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
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
TheVibeCoder has quit [Ping timeout: 244 seconds]
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
iive has quit [Quit: They came for me...]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 252 seconds]
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
_whitelogger has joined #ffmpeg-devel
mkver has quit [Ping timeout: 244 seconds]
mkver has joined #ffmpeg-devel
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
mkver has quit [Read error: Connection reset by peer]
<compnn>
kierank, because money laundering is illegal. so converting funds from one thing to another is bad
<compnn>
without following proper banking laws
<compnn>
oooooh how i dislike discord
<steven-netint>
Who is the best to contact for code previously developed by elenril?
<steven-netint>
I have a one line change for something he worked on after n7.1 in ffmpeg_filter.c
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 244 seconds]
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
jamrial has quit []
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
Quackdoc has joined #ffmpeg-devel
<Marth64>
thardin - ahhh just realized you meant in context of HTML, my bad.
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
mkver has quit [Remote host closed the connection]
mkver has joined #ffmpeg-devel
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
<fflogger>
[editedticket] atayeem: Ticket #11640 ([avfilter] Specific combination of timeclamp, fps, and count on showcqt causing crash) updated https://trac.ffmpeg.org/ticket/11640#comment:3
Quackdoc has left #ffmpeg-devel [#ffmpeg-devel]
mkver has quit [Ping timeout: 276 seconds]
mkver has joined #ffmpeg-devel
<compnn>
steven-netint, just post question about the code/patch to -devel
<compnn>
it is a discussion mailing list after all :)
<steven-netint>
👍
<steven-netint>
another question, for a 42k line patch with a whole bunch of new HW codecs/filters, would you guys prefer I put it on a forgejo branch first or just send it in mail-list?
System_Error has quit [Remote host closed the connection]
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
mkver has quit [Ping timeout: 260 seconds]
mkver has joined #ffmpeg-devel
<kierank>
Mailing list
<kierank>
Ban derek
<compnn>
mmm right on time. i was just getting bored again. and now i'm all fired up!
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
<compnn>
steven-netint, mailing list. having it in a repo or as one big .patch file is also useful for testing, too
<compnn>
for when someone doesnt get mail 15/100 in their inbox :D
TheVibeCoder has joined #ffmpeg-devel
<compnn>
TheVibeCoder, how come you didnt mine bitcoins ?
<steven-netint>
compnn: thanks for the advice
<thardin>
kepstin: slow even compared to a decade old machine?
<thardin>
compnn: it's only money laundering if the money comes from an illegal source
<compnn>
thardin, said another way, its money laundering unless you're a bank
<compnn>
or 'currency exchange'
<thardin>
no, banks can still engage in money laundering. Swedbank did in the baltics
<TheVibeCoder>
compnn: i have nvidia GPU that i do not use at all
<compnn>
TheVibeCoder, tnt 2 ?
<thardin>
it's just that at the moment if you come in with 1,000€ in cash the bank asks a lot of questions. if on the other hand you come in with literal garbage bags of ~10,000,000€ cash then the bank doesn't ask questions
<thardin>
which is what happened in the swedbank case
<kepstin>
thardin yes, although somewhat workload dependant. it's often faster to cross-compile stuff on a decade-old machine than build it natively on a spacemit k1 or m1 chip :/
<CounterPillow>
all the risc-v socs on the market right now are complete and utter trash and you shouldn't fall for them.
