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
<fflogger> [editedticket] Kentzo: Ticket #10741 ([avcodec] aac_at quality only settable with q:a, not global_quality:a) updated https://trac.ffmpeg.org/ticket/10741#comment:2
iive has quit [Quit: They came for me...]
Kei_N has quit [Ping timeout: 248 seconds]
rvalue has quit [Quit: 'forgot to sleep']
Kei_N has joined #ffmpeg-devel
deus0ww has quit [Quit: Textual IRC Client: www.textualapp.com]
blb has quit [Quit: brb]
thilo has quit [Ping timeout: 276 seconds]
cone-795 has joined #ffmpeg-devel
<cone-795> ffmpeg Michael Niedermayer release/7.1:f44a7207f018: avcodec/cfhdenc: Clear dwt_tmp
<cone-795> ffmpeg Michael Niedermayer release/7.1:d32543f52057: avformat/rmdec: check that buf if completely filled
<cone-795> ffmpeg Michael Niedermayer release/7.1:0da3deb079a6: avcodec/hapdec: Clear tex buffer
<cone-795> ffmpeg Michael Niedermayer release/7.1:14af28c2f5fa: avformat/mxfdec: Check that key was read sucessfull
<cone-795> ffmpeg Michael Niedermayer release/7.1:d322ffdef526: avformat/jpegxl_anim_dec: clear buffer padding
<cone-795> ffmpeg Michael Niedermayer release/7.1:5129ae45bf1c: avcodec/get_buffer: Use av_buffer_mallocz() for audio same as its done for video
<cone-795> ffmpeg Michael Niedermayer release/7.1:553517868db1: avfilter/vf_addroi: Add missing NULL termination to addroi_var_names[]()
<cone-795> ffmpeg Michael Niedermayer release/7.1:601616f2de42: avfilter/vf_grayworld: Use the correct pointer for av_log()
<cone-795> ffmpeg Michael Niedermayer release/7.1:b827ac49b770: avfilter/af_pan: Fix sscanf() use
<cone-795> ffmpeg Michael Niedermayer release/7.1:658029334e07: swscale/output: Fix undefined overflow in yuv2rgba64_full_X_c_template()
<cone-795> ffmpeg Michael Niedermayer release/7.1:da1e316a1f3d: avcodec/osq: Fixes several undefined overflows in do_decode()
<cone-795> ffmpeg Michael Niedermayer release/7.1:585c799b1088: avformat/iamf_parse: Check output_channel_count
<cone-795> ffmpeg Michael Niedermayer release/7.1:73fd3789955a: avcodec/aacsbr_template: Clear n_q on error
<cone-795> ffmpeg Michael Niedermayer release/7.1:5e02e1ba96ef: avcodec/vc1dec: Clear block_index in vc1_decode_reset()
<cone-795> ffmpeg Michael Niedermayer release/7.1:41f805228415: avformat/mov: Factorize sanity check out
<cone-795> ffmpeg Michael Niedermayer release/7.1:0bb8a38c10a1: avformat/mov: perform sanity checks for heif before index building
<cone-795> ffmpeg Michael Niedermayer release/7.1:5fed7010f858: avcodec/ffv1enc: Fix handling of 32bit unsigned symbols
<cone-795> ffmpeg Michael Niedermayer release/7.1:ece76d7dd466: avcodec/mpegvideo_enc: Check FLV1 resolution limits
<cone-795> ffmpeg Michael Niedermayer release/7.1:75385a729f6c: avfilter/vf_v360: Fix NULL pointer use
<cone-795> ffmpeg Michael Niedermayer release/7.1:ac9c710e1808: avutil/avstring: dont mess with NULL pointers in av_match_list()
<cone-795> ffmpeg Michael Niedermayer release/7.1:c3c7ecfe48d4: avformat/dashdec: Check whitelist
<cone-795> ffmpeg Michael Niedermayer release/7.1:5ee56ebd3048: avcodec/ffv1enc: Fix RCT for GBR colorspace
