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 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
System_Error has joined #ffmpeg-devel
bhaskar has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
IndecisiveTurtle has quit [Ping timeout: 248 seconds]
System_Error has joined #ffmpeg-devel
iive has quit [Quit: They came for me...]
ChippyTech has joined #ffmpeg-devel
minimal has quit [Quit: Leaving]
BradleyS has quit [Quit: quit]
bhaskar has quit [Quit: Konversation terminated!]
BradleyS has joined #ffmpeg-devel
ChippyTech has quit [Quit: Client closed]
thilo has quit [Ping timeout: 244 seconds]
BradleyS has quit [Quit: quit]
thilo has joined #ffmpeg-devel
thilo has quit [Changing host]
thilo has joined #ffmpeg-devel
BradleyS has joined #ffmpeg-devel
cone-107 has joined #ffmpeg-devel
<cone-107>
ffmpeg Lynne master:5cf4f145adde: ffv1dec: set f->state_transition for default range coder table
<cone-107>
ffmpeg Lynne master:542a567d504b: ffv1enc_vulkan: support default range coder tables
<cone-107>
ffmpeg Lynne master:b6bf568a44ac: vulkan: don't query video queue properties unless the extension is enabled
<aaabbb>
what kind of performance are you getting with ffv1 vulkan?
<Lynne>
can't properly benchmark, since I can't *not* let the frame get autodownloaded
<Lynne>
I get "[vf#0:0 @ 0x559ff562c3c0] Reconfiguring filter graph because hwaccel changed" and then "Impossible to convert between the formats supported by the filter 'Parsed_null_0' and the filter 'auto_scale_0'"
<Lynne>
what the hell, FFHWAccel.init gets called on every single frame
<Lynne>
WHY
ChippyTech has joined #ffmpeg-devel
Warcop has joined #ffmpeg-devel
ChippyTech has quit [Quit: Client closed]
<aaabbb>
bug in the resampler? aresample=resampler=soxr:cutoff=X seems to do nothing (unlike resampler=swr), cutoff frequency is not applied
<Lynne>
no
<Lynne>
I can't figure out ff_get_format()
<Lynne>
codecs do their own caching of the pixel format and I have no idea why
<aaabbb>
no i mean it might be a separate thing (unrelated to vulkan convo)
<Lynne>
no
<Lynne>
ah
<aaabbb>
"ffmpeg -f lavfi -i anoisesrc=d=10:r=44100:seed=1,aresample=resampler=$r:cutoff=$n -f md5 -" with r=swr, the md5 is different even for n=0.600 and n=0.601. but with r=soxr, even n=0.4 and n=0.8 return the same md5. maybe i'm misunderstanding but this seems like a bug in aresample?
<aaabbb>
(also add :osr=48000 to that filter, forgot that when typing it out on irc)
System_Error has quit [Remote host closed the connection]
<Lynne>
fixed it
<Lynne>
speed is massive
<aaabbb>
Lynne: enough for real-time 4k?
<Lynne>
lol, and then some
<aaabbb>
haha awesome
<Lynne>
what do you expect, the vulkan encoder can do 60fps on 4k
<aaabbb>
similar compression efficiency as software ffv1?
System_Error has joined #ffmpeg-devel
<Lynne>
within a few percent
<aaabbb>
nice then i guess any time real space is needed, it could be 2pass'd after with software ffv1
<Lynne>
I get 85fps on 4k yuv420p on a 6900xt
<Lynne>
hahaha I get 280fps on 4k yuv420p on a 6000 ada
<Lynne>
best all my CPU cores can do is 56fps
<aaabbb>
what about 4k yuv422p?
<compnn>
Lynne, what would be the maximum bandwidth across the pcie ?
<compnn>
probably something we could never reach but i'm curious
<compnn>
128gb/s
<compnn>
yeah nevermind
<Lynne>
the stream I'm testing with is around 1.3 Gbps
jamrial has quit []
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
ramiro has quit [Ping timeout: 260 seconds]
ramiro has joined #ffmpeg-devel
BradleyS has quit [Quit: quit]
BradleyS has joined #ffmpeg-devel
BradleyS has quit [Quit: quit]
zsoltiv_ has quit [Ping timeout: 245 seconds]
zsoltiv has quit [Ping timeout: 248 seconds]
BradleyS has joined #ffmpeg-devel
^Neo has quit [Ping timeout: 252 seconds]
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 244 seconds]
BradleyS has quit [Quit: quit]
BradleyS has joined #ffmpeg-devel
cone-107 has quit [Quit: transmission timeout]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
rvalue- has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 248 seconds]
rvalue- is now known as rvalue
bhaskar has joined #ffmpeg-devel
tufei__ has quit [Remote host closed the connection]
tufei__ has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
bhaskar has quit [Quit: Konversation terminated!]
j45_ has joined #ffmpeg-devel
j45 has quit [Ping timeout: 248 seconds]
j45_ is now known as j45
j45 has joined #ffmpeg-devel
j45 has quit [Changing host]
System_Error has joined #ffmpeg-devel
psilokos has quit [Read error: Connection reset by peer]
psilokos has joined #ffmpeg-devel
mkver has quit [Ping timeout: 260 seconds]
mkver has joined #ffmpeg-devel
mkver has quit [Ping timeout: 272 seconds]
mkver has joined #ffmpeg-devel
tufei_ has joined #ffmpeg-devel
tufei__ has quit [Ping timeout: 264 seconds]
mkver has quit [Ping timeout: 248 seconds]
mkver has joined #ffmpeg-devel
<mkver>
who needs a vp6 encoder?
