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
iive has quit [Quit: They came for me...]
secondcreek has quit [Quit: secondcreek]
minimal has quit [Quit: Leaving]
xvaclav has quit [Remote host closed the connection]
s55 has quit [Ping timeout: 276 seconds]
s55 has joined #ffmpeg-devel
Guest58 has joined #ffmpeg-devel
Guest58 has quit [Client Quit]
libra811 has joined #ffmpeg-devel
BradleyS has quit [Read error: Connection reset by peer]
BradleyS_ has joined #ffmpeg-devel
BradleyS_ is now known as BradleyS
kasper93 has joined #ffmpeg-devel
kasper93 has quit [Client Quit]
kasper93 has joined #ffmpeg-devel
ramiro has quit [Ping timeout: 252 seconds]
ramiro has joined #ffmpeg-devel
jamrial has quit []
MisterMinister has quit [Ping timeout: 248 seconds]
NotWarcop has joined #ffmpeg-devel
zsoltiv has quit [Read error: Connection reset by peer]
Warcop has quit [Ping timeout: 260 seconds]
zsoltiv__ has joined #ffmpeg-devel
zsoltiv_ has quit [Ping timeout: 252 seconds]
zsoltiv has joined #ffmpeg-devel
libra811 has quit [Quit: Client closed]
BradleyS has quit [Quit: quit]
BradleyS has joined #ffmpeg-devel
secondcreek has joined #ffmpeg-devel
bsFFFFFF has joined #ffmpeg-devel
bsFFFFFF has quit [Ping timeout: 260 seconds]
bsFFFFFF has joined #ffmpeg-devel
zsoltiv has quit [Ping timeout: 252 seconds]
zsoltiv_ has joined #ffmpeg-devel
zsoltiv__ has quit [Ping timeout: 265 seconds]
zsoltiv has joined #ffmpeg-devel
zsoltiv has quit [Ping timeout: 244 seconds]
zsoltiv__ has joined #ffmpeg-devel
zsoltiv_ has quit [Ping timeout: 244 seconds]
zsoltiv has joined #ffmpeg-devel
TheVibeCoder has joined #ffmpeg-devel
blb has quit [Ping timeout: 248 seconds]
blb has joined #ffmpeg-devel
secondcreek has quit [Remote host closed the connection]
secondcreek has joined #ffmpeg-devel
drv has quit [Quit: No Ping reply in 180 seconds.]
drv has joined #ffmpeg-devel
xvaclav has joined #ffmpeg-devel
rvalue- has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 252 seconds]
rvalue- is now known as rvalue
pelotron has quit [Ping timeout: 245 seconds]
pelotron has joined #ffmpeg-devel
<Lynne>
TheVibeCoder: first dequantization clue, there's an offset involved
<Lynne>
coeffs which were quantized to 0 appear non-zero after dequant
<Lynne>
DCs are also definitely offset
<TheVibeCoder>
offset values are consistent?
<Lynne>
let me iterate over all zero coeffs
<Lynne>
nope, DCs are very very weird
<Lynne>
seems like the DCs are offset, but the ACs are not offset
jdarnley has joined #ffmpeg-devel
<Lynne>
DC offset varies, quant zero DC -> dequant ~21 (which happens to be the average offset needed to make the DC values approximately match)
<Lynne>
the average DC offset seems identical between non-hq and hq
<pross>
git grep draw_line :(
bsFFFFFF has quit [Ping timeout: 244 seconds]
bsFFFFFF has joined #ffmpeg-devel
bsFFFFFF has quit [Ping timeout: 276 seconds]
bsFFFFFF has joined #ffmpeg-devel
<kierank>
F to pay respects
indecisiveturtle has quit [Quit: indecisiveturtle]
<beastd>
ok, learn more aboout gitlab: "01 Software faster 02 AI-powered 03 Privacy-first AI ... 10 Open core"
<TheVibeCoder>
AI - first | Privacy - second
<beastd>
TheVibeCoder: nice take, i had the same image in my mind :)
<beastd>
Still trying to find the source via their official website (the number "10 Open core" implies to me parts of it's core software are open source)
<beastd>
Tutorial: Make a GitLab contribution ... Before you begin ... Once your community forks access request is approved you can start using GitLab Duo, our AI-native features including Code Suggestions, Chat, Root Cause Analysis, and more.
