michaelni changed the topic of #ffmpeg-devel to: Welcome to the FFmpeg development channel | Questions about using FFmpeg or developing with libav* libs should be asked in #ffmpeg | This channel is publicly logged | FFmpeg 7.1.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
<fflogger>
[editedticket] MasterQuestionable: Ticket #11525 ([avdevice] Better frame dropping logic for live streams to avoid unjustifiable freezing?) updated https://trac.ffmpeg.org/ticket/11525#comment:22
jamrial has quit [Read error: Connection reset by peer]
jamrial has joined #ffmpeg-devel
HarshK23 has quit [Quit: Connection closed for inactivity]
IndecisiveTurtle has quit [Ping timeout: 265 seconds]
kasper93 has quit [Remote host closed the connection]
jamrial has quit [Read error: Connection reset by peer]
jamrial_ has joined #ffmpeg-devel
averne has quit [Ping timeout: 260 seconds]
averne_ has joined #ffmpeg-devel
averne_ is now known as averne
thilo has quit [Ping timeout: 244 seconds]
thilo has joined #ffmpeg-devel
thilo has quit [Changing host]
thilo has joined #ffmpeg-devel
derpydoo has joined #ffmpeg-devel
jamrial_ has quit [Read error: Connection reset by peer]
jamrial_ has joined #ffmpeg-devel
cone-938 has quit [Quit: transmission timeout]
^Neo has quit [Ping timeout: 244 seconds]
System_Error has quit [Remote host closed the connection]
abdu18 has quit [Ping timeout: 240 seconds]
System_Error has joined #ffmpeg-devel
LainExperiments has joined #ffmpeg-devel
LainExperiments has quit [Quit: Client closed]
LainExperiments has joined #ffmpeg-devel
jamrial_ has quit []
LainExperiments3 has joined #ffmpeg-devel
LainExperiments has quit [Ping timeout: 240 seconds]
martinr1 has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
martinr1 has quit [Ping timeout: 252 seconds]
LainExperiments3 has quit [Quit: Client closed]
LainExperiments has joined #ffmpeg-devel
LainExperiments4 has joined #ffmpeg-devel
LainExperiments has quit [Ping timeout: 240 seconds]
martinr1 has joined #ffmpeg-devel
martinr1 has quit [Ping timeout: 265 seconds]
LainExperiments4 has quit [Ping timeout: 240 seconds]
martinr1 has joined #ffmpeg-devel
LainExperiments has joined #ffmpeg-devel
martinr1 has quit [Ping timeout: 244 seconds]
LainExperiments has quit [Ping timeout: 240 seconds]
System_Error has quit [Remote host closed the connection]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
LainExperiments has joined #ffmpeg-devel
<fflogger>
[editedticket] catap: Ticket #11525 ([avdevice] Better frame dropping logic for live streams to avoid unjustifiable freezing?) updated https://trac.ffmpeg.org/ticket/11525#comment:23
Guest28 has joined #ffmpeg-devel
LainExperiments has quit [Quit: Client closed]
IndecisiveTurtle has joined #ffmpeg-devel
Grimmauld has joined #ffmpeg-devel
Grimmauld has quit [Quit: Client closed]
Grimmauld has joined #ffmpeg-devel
auri has joined #ffmpeg-devel
martinr1 has joined #ffmpeg-devel
martinr1 has quit [Ping timeout: 248 seconds]
ngaullier has joined #ffmpeg-devel
ngaullier has quit [Remote host closed the connection]
Guest28 has quit [Quit: Client closed]
ngaullier has joined #ffmpeg-devel
Anthony_ZO has joined #ffmpeg-devel
martinr1 has joined #ffmpeg-devel
System_Error has quit [Ping timeout: 264 seconds]
System_Error has joined #ffmpeg-devel
martinr1 has quit [Ping timeout: 246 seconds]
DVedaa has quit [Read error: Connection reset by peer]
DVedaa has joined #ffmpeg-devel
HarshK23 has joined #ffmpeg-devel
abdu18 has joined #ffmpeg-devel
derpydoo has quit [Quit: derpydoo]
abdu18 has quit [Ping timeout: 240 seconds]
lazynorm has joined #ffmpeg-devel
j45 has quit [Ping timeout: 276 seconds]
j45 has joined #ffmpeg-devel
j45 has joined #ffmpeg-devel
Grimmauld has quit [Quit: Client closed]
lazynorm has quit [Quit: Client closed]
faidz has joined #ffmpeg-devel
^Neo has joined #ffmpeg-devel
^Neo has joined #ffmpeg-devel
Grimmauld has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
<averne>
Lynne: I'm doing this here since the GSoC'25 page lists this place as contact point for the VK Prores decoder project. As I wrote a couple days ago I'm interesting in enrolling in it, and as for the qualification tasks I've implemented a basic VK iDCT shader which you can find here: https://github.com/averne/Vulkan-iDCT
<averne>
Is there anything else I need to do/know before drawing up the proposal?
ccawley2011 has joined #ffmpeg-devel
faidz has quit [Quit: #gsoc]
<mkver>
michaelni: Are these new fourccs for raw pixel formats your creation or do they come from some spec?
