[Mplayer-cvslog] Re: CVS: main/libmpdemux demux_ogg.c,1.20,1.21

Balatoni Denes pnis at coder.hu
Sat Dec 7 01:21:11 CET 2002


Hi!

The source is quite distorted (badly tuned tv card)  but then playback is more 
like plain noise. I don't know what causes this, but *32 seemed to be good, 
and *128 not good. And anyway it might not be that important for people with 
normal machines, and those who need this might not care that there are only 
32 different gains (between gain 0 and 1). Maybe I misheard it, the source 
was definietly a bad example, decide what you think is the right tradeoff 
(myself, I am content with only 32 steps for gain)

bye
Denes



On 2002. december 6. 01:15, Rémi Guyomarch wrote:
> On Thu, Dec 05, 2002 at 09:49:19PM +0100, Balatoni Denes wrote:
> > Hi!
> >
> > Btw. with ov->rg_scale*128 it is quite distorted for what I have tried.
> > How about making it just *32 (to be sure) ? Should I Provide  a patch?
>
> Are you sure the source itself isn't distorded ?
>
> Checking the code, I don't think there's a possibility of overflow
> with 7 bits of precision in rg_scale..




More information about the MPlayer-cvslog mailing list