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
<JEEB>
I really liked during the HEVC development process that the group just told the broadcasters that they don't require specific tools for interlaced coding
<JEEB>
when the broadcasters attempted to start some interlacism working group for coding changes
<JEEB>
"nope, you get good enough compression benefits VS AVC already, you just get the field pic VS frame pic field. enjoy it"
<JEEB>
but yea, since interlaced HEVC in the wild is a thing, if someone did the work to output the fields combined, that's good. means that it can actually be deinterlaced afterwards with our standard filter chains :P
<JEEB>
meanwhile I recall the kodi people mentioning some subtitle duration patch from a few years ago posted on the ML, anyone know which one it is?
<JEEB>
something related to unknown duration IIRC
compn has quit [Remote host closed the connection]
compn has joined #ffmpeg-devel
ngaullier has joined #ffmpeg-devel
rvalue- has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 255 seconds]
rvalue- is now known as rvalue
Krowl has joined #ffmpeg-devel
Krowl has quit [Client Quit]
Cheetahze has joined #ffmpeg-devel
deus0ww has quit [Ping timeout: 260 seconds]
deus0ww has joined #ffmpeg-devel
putacho has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
microchip_ is now known as putacho
ccawley2011 has joined #ffmpeg-devel
<kurosu>
JEEB: Ateme's patch is maybe 4yo now, but obviously I won't discuss specifics of either side (ffhevc, or standardization)
<JEEB>
had missed it and also the folk who were offering € for implementing the merging
<JEEB>
I think at some point there was an idea of handling field pictures in AVFrame, but I don't think that ever went anywhere so we might as well normalize the field picture output to be frame picture based
<JEEB>
because in theory it's simpler if you don't have to have the merging logic
<JEEB>
but then the rest of the framework doesn't really work with this :P
<JEEB>
and that field picture infra was never written since the HEVC decoder was originally done
kierank has joined #ffmpeg-devel
<kierank>
[FFmpeg-devel] [PATCH 2/2] avcodec/x86/diracdsp: Migrate last remaining MMX function to SSE2
<kierank>
External is LGTM but I don't have the email because of our amazing mailing list
Cheetahze has quit [Quit: Connection closed for inactivity]
Krowl has quit [Ping timeout: 260 seconds]
sudden has quit [Ping timeout: 252 seconds]
Krowl has joined #ffmpeg-devel
mkver has quit [Remote host closed the connection]
mkver has joined #ffmpeg-devel
sudden has joined #ffmpeg-devel
<jamrial>
wbs: i could reproduce the nv12/24 related failures now, but not the yuv444p12-yuv444p one
<jamrial>
this revealed bugs in sws
<jamrial>
maybe i should disable the failing tests until someone looks at it
cone-027 has joined #ffmpeg-devel
<cone-027>
ffmpeg Araz Iusubov master:f63f1641255a: avcodec/amfenc: GOP size check
ngaullier has quit [Ping timeout: 255 seconds]
<cone-027>
ffmpeg James Almer master:2eb9c35010fe: x86/swscale: disable AVX2 yuv2nv12cX functions if accurate_rnd is requested
<BBB>
elenril: does the code to expose reconstructed images from e.g. libx264 back into a libavfilter pipeline (e.g. for metric calculations) already exist?
<BBB>
elenril: you mentioned that in your vdd talk
<JEEB>
> tools/enc_recon_frame_test is a thing at least
Krowl has quit [Read error: Connection reset by peer]
paulk has joined #ffmpeg-devel
<BBB>
thilo: your message (in "GSoC Mentor Summit Reimbursement Request") appears to have been sent 3x?
<BBB>
elenril: why are recon frames made writable?
<BBB>
cool API btw, I quite like the outside of it (receive_frame on an encoder!), and the encoder hook seems easy enough
<JEEB>
:)
Krowl has joined #ffmpeg-devel
j45_ has joined #ffmpeg-devel
j45 has quit [Ping timeout: 260 seconds]
j45_ is now known as j45
j45 has joined #ffmpeg-devel
j45 has quit [Changing host]
cone-027 has quit [Quit: transmission timeout]
SuperFashi has quit [Quit: No Ping reply in 180 seconds.]