<CounterPillow>
(they are also not in any way more "open" than other SoCs, the only thing different is that the SoC vendor doesn't pay an ISA license if they make their own core design, which they're not making anyway since all of them license stuff from SiFive)
<kepstin>
they're not all sifive licensed cores, spacemit seem to have made their own. or maybe just used rocket. XuanTie has some core IP that has been used in a few chips (they're the ones who have that annoying pre-standardization vector support on some chips)
<kierank>
TheVibeCoder: go away
<kierank>
do not upset leadership
mkver has quit [Ping timeout: 245 seconds]
mkver has joined #ffmpeg-devel
av500 has joined #ffmpeg-devel
mkver has quit [Ping timeout: 265 seconds]
jamrial has joined #ffmpeg-devel
averne has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
HarshK23 has quit [Quit: Connection closed for inactivity]
Richardcavell_ has quit [Ping timeout: 268 seconds]
<thardin>
CounterPillow: it's also about network effects
mkver has quit [Ping timeout: 272 seconds]
mkver has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
cone-426 has quit [Quit: transmission timeout]
<Lynne>
haasn: shit, I keep thinking that ff_vk_init gets called once and it inits everything
minimal has joined #ffmpeg-devel
Richardcavell_ has joined #ffmpeg-devel
<beastd>
steven-netint: I might be too late (didn't read mails yet), but actually I would say put it up on our forgejo test instance (or some other public git) first. then send a mail explaining motivation and scope to the ML with the link to the git. Also best actively request feed. (ping after a week or so if it doesn't get attention)
<kierank>
the other issue is that to buy this hardware requires an nda
* kierank
tried once
<TheVibeCoder>
which hw?
<kierank>
netint
MisterMinister has quit [Ping timeout: 276 seconds]
<Lynne>
TheVibeCoder: how good is the prores raw decoder in librempeg?
<TheVibeCoder>
Lynne: version >= 1 not supported, tried to add support but something is not correct with tile layout
<TheVibeCoder>
half of horizontal data is missing somehow..
<TheVibeCoder>
have samples, if you wish to work on it
<Lynne>
I'd like to get that working and write a vulkan decoder
<Lynne>
do you have the wip version 1 code somewhere?
<Lynne>
guessing no spec, so its RE'd, right?
<TheVibeCoder>
i have some bad patch
<TheVibeCoder>
yes
<TheVibeCoder>
only RE
<TheVibeCoder>
currently it skips some fields, because i dunno what such fields represent
<TheVibeCoder>
so its still mostly guessing game
<TheVibeCoder>
Lynne: https://0x0.st/8lct.diff it does not apply cleanly but you get idea what changed
<TheVibeCoder>
maybe tile layout changed
<TheVibeCoder>
each tile have 3 fields, that says offsets of each component
<TheVibeCoder>
the 4th component is size of whole tile minus sum of above mentioned sizes of three components
<TheVibeCoder>
tile seems to always start with 0x400f hex
<Lynne>
thanks, I'll take a look
<Lynne>
does at least one tile decode?
<TheVibeCoder>
it decodes something badly
<TheVibeCoder>
and lots of underreads
<TheVibeCoder>
not much of overreads
<Lynne>
guess they added new fields per tile
<TheVibeCoder>
or tile WxH size are not 128x16
<TheVibeCoder>
for example, version #1 file, dimensions 5760x3040 have 4370 tiles
<TheVibeCoder>
which means 190 horizontal tiles X 23 vertical tiles
darkmushroom has joined #ffmpeg-devel
<Lynne>
not bad for a GPU decoder
hpkn has quit [Remote host closed the connection]
hpkn has joined #ffmpeg-devel
usagi_mimi has quit [Ping timeout: 248 seconds]
System_Error has quit [Ping timeout: 244 seconds]
mkver has quit [Ping timeout: 245 seconds]
Richardcavell_ has quit [Ping timeout: 260 seconds]
paulk has quit [Ping timeout: 268 seconds]
paulk has joined #ffmpeg-devel
paulk has joined #ffmpeg-devel
paulk has quit [Changing host]
ccawley2011 has quit [Ping timeout: 245 seconds]
System_Error has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
IndecisiveTurtle has joined #ffmpeg-devel
<thardin>
ffabrice help me I'm actually considering what it *would* take to bring edit list support to the ffmpeg CLI
<thardin>
assuming someone is willing to pay for it
<thardin>
it's so completely orthogonal to how program arguments are given at the moment. plus the whole issue of the command line only supporting so many arguments
<JEEB>
well, you would open the file. then you would receive the side data
<JEEB>
at that point you have the time line
<thardin>
sure, but what if the user wants to specify the timeline themselves?
