[00:02:05 CET] <jamrial> jkqxz: if you're also pushing them to release/4.1 yourself, do the same with a4fb2b1150 and e9d2e3fdaa while at it [00:03:47 CET] <jkqxz> Ok, sure. [00:07:00 CET] <cone-309> ffmpeg 03Mark Thompson 07master:6bdb7712ae02: cbs_av1: Fix header writing when already aligned [00:07:01 CET] <cone-309> ffmpeg 03Mark Thompson 07master:f5894178fb80: cbs_av1: Support redundant frame headers [00:17:37 CET] <cone-309> ffmpeg 03Mark Thompson 07release/4.1:066ff02621ce: configure: Add missing IVF muxer BSF dependency [00:17:38 CET] <cone-309> ffmpeg 03Mark Thompson 07release/4.1:ec1b5216fc65: configure: Add missing V4L2 M2M decoder BSF dependencies [00:17:39 CET] <cone-309> ffmpeg 03Mark Thompson 07release/4.1:af3fccfeff74: cbs_av1: Fix header writing when already aligned [00:17:40 CET] <cone-309> ffmpeg 03Mark Thompson 07release/4.1:10506de9ad1f: cbs_av1: Support redundant frame headers [01:09:27 CET] <cone-309> ffmpeg 03Mark Thompson 07n4.1:HEAD: cbs_av1: Support redundant frame headers [01:25:13 CET] <cone-309> ffmpeg 03Mark Thompson 07n4.2-dev:HEAD: cbs_av1: Support redundant frame headers [02:35:21 CET] <cone-309> ffmpeg 03James Almer 07master:beaa350e2416: avcodec: libdav1d AV1 decoder wrapper. [10:51:55 CET] <cone-871> ffmpeg 03Linjie Fu 07master:87368884a52b: lavc/qsvdec: flush buffered data before reinit [10:51:56 CET] <cone-871> ffmpeg 03Zhong Li 07master:d96ae9d5ea1f: lavc/qsv_overlay: suppress code scan complain [12:40:56 CET] <funman> sys.path.append(os.path.join(sys.MEIPOASS, 'out00-PYZ.pyz')) [12:42:55 CET] <JEEB> what's the best way to figure out which part is blocking with f.ex. ffmpeg.c? [12:43:28 CET] <JEEB> as in, is it filter chain with splits etc, or is it encoding or what [12:43:45 CET] <JEEB> what are good tools to profile such stuff? [12:44:05 CET] <nevcairiel> i would go low-tech and test the invidiual steps individually [12:44:14 CET] <nevcairiel> measure their speed [12:44:41 CET] <JEEB> it's probablematic with live inptus [12:44:50 CET] <JEEB> I could of course dump and check [12:45:09 CET] <JEEB> but then you're not limited in input speed unless I use something like multicat [12:45:22 CET] <JEEB> which outputs MPEG-TS into UDP with speed control [12:45:32 CET] <nevcairiel> of course, but if filtering and encoding is like 10x realtime, its not going to limit you [12:46:27 CET] <JEEB> true. it's just that I have a hunch that I have a case of a filter chain getting stuck or so on one of the outputs or something like that [12:46:42 CET] <JEEB> but yea, I think that makes sense. dump some and see the steps separately [12:46:48 CET] <JEEB> also the overall load is probably one thing [14:06:11 CET] <Compn> ahhh [14:06:16 CET] <Compn> we still don't have qdmc decoder ? [14:06:54 CET] <Compn> no wait we do [14:06:55 CET] <Compn> hmm [14:07:53 CET] <Compn> is my vlc so old it does not have this mapped ? [14:08:45 CET] <durandal_1707> yes [14:09:58 CET] <Compn> durandal_1707 : btw i have imm5 sample (and other samples) for you [14:10:32 CET] <durandal_1707> i accept only money [14:17:43 CET] <Compn> sent you link [14:18:02 CET] <durandal_1707> link is not money, ignored [14:18:21 CET] <Compn> j-b iirc said he would pay you for imm5 decoder [14:18:30 CET] <Compn> there is imm5 sample in there [14:18:44 CET] <durandal_1707> what is in sample? [14:19:06 CET] <Compn> boring video , security cam or dash cam [23:07:03 CET] <durandal_1707> i'm retiring [23:13:18 CET] <durandal_1707> nobody cares, eh! [23:14:35 CET] <jamrial> i do! [23:14:36 CET] <j-b> retiring from? [23:16:14 CET] <Compn> durandal_1707 : retiring from ffmpeg ? [00:00:00 CET] --- Wed Nov 7 2018
participants (1)
-
burek