HarshK23 has joined #ffmpeg-devel
SuperFashi has joined #ffmpeg-devel
NotWarcop has joined #ffmpeg-devel
Warcop has quit [Ping timeout: 246 seconds]
Traneptora has quit [Quit: Quit]
<courmisch>
Sebastinas: GCC bug
ccawley2011_ has joined #ffmpeg-devel
<ramiro>
another|: yes, I use rtmidi for midi input
ccawley2011 has quit [Ping timeout: 245 seconds]
<another|>
ramiro: And you use midi to change parameters?
<another|>
ramiro: maybe check out https://midimonster.net/ It can translate between all kinds of control protocols
<ramiro>
another|: yes, the parameters are changed in realtime using midi and quickjs.
<ramiro>
another|: midimonster looks amazing. I'll have a look when I'm back from vacation.
ngaullier has quit [Ping timeout: 276 seconds]
Krowl has quit [Read error: Connection reset by peer]
ccawley2011 has joined #ffmpeg-devel
ccawley2011_ has quit [Ping timeout: 276 seconds]
System_Error has joined #ffmpeg-devel
osvein has quit [Ping timeout: 248 seconds]
<jamrial>
wbs: can you reproduce the remaining failure locally?
Traneptora has joined #ffmpeg-devel
osvein has joined #ffmpeg-devel
<BBB>
does ffmpeg.exe actually use AV_CODEC_CAP_ENCODER_RECON_FRAME?
<BBB>
I'm trying to see how it's hooked up but my version (7.0) doesn't seem to use it (yet?)
<BBB>
or maybe it's really well-hidden :)
<JEEB>
not sure if ffmpeg.c has it integrated
<wbs>
jamrial: yes
tyzef has joined #ffmpeg-devel
<JEEB>
yea, even in master if you grep for that under fftools it gives no hits
<BBB>
right
<BBB>
I can imagine maybe it's auto-detected?
<Daemon404>
some ML mails seems to simply not be making it to me (and others)
<Daemon404>
not in spam folder either
ccawley2011_ has joined #ffmpeg-devel
<Daemon404>
either that, or a very very large delay for random messages
ccawley2011 has quit [Ping timeout: 260 seconds]
tyzef has quit [Quit: WeeChat 3.8]
* Daemon404
cannot reply to mails that never arrive
<BtbN>
Daemon404: are you using gmail? I just looked at the postfix log, and gmail is refusing to talk to the server completely now
<Daemon404>
yes
<Daemon404>
cool so the mailing list cannot function as a mailing list
<BtbN>
yeah, the ML can't talk to gmail anymore it seems, cause google-reasons
<Daemon404>
good shit
mkver has quit [Ping timeout: 248 seconds]
<BtbN>
421-4.7.29 To learn more about Gmail requirements for bulk senders, visit 421 4.7.29 https://support.google.com/a?p=sender-guidelines. 38308e7fff4ca-2ff178ec661si7644041fa.135 - gsmtp (in reply to end of DATA command)
<Daemon404>
im pretty sure "nobody can use the most widely used email provider in the world" is our problem
<Daemon404>
even if it isnt our fault
<BtbN>
Well, they're not telling us what's wrong tho
<BtbN>
Just point to that document and "go figure"
<BtbN>
Pretty sure the ML does already use SPF, DKIM and DMARC
<Daemon404>
there's definitely stuff there we've never checked
<Daemon404>
the From: header thing seems maybe relevant
<BtbN>
Beyond "Use SPF/DKIM/DMARC" and "be a good guy and don't spam", I don't see much there
<Daemon404>
do we have ARC?
<BtbN>
The ML already does that. That's why so many people have mangled From-Lines
<Daemon404>
i never understood why only *some* do
<Daemon404>
it should be all or none, no?
<BtbN>
Only if that domain is configured to the strict mode
<BtbN>
I might have unclogged it?
<Daemon404>
what did you do
<BtbN>
Postfix wasn't configured to use tls for outgoing smtp connections
<Daemon404>
i just got my mising emails
<Daemon404>
youre joking...
<Daemon404>
that is insane
<another|>
git.videolan.org return 502 ?
