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> the array index is just the n-th bit that was 1, and the value its index
<BtbN> So that paste just means the 11th mask flag was 1
<BtbN> And I'm not sure if it even makes sense for so many of the bits to be 1
<jamrial> not possible, the samples created by nvenc with alpha should only have the fourth bit set
<jamrial> meaning, idx 3
<jamrial> rest should be 0
<BtbN> Yeah, something is wrong then
<jamrial> you can check a reduced implementation in lavc/hevc/ps.c
<jamrial> it makes a lot of assumptions and only works for multiview, but it at least parses that part
<BtbN> I think my issue must be somewhere before that
<BtbN> Like, the reader must be completely misaligned from the get-go
<Thul> BBB: Forgejo is a fork of Gitea, in case that does ring a bell for you.
Thul is now known as Thulinma
<BtbN> What makes me think that I can't be totally off is that disabling the alpha flag to nvenc swaps the vps_extension_flag to 0
<jamrial> can you push your current code to your github?
<jamrial> for 64 bits you need skip_bits_long i think
<BtbN> ah. Though that shouldn't matter. There is no timing info
MrZeus__ has joined #ffmpeg-devel
<BtbN> There's always the chance that nvenc generates bogous data
<BtbN> But hm
<BtbN> jamrial: I've done something silly, and now the correct data comes out. align_get_bits(gb); skip_bits(gb, 1); align_get_bits(gb);
<BtbN> So the parser must be missing at least two bits before it gets to the align_get_bits part
<BtbN> (based on the positions I logged)
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 260 seconds]
witchymary has joined #ffmpeg-devel
MisterMinister has quit [Ping timeout: 244 seconds]
<jamrial> it was not calling hvcc_parse_ptl() because of the condition being wrong
<BtbN> ah
<BtbN> funny that it perfectly aligned with aligning again
<BtbN> I was currently double-checking the ptl parse function itself against the spec...
witchymary has quit [Remote host closed the connection]
arch1t3cht4 has joined #ffmpeg-devel
Marth64 has joined #ffmpeg-devel
arch1t3cht has quit [Ping timeout: 245 seconds]
arch1t3cht4 is now known as arch1t3cht
witchymary has joined #ffmpeg-devel
iive has quit [Quit: They came for me...]
witchymary has quit [Remote host closed the connection]
witchymary has joined #ffmpeg-devel
<BtbN> It's too late to try and understand the splitting_flag case. That's for tomorrow.
MrZeus__ has quit [Read error: Connection reset by peer]
MrZeus__ has joined #ffmpeg-devel
MrZeus__ has quit [Read error: Connection reset by peer]
thilo has quit [Ping timeout: 272 seconds]
thilo has joined #ffmpeg-devel
thilo has quit [Changing host]
thilo has joined #ffmpeg-devel
MrZeus has joined #ffmpeg-devel
MrZeus has quit [Ping timeout: 252 seconds]
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 260 seconds]
<fflogger> [newticket] Xorgon: Ticket #11336 ([ffmpeg] Pipe output produces file with incorrect duration metadata) created https://trac.ffmpeg.org/ticket/11336
Marth64 has quit [Quit: Leaving]
Mirarora has quit [Quit: Mirarora encountered a fatal error and needs to close]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 246 seconds]
Mirarora has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
Marth64 has joined #ffmpeg-devel
MisterMinister has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
MrZeus has joined #ffmpeg-devel
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 246 seconds]
System_Error has joined #ffmpeg-devel
MrZeus has quit [Ping timeout: 252 seconds]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 248 seconds]
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 246 seconds]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 244 seconds]
System_Error has quit [Remote host closed the connection]
jamrial has quit []
System_Error has joined #ffmpeg-devel
HarshK23 has quit [Quit: Connection closed for inactivity]
Martchus_ has quit [Ping timeout: 260 seconds]
Martchus has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 246 seconds]
Martchus has joined #ffmpeg-devel
<fflogger> [editedticket] MasterQuestionable: Ticket #11336 ([avformat] Pipe output for AVI resulted incorrect duration metadata) updated https://trac.ffmpeg.org/ticket/11336#comment:1
Martchus has quit [Ping timeout: 260 seconds]
cone-431 has joined #ffmpeg-devel
<cone-431> ffmpeg Gyan Doshi master:914f89dbc617: avformat/movenc: log stream index for invalid pkt duration
Martchus has joined #ffmpeg-devel
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 252 seconds]
Marth64 has quit [Quit: Leaving]
Sean_McG has quit [Quit: Lost terminal]
Martchus_ has quit [Read error: Connection reset by peer]
Martchus has joined #ffmpeg-devel
Martchus_ has joined #ffmpeg-devel
HarshK23 has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 265 seconds]
<nevcairiel> fate-color_utils fails over here (and here https://fate.ffmpeg.org/report.cgi?time=20241207052051&slot=x86_32-mingw-w64-dll-windows-native) anyone any thoughts?
