[09:11] *** aballier has quit IRC (Ping timeout: 260 seconds)
[11:03] <larkfisherman> quick question :)  I'm hacking around with video4linux, and I'm wondering what is the use case for V4L2_PRIORITY_RECORD, except for simply not allowing external tools to modify device controls?
[11:04] <larkfisherman> I'm talking about video capturing devices, mainly.
[11:05] <larkfisherman> Is it be used to prevent some device issues/firmware bugs/something else, or is it completely unrelated to that?
[11:06] <hverkuil> larkfisherman: it prevents other applications from making changes. I.e. if an application sets the prio to RECORD, then other applications can only query the video device, but not make changes.
[11:07] <hverkuil> By default, if one application is capturing video, then another application can still change e.g. the brightness. Raising the priority to RECORD will prevent that.
[11:13] <larkfisherman> I see. So it's not used for preventing (driver/firmware/etc) issues? Is video4linux always safe when used from two applications simultaneously, in the way you described (i.e. one application capturing video, other application changing manual focus at the same time)??
[11:13] <larkfisherman> s/??/?/
[11:16] <hverkuil> Yes, it's safe.
[11:16] <larkfisherman> Okay, thank you :)
[11:17] *** larkfisherman has left 
[16:00] *** benjiG has left 
[20:28] <ndufresne> svarbanov: tfiga: Remember the discussion about H.264 IDR with B-Frames, and how this could cause an invalid stream ?
[20:29] <ndufresne> svarbanov: tfiga: There is a no_output_of_prior_pics_flag part of the ref_pic marking
[20:30] <ndufresne> This flag is set on user requested IDR to avoid this problem
[22:29] *** Kwiboo has quit IRC (*.net *.split)
[22:29] *** mmattice has quit IRC (*.net *.split)
[22:29] *** syoung has quit IRC (*.net *.split)
[22:29] *** ebarretto has quit IRC (*.net *.split)
[22:29] *** larsc has quit IRC (*.net *.split)
[22:29] *** koike has quit IRC (*.net *.split)
[22:29] *** juvenal has quit IRC (*.net *.split)
[22:29] *** smartin has quit IRC (*.net *.split)
[22:29] *** RzR has quit IRC (*.net *.split)
[22:29] *** r0kc4t_ has quit IRC (*.net *.split)
[22:29] *** lexano has quit IRC (*.net *.split)
[22:29] *** rektide has quit IRC (*.net *.split)
[22:29] *** elGamal has quit IRC (*.net *.split)
[22:29] *** andrzej_p has quit IRC (*.net *.split)
[22:29] *** robertfoss has quit IRC (*.net *.split)
[22:29] *** ChanServ has quit IRC (*.net *.split)
[22:29] *** kbingham has quit IRC (*.net *.split)
[22:29] *** marvs has quit IRC (*.net *.split)
[22:29] *** harrow has quit IRC (*.net *.split)
[22:29] *** uajain has quit IRC (*.net *.split)
[22:29] *** styl1te has quit IRC (*.net *.split)
[22:29] *** ndufresne has quit IRC (*.net *.split)
[22:29] *** padovan has quit IRC (*.net *.split)
[22:33] *** barjavel.freenode.net sets mode: +o ChanServ
[23:10] <tfiga> ndufresne: aha :)
[23:11] <tfiga> now we *only* need to make sure the encoders support that ;)
[23:43] <ndufresne> tfiga: yeah, I can only check for CODA, pretty much the only chip I got documentation for in the stateful side
[23:44] <ndufresne> well, Hantro too, but that's not mainline yet
[23:53] <ndufresne> tfiga: so on CODA, GOP is handled by the firmware on the bit processor (in fact only a fixed set of pattern is possible), and optionally you can flag a frame to force that frame to be encoded as an IDR, I'll need to test, but I when we force an IDR, we endup with this ref_pic_marking flag being set