<TheVibeCoder>
F.
<TheVibeCoder>
because forgejo is not full with AI crap
<kierank>
doesn't even have non-experimental CI
<beastd>
https://gitlab.com/gitlab-org/gitlab <-- finally found what seems to be source repo of the open core, after painfully reading thru their website and following my best intuition thru lists and links and lists... (also crtl-f was a life saver)
<beastd>
also in every footer there is Commnity column with a direct entry Code which also yields the same result: https://codeberg.org/forgejo/forgejo
NotWarcop has quit [Remote host closed the connection]
iive has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
Traneptora has quit [Quit: Quit]
bsFFFFFF has quit [Read error: Connection reset by peer]
bsFFFFFF has joined #ffmpeg-devel
bsFFFFFF has quit [Ping timeout: 244 seconds]
bsFFFFFF has joined #ffmpeg-devel
<Lynne>
oh hey, now that I think about it
<Lynne>
the version 0 sample is from an hdmi recorder
<Lynne>
those are raw-raw, not just raw, fully linear
<Lynne>
that's why camera manufacturers give you a raw gamut to whateverlogamut they have
<TheVibeCoder>
prores raw are raw raw raw raw linear 12bit
<Lynne>
not linear
<Lynne>
they're whatever gamut they are
<BtbN>
michaelni: how do I actually reproduce the infinite loop with the sample? No matter what I try, it's just flat out unparseable and broken, but not infinite looping
bsFFFFFF has quit [Quit: bsFFFFFF]
indecisiveturtle has quit [Ping timeout: 265 seconds]
<TheVibeCoder>
Lynne: than its 100% lossy
<Lynne>
TheVibeCoder: ?
<TheVibeCoder>
proresraw
<Lynne>
yeah, its lossy
<JEEB>
same as with ye olde prores, after all
<TheVibeCoder>
the remove LOSSLESS flag you added to codec_decs.c
indecisiveturtle has joined #ffmpeg-devel
<Lynne>
nice catch, I didn't even notice it
cone-050 has joined #ffmpeg-devel
<cone-050>
ffmpeg Marton Balint master:71468e85ae72: avfilter/framequeue: add support for limiting and tracking buffered frames in the queues
<kasper93>
main() is provided by -fsanitizer=fuzzer
<kasper93>
LLVMFuzzerTestOneInput() is entry point in user code
<BtbN>
no idea how to work this stuff. I also wonder why the sample it generated doesn't cause any issues when used with the CLI tools
<kasper93>
`--libfuzzer=-fsanitize=fuzzer --enable-ossfuzz` for configure is needed for proper build
<BtbN>
I don't even have llvm/clang installed...
<BtbN>
that's gonna take a few hours
<kasper93>
you can use docker with prebuilt
<BtbN>
Why do I need to jump through so many hoops to get a sample to do what it's supposed to do? my god
<BtbN>
And how real is the thing it found even, when outside of its framework, the issue does not exist?
<kasper93>
ah, nevermind, it needs to be build anyay
<kasper93>
can't tell you without looking at sample, but those fuzzer tools set a lot of parameters based on the binary reproducer file
<kasper93>
it's not like it's your input
<BtbN>
Ah, it isn't just a video file?
<BtbN>
That would explain it
<BtbN>
The issue is relatively obvious once pointed out, so I have no doubt it's real. And I think my patch fixes it
<BtbN>
"-fsanitize-coverage=trace-pc-guard is no longer supported by libFuzzer. Please either migrate to a compiler that supports -fsanitize=fuzzer or use an older version of libFuzzer"
<BtbN>
well
<BtbN>
turns out I do have llvm-20 installed. But it seems to be too new?