<BtbN>
Well, it's a public ML. So it's not exactly an issue to send stuff in plain
<Daemon404>
another|, seems so
<Daemon404>
it actually times out for me
<Daemon404>
not even 502
<another|>
either way it's unreliable
<Daemon404>
youll have to ping thresh in #videolan
<Daemon404>
code.videolan.org works btw :)
<Daemon404>
(where gitlab is)
<another|>
yep
<BtbN>
It's still rejecting a few mails, claiming SPF errors. But accepts like 95% of the other mails to gmail
<BtbN>
no idea what the issue is
<Daemon404>
love me some email monopoly
<Daemon404>
granted, mostly because everything else is worse
<llyyr>
ah that's why I'm missing some emails that are in the ffmpeg-devel archive
<BtbN>
gmail is simply not even an option to consider
<BtbN>
They openly state that they read the mails to track and profile you
MrZeus has joined #ffmpeg-devel
<Daemon404>
like it or not, you cant ignore gmail users
<BtbN>
If Google makes it impossible to send to them, eventually we'll have to or also use gmail
<BtbN>
that's what they want
<Daemon404>
or work through the list a bit deeper
<Daemon404>
... or use gitlab
* Daemon404
runs
<BtbN>
I have simply given up on sending mails to google or microsoft on my personal mail server
<BtbN>
not my problem
<BtbN>
Server is set up fine, if they want to play stupid games, I just ignore them
<Daemon404>
certainly a principled approach.
<BtbN>
Specially Microsoft every half year decided that my server is "in bad standing", cause I don't send enough mail
<BtbN>
then I had to contact support and get that fixed
<BtbN>
No idea what the logic there is. Do they want me to spam
<BtbN>
Google does something similar. Where they calculate a "spam score" for every server. And if the sample size is too small, they default to "must be spam", and you just get blocked
<Daemon404>
i am aware
<BtbN>
Probably THAT is at least not an issue for the ML
<Daemon404>
i mean
<Daemon404>
a ton of ml mail also gets marked as spam for me.
<Daemon404>
so i wouldnt be so sure
<Daemon404>
all sorts of missing mails are slowly showing up now
System_Error has quit [Remote host closed the connection]
<BBB>
Daemon404: so I'm sure you're using -dec at $work, but are you actually using recon_frame? or is this a separate decode pass?
<Daemon404>
BBB, we dont use avcodec for encoding
<BBB>
ah
<Daemon404>
so we get the recon frames directly from the encoding libs
<BBB>
understood
System_Error has joined #ffmpeg-devel
<BBB>
yeah I'm pretty sure recon_frame is not used in ffmpeg, if I put garbage in it, ffmpeg's -dec output is unmodified
<BBB>
but enc_recon_frame_test gives a mismatch (as expected) if I put garbage in recon_frame
<BBB>
elenril: is this a PATCH_WELCOME situation? or is there some ffmpeg cli param I should use?
rvalue- has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 276 seconds]
rvalue- is now known as rvalue
Traneptora has quit [Quit: Quit]
<JEEB>
BBB: I mean there are no references to it in fftools so I'd expect PATCH_WELCOME
<BBB>
I was hoping it'd be something like -flags +recon_frame in the encoder and some magic in libavcodec/...
<BBB>
but I can't see the magic anywhere :-p it's too magical for me
<wbs>
wasn't this mentioned as future development that will become possible in the future (and which might exist in some not yet upstreamed branch of his)?
ccawley2011 has joined #ffmpeg-devel
ccawley2011_ has quit [Ping timeout: 252 seconds]
<BBB>
ah. ok that's possible. I don't recall
MrZeus has quit [Read error: Connection reset by peer]
ccawley2011_ has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 246 seconds]
ccawley2011 has joined #ffmpeg-devel
ccawley2011_ has quit [Ping timeout: 252 seconds]
esu has quit [Remote host closed the connection]
esu has joined #ffmpeg-devel
Traneptora has joined #ffmpeg-devel
ccawley2011_ has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 248 seconds]
ccawley2011_ has quit [Read error: Connection reset by peer]
<BtbN>
If you didn't read #videolan, git.videolan.org is down/slow cause of intense ddos like scraping by hundreds and hundreds of different IPs