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
durandal_1707 has quit [Ping timeout: 256 seconds]
durandal_1707 has joined #ffmpeg-devel
epony has quit [Remote host closed the connection]
epony has joined #ffmpeg-devel
dellas has quit [Remote host closed the connection]
dellas has joined #ffmpeg-devel
thilo has quit [Ping timeout: 256 seconds]
navi has quit [Quit: WeeChat 4.0.4]
navi has joined #ffmpeg-devel
thilo has joined #ffmpeg-devel
thilo has quit [Changing host]
thilo has joined #ffmpeg-devel
jarthur_ has joined #ffmpeg-devel
jarthur has quit [Ping timeout: 260 seconds]
navi has quit [Quit: WeeChat 4.0.4]
navi has joined #ffmpeg-devel
<ramiro>
another|: I'm thinking about going. I live quite close.
<ramiro>
elenril: you'll be giving a talk, is that right?
dellas has quit [Remote host closed the connection]
navi has quit [Quit: WeeChat 4.0.4]
lemourin has quit [Read error: Connection reset by peer]
lemourin has joined #ffmpeg-devel
jarthur_ is now known as jarthur
Kei_N_ has joined #ffmpeg-devel
Kei_N has quit [Ping timeout: 240 seconds]
jamrial has quit []
MetaNova has quit [Ping timeout: 264 seconds]
MetaNova has joined #ffmpeg-devel
\\Mr_C\\ has quit [Remote host closed the connection]
<Marth64>
good news: dvd chapters is fixed and accurate. bad news: i have to read the whole disc first (to get accuracy), so it's slow.
<Marth64>
x_x
<Marth64>
well, the whole title actually. need to process the nav packets
Marth64 has quit [Quit: Leaving]
Marth64 has joined #ffmpeg-devel
chainik has quit [Quit: (╯°□°）╯︵ ┻━┻]
chainik has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
AbleBacon has quit [Read error: Connection reset by peer]
<elenril>
ramiro: yes
<elenril>
Marth64: how do other players do it?
HarshK23 has quit [Quit: Connection closed for inactivity]
cone-580 has joined #ffmpeg-devel
<cone-580>
ffmpeg Anton Khirnov master:fcddd233fef4: fftools/ffplay: make option strings dynamically allocated
<cone-580>
ffmpeg Anton Khirnov master:8aed3911fc45: fftools/ffmpeg: make attachment filenames dynamically allocated
<cone-580>
ffmpeg Anton Khirnov master:4ce1987abca4: fftools/ffprobe: make option strings dynamically allocated
<cone-580>
ffmpeg Anton Khirnov master:6d17991b7e1b: fftools/cmdutils: add option syntax for loading arbitrary arguments from a file
<cone-580>
ffmpeg Anton Khirnov master:a8bc79c3fd96: fftools/ffmpeg: deprecate -filter_script
<cone-580>
ffmpeg Anton Khirnov master:c316c4c77b1b: fftools/ffmpeg: deprecate -filter_complex_script
<cone-580>
ffmpeg Anton Khirnov master:08bebeb1bed3: Revert "all: Don't set AVClass.item_name to its default value"
<cone-580>
ffmpeg Anton Khirnov master:d8173cfcafac: fftools/cmdutils: hide some information listing options from basic help
<cone-580>
ffmpeg Anton Khirnov master:03aedbdd4090: fftools/cmdutils: add a flag for per-stream options
<cone-580>
ffmpeg Anton Khirnov master:67529ad8a478: fftools/cmdutils: drop alt_flags parameter from show_help_options()
<cone-580>
ffmpeg Anton Khirnov master:43323c380206: fftools/ffmpeg_opt: add more structure to long help output
<cone-580>
ffmpeg Anton Khirnov master:68af3e37ccec: fftools/cmdutils: indicate specifiers in option syntax in help output
<cone-580>
ffmpeg Anton Khirnov master:0b95e71aa046: fftools/ffmpeg_opt: briefly mention stream specifier syntax in help output
<cone-580>
ffmpeg Anton Khirnov master:9cb52927b9d5: fftools/cmdutils: surround option arguments by <> in help output
<cone-580>
ffmpeg Anton Khirnov master:344d3a7d4529: fftools/ffmpeg_opt: improve wording in option descriptions
<another|>
Gooooooooood Morning!
jarthur has quit [Quit: jarthur]
<haasn>
thanks for reminding me I still need tickets
<another|>
you mean sccomodation and train tickets?
