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 6.1.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
michaelni has quit [Ping timeout: 260 seconds]
rvalue has quit [Ping timeout: 246 seconds]
iive has quit [Quit: They came for me...]
MisterMinister has quit [Ping timeout: 260 seconds]
rvalue has joined #ffmpeg-devel
<cone-569>
ffmpeg James Almer master:56d630e6c2ae: avformat/iamf_writer: update extradata from packet side data
<cone-569>
ffmpeg James Almer master:44e5749462e0: avformat/iamfenc: ensure updated extradata is written
michaelni has joined #ffmpeg-devel
Marth128 has joined #ffmpeg-devel
Marth64 has quit [Killed (NickServ (GHOST command used by Marth128!~Marth64@188.215.95.221))]
Marth128 is now known as Marth64
michaelni has quit [Ping timeout: 272 seconds]
michaelni has joined #ffmpeg-devel
lexano has quit [Ping timeout: 260 seconds]
michaelni has quit [Ping timeout: 264 seconds]
michaelni has joined #ffmpeg-devel
thilo has quit [Ping timeout: 255 seconds]
thilo has joined #ffmpeg-devel
thilo has quit [Changing host]
thilo has joined #ffmpeg-devel
michaelni has quit [Ping timeout: 246 seconds]
jarthur_ has joined #ffmpeg-devel
jarthur has quit [Killed (NickServ (GHOST command used by jarthur_))]
jarthur_ is now known as jarthur
\\Mr_C\\ has joined #ffmpeg-devel
michaelni has joined #ffmpeg-devel
mkver has quit [Ping timeout: 256 seconds]
Gramner has quit [Remote host closed the connection]
cone-569 has quit [Quit: transmission timeout]
jamrial_ has quit []
michaelni has quit [Ping timeout: 264 seconds]
michaelni has joined #ffmpeg-devel
AbleBacon has quit [Read error: Connection reset by peer]
Martchus_ has quit [Ping timeout: 246 seconds]
Martchus has joined #ffmpeg-devel
beastd has quit [Ping timeout: 264 seconds]
beastd has joined #ffmpeg-devel
Marth64 has quit [Remote host closed the connection]
<rajivharlalka>
Couldn't get a reply yesterday, can anyone help me out with my issue.
<rajivharlalka>
> 11:02 AM <rajivharlalka> @JEEB I had been working on this bug (https://trac.ffmpeg.org/ticket/10692) and could understand the reason for assertion. Was able to log the nsamples and zeroes to find out that when the asendcmd starts atempo from 0.00 sec , the number of samples to be set to 0 gets equal to the number of the total samples in the buffer, which is unexpected. Though I couldn't understand why this was
<rajivharlalka>
happening only when the filter started from 0.00 sec.
justache has quit [Remote host closed the connection]
justache has joined #ffmpeg-devel
michaelni has quit [Ping timeout: 246 seconds]
michaelni has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
kurosu has joined #ffmpeg-devel
psykose has quit [Remote host closed the connection]
psykose has joined #ffmpeg-devel
cone-266 has joined #ffmpeg-devel
<cone-266>
ffmpeg Andreas Rheinhardt master:a62c8558ae16: avutil/hwcontext_dxva2: Allocate pub and priv frames hwctx together
<cone-266>
ffmpeg Andreas Rheinhardt master:32fd27df2fc4: avutil/hwcontext_d3d11va: Allocate pub and priv frames hwctx together
deus0ww has joined #ffmpeg-devel
psykose has quit [Ping timeout: 264 seconds]
<cone-266>
ffmpeg Andreas Rheinhardt master:028907b08ad9: avutil/hwcontext_qsv: Allocate public and priv device hwctx together
<cone-266>
ffmpeg Andreas Rheinhardt master:91745ac2334f: avutil/hwcontext_qsv: Allocate pub and priv frames hwctx together
<elenril>
but we can't have anonymous unions in public headers
<mkver>
Yeah, so let's just drop this.
<elenril>
ok
<elenril>
could also use __attribute__ ((packed)) on element counts, to make sure there is no padding
<mkver>
You want to use non-standard extensions to address my spec-compliance concerns?
<elenril>
dunno, you could have an arch with 32bit pointers and 64bit ints hidden somewhere
zsoltiv_ has quit [Ping timeout: 260 seconds]
<courmisch>
isn't that what x32 is?
<elenril>
does it have a 64-bit int?
<elenril>
a plain int, not int64_t
<elenril>
Received: Mar 2024 12:32:44 +0200 (EET)
<elenril>
Date: Tue, 05 Mar 2024 11:34:33 +0100
<elenril>
this just in: mkver is from the future
<courmisch>
elenril: are you confusing 386 and x32?
<elenril>
am I?