<mkver>
jamrial: Can you hold off a bit with the bump, please?
<Lynne>
averne: you need to do the actual iDCT that you'll use
<Lynne>
simple_idct_template.c
<ramiro>
haasn: in ff_sws_linear_mask, shouldn't the condition be (c.m[i][j].num && av_cmp_q(c.m[i][j], Q(i == j)))? on yuv2rgb I got an extra multiplication by zero from the mask.
<haasn>
ramiro: the mask specifies which components differ from an identity matrix
DVedaa has quit [Read error: Connection reset by peer]
<haasn>
Consider than explicitly zeroing out a channel is different from leaving that channel untouched, which is the “default” behavior
<haasn>
But we could definitely improve this
<haasn>
Otherwise you’d get an explicit multiply-by-one
DVedaa has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
<averne>
Lynne: I'm not sure I follow you here, I followed the iDCT definition given in the Prores SMPTE paper. Unless you mean the shader should output 10/12-bit integer data?
<averne>
JEEB: that's the document I was looking at, yeah
<averne>
Lynne: Indeed there's a lot of room for fun stuff here, I was planning to dedicate a chunk of the timeline to exploration in the proposal. Maybe coop matrices could be used too
<averne>
In any case, if you're satisfied with my shader in its current state I'll move forward with the proposal. Is it expected that tutors review it before submission?
<Lynne>
sure, the code is fine
<ramiro>
haasn: ok, I check the mask just to select which implementation I should use, then check m.num for each element if I should emit the instruction or not.
<ramiro>
it's progressing really well, I'm getting more and more conversions with a lot of speedup.
<ramiro>
haasn: the few slower ones are ones that I still think should be optimized away to memcpy/memset but that aren't. like yuv444p->yuv444p and yuv444p->yuva444p
<haasn>
neat
<haasn>
I think I will revive the dedicated backend for memcpy/memset
<haasn>
yuv444p->yuv444p? wdym?
<haasn>
that one should be a noop in any case
<haasn>
ramiro: my x86 backend is coming along nicely as well
<ramiro>
haasn: for some reason it isn't (:
<ramiro>
I haven't checked why
<haasn>
it's actually kinda funny how much better avx512 is at pixel type conversion
<haasn>
may end up writing dedicated avx512 variants of those functions down the line
<haasn>
(though ideally a dedicated avx512 backend could use zmm instead of ymm)
<Lynne>
michaelni: done, could you add me?
<ramiro>
oh, that's odd. indeed yuv444p->yuv444p is not using the asmjit backend, but it is slower (?)
<michaelni>
Lynne, added
abdu18 has joined #ffmpeg-devel
<haasn>
ramiro: I mean, there's literally a (src.fmt == dst.fmt) check
<haasn>
to turn the whole SwsGraph into a noop
<haasn>
how are you testing? try with -v debug
abdu18 has quit [Quit: Client closed]
abdu18 has joined #ffmpeg-devel
Grimmauld has quit [Quit: Client closed]
jamrial has quit [Read error: Connection reset by peer]
<jamrial_>
oh, it calls pl_av_stream_get_side_data(), which should follow what you said
<jamrial_>
yeah
<kasper93>
but if it is compiled without Wundef it is possible that missing header causes the check to fail
Marth64 has quit [Remote host closed the connection]
<jamrial_>
another|: you're most likely using a very old libplacebo version
martinr1 has joined #ffmpeg-devel
martinyr has quit [Ping timeout: 248 seconds]
rvalue- has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 246 seconds]
rvalue- is now known as rvalue
<another|>
jamrial_: 7.349.0-4
<another|>
not very old
martinr1 has quit [Ping timeout: 252 seconds]
<another|>
haasn: doesn't seem to fix it
Grimmauld has quit [Quit: Client closed]
IndecisiveTurtle has joined #ffmpeg-devel
<another|>
haasn: tried it again in a clean container. nope
martinr1 has joined #ffmpeg-devel
martinr1 has quit [Ping timeout: 260 seconds]
<jamrial_>
another|: is the correct libavformat/version_major.h being included? No reason for the "LIBAVFORMAT_VERSION_INT >= AV_VERSION_INT(60, 15, 100)" check to fail with git master
<jamrial_>
you may have a system package header being included instead of the git master one
<kasper93>
this is always a pain with libplacebo cyclic dependency, you need to build ffmpeg first without libplacebo, install and rebuild...
martinr1 has joined #ffmpeg-devel
minimal has joined #ffmpeg-devel
<another|>
jamrial_: no ffmpeg installed in the container so I assume it takes from git?
