[LAU] Segfault using zita-j2a

Jannis Achstetter jannis_achstetter at web.de
Fri Oct 26 10:27:32 UTC 2012


Am 26.10.2012 09:37, schrieb Fons Adriaensen:
> On Thu, Oct 25, 2012 at 11:37:30PM +0200, Jannis Achstetter wrote:
>  
>> Okay, the different classes explain the different behavoirs then because
>> zretune crashes instantly:
> 
> If you install libzita-resampler from source, try this:
> 
> int VResampler::process (void)
> {
> -    unsigned int   k, hl, np, in, nr, nz, i, n, c;
> +    unsigned int   k, np, in, nr, n, c;
> +    int            i, hl, nz;
>      double         ph, dp, dd;
>  
> In line 217 the index [i - hl] goes negative. If it is extended
> to 64 bit as unsigned that would explain the segfault.

That patch indeed fixes the problem and
zita-j2a works like a charm now. Nice :) And as expected it uses less
CPU than alsa_in and alsa_out.

I tried alsa_loopback from zita-alsa-pcmi this morning (as
"./alsa_loopback hw:1 hw:1 48000 2048 10"). No errors, no crashes, but
no output on the soundcard. Last line printed is "synced". However, the
documentation is sparse so I didn't know what all these parameters are
for and guessed most of them.

The ajbridge has the same problem. There's the quality-parameter that
defaults to "48" but that's all I get to know about it. No value range,
no indication whether 0 is better or worse than 1000. You could add some
hints to the readme ;)

But thanks for all your work you contribute to linux-audio and your
support :)

	Jannis



More information about the Linux-audio-user mailing list