<courmisch>
x32 is x86-64 ISA with 32-bit addresses
<elenril>
I didn't know about any arch where a plain int is 64bit
<elenril>
though I'm unsure about itanium
<courmisch>
it's not clear that you specifically meant 'int' as opposed to some integer type
<mkver>
cray computers support(ed?) only one byte and eight byte integer types.
rajivharlalka has quit [Quit: Connection closed for inactivity]
<elenril>
courmisch: it's clear from the context
<elenril>
but of course who cares about the context when you can troll instead
<courmisch>
that's pretty much insulting TBH
courmisch has left #ffmpeg-devel [#ffmpeg-devel]
Gramner has joined #ffmpeg-devel
unlord has quit [Ping timeout: 260 seconds]
unlord has joined #ffmpeg-devel
kurosu has quit [Quit: Connection closed for inactivity]
<cone-266>
ffmpeg Andreas Rheinhardt master:f0abb44fbf26: avformat/aviobuf: Move code specific to URLContexts to avio.c
<cone-266>
ffmpeg Andreas Rheinhardt master:57e20dd6b6ee: avformat/avio: Avoid function pointer casts
<cone-266>
ffmpeg Andreas Rheinhardt master:c856e4c54697: avformat/avio: Avoid indirection in ffio_fdopen()
<cone-266>
ffmpeg Andreas Rheinhardt master:fed46d770627: avformat/avio: Avoid av_strdup(NULL)
_av500_ has joined #ffmpeg-devel
av500 has quit [Ping timeout: 256 seconds]
_av500_ is now known as av500
jamrial has joined #ffmpeg-devel
mkver has quit [Ping timeout: 264 seconds]
iive has joined #ffmpeg-devel
lexano has joined #ffmpeg-devel
<cone-266>
ffmpeg Zhao Zhili master:b0e17e0f959b: avcodec/mediacodecenc: Add a hint message to use nv12 pix_fmt
Livio has joined #ffmpeg-devel
Livio has quit [Ping timeout: 268 seconds]
mkver has joined #ffmpeg-devel
hamzah has joined #ffmpeg-devel
<cone-266>
ffmpeg Andreas Rheinhardt master:d085f341d66e: avformat/hlsenc: Redo checking for strftime %s support to avoid warnings
<cone-266>
ffmpeg Andreas Rheinhardt master:440f8186ef9b: avformat/dashenc, hlsenc: Return 0 on succes from write_header
MikhailAMD has joined #ffmpeg-devel
<cone-266>
ffmpeg James Almer master:988e3a061a91: avformat/iamf_writer: clear extradata_size on extradata allocation failure
<cone-266>
ffmpeg James Almer master:79c6ba900792: avformat/iamf_writer: constify some variables
<cone-266>
ffmpeg James Almer master:ab15c04dee5a: avformat/avformat: add a function to return the name of stream groups
AbleBacon has joined #ffmpeg-devel
<Lynne>
still find it funny of all things gitlab could be fixing, they're doing an AI chatbot
<elenril>
ever heard about the notion of hype?
<elenril>
i'm very much looking forward to the bubble bursting
marcj has quit [Ping timeout: 264 seconds]
<llyyr>
can they make gitlab any slower?
<mkver>
courmisch, elenrl: Why did you not deprecate av_vdpau_alloc_context in e3e158e81f? (Right now, we have two allocators for AVVDPAUContext, both of which should not be used.)
<mkver>
elenril of course, not elenrl
<elenril>
don't think there's any better reason than laziness
<Lynne>
elenril: bubbles like that don't burst, they fizzle out, and thay fizzle is very well masked and mitigated
<Lynne>
no pity for us, popcorn fans
<elenril>
nah
<elenril>
the stock market does see it
<bencoh>
something tells me we'll have to wait some to see it burst though
<bencoh>
(quite some time, even)
<elenril>
I expect a major crash before the end of 2025
<elenril>
possibly even this year
<elenril>
there's been too much hype and too little to show for it
<Lynne>
because of AI and quantum computers?
<elenril>
ai mainly, nobody cares about quantum computers
<Lynne>
it's okay, room temperature superconductors will definitey come and there will just be a huge peak
<elenril>
tothemoon.tiff
<Daemon404>
i cant wait for my ffmpeg stocks to skyrocket because of nvenc support and ai backends
<Lynne>
as much as everything of interest uses cuda, I think that manufacturers go out of their way to avoid using any standards they themselves are developing, like sycl
<cone-266>
ffmpeg James Almer master:3769968afdae: fftools/ffprobe: fix printing side data components and pieces
<cone-266>
ffmpeg James Almer master:7e4334e16aaf: fftools/ffprobe: add support for Stream Groups
<cone-266>
ffmpeg James Almer master:e3671e6441ae: fate/iamf: print stream group information
<cone-266>
ffmpeg James Almer master:3635c1ee97fa: fate/mov: print stream group information for iamf tests
derpydoo has joined #ffmpeg-devel
kurosu has joined #ffmpeg-devel
microchip__ has joined #ffmpeg-devel
microchip_ has quit [Killed (NickServ (GHOST command used by microchip__))]
microchip__ is now known as microchip_
<cone-266>
ffmpeg James Almer master:1a34eb3ffc32: avformat/mov: fix setting dependent disposition on ambisonic IAMF streams
<mkver>
jamrial: Besides that the only difference is that you fixed (i.e. removed) get_channel_name in lavu/channel_layout.c, as I told you, whereas I simply kept an old version of your patch.