<cone-795> ffmpeg Michael Niedermayer release/7.1:30bef79c6929: avformat/vqf: Check avio_read() in add_metadata()
<cone-795> ffmpeg Michael Niedermayer release/7.1:c82813c96608: avformat/vqf: Propagate errors from add_metadata()
<cone-795> ffmpeg Michael Niedermayer release/7.1:f710d5af0a74: avcodec/cbs_vp9: Initialize VP9RawSuperframeIndex
<cone-795> ffmpeg Michael Niedermayer release/7.1:1646840988ed: avformat/wtvdec: Initialize buf
<cone-795> ffmpeg Michael Niedermayer release/7.1:43eaca54bc88: avformat/ipmovie: Check signature_buffer read
<cone-795> ffmpeg Michael Niedermayer release/7.1:5f337bd79837: avformat/iamf_reader: Initialize padding and check read in ff_iamf_read_packet()
<cone-795> ffmpeg Michael Niedermayer release/7.1:293e4ddd986f: avcodec/huffyuvdec: Initialize whole output for decode_gray_bitstream()
<cone-795> ffmpeg Michael Niedermayer release/7.1:a7a6d92e8418: avformat/mxfdec: Check avio_read() success in mxf_decrypt_triplet()
<cone-795> ffmpeg Michael Niedermayer release/7.1:dc07f98934ed: Changelog: update
thilo has joined #ffmpeg-devel
Mirarora has quit [Quit: Mirarora encountered a fatal error and needs to close]
blb has joined #ffmpeg-devel
Mirarora has joined #ffmpeg-devel
Mirarora has quit [Quit: Mirarora encountered a fatal error and needs to close]
Mirarora has joined #ffmpeg-devel
Mirarora has quit [Remote host closed the connection]
Mirarora has joined #ffmpeg-devel
^Neo has quit [Ping timeout: 272 seconds]
jamrial has quit []
mkver has quit [Ping timeout: 248 seconds]
cone-795 has quit [Quit: transmission timeout]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 265 seconds]
<fflogger> [editedticket] Gyan: Ticket #11428 ([undetermined] SIGFPE in xpsnr filter) updated https://trac.ffmpeg.org/ticket/11428#comment:1
zsoltiv has quit [Ping timeout: 252 seconds]
zsoltiv_ has quit [Ping timeout: 244 seconds]
desmond has joined #ffmpeg-devel
desmond has quit [Client Quit]
desmond has joined #ffmpeg-devel
desmond has quit [Client Quit]
desmond has joined #ffmpeg-devel
desmond has quit [Client Quit]
dliu has joined #ffmpeg-devel
Traneptora has joined #ffmpeg-devel
dliu has quit [Quit: gotta go]
HarshK23 has joined #ffmpeg-devel
ngaullier has joined #ffmpeg-devel
ngaullier has quit [Ping timeout: 264 seconds]
<fflogger> [editedticket] Gyan: Ticket #11428 ([undetermined] SIGFPE in xpsnr filter) updated https://trac.ffmpeg.org/ticket/11428#comment:2
ngaullier has joined #ffmpeg-devel
ngaullier has quit [Ping timeout: 272 seconds]
ngaullier has joined #ffmpeg-devel
<fflogger> [editedticket] DodoGTA: Ticket #7916 ([avformat] Incorrect seeking for module formats) updated https://trac.ffmpeg.org/ticket/7916#comment:4
<fflogger> [newticket] jozefchutka: Ticket #11429 ([undetermined] mp3float Header missing vs. Conversion failed!) created https://trac.ffmpeg.org/ticket/11429
ngaullier has quit [Remote host closed the connection]
ngaullier has joined #ffmpeg-devel
ngaullie has joined #ffmpeg-devel
ngaullier has quit [Ping timeout: 252 seconds]
thesynthax has joined #ffmpeