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
<BtbN>
We blocked /log, cause the git log is expensive
<Marth64>
sorry :(
<Marth64>
oh wow that makes sense
<BtbN>
hopefully an easy fix by adding a trailing /
<Marth64>
(random suggestion) I noticed it also spills out which version of nginx its running. if you're in nginx conf's, you can `server_tokens off;` to hide that if you choose
<michaelni>
and i only notice it now when the page should already be in shape
<Marth64>
michaelni, in spirit I want to and have thought of it. but I don't want to overcommit and let anyone down :) thank you for asking
<michaelni>
sure, thx anyway
<Marth64>
thank you likewise
<Marth64>
the Closed Captions work would have been good.. I would have to very quickly come up with a plan though and align with some folks who contacted me willing to help as well, and I can't rush a good strategy today itself
<BtbN>
I'm writing a thesis myself this GSOC period, and we are moving at work. So mentoring GSOC is sadly too much
<michaelni>
Lynne, you mentored some projects last year, do you want to mentor something in gsoc 2025 ?
<michaelni>
BtbN, sure, thx
<Marth64>
deadline is today right?
<michaelni>
haasn, you where backup mentor in 2023, are you interrested in mentoring something in 2025 ?
<Marth64>
(for plan)
<michaelni>
Marth64, for the page technically, yes
<michaelni>
google will review in the next days we dont know when
<michaelni>
so we statistically probably stil have a little time
<michaelni>
if you know anyone who maybe wants to mentor something, please ask them!
<michaelni>
wbs, are you interrested in mentoring something in gsoc ? :)
<Marth64>
will do
<Marth64>
if we truly have nothing, possibly we can assign CLI fixes and that could be low touch to mentor (for example: fix the ffpresets issue, or add ability to set container crop flags).
<Marth64>
other option is a broader "help us fix closed captions" plan, with some text I can repurpose, and if I can get some help (maybe devenheitmueller ?) we can tag team this. but a strategy has to be made
<haasn>
michaelni: until when is the project submission deadline?
<haasn>
I might have some swscale simd work lined up soon (tm)
<michaelni>
we have submitted ffmpeg already weeks ago but our page is in bad shape
<haasn>
After I land the first part of my rewrite-everything series
<Marth64>
^^ haasn's work sounds lucrative and interesting for college students
<michaelni>
cosminaught, you where mentor and backup mentor in gsoc2024, are you interrested in being (backup)mentor in gsoc2025 again ?
<cosminaught>
depends on the project, but certainly interested
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
<kierank>
I see SDR is back...
<fflogger>
[newticket] MonkeyPaw: Ticket #11464 ([avfilter] Subtitles extracted out of order 608 TS [out+subcc]) created https://trac.ffmpeg.org/ticket/11464
<Marth64>
another one ....... *facepalm*
<Marth64>
(re: 608 ticket)
<Marth64>
re: SDR. I don't understand the hate for it personally.
<Marth64>
Idk just depends on what its final scope is. If it is just another input device, it could open new possibilities too. A/V can be exchanged over RF after all. For example ld-decode or its VHS variant. But I am not an SDR expert, so I wouldn't be able to answer to complexity of implementation.
thilo has joined #ffmpeg-devel
<Marth64>
(off topic but the ld-decode stuff is really cool)
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
abdu97 has joined #ffmpeg-devel
Teukka has quit [Read error: Connection reset by peer]
Teukka has joined #ffmpeg-devel
Teukka has quit [Changing host]
Teukka has joined #ffmpeg-devel
LordSDRLeader has joined #ffmpeg-devel
<LordSDRLeader>
i have plans to write light-weight clone of mpv-like player, main feature is to provide gapsless audio experience with audio filtering, and also it will no be that bloated with various stuff like current mpv
_whitelogger has quit [Ping timeout: 260 seconds]
_whitelogger_ has joined #ffmpeg-devel
<kierank>
LordSDRLeader: will it support SDR?
<LordSDRLeader>
kierank: it will only use libav* from librempeg
LordSDRLeader has quit [Quit: Client closed]
LordSDRLeader has joined #ffmpeg-devel
<kierank>
Realtime playback is hard problem though
<kierank>
Beware
<kierank>
Vlc spent decades trying to rewrite clock code
<LordSDRLeader>
first gets audio, than video if there is enough interest from community
<LordSDRLeader>
also 1.2.0 of lavfi-preview just released: https://github.com/richardpl/lavfi-preview/releases/tag/v1.2.0 , though for best experience you need librempeg build instead of ffmpeg, as ffmpeg have filters that are lacking array options interface so every interaction is clumsy and needs manually entering strings instead of using/dragging numbers inside
<kierank>
Audio is hardest part
<LordSDRLeader>
widgets
<LordSDRLeader>
for audio you just send samples to AO device
<kierank>
Lol
<kierank>
For real-time this will leak memory quickly
<kierank>
And it's hard problem to not
jarthur has quit [Quit: jarthur]
<LordSDRLeader>
lavfi-preview uses openal AO and it does not leak memory
<LordSDRLeader>
it should work like file, if something drops it will exit, instead of looping forever
<galad>
Is there a common list of metadata identifiers somewhere, or does each muxer/demuxer have its own? I want to complete the support of itunes style metadata in mov/mp4, I guess I'll have to make up some up arbitrary identifier strings
<JEEB>
yea that's the problem with string=string stuff
rix has quit [Ping timeout: 252 seconds]
rix has joined #ffmpeg-devel
abdu97 has quit [Ping timeout: 240 seconds]
abdu has joined #ffmpeg-devel
LordSDRLeader has quit [Quit: Client closed]
LordSDRLeader has joined #ffmpeg-devel
LordSDRLeader has quit [Quit: Client closed]
LordSDRLeader has joined #ffmpeg-devel
^Neo has joined #ffmpeg-devel
^Neo has quit [Changing host]
^Neo has joined #ffmpeg-devel
LordSDRLeader has quit [Quit: Client closed]
LordSDRLeader has joined #ffmpeg-devel
abdu has quit [Ping timeout: 240 seconds]
^Neo has quit [Ping timeout: 252 seconds]
Marth64 has quit [Remote host closed the connection]
LordSDRLeader has quit [Quit: Client closed]
LordSDRLeader has joined #ffmpeg-devel
witchymary has quit [Remote host closed the connection]
<thardin>
galad: I had the same thought
<thardin>
metadata.texi does not have such info, but perhaps we could add it? or a new file like metadata_keys.texi
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
^Neo has joined #ffmpeg-devel
^Neo has quit [Changing host]
^Neo has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
<LordSDRLeader>
will more users use lavfi-preview if i add encoding/muxing/transcoding support to it?
pross has quit [Ping timeout: 265 seconds]
<LordSDRLeader>
need more users | engagement | actions
paulk has quit [Ping timeout: 268 seconds]
paulk has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
ccawley2011_ has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 252 seconds]
<ePirat>
LordSDRLeader, whats lavfi-preview?
<galad>
thardin: that would be nice, there are already multiple identifier for the same metadata, for example @too is mapped "encoder" instead of "encoding_tool", and there in some place "encoder" is used instead of "encoded_by", but any change will surely break something out there…
<LordSDRLeader>
ePirat: GUI for interacting with filters, i linked above github link to release binaries
<thardin>
galad: oh yeah, this is why discipline is important with these things
HarshK23 has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
Marth64 has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
uau_ is now known as uau
LordSDRLeader has quit [Quit: Client closed]
ccawley2011_ has quit [Ping timeout: 246 seconds]
ccawley2011_ has joined #ffmpeg-devel
abdu has joined #ffmpeg-devel
LordSDRLeader has joined #ffmpeg-devel
odrling has quit [Remote host closed the connection]
odrling has joined #ffmpeg-devel
<Lynne>
I'm surprised itunes metadata is that massive
<Lynne>
synopsis, art director, executive producer, phonogram rights
<Marth64>
yeah it's intense
<LordSDRLeader>
stop torrenting music
<galad>
that most of the original list, before they decided to put all the new tags in an xml blob
ccawley2011__ has quit [Ping timeout: 252 seconds]
ccawley2011_ has joined #ffmpeg-devel
thardin has joined #ffmpeg-devel
ccawley2011__ has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 252 seconds]
ccawley2011_ has quit [Ping timeout: 252 seconds]
<Lynne>
what's up with the spam on IRC lately? #dri-devel gets a spammer once every few hours these days
<Marth64>
my qemu keeps crashing, sorry for my join/part spam, i have a plan to fix it but had not gotten a chance
ccawley2011__ has quit [Ping timeout: 252 seconds]
<LordSDRLeader>
define spam
System_Error has quit [Remote host closed the connection]
<psykose>
dri-devel guy has been around for like a decade on/off
<psykose>
makes a new account every time
<Lynne>
nah, there's some NN-based spam too
System_Error has joined #ffmpeg-devel
LordSDRLeader has quit [Quit: Client closed]
LordSDRLeader has joined #ffmpeg-devel
abdu has quit [Quit: Client closed]
abdu has joined #ffmpeg-devel
LordSDRLeader has quit [Quit: Client closed]
LordSDRLeader has joined #ffmpeg-devel
tufei__ has quit [Remote host closed the connection]
tufei__ has joined #ffmpeg-devel
LordSDRLeader has quit [Quit: Client closed]
LordSDRLeader has joined #ffmpeg-devel
<frankplow>
The forgejo should be available at code.ffmpeg.org right?
<Lynne>
yes, I guess BtbN is doing maintenance atm
<BtbN>
I rebootet it earlier. It should be long back up
<frankplow>
I just get a blank page
<BtbN>
it didn't start...
<BtbN>
forgejo failed to start cause it started faster than mariadb, and it concluded exiting forever is the right reaction to that...
<LordSDRLeader>
heh
<nevcairiel>
sounds like your containers need a dependency
<Marth64>
happy to peer review docker compose files if needed
<Marth64>
sounds like you got it though
<kierank>
LordSDRLeader: lool
<BtbN>
nevcairiel: containers do not have dependencies.
<BtbN>
they are a docker compose concept, only compose honors them, not docker on a reboot
<BtbN>
docker devs say: The application should handle that, not we.
<JEEB>
didn't swarm do deps?
<JEEB>
if not then systemd around docker containers?
<BtbN>
I don't see how that would work
<BtbN>
forgejo officially supports docker, so it has to support dealing with out-of-order startup imo
<LordSDRLeader>
fill bug reports
<Marth64>
with docker compose you can address this. there is a depends_on option and you can also couple it with a health check. I agree in spirit a good software should handle out of order startup
<psykose>
you could use restart=always on the forgejo container as a hack :P
<BtbN>
Marth64: docker compose is not re-run on reboot though
<BtbN>
those dependencies become meaningless once the containers are initially started.
<BtbN>
They're not an inherent feature of docker, but just a compose thing, that only "compose up" honors.
<BtbN>
psykose: I don't like that, cause it enables potential crashy attacks to be re-tried rapidly
<psykose>
makes sense
<Marth64>
BtbN: yes it can be done with this param, restart: unless-stopped
<BtbN>
That's already set and does not help.
<Marth64>
hmm. are you on older docker? (e.g. are you typing "docker-compose" instead of "docker compose")
<BtbN>
And it also still has the attack amplification problem.
<BtbN>
No
<BtbN>
this is not a compose problem, docker just has no concept of dependencies.
<Marth64>
Hmmm. I'll look back at my stuff. Possible I'm wrong and my stuff is working by luck (or held up by the health check in addition to the depends_on)
<Marth64>
I get what you're saying now about only "compose up" honors. I thought I had solved this once so will check.
srikanth has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
abdu78 has quit [Ping timeout: 240 seconds]
srikanth has joined #ffmpeg-devel
eslam has joined #ffmpeg-devel
cone-149 has joined #ffmpeg-devel
<cone-149>
ffmpeg James Almer master:eb3d507b7f00: avdevice/alsa: simplify passing ff_alsa_open a channel layout
Eslam38 has joined #ffmpeg-devel
eslam has left #ffmpeg-devel [#ffmpeg-devel]
\\Mr_C\\ has joined #ffmpeg-devel
<ePirat>
BtbN, we did the same for VideoLAN gitlab in the beginning but in recent time spam got so bad that forced us to do full approval only registration
<BtbN>
With full approval only registration you completely kill off contribution though :/
<ePirat>
not like there was much choice sadly
<srikanth>
as a new developer to ffmpeg project though, creating a PR was very straigh forward using the Forgejo instance
<cone-149>
ffmpeg James Almer master:22aa71d4da37: avcodec/libtheora: fix setting keyframe_mask
<cone-149>
ffmpeg James Almer master:a9e8586e3fa8: avcodec/ac3dec: move some fields up in AC3DecodeContext
<cone-149>
ffmpeg James Almer master:045a8b15b19e: avcodec/ac3dec: add a flush callback for the ac3 and eac3 decoders
<BtbN>
srikanth: If you had to first have an admin approve your account before being able to do anything, it'd be super annoying though
<BtbN>
Cause then you'd STILL need to ML or IRC to get in touch
<BtbN>
Also, keep in mind the Forgejo instance is just a test. PRs cannot be actually merged there
<BtbN>
but you're the first to test the AGit workflow :D