<mkver>
Btw: I could push the d3d12va and videotoolbox patches immediately if someone tested them.
rajivharlalka has joined #ffmpeg-devel
Livio has joined #ffmpeg-devel
<BtbN>
if they were on the branch I compiled the other day, they at least build.
Sean_McG has joined #ffmpeg-devel
<JEEB>
what would be required for a test?
<JEEB>
that videotoolbox h264 works or so?
<JEEB>
(decoder)
<mkver>
I already know that they build.
<mkver>
Or rather I know this for d3d12va.
<mkver>
JEEB: Yes, that would be enough.
<BtbN>
oh right, VT is OSX stuff
Marth64 has joined #ffmpeg-devel
<Sean_McG>
hi, I know I talked about adding a '-std=c17' FATE node but after a quick peek, 'configure' won't let me do that as it tests and adds '-std=c11' explicitly
<mkver>
You can also leave "avutil/hwcontext: Allocate AVHWFramesCtx jointly with its internals" and "avutil/hwcontext: Allocate AVHWDevCtx jointly with its internals" out, as they change hwcontext.h and therefore may lead to a rebuild of everything with avcodec.h in it.
<JEEB>
it's OK, I haven't built in a while anyways so with this update it would be a full rebuild anyways :)
<JEEB>
alright, configure rolling
<BtbN>
a full builds only takes like 30 seconds anyway
rvalue has quit [Ping timeout: 252 seconds]
<JEEB>
a bit longer on this thing but it already finished :)
<JEEB>
mkver: g229aa57d4a from that branch seems to decode just fine with videotoolbox_vld pix_fmt and hwaccel videotoolbox :) w:1280 h:720 pixfmt:videotoolbox_vld tb:1/24000 fr:24000/1001 sar:0/1 csp:bt709 range:tv
rvalue has joined #ffmpeg-devel
<mkver>
JEEB: Thanks. Then I will apply the videotoolbox commit.
<cone-266>
ffmpeg Andreas Rheinhardt master:68f48d5c3021: avutil/hwcontext_videotoolbox: Allocate pub and priv frames hwctx together
<Sean_McG>
I gave my Intel Mac away a few months ago
<Sean_McG>
I'd mothballed it for quite some time
<Sean_McG>
my dad might get one of the M3 iMacs -- he has a 2009 Intel Mac that all of a sudden he is finding "slow" *laughs*
elastic_dog has quit [Ping timeout: 256 seconds]
ngaullier has quit [Ping timeout: 255 seconds]
Marth64 has joined #ffmpeg-devel
<Marth64>
hopefully fixed my connection so it stops flapping
derpydoo has quit [Quit: derpydoo]
elastic_dog has joined #ffmpeg-devel
Sean_McG has quit [Quit: leaving]
tufei has joined #ffmpeg-devel
* microchip_
farts
<microchip_>
excuse me :/
Sean_McG has joined #ffmpeg-devel
qeed has quit [Quit: qeed]
qeed has joined #ffmpeg-devel
psykose has joined #ffmpeg-devel
devinheitmueller has joined #ffmpeg-devel
rajivharlalka has quit [Quit: Connection closed for inactivity]
derpydoo has joined #ffmpeg-devel
hamzah has joined #ffmpeg-devel
dionisis has joined #ffmpeg-devel
<Marth64>
mkver, JEEB: I am testing now on an M1 as well (the videotoolbox change)
<Marth64>
works fine
<mkver>
As expected
<Marth64>
I haven't found anything outrageous in doing bunch of muxing the past few days. current master rocks
<Marth64>
current-ish*
<Marth64>
will test nvenc later
<j-b>
good morming
<Marth64>
is there a therapy group, forum, or IRC channel for help with ASS subtitles styling?
<Marth64>
good morning!
<Sean_McG>
*laughs*
<elenril>
Marth64: doom9?
<elenril>
also #darkhold used to be the place, but probably not anymore
<JEEB>
the people doing aegisub forks most likely, and #libass probably?
<JEEB>
I wonder which fork is the most sane one, would be interesting to attempt to integrate proper video rendering to it
<elenril>
did aegisub itself die?
<Sean_McG>
do people even use that anymore? isn't anime all just stream-rips now?