Martchus has joined #ffmpeg-devel
<nevcairiel> (also fmpeg-error-rate-fail keeps failing)
Martchus_ has quit [Ping timeout: 244 seconds]
Martchus has quit [Ping timeout: 246 seconds]
Martchus has joined #ffmpeg-devel
cone-431 has quit [Quit: transmission timeout]
cone-343 has joined #ffmpeg-devel
<cone-343> ffmpeg Rémi Denis-Courmont release/6.1:30cd0d7bd060: riscv: test for assembler support
Martchus has quit [Ping timeout: 276 seconds]
Martchus has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 265 seconds]
Martchus has joined #ffmpeg-devel
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 252 seconds]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
<cone-343> ffmpeg Frank Plowman master:34c6ad0a077a: lavc/vvc: Use a bitfield to store MIP information
klaxa has quit [Remote host closed the connection]
klaxa has joined #ffmpeg-devel
<nevcairiel> lowering the fail threshold in the color_utils trc test to 1e-7 instead of 1e-8 makes it pass, its such a minute error that its not even printed, and the test data is specified only to 6 digits in this test .. seems like an acceptable change?
<nevcairiel> haasn: ^^ ? :)
___nick___ has joined #ffmpeg-devel
___nick___ has quit [Client Quit]
___nick___ has joined #ffmpeg-devel
<haasn> Oh yes, that patch LGTM
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 252 seconds]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 265 seconds]
j45_ has joined #ffmpeg-devel
j45 has quit [Ping timeout: 276 seconds]
j45_ is now known as j45
j45 has quit [Changing host]
j45 has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 244 seconds]
Martchus has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
cone-343 has quit [Quit: transmission timeout]
MyNetAz has quit [Read error: Connection reset by peer]
MyNetAz has joined #ffmpeg-devel
<BtbN> Good god, what were they thinking with that splitting_flag stuff. "Let's save 10 bits, and introduce the most arcane magic to achieve it"
<BBB> courmisch: re: your comment about backporting the rv40dsp riscv asm patch, I think you're in the best position to suggest a course of action
<BBB> courmisch: if you want to merge it as-is and then later change it to use the macros, I'm fine with that, if you think that's best
cone-096 has joined #ffmpeg-devel
<cone-096> ffmpeg Hendrik Leppkes master:5215ec677c55: avutil/tests/color_utils: reduce accuracy threshold to pass to 1e-7
<jamrial> BtbN: i recommend you for your sanity to just bail out if vps_base_layer_internal_flag == 0
<BtbN> Why that flag? Doesn't that just skip parsing the profile stuff?
<BtbN> I didn't plan to parse any further than the dimension_id[ i ][ j ] fields. They contain all the info that's needed to find the alpha layer
<jamrial> BtbN: i think it signals that the base layer and upper layers are in separate bitstreams or so
mkver has joined #ffmpeg-devel
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 252 seconds]
witchymary has quit [Remote host closed the connection]
witchymary has joined #ffmpeg-devel
lexano has quit [Quit: Leaving]
elvis_a_presley has quit [Quit: smoke-bomb ; grapple-hook]
elvis_a_presley has joined #ffmpeg-devel
iive has joined #ffmpeg-devel
<Traneptora> kasper93: do you have commit access to ffmpeg? if not, pong me when you have an updated commit message (re: af_ashowinfo) and I can push
lexano has joined #ffmpeg-devel
<fflogger> [newticket] pablo: Ticket #11337 ([undetermined] asegment with relative timestamps creates segments with wrong duration) created https://trac.ffmpeg.org/ticket/11337
<fflogger> [editedticket] MasterQuestionable: Ticket #11337 ([avfilter] "asegment" with relative timestamps created segments of wrong duration) updated https://trac.ffmpeg.org/ticket/11337#comment:1
<fflogger> [editedticket] thebombzen: Ticket #11337 ([avfilter] "asegment" with relative timestamps created segments of wrong duration) updated https://trac.ffmpeg.org/ticket/11337#comment:2
<Traneptora> I lied, I missed v2
<Traneptora> thanks
cone-096 has quit [Quit: transmission timeout]
cone-688 has joined #ffmpeg-devel
<cone-688> ffmpeg Kacper Michajłow master:790270276225: avfilter/af_ashowinfo: fix scaling factor of replaygain peak values
<Traneptora> is it possible to change trac's username?
<Traneptora> trac is still using my old thebombzen username, but I'm using Traneptora now
Martchus_ has quit [Ping timeout: 264 seconds]
Martchus has joined #ffmpeg-devel
witchymary has quit [Remote host closed the connection]
<thardin> good question. I might be inclined to change username as well
Marth64 has joined #ffmpeg-devel
<BtbN> not really, no
witchymary has joined #ffmpeg-devel
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 252 seconds]
zenmov has quit [Ping timeout: 245 seconds]
zenmov has joined #ffmpeg-devel
<BtbN> jamrial: https://github.com/BtbN/FFmpeg/commit/1d88055fd2f657166a2dd53e1b1d8702b9664e44 this should be the final version of it. No idea what the point of the splitting_flag==1 logic is. I implemented it by following the spec to the letter, so it's likely correct. But I don't have samples to test.
<jamrial> BtbN: seems to work with the output of nvenc (after forcing it to not bother with nvEncGetSequenceParams and just export the PS in the first packet)
<BtbN> I need to verify if it really just works for me now
<BtbN> It looks like it just works, but I got nothing to check with
<jamrial> BtbN: http://pastie.org/p/4JpxZRJedG0eCfFcsTngMq tmp is ugly :p
<BtbN> In theory vps_max_layer_id and vps_num_layer_sets_minus1 aren't even needed
<BtbN> but I'm not sure how reliable execution order is when writing an equation that calls functions
<BtbN> apparently it's implementation defined. So can't do that.
Everything has joined #ffmpeg-devel
<jamrial> BtbN: the alpha parameter sets are not in hvcC because nvEncGetSequenceParams() still wont output them
<jamrial> so they are added to the first packet, and the muxer should extract them from it
<BtbN> hm, so now it's on Nvidia to fix that I guess
<jamrial> yeah
<BtbN> But it does emit _some_ alpha info
<BtbN> otherwise the muxer wouldn't find the alpha layer
cone-688 has quit [Quit: transmission timeout]
___nick___ has quit [Ping timeout: 244 seconds]
kasper93_ has joined #ffmpeg-devel
kasper93 has quit [Ping timeout: 272 seconds]
<BtbN> If I add that hack, all my browsers refuse to play the file completely. Which I guess means success?
<jamrial> BtbN: yeah, that one has all the PS in hvcC and alpha slices in packets
<jamrial> so now it's up to nvidia
<BtbN> they said they're all leaving for christmas holiday already
<BtbN> so not before next year
<BtbN> I'd say I'll try to figure out how to do it with x265 in the mean time. But my god their code is such a mess
PAUL007 has joined #ffmpeg-devel
TheSashmo has joined #ffmpeg-devel
PAUL007 has quit [Quit: Client closed]
<fflogger> [editedticket] Balling: Ticket #7799 ([avcodec] hevc_mp4toannexb filter inserts extradata wrong position (hevc_cuvid ffplay too)) updated https://trac.ffmpeg.org/ticket/7799#comment:27
cone-768 has joined #ffmpeg-devel
<cone-768> ffmpeg Paul B Mahol master:8050a7955e37: avformat/mm: fix packets pts generation and add seek support
darkapex has quit [Remote host closed the connection]
darkapex has joined #ffmpeg-devel
<cone-768> ffmpeg Osamu Watanabe master:c893dcce312a: avcodec/jpeg2000: Fix undefined behaviour in left shift operations