HarshK23 has joined #ffmpeg-devel
<elenril>
just bring a tent
<another|>
!google weather february brussels
<elenril>
so I'm trying to move bitstream filters into libavcodec/bsf/, but want to keep libavcodec/ in include search path
<elenril>
but something like $(OBJS_BSF): CPPFLAGS += -Ilibavcodec/ does not seem to work
<elenril>
any ideas why
<haasn>
another|: I hope not
<haasn>
unless videolan hotel burns down
<another|>
at least it'll be warm then
<courmisch>
several instances of the conference featured snow
<courmisch>
I don't recommend camping
<courmisch>
did anyone receive news of the hotel?
<durandal_1707>
why use multiple directories at all?
<durandal_1707>
fflabs have no more useful ideas?
<durandal_1707>
also you work on something only if paid
patricia_ has joined #ffmpeg-devel
patricia_ has left #ffmpeg-devel [#ffmpeg-devel]
<durandal_1707>
i will never merge 08bebeb1bed3feee99d30cd6066d3466291dc8ed
<durandal_1707>
librempeg have already more decoders, muxers and filters than ffmpeg
<durandal_1707>
That's just so typical of the newfangled digital bureaucracy: it's far easier to change the workflow than it is to fix the software.
jamrial has joined #ffmpeg-devel
<cone-580>
ffmpeg Haihao Xiang master:6414ecd85754: doc/ffmpeg: update the documentation about qsv device creation
<cone-580>
ffmpeg Haihao Xiang master:6e4d85a94ef0: doc/ffmpeg: update the documentation about vaapi device creation
<cone-580>
ffmpeg Stefano Sabatini master:261fd5c841fe: doc/muxer: lowercase raw in raw PCM muxers section
<cone-580>
ffmpeg James Almer master:9da0594f1d63: avformat/iamfenc: don't write empty packets
dellas has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
navi has joined #ffmpeg-devel
ccawley2011 has quit [Read error: Connection reset by peer]
dellas has quit [Remote host closed the connection]
Krowl has joined #ffmpeg-devel
cone-580 has quit [Quit: transmission timeout]
dellas has joined #ffmpeg-devel
<Marth64>
elenril: DVD format does have a table of chapters durations ("programs" within the context of a PGC). I can get this from dvdnav or directly from the IFO, easy. But, it is imprecise (as is the overall duration). The common players use this table for visual/GUI purposes but not for accurate indexing. The chapter points are only reached accurately by seeking to them. You can see this, for example by playing certain DVDs on VLC and noticing that the
<Marth64>
I can do this same seeking.. BUT, then, I lose the ability to build an accurate index, because discontinuities can occur during playback when considering the entire stream (I need to keep track the duration of each segment). The challenge is, with ffmpeg, this has to be done during the read header phase.
<elenril>
demuxers are not required to guarantee accurate anything
<elenril>
they should export what's stored in the file, that's it
<elenril>
callers can build accurate indexing on top of that, if they wish
<elenril>
e.g. ffms2
<Marth64>
Well, I was hoping to settle on a middle ground.. Leave the option to read the table as a default, and get the user through muxing quickly. Then provide a "-accurate_chapters", the slow option by choice.
<Marth64>
Because the inaccurate times are annoying, they cut just before new scenes (the actual points)
<Marth64>
I already have the code for both so it's not a big time loss
psykose has quit [Remote host closed the connection]
<Marth64>
I'll send that up in the next 2 days along with a few other fixes. I've reached about the end with this work. Everything works well now for the discs I have been testing with (~100).
<Marth64>
I need to move on from this cursed work, want to do subtitle things\
<Marth64>
j-b was right, there is no clean way to deal with DVD, it's horror
<Traneptora>
that line looks straight outa wm4
<elenril>
>move on from this cursed work
<elenril>
>do subtitle things
<elenril>
does not compute
<Marth64>
from one cursed work to another
<courmisch>
arg toasted
<Marth64>
fudge dvd
mkver has quit [Remote host closed the connection]
<courmisch>
on a trip in .my InTheWings and I found legit free Video CDs
mkver has joined #ffmpeg-devel
<courmisch>
though it was religious propaganga from the local temple
<elenril>
now that is cursed
<Traneptora>
VCDs, damn
<Traneptora>
do they expect people to even be able to play them?