<Marth64>
SubtitleEdit is what i've been using
<Marth64>
will try these resources thx
<JEEB>
Sean_McG: they most definitely do
hamzah has quit [Ping timeout: 250 seconds]
<JEEB>
also various annie may streaming sites did utilize ASS as source format
<JEEB>
heck, bandai I think sponsored support for some format for ingest was it?
<JEEB>
elenril: basically a case of a critical mass of maintainers not having the time & resource to prod it. then I think someone got access, and then $dayjob happened hard for that person too
<JEEB>
(something something build system rewrites etc)
<elenril>
why do people wank so much over build system rewrites
<Sean_McG>
gotta go fast?
<JEEB>
I do agree that for windows the aegisub build system had two sides: one the positive side you got very specific bundled versions of dependencies and you could directly build them *all* in a single MSVS project. but then on the other hand... I am not sure how many people actually knew how to update those dependencies :D
<JEEB>
so reworking the windows side made sense
<JEEB>
the autoconf for *nix... now there I saw less showstoppers
<JEEB>
anyways, should one day figure out which fork is the one
Livio has joined #ffmpeg-devel
derpydoo has quit [Quit: derpydoo]
<Marth64>
I just wanted to consolidate to one format, ASS fits the bill. Except I can't figure out how make the opaque background shadow box have some small padding.
<Marth64>
But yeah it's my problem, I will go ask around/research some more. Thank you!
<Marth64>
WebVTT would have been the true goal but it's not supported well
<Lynne>
lol, even google abandoned webvtt
<elenril>
all subtitle formats suck
<Lynne>
^^
<elenril>
someone needs to make latex-based subtitles
<Lynne>
all, except kate
<Lynne>
and that's just because no one's used it to know why it sucks
<Marth64>
I thought TTML would changes things for the better. It's horrifying
<Marth64>
-s
<Marth64>
ASS wins
<Marth64>
for home use, at least.
<JEEB>
I think NHK plus in Japan first went with TTML, but then after they found out that there is a JS renderer for the Japanese broadcast captions format, they just started stuffing them in id3 messages in HLS
<elenril>
>XML-based
<elenril>
nothing good can come from that
<Sean_McG>
+1
<JEEB>
(or something similar, the TTML -> ARIB broadcast caption switch is what is the main point)
<cone-266>
ffmpeg James Almer master:1f73b52850b5: fate/mov: print stream group information for avif/heic tests
<cone-266>
ffmpeg James Almer master:708175616c5e: fate/mov: use framecrc for the remaining avif/heic tests
<cone-266>
ffmpeg James Almer master:924c7cd74576: fate/mov: add a heic test with a derived image using a single item twice
System_Error has quit [Remote host closed the connection]
jarthur has joined #ffmpeg-devel
<Marth64>
I don't know what kind of sorcery it does, if any at all, but VLC renders Arial subs really nicely
<Marth64>
in srt
<Marth64>
</subtitle-tangent>
System_Error has joined #ffmpeg-devel
<Sean_McG>
Tag malformed, missing start.
<Sean_McG>
:)
<Marth64>
*Aborted (core dumped)*
Marth128 has joined #ffmpeg-devel
Marth64 has quit [Killed (NickServ (GHOST command used by Marth128!~Marth64@85.237.194.74))]
Livio has quit [Ping timeout: 246 seconds]
Livio has joined #ffmpeg-devel
<pal>
@Marth64 care to elaborate on your strong feelings re: TTML?
Livio has quit [Ping timeout: 264 seconds]
<Marth128>
pal: ehh, could have phrased it better, not targeted to TTML. I meant more the diversity of formats out there is horrifying. I think TTML itself is what I'd wish I could use, XML doesn't bother me. but it is a bit too flexible and the devices/players I use don't handle that well.
<Marth128>
probably because they all use libass, so they natively work best with ass
<Marth128>
it's reasonable as an interchange format
Marth64 has joined #ffmpeg-devel
Marth128 has quit [Ping timeout: 260 seconds]
<cone-266>
ffmpeg James Almer master:d34e9a81b1d2: fate/iamf: print stream disposition
<Lynne>
Error: operand type mismatch for `shr'
<Lynne>
have we had any recent fixes for inline asm?
<Lynne>
old revisions (less than a year ago) don't compile now
<pal>
Marth128: yes to your points. I would say that support for timed text in end user devices could in general be improved, i.e. not only a TTML issue
<pal>
... as a result we see a lot of centered white on black text, regardless of transmission or authoring format
darkapex has quit [Remote host closed the connection]
darkapex has joined #ffmpeg-devel
unlord has quit [Changing host]
unlord has joined #ffmpeg-devel
Marth64 has quit [Ping timeout: 264 seconds]
<jdarnley>
Lynne: yes there was one to allow new tools to use the inline asm