<another|>
I'm trying to build placebo from src
twelve has quit [Remote host closed the connection]
<another|>
okay, tried placebo from git w/ w/o patch. all fails
martinr1 has quit [Ping timeout: 248 seconds]
<jamrial_>
another|: given what kasper93 said, try first compiling ffmpeg with no libplacebo support, install it, then compile libplacebo
twelve has joined #ffmpeg-devel
abdu64 has joined #ffmpeg-devel
<another|>
jamrial_: I do not see how that would make a difference. It doesn't seem like placebo has libav* as a build dep
<another|>
or only for tests, demos
<jamrial_>
the header with the related failing call is installed and depends on libav*
<another|>
yes, but that's just a header. It's copied verbatim. It doesn't require ffmpeg at build time
martinr1 has joined #ffmpeg-devel
<another|>
that patch is broken but I'll manage
martinr1 has quit [Ping timeout: 248 seconds]
<another|>
jamrial_: that fixes it
<jamrial_>
did it at no point print a warning about LIBAVFORMAT_VERSION_INT not being defined?
<jamrial_>
as in, without the patch
<another|>
nope
cone-936 has quit [Quit: transmission timeout]
<jamrial_>
haasn: ^
Grimmauld has joined #ffmpeg-devel
Grimmauld has quit [Quit: Client closed]
faidz has joined #ffmpeg-devel
martinr1 has joined #ffmpeg-devel
<Xe>
new anubis release, it has a feature you may care about: specifying the cookie domain so you can have people that pass the check for trac don't have to pass it for code and vice versa: https://github.com/TecharoHQ/anubis/releases/tag/v1.15.0
<Xe>
you'd also need to set the same ed25519 key on both services
martinr1 has quit [Ping timeout: 244 seconds]
Mirarora has joined #ffmpeg-devel
<kasper93>
-Wundef would help, but even then it should complain about deprecated use for over a year, no?
s55 has quit [Quit: Bye]
s55 has joined #ffmpeg-devel
<faidz>
pross: Hi. Is anyone working on the VP6 encoder project for GSoC? I'm interested in working on it.
abdu64 has quit [Ping timeout: 240 seconds]
<llyyr>
hwcontext_vulkan uses AV_VK_FRAME_FLAG_CONTIGUOUS_MEMORY which just got removed lol
<llyyr>
Lynne: ^
<llyyr>
you can actually probably just remove those lines
<IndecisiveTurtle>
Lol also got a build error regarding this flag after a rebase
<llyyr>
sent a patch
ccawley2011_ has joined #ffmpeg-devel
ccawley2011_ has quit [Remote host closed the connection]
ccawley2011_ has joined #ffmpeg-devel
ccawley2011__ has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 244 seconds]
ccawley2011__ has quit [Remote host closed the connection]
ccawley2011__ has joined #ffmpeg-devel
ccawley2011_ has quit [Ping timeout: 265 seconds]
ccawley2011__ has quit [Ping timeout: 265 seconds]
Anthony_ZO has quit [Remote host closed the connection]
Mirarora has quit [Quit: Mirarora encountered a fatal error and needs to close]
martinr1 has joined #ffmpeg-devel
twelve has quit [Remote host closed the connection]
<BtbN>
jamrial_ haasn https://bpa.st/7XXQ something seems to have broken with libplacebo in the major bump?
<jamrial_>
BtbN: yes, see above
<jamrial_>
the header never included version.h, so LIBAVFORMAT_VERSION_INT was never defined (you can see the warning there)
<jamrial_>
haasn, kasper93: can you apply the fix?
twelve has quit [Remote host closed the connection]
<fflogger>
[editedticket] MasterQuestionable: Ticket #11531 ([ffmpeg] [Regression] AAC audio in RTSP no more muxed with "-c copy") updated https://trac.ffmpeg.org/ticket/11531#comment:5
zulleyy3 has quit [Ping timeout: 252 seconds]
AMM has quit [Ping timeout: 252 seconds]
AMM has joined #ffmpeg-devel
zulleyy3 has joined #ffmpeg-devel
<Lynne>
llyyr: thanks
cone-325 has joined #ffmpeg-devel
<cone-325>
ffmpeg llyyr master:8c7b00ba3ac4: avutil/hwcontext_vulkan: stop checking for deprecated and removed flag