<TheVibeCoder>
NLE
<JEEB>
yea that's a common problem with a lot of side data
<JEEB>
there is no single function that takes in a string
<JEEB>
and then would parse that as stuff for side data
<thardin>
that might actually be a good thing in this case
<thardin>
one could of course write a new fftool to do it, specifically tailored for the job. but at that point it's just reinventing melt
<JEEB>
yea
ccawley2011 has quit [Read error: Connection reset by peer]
<JEEB>
anyways, we definitely need that string to side data thing for various stuff, since essentially a lot of stuff is right now not override'able
<TheVibeCoder>
melt have edit list support?
<thardin>
melt is an NLE
<thardin>
you can give an edit list (or an entire timeline with transitions, effects and everything) as XML
<TheVibeCoder>
melt is poor man NLE
<JEEB>
one of the rare examples of something being override'able is the display matrix side data, since that has separate options in ffmpeg cli
<thardin>
melt is the CLI side of editors like kdenlive
ccawley2011 has joined #ffmpeg-devel
<TheVibeCoder>
add seeking support within filters and you get much better edit list support than melt
<thardin>
it's highly WISYWYG, seekable and so on. unlike lavfi
<thardin>
I objected to lavfi's design back in the day because it makes filtering random frames essentially impossible
<TheVibeCoder>
lies
<thardin>
melt is far more sane with how it does things
<TheVibeCoder>
melt is using frei0r and bunch of stull is limited to 8-bit only
<TheVibeCoder>
*stuff
<thardin>
you're 8-bit
cone-003 has joined #ffmpeg-devel
<cone-003>
ffmpeg James Almer master:a7beac704b1c: avformat/iamf_parse: prevent overreads in update_extradata
<cone-003>
ffmpeg James Almer master:9b5abbf38726: avformat/iamf_writer: ensure each layer's channel layout contains all channels from the previous one
<cone-003>
ffmpeg James Almer master:534eb7260a81: tests/iamf: reorder muxed streams
<cone-003>
ffmpeg James Almer master:13e81dbd2724: avformat/iamf_parse: try to retype the channel layout for ambisonics_mode == 0
<cone-003>
ffmpeg James Almer master:e5f23a3c5efe: tests/iamf: rename BACK to SIDE filterchain labels in the 5.1.4 iamf tests
<cone-003>
ffmpeg James Almer master:cd2461e627cc: avformat/iamf: fix setting channel layout for Scalable layers
<cone-003>
ffmpeg James Almer master:a3a7b8edc94c: avformat/iamf_writer: factor out getting loudspeaker_layout values
<cone-003>
ffmpeg James Almer master:75c514e468ed: avformat/iamf_writer: add extra constrains for Parameter Sets in Audio Elements
<cone-003>
ffmpeg James Almer master:38f9fbe30d7f: avformat/iamf_writer: reindent after previous commit
<cone-003>
ffmpeg James Almer master:cbb72e6ab646: avformat/iamf_writer: use named constants in more places
<thardin>
>mlt_image_yuv422p16
<thardin>
ding dong ur wrong
<TheVibeCoder>
la, la, la, you can't see me..
<thardin>
anyway, I'll see if either of the usual suspects are interested in funding EDL support. either way I don't think I can spare the time for it until next year
<thardin>
if not then I guess mov.c and mxfdec.c will languish for a few years more with piss-poor performance over the network
<cone-003>
ffmpeg James Almer release/7.1:2c4ca52c8aed: avformat/iamf_parse: increase PutBytes buffer when writing AAC extradata
<cone-003>
ffmpeg James Almer release/7.1:6400860b9d4b: avformat/iamf_parse: prevent overreads in update_extradata
<TheVibeCoder>
XML/json/yaml/whetever like AVString
ccawley2011_ has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 265 seconds]
ccawley2011_ has quit [Read error: Connection reset by peer]