wyatt8740 has quit [Ping timeout: 260 seconds]
^Neo has joined #ffmpeg-devel
^Neo has joined #ffmpeg-devel
^Neo has quit [Changing host]
jamrial has joined #ffmpeg-devel
<JEEB>
(45
<JEEB>
whoops
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
wyatt8740 has joined #ffmpeg-devel
IndecisiveTurtle has joined #ffmpeg-devel
IndecisiveTurtle has quit [Client Quit]
tanay-man has quit [Remote host closed the connection]
<haasn>
make libswscale/tests/swscale && libswscale/tests/swscale -src gray -dst grayf32l
<haasn>
le*
microlappy has joined #ffmpeg-devel
microlappy has quit [Client Quit]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
MyNetAz has quit [Remote host closed the connection]
MyNetAz has joined #ffmpeg-devel
microlappy has joined #ffmpeg-devel
microlappy has quit [Quit: Konversation terminated!]
microlappy has joined #ffmpeg-devel
microlappy has quit [Client Quit]
<ePirat>
Why does initing a hwdevice needs to allocate a pool, thats somewhat inconvenient on macOS as I need to know a pixel format and sizes beforehand for no good reason?
<ePirat>
frameworks like Videotoolbox or ScreenCaptureKit manage their own pool which in case of the latter you can not even access as it lives in the WindowServer
<ePirat>
so allocating a pool that sits around doing nothing just to be able to pass around the hw frames seems weird
<fflogger>
[editedticket] teckchris: Ticket #11468 ([undetermined] HEVC parsing breaks on some files "Error processing packet in decoder: File exists") updated https://trac.ffmpeg.org/ticket/11468#comment:1
mkver has quit [Ping timeout: 248 seconds]
Warcop has quit [Remote host closed the connection]
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
wyatt8740 has quit [Ping timeout: 252 seconds]
jdek has joined #ffmpeg-devel
iive has joined #ffmpeg-devel
<fflogger>
[editedticket] MasterQuestionable: Ticket #11468 ([avcodec] [Regression] HEVC decoding error for those of image based subtitles) updated https://trac.ffmpeg.org/ticket/11468#comment:2
wyatt8740 has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
Flat_ has joined #ffmpeg-devel
Flat has quit [Ping timeout: 260 seconds]
ccawley2011 has quit [Ping timeout: 252 seconds]
<steven-netint>
ePirat, I'm curious about that too. Its a a bit inflexible to require application control memory on hardware. Maybe its a legacy convention established by the first vendor HWcontexts
<ePirat>
I am not that familiar with other hwaccels, I guess it makes more sense on other more low-level ones.
<ePirat>
(like when you are dealing with Metal directly for example you would probably maintain your own texture cache)
<Lynne>
allocate a pool? what?
<ePirat>
Lynne, AVHWFramesContext is essentially a frame pool abstraction from what I can see
<Lynne>
yes, and?
<ePirat>
so when I want to hand out VIDEOTOOLBOX hwframes, I need to create one even though its entirely useless
<Lynne>
it isn't useless, it pools frames
<ePirat>
it doesnt?
<ePirat>
how would it pool frames if it does not get them from the framepool it abstracts
<Lynne>
of course it does, that's where all filters, decoders and encoders get their frames
<Lynne>
what framepool it abstracts? it *is* the frame pool
<BtbN>
If you want to supply externally provided frames, you still need to prove a hwframesctx, that abstracts that source
<Lynne>
yup, for external frames its basically a descriptor of what the frames contain
ccawley2011 has joined #ffmpeg-devel
<ePirat>
right, IIUC what it does in case of Videotoolbox, it gets the AVFrame and immediately releases the hwframe in there and replaces it with its own, which seems absurd
<Lynne>
yeah, you don't need to do any of this
<ePirat>
so whats the correct way to do this
<Lynne>
just ref the context onto the avframe->hw_frame_ctx
<ePirat>
Lynne, right but I still need a hwframe context which is essentially a pool, and the hwaccel abstraction in ffmpeg allocates frames from the pool for me?
<Lynne>
yes, but it doesn't alloc any frames or data apart from the bufferref and the structs
<Lynne>
if you don't want to allocate frames from it, literally just don't
ccawley2011 has quit [Ping timeout: 276 seconds]
<ePirat>
but it does call my contexts get_buffer implementation, no?
<ePirat>
which does the av_buffer_pool_get()
<Lynne>
no
MetaNova has quit [Excess Flood]
MetaNova has joined #ffmpeg-devel
<ePirat>
Lynne, oh I see, thanks
<ePirat>
Lynne, in the sample code you linked, why do you pass f to av_buffer_create?
<fflogger>
[newticket] emcodem: Ticket #11480 ([undetermined] Support or ignore Iphone16 spatial Apple apac audio) created https://trac.ffmpeg.org/ticket/11480