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
<fflogger>
[newticket] lcyf-fizz: Ticket #11483 ([ffmpeg] Segmentation fault on function mov_read_trak) created https://trac.ffmpeg.org/ticket/11483
<fflogger>
[newticket] lcyf-fizz: Ticket #11484 ([ffmpeg] Segmentation fault on function av_log_default_callback) created https://trac.ffmpeg.org/ticket/11484
<llyyr>
Lynne: when building mpv with threadsanitizer and using hwdec=vulkan, I get a bunch of the following warnings. Is this noise and I should ignore it? ThreadSanitizer: lock-order-inversion (potential deadlock) /home/llyyr/workspace/ffmpeg/./libavutil/vulkan.c:731:5 in ff_vk_exec_add_dep_frame
<Lynne>
I don't know whether it's noise or not, but I seem to remember it may be false
<llyyr>
i se
<compnn>
wow i forgot all about lowres
thilo has quit [Ping timeout: 248 seconds]
thilo has joined #ffmpeg-devel
cone-811 has quit [Quit: transmission timeout]
sparsh-n has quit [Quit: Client closed]
_av500_ has joined #ffmpeg-devel
av500 has quit [Ping timeout: 260 seconds]
cone-425 has joined #ffmpeg-devel
<cone-425>
ffmpeg Michael Niedermayer release/7.1:9dea077922bc: avformat/hls: Print input format in error message
<cone-425>
ffmpeg Michael Niedermayer release/7.1:b753bac08f68: avformat/hls: Be more picky on extensions
<cone-425>
ffmpeg Michael Niedermayer release/7.1:880007e75cf5: avformat/hls: .ts is always ok even if its a mov/mp4
<cone-425>
ffmpeg Michael Niedermayer release/7.1:c74740fcd543: avcodec/h263dec: Check against previous dimensions instead of coded
<cone-425>
ffmpeg Michael Niedermayer release/7.1:4018b915a63f: avcodec/vvc/refs: fix negative pps_scaling_win offsets
<cone-425>
ffmpeg Michael Niedermayer release/7.1:6c4e56f07d1a: libavformat/hls: Be more restrictive on mpegts extensions
<cone-425>
ffmpeg Michael Niedermayer release/7.1:43ba995a31d8: avformat/hls: Fix twitter
<cone-425>
ffmpeg Michael Niedermayer release/7.1:c388bc1774f0: avformat/mxfdec: Check edit unit for overflow in mxf_set_current_edit_unit()
<cone-425>
ffmpeg Michael Niedermayer release/7.1:6acfaa1084fb: avformat/wavdec: Fix overflow of intermediate in block_align check
<cone-425>
ffmpeg Michael Niedermayer release/7.1:2e7364727caa: avformat/mlvdec: fix size checks
<cone-425>
ffmpeg Michael Niedermayer release/7.1:0e5b6a715627: avcodec/aac/aacdec: Clear SFO on error
<cone-425>
ffmpeg Michael Niedermayer release/7.1:bf8c0be97135: avcodec/aac/aacdec_usac: Simplify decode_usac_scale_factors()
<cone-425>
ffmpeg Michael Niedermayer release/7.1:6646dd2825ef: avcodec/aac/aacdec_lpd: Limit get_unary()
<cone-425>
ffmpeg Michael Niedermayer release/7.1:d3929313917f: avformat/iff: Check that we have a stream in read_dst_frame()
<cone-425>
ffmpeg Michael Niedermayer release/7.1:3e16202c39b6: Changelog: update
sparsh-n has joined #ffmpeg-devel
jamrial has quit [Read error: Connection reset by peer]
jamrial_ has joined #ffmpeg-devel
sr55 has joined #ffmpeg-devel
s55 has quit [Ping timeout: 252 seconds]
jamrial_ has quit []
c1480 has joined #ffmpeg-devel
^Neo has quit [Ping timeout: 260 seconds]
s55 has joined #ffmpeg-devel
sr55 has quit [Ping timeout: 252 seconds]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 244 seconds]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
bhaskar has quit [Read error: No route to host]
bhaskar has joined #ffmpeg-devel
cone-425 has quit [Quit: transmission timeout]
bhaskar has quit [Quit: Konversation terminated!]
zeezie23 has quit [Ping timeout: 260 seconds]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
sparsh-n has quit [Quit: Client closed]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
_av500_ is now known as av500
HarshK23 has joined #ffmpeg-devel
Teukka has quit [Read error: Connection reset by peer]
<fflogger>
[newticket] krs_vital: Ticket #11485 ([ffmpeg] Unable to choise GPU on Nvidia A16 card with 4 chips on board) created https://trac.ffmpeg.org/ticket/11485
LainExperiments has joined #ffmpeg-devel
LainExperiments5 has joined #ffmpeg-devel
LainExperiments has quit [Ping timeout: 240 seconds]
rvalue has quit [Ping timeout: 248 seconds]
LainExperiments5 has quit [Ping timeout: 240 seconds]
<j-b>
thardin: FFMPEG fast seek behavior works fine if a fragmented MP4 file has an SIDX box right after the MOOV header, but we get a lot of uploaded fMP4 files that only have a mfra box and no SIDX boxes. These are mostly meeting recordings or screen capture outputs.
cone-901 has quit [Quit: transmission timeout]
realies has quit [Ping timeout: 244 seconds]
System_Error has quit [Remote host closed the connection]
<ExtremeLeader>
dear Leadership:
ExtremeLeader has quit [Quit: Client closed]
System_Error has joined #ffmpeg-devel
realies has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
ccawley2011_ has joined #ffmpeg-devel
abdu69 has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 252 seconds]
LainExperiments has quit [Quit: Client closed]
Guest90 has joined #ffmpeg-devel
<Guest90>
hi
Guest90 has quit [Client Quit]
LainExperiments has joined #ffmpeg-devel
LainExperiments9 has joined #ffmpeg-devel
cone-251 has joined #ffmpeg-devel
<cone-251>
ffmpeg Ulrik master:95314cd7c5c0: avformat/flacdec: Return correct error-codes on read-failure
<Traneptora>
ffmpeg Peter Zebühr master:e24920375cd2: Make mime-type award a bonus probe score
<Traneptora>
this one's also missing a prefix like avformat/probe: foo or something
<RadioLeader>
years passed and still no sign of new version of ac-4 specification in pdf
<another|>
Lynne: With ANMR gone the indexes for `aac_coder` has shifted. I expect this'll break some people setup where they erroneously use the int instead of the name.
<another|>
s/has/have/
<Lynne>
the option itself will be gone soon enough
<another|>
well, not break but using fast instead of twoloop
<RadioLeader>
using ints instead of names is not allowed
RadioLeader has quit [Quit: Client closed]
Tanay has quit [Quit: Leaving]
BaggyBro has joined #ffmpeg-devel
<BaggyBro>
Hi, I’m new here and interested in contributing to FFmpeg!
<ePirat>
Welcome
<BaggyBro>
Thanks! Where should I start as a beginner? Any recommended tasks or docs to read?
<ePirat>
BaggyBro, but the obvious first step would be to clone the repo and build from source if you havent done that yet
<BaggyBro>
yes surely I'll do so
Everything has joined #ffmpeg-devel
<welder>
I'm reading through neon implementation and template of fft. What does FF_TX_PRESHUFFLE do? It indicates that the routine expects the even elements followed by odd ones instead of interleaved?
ccawley2011_ has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
ccawley2011__ has quit [Ping timeout: 244 seconds]
ccawley2011_ has quit [Ping timeout: 260 seconds]
<cone-251>
ffmpeg Chris Hodges master:330c8f8b936d: avformat: add AV1 RTP depacketizer and packetizer
BaggyBro has quit [Ping timeout: 240 seconds]
Everything has quit [Quit: leaving]
Everything has joined #ffmpeg-devel
DauntlessOne4 has quit [Remote host closed the connection]
<fflogger>
[editedticket] axfelix: Ticket #11352 ([avcodec] On the state of FFmpeg's ffv1_vulkan encoder implementation on Intel & NVIDIA) updated https://trac.ffmpeg.org/ticket/11352#comment:2
ccawley2011__ has quit [Ping timeout: 260 seconds]
paulk-bis has joined #ffmpeg-devel
paulk has quit [Read error: Connection reset by peer]
Guest40 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
Guest40 has joined #ffmpeg-devel
zeezie23 has quit [Quit: Leaving]
zeezie01 has joined #ffmpeg-devel
IndecisiveTurtle has quit [Ping timeout: 268 seconds]
<cone-251>
ffmpeg Timo Rothenpieler master:99e2af4e7837: avcodec/cuviddec: correctly handle buffer size and status when deinterlacing
<BtbN>
Forgejo instancen can mail again now. Very weird issues. It seems like an issue with leveldb's queue binary format being slightly different between x86_64 and aarch64, so when migrating, the queue DB had to be deleted. Or something else broke, and the queue-reset fixed it.
<kierank>
gotta love the forgejo jank
<BtbN>
yes, cause leveldb is clearly "Forgejo Jank"...
<kierank>
forgejo is inevitably gonna be run with a bunch of scripts that only one person understands
<kierank>
whereas we could have videolan with an actual responsive sysadmin
<kierank>
and actual infrastruture
<BtbN>
Are you trying to insult me with that or why do you keep going on about responsible and actual sysadmins elsewhere?
<kierank>
no, I appreciate what you are doing
<kierank>
but it mirrors exactly the problem we have in ffmpeg right now
<kierank>
that only one person understands how things work
<kierank>
and there is no transparency
<BtbN>
There's at least 3 active sysadmins of current FFmpeg infra, and two kinda retired ones. All of which can maintain what currently exists.
<kierank>
lol
<kierank>
the ml is broken like every other day
<BtbN>
Is it now?
<BtbN>
And it mainly breaks because Google and Microsoft decided to implement new great anti-"spam" measures
<kierank>
we could have a proper setup in videolan, not some random server that nobody knows where it is
<kierank>
and aaarch64 for some reason?
<BtbN>
It's not a random server, and everyone knows where it is...
<kierank>
lol
<kierank>
this is exactly like x265 choosing hg to be contrarian
<BtbN>
I feel like you're hunting non-existent windmills
<kierank>
I wasn't even able to sign up to forejgo succesffuly
<kierank>
and neither was frank
<BtbN>
Yes, cause of aforementioned issue
<BtbN>
And last I checked you signed up fine?
<kierank>
no
<kierank>
never got the email
<BtbN>
So I manually approved you
<kierank>
needed manual enabling
<kierank>
omg this is just insane
<BtbN>
And then worked on fixing the issue
<BtbN>
It's a test instance for a reason
<kierank>
this is all stuff created by forejo
<kierank>
and then you are unable to take criticism of that
<kierank>
in videolan gitlab there isn't this
<BtbN>
It's stuff created by me migrating it the way I did
<BtbN>
If you did the same to your beloved Gitlab, good chance something similar would happen
<kierank>
exactly, some jank thing one person has setup and then we will need to depend on for ten years
<kierank>
this is not how a mature open source project should operate
<BtbN>
Nothing jank about it.
<kierank>
omg
<BtbN>
Repeting it over and over again does not suddenly make it true.
<kierank>
saying "i'm fixing it" doesn't make it not true
<kierank>
it's like all the sane people left this project
<BtbN>
You realize how many weird and critical bugs Gitlab has on the regular, right?
<BtbN>
On the Jank-Scale, Gitlab very much beats Forgejo for me
<jamrial>
kierank: nothing is ever going to be definite. we can move now to forgejo in our own server an later to whatever in some other server
<kierank>
we need gitlab for proper CI
<kierank>
on many targets, with logs
<BtbN>
Forgejos CI is not "proper CI"?
<kierank>
no
<BtbN>
Why?
<jamrial>
instead of complaining it's not done the way you like or want from the start, at least don't try to block an actual move actually happening for once
<kierank>
in a few years it'll be broken and unmaintained like trac
<kierank>
it's very obvious to see that
<kierank>
when forjego eventually forks
<kierank>
I know this from experience
<kierank>
having moved from gogs to gitea and now being told we have to move to forjego
<BtbN>
seems like you're doing fine if you can just move on like that?
<kierank>
no it's a nightmare
<kierank>
having to do these migrations
<kierank>
with random scripts and stuff
<BtbN>
You can also just stay on Gitea, it's far from dead
<RadioLeader>
gitlab is backed by corpos
<kierank>
that's not a mature CI system
<jamrial>
kierank: in the upcoming vote, you can vote gitlab
<jamrial>
and we can also run a gitlab test instance after this forgejo one
<kierank>
all the people who understand the reasons why forjego are bad have left
<kierank>
it's just pure contrarianism
<BtbN>
Gotta go explain how bad Forgejo is to an increasing number of Projects as well then.
<RadioLeader>
i miss old, great, fame days of mplayer
<kierank>
ffmpeg is a big project run like someone's hobby
<BtbN>
The literal only issue I have with Forgejo is that the name is slightly odd
<kierank>
videolan is like a different universe
<RadioLeader>
videolan is dead
<BtbN>
Well, we have a significantly higher release output than videolan, so we can't be doing things THAT wrong :D
<RadioLeader>
have more dead and unmaintained projects than google
<jamrial>
doubt it
<kierank>
absolutely wild, after all this argument about infrastructure, we just continue going on
<jamrial>
kierank: mirrored in our own server, wiht development on an ml not in videolan. moving to gitlab would mean development would happen exclusively there
<jamrial>
kierank: i'm with you, i'd rather move to videolan's infra, but that wish is what has kept the project in its current git+ml state for so long
<kierank>
if we have a vote and that wins
<kierank>
then what
<RadioLeader>
goto gstreamer;
<BtbN>
Then we gotta convince michael really
<jamrial>
then it should happen, assuming the democracy is real and not just a facade
<kierank>
RadioLeader: personally librempeg is looking better than ffmpeg
<RadioLeader>
deMiNicracy is real
<BtbN>
Honestly, I'm not too big a fan of the videolan gitlab. It's quite slow and signing up there requires you to know an admin to activate your account. Which at the moment nearly completely locks out outside contribution to Videolan.
<BtbN>
At least from new contributors
<kierank>
huh, but needing you to activate my account was somehow ok?
<jamrial>
during a test instance...
<BtbN>
You realize the difference between the mail system being broken and it being intentionally set up like that?
<kierank>
but this, but that
<kierank>
Forgejo is such brittle stuff
<kierank>
but it's what this project deserves
<jamrial>
kierank: please, nothing has been decided. you're acting like shit is done and there's no going back
<BtbN>
Do you even hear yourself right now? Sorry to say this, but you do not sound reasonable to me.
<kierank>
I'm trying to speak for those who left like derek
<jamrial>
speak what? complaining about a decision that has not been made?
<kierank>
that the push to forejgo out of pure contrarianism is bad for the long term of the project
<BtbN>
It's a push to _anything_ new
<BtbN>
I literally don't care anymore to WHAT we move, as long as we finally do move
<BtbN>
Anything is better than the current ML
<jamrial>
there's about 50 people that can vote, i'm sure that if the push is contrarianism from a small amount of people, then that choice will lose
<kierank>
the people that left won't vote
<kierank>
this is pure selection bias
<kierank>
"The inmates running the asylum" so to speak
<kierank>
(that's a figure of speech)
<jamrial>
last committe votes got about 2/3rds of eligible people voting
<kierank>
not saying anyone is an inmate
<jamrial>
derek could have voted, but he explicitly asked not to be included in the list
<jamrial>
anyone here who contributed to dav1d will probably choose gitlab because they are used to it
DauntlessOne4 has joined #ffmpeg-devel
<another|>
gitlab is okay. it has it's problems.
<another|>
I think ffmpeg would fare better with gitlab than forgejo but anything is better at this point
<another|>
Then again: I have not used forgejo and gitea only briefly, years ago
<BtbN>
For FFmpeg needs, I see no difference in what Forgejo and Gitlab offers. Gitlab almost seems bloated for our needs.
<BtbN>
And Gitlab really has quite a poor track record for security issues
<RadioLeader>
i think that part is just because nobody looking at Forgejo
<BtbN>
If you used Codeberg before. That's Forgejo.
<BtbN>
And if you used Github before, Forgejo/Gitea is pretty dang close to its UI
<RadioLeader>
so codeberg security is rock solid?
iive has joined #ffmpeg-devel
<BtbN>
It could just be an issue of more people looking at Gitlab
<RadioLeader>
yes, that i mean
<BtbN>
But Gitlab is just really dang big and complex, while Forgejo is all in all still a relatively small codebase
<RadioLeader>
yea
<BBB>
jamrial: you're right about that last one :)
<BtbN>
Gitlab is also Ruby iirc, which is not exactly very efficient or fast
<BtbN>
So hosting it requires quite a bit more resources
<BBB>
but I giggled at paul's deminicracy
Guest40 has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<BBB>
you guys are amazing
<BBB>
I think we've all got good spirits about anything-but-ml
<RadioLeader>
watch it! FFmpeg is MiNi's DoMiNion
<RadioLeader>
can I ask Mike to give me blog place at leader.multimedia.cx ?
<RadioLeader>
i have no idea what to write, but who knows?
Everything has quit [Quit: leaving]
System_Error has quit [Remote host closed the connection]
abdu1 has joined #ffmpeg-devel
abdu69 has quit [Ping timeout: 240 seconds]
System_Error has joined #ffmpeg-devel
RadioLeader has quit [Quit: Client closed]
zeezie01 has quit [Ping timeout: 252 seconds]
DauntlessOne4 has quit [Remote host closed the connection]
Guest40 has joined #ffmpeg-devel
HarshK23 has quit [Quit: Connection closed for inactivity]