The XA ADPCM decoder resets the last sample predictors to 0 after every frame. This is wrong, and causes loud clicking constantly in all my samples. Attached patch fixes it. I can't tell if seeking is affected, since the STR demuxer can't seek in ffplay (not sure why) and can't be remuxed (no codec_tag).
On Dec 5, 2007, at 11:32 AM, Alexander Strange wrote:
The XA ADPCM decoder resets the last sample predictors to 0 after every frame. This is wrong, and causes loud clicking constantly in all my samples. Attached patch fixes it.
I can't tell if seeking is affected, since the STR demuxer can't seek in ffplay (not sure why) and can't be remuxed (no codec_tag).
Samples where it's audible: http://samples.mplayerhq.hu/game-formats/psx-str/lunar2.str (when they talk) http://astrange.ithinksw.net/ggg.str http://astrange.ithinksw.net/lain.str Both of those are in mphq incoming already.
On Wed, Dec 05, 2007 at 11:32:53AM -0500, Alexander Strange wrote:
The XA ADPCM decoder resets the last sample predictors to 0 after every frame. This is wrong, and causes loud clicking constantly in all my samples. Attached patch fixes it.
patch is probably ok
I can't tell if seeking is affected, since the STR demuxer can't seek in ffplay (not sure why) and can't be remuxed (no codec_tag).
i think the str demuxer needs some cleanup work ... hell there are even printf() in there ... [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB When the tyrant has disposed of foreign enemies by conquest or treaty, and there is nothing more to fear from them, then he is always stirring up some war or other, in order that the people may require a leader. -- Plato
On Dec 8, 2007, at 6:13 PM, Michael Niedermayer wrote:
On Wed, Dec 05, 2007 at 11:32:53AM -0500, Alexander Strange wrote:
The XA ADPCM decoder resets the last sample predictors to 0 after every frame. This is wrong, and causes loud clicking constantly in all my samples. Attached patch fixes it.
patch is probably ok
Ping; does anyone else have a problem with it? I think it wouldn't be any worse than missing the bit reservoir in mp3, in a perfect world where it could seek in this format in the first place. I'm looking at a crash (will add a ticket after I wake up) and don't want to have more than one patch against this floating around.
On Thu, Dec 27, 2007 at 03:20:24AM -0500, Alexander Strange wrote:
On Dec 8, 2007, at 6:13 PM, Michael Niedermayer wrote:
On Wed, Dec 05, 2007 at 11:32:53AM -0500, Alexander Strange wrote:
The XA ADPCM decoder resets the last sample predictors to 0 after every frame. This is wrong, and causes loud clicking constantly in all my samples. Attached patch fixes it.
patch is probably ok
Ping; does anyone else have a problem with it? I think it wouldn't be any worse than missing the bit reservoir in mp3, in a perfect world where it could seek in this format in the first place.
I'm looking at a crash (will add a ticket after I wake up) and don't want to have more than one patch against this floating around.
Applied. Diego
On Thu, Dec 27, 2007 at 03:20:24AM -0500, Alexander Strange wrote:
On Dec 8, 2007, at 6:13 PM, Michael Niedermayer wrote:
On Wed, Dec 05, 2007 at 11:32:53AM -0500, Alexander Strange wrote:
The XA ADPCM decoder resets the last sample predictors to 0 after every frame. This is wrong, and causes loud clicking constantly in all my samples. Attached patch fixes it.
patch is probably ok
Ping; does anyone else have a problem with it? I think it wouldn't be any worse than missing the bit reservoir in mp3, in a perfect world where it could seek in this format in the first place.
Dropping the bit reservoir once at the beginning of decoding mp3 will not cause any major problem. Dropping it on every frame would create horrible ear-grating artifacts, perhaps clicking or otherwise. Rich
participants (4)
-
astrange@ithinksw.com -
dalias@aerifal.cx -
diego@biurrun.de -
michaelni@gmx.at