Baek Chang wrote, On 09/16/2010 02:26 AM:
> looks like in the remap_channels method, calling oil_vectoradd_s16,
> the constant to multiply by is set to 1.
>
> if (r->map_table[oc][ic] >= 1.0) {
> static const int16_t one = 1;
>
> oil_vectoradd_s16(
> (int16_t*) dst + oc, o_skip,
> (int16_t*) dst + oc, o_skip,
> (int16_t*) src + ic, i_skip,
> (int) n_frames,
> &one, &one);
>
> }
>
> shouldn't there be a divide by 2 somewhere if we are adding two
> channels together?
>
>
> On Wed, Sep 15, 2010 at 2:54 PM, Baek Chang
> <baeksan at ccrma.stanford.edu <mailto:baeksan at ccrma.stanford.edu>> wrote:
>
> I am seeing warp around happening due to overflow when doing the
> following.
>
> parec -r
> --device=alsa_output.pci_8086_269a_sound_card_0_alsa_playback_0.monitor
> output.raw --format=s16le --rate=44100 --channels=1
> paplay --format=s16le --rate=44100 --channels=2
>
>
> if i use parec with the same configuration as the sink that I am
> playing the wave file on, then there is no overflow/wraparound. It
> seems to be a problem with the remapping of the channels from stereo
> to mono. Is there an additional gain stage applied in this remapping
> that might be causing this?
>
> I am using pulseaudio 0.9.14
>
Note that the latest version is 0.9.21 and that the code you are
referring to seems to have been completely rewritten.
/Mads
