted;
should fix the issue above and raises a concern regarding the way 
PulseAudio does signed int to float conversion

I think the way libavresample does it is way better/saner; or is there a 
particular reason for the way PA performs these conversions?

I'm working to see if the s16->float NEON code can be fixed (or made 
compatible with PA)

regards, p.

-- 

Peter Meerwald
+43-664-2444418 (mobile)