<Marth64>
I can almost visualize the blocky crusty mpeg-1 visuals of it
HarshK23 has quit [Quit: Connection closed for inactivity]
<Traneptora>
mpeg-1 compression on a 650 Meg medium
<Traneptora>
hm.
<Traneptora>
at least DVDs are large enough that you can have like 6 Mbps a 720x480
<Marth64>
think it's like 352x320 for ntsc. abysmal, but great for 1996
<Traneptora>
did it actually beat VHS though, realistically
<Marth64>
I preferred the warm/cozy feel of analog VHS at that time :D so not to me
<jamrial>
I remember you could burn a VCD player for the Dreamcast
<Marth64>
<3 dreamcast
<Traneptora>
I grew up on VHSs but I never considered them cozy
<Traneptora>
in fact, I always hated that VHS tapes would have adverts in the beginning for movies that already existed
<Traneptora>
like "why is there a preview for a movie in this? did they not think that this tape is going to last longer than 2 years"
<Traneptora>
and unlike with DVDs you actually had to fast forward
<kierank>
scarcity
<Traneptora>
though the tacticle sensation of pushing a VHS into a VCR was pretty nice
<jamrial>
until it wouldn't go all the way in and you'd press the eject button in panic hoping it would spit it out without breaking it
<Marth64>
all points granted. but I'll say at least with VHS the fast forward was real...DVD literally calls FF "trick play"...VHS you can actually see the motion as you cruise fast forward
<Traneptora>
well when I was a child that panic wouldn't be there I'd just break it and not know it
epony has quit [Remote host closed the connection]
<Marth64>
off to get coffee and try to track down a DVD with SDDS. y'all have a good day
mkver has quit [Ping timeout: 252 seconds]
aljazmc has joined #ffmpeg-devel
<Lynne>
worst with vhs was tape tracking
mkver has joined #ffmpeg-devel
aljazmc has quit [Remote host closed the connection]
epony has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
<Lynne>
durandal_1707: no, you'll need a lot more digits to express the planck constant accurately enough to do reliable calculations with it
<Traneptora>
Lynne: we require sizeof(int) to be 4 for asm? why? can't we just use uint32_t which is specified to be 32-bit
<Lynne>
but hey, if we ever do get precision like that, it'll be guaranteed that we'll be able to accuratemly compute everything physically applicable
<Lynne>
we have so many ints through the codebase, even our stride is an int
<Traneptora>
yea, but all the entry and exit points for asm can just be guarded or somth
<Traneptora>
that said I don't like that we use int and not ptrdiff_t for strides
<Traneptora>
or that we use int and not size_t for buffer sizes
<Lynne>
we don't even guard the asm init functions properly and rely on DCE
<Traneptora>
oh well that sounds like the issue then
<Lynne>
you'd be the 5th or 6th person to try to fix that if you do want to
<Lynne>
no one cares much about it because it's only an issue on msvc
<Lynne>
gcc and clang will do DCE even on -O0
<Lynne>
I've proposed changing the stride to 64 bits, but there's a lot of asm to fix
<Lynne>
we should probably do this while we still have less asm than dav1d
<durandal_1707>
perfect job task for FFlabs employee
ccawley2011 has quit [Read error: Connection reset by peer]
<durandal_1707>
what is meme hiearchy in FFmpeg and/or FFlabs?
<Lynne>
how would the ABI even look like for 64bit soft ints on 32-bit systems?
<Lynne>
a function to sanitize 64bit strides to 32bits on each function call?
<kierank>
21:53:08 <durandal_1707> perfect job task for FFlabs employee
<kierank>
I'm sure big corps are desperate for this feature
<nevcairiel>
stride should be a register-sized variable like ptrdiff_t or size_t, you avoid ABI nonsense with 64-bit ints on 32-bit systems that way and just always have a naturally register sized variable, going 64 on all systems is just extra pain for no gain
<Lynne>
oh, right, that was an option too
<kierank>
I forget what the objection to ptrdiff_t was
<jamrial>
there wasn't. ptrdiff_t is used in a lot of functions for strides
<kierank>
there was one thing that I forget
<kierank>
that someone objected about that was semi-reasonable