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 7.0.2 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
thilo has quit [Ping timeout: 245 seconds]
thilo has joined #ffmpeg-devel
lexano has quit [Ping timeout: 255 seconds]
arch1t3cht0 has joined #ffmpeg-devel
arch1t3cht has quit [Ping timeout: 276 seconds]
arch1t3cht0 is now known as arch1t3cht
AbleBacon has joined #ffmpeg-devel
IndecisiveTurtle has quit [Ping timeout: 264 seconds]
cone-165 has quit [Quit: transmission timeout]
mkver has quit [Ping timeout: 260 seconds]
jamrial has quit []
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 248 seconds]
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
microchip__ has joined #ffmpeg-devel
microchip_ has quit [Ping timeout: 244 seconds]
microchip__ is now known as microchip_
System_Error has quit [Remote host closed the connection]
j45_ has joined #ffmpeg-devel
j45 has quit [Ping timeout: 260 seconds]
j45_ is now known as j45
j45 has quit [Changing host]
j45 has joined #ffmpeg-devel
AbleBacon has quit [Read error: Connection reset by peer]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
cone-455 has joined #ffmpeg-devel
<cone-455>
ffmpeg Fei Wang master:be7ab635522a: lavc/qsvdec: Add vvc_mp4toannexb bsf for QSV VVC decoder
System_Error has joined #ffmpeg-devel
ChanServ has quit [shutting down]
ChanServ has joined #ffmpeg-devel
bilboed has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
IndecisiveTurtle has joined #ffmpeg-devel
derpydoo has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
ccawley2011 has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
cone-455 has quit [Quit: transmission timeout]
IndecisiveTurtle has quit [Ping timeout: 260 seconds]
<BtbN>
I can't easily test that right now, just saw CI explode with that error
<another|>
please don't use pastebin.com especially not for code. it changes all new lines to crlf
<Lynne>
lol
<jamrial>
BtbN: eh, i'll push it. it should work
<cone-064>
ffmpeg James Almer master:211c88b9d5f1: avfilter/f_zmq: fix graph argument
<elenril>
jamrial: thanks
<elenril>
guess I should install zmq
* Lynne
stares at hw_base_encode
<Lynne>
why do I get the feeling I'll be maintaining or rewriting this?
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
IndecisiveTurtle has quit [Ping timeout: 252 seconds]
<BtbN>
jamrial: it built successfully
CATS has quit [Quit: Lost terminal]
darkapex_ has quit [Remote host closed the connection]
darkapex_ has joined #ffmpeg-devel
<elenril>
also, wtf is a zmq filter and why does it exist
<JEEB>
IIRC IPC to send messages to filter chain?
<elenril>
which obviously needs to be a filter
<elenril>
and we need a filter for any kind of IPC out there
rvalue- has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 258 seconds]
<BtbN>
My understanding is that you just throw that filter into the chain, and the commands go to what's behind it
<BtbN>
It's a filter so you can place it accordingly I guess
<elenril>
still sounds like something that should be in the caller
<BtbN>
the caller?
<elenril>
lavfi caller
<BtbN>
Could probably move it into ffmpeg.c, but I guess this way you can also use it in the likes of OBS
rvalue- is now known as rvalue
<Lynne>
the lavfi command API is straightforward, though
<elenril>
and super broken
Krowl has quit [Read error: Connection reset by peer]
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 276 seconds]
<Lynne>
wow
<Lynne>
nvidia's 30x0 GPUs have buggy hardware decoders
<Lynne>
once in a blue moon, the decode output buffers might end up in an address where they can't be reached by the decoder
<Lynne>
they "fixed" that for vulkan by requiring that the *output* images have to be layered, e.g. allocated all at the same time to avoid fragmenting them
<Lynne>
which I didn't even know its possible
<Lynne>
obviously we can't change our API, I don't mind it, so we'll have to do gpu memcpys nvdec-style in our decoder for those particular cards
cone-064 has quit [Quit: transmission timeout]
Traneptora has quit [Quit: Quit]
<BtbN>
Lynne: nvdec/cuviddec has the exact same issue
<BtbN>
you need to tell it the amount of output pictures in advance, and the max is 32
<BtbN>
It has that exact limitation for much much longer though
<Lynne>
right, but wasn't that an API limitation?
Traneptora has joined #ffmpeg-devel
<BtbN>
The entire API is designed like that, yeah
<BtbN>
but not like it couldn't be changed
<BtbN>
the current way is hyper annoying, since you need to know the exact amount of output frames you need ahead of time
<BtbN>
and depending on what you're doing, 32 also is far from enough
<BtbN>
nvenc alone with reordering and look-ahead stuff can exhaust it
<BtbN>
If you guess wrong in either direction, it sucks
<BtbN>
cause you either waste vram or crash
jarthur has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
lemourin has quit [Read error: Connection reset by peer]
lemourin has joined #ffmpeg-devel
AbleBacon has joined #ffmpeg-devel
* Sean_McG
peeks in
Krowl has quit [Read error: Connection reset by peer]
<Lynne>
haasn: til, requiresDedicatedAllocation == "prefer", prefersDedicatedAllocation == spam letter titled "would you kindly consider using dedicated allocation?"
<Lynne>
the only place where dedicated allocation is actually required is when importing external images (dmabufs/d3d1x/android)
<Lynne>
classic overengineering which requires you to sphagetti contexts around
<haasn>
huh really?
<Lynne>
yeah, its one of those double-safety features of the spec
<Lynne>
the spec says that you can't pass unknown structs to drivers, and drivers are also required to ignore unknown structs
<Lynne>
the spec says that requiresDedicatedAllocation == require, but then they hard-specify where dedicated allocation is actually required in the spec language for each external memory type
<haasn>
I guess that would explain why the best practices layer happily complains about you giving dedicated (small) allocations to an image with requiresDedicatedAllocation = true
<Lynne>
only android and qnx specify that they require dedicated allocation when importing external memory
Sean_McG has quit [Quit: leaving]
mkver has joined #ffmpeg-devel
System_Error has quit [Ping timeout: 260 seconds]
compnnn has joined #ffmpeg-devel
compnn has quit [Read error: Connection reset by peer]
System_Error has joined #ffmpeg-devel
cubicibo has joined #ffmpeg-devel
cubicibo has quit [Client Quit]
IndecisiveTurtle has joined #ffmpeg-devel
witchymary has joined #ffmpeg-devel
ccawley2011 has quit [Read error: Connection reset by peer]
System_Error has quit [Remote host closed the connection]
cone-180 has quit [Quit: transmission timeout]
System_Error has joined #ffmpeg-devel
IndecisiveTurtle has quit [Ping timeout: 265 seconds]
darkapex_ has quit [Remote host closed the connection]
darkapex_ has joined #ffmpeg-devel
lexano has quit [Ping timeout: 255 seconds]
System_Error has quit [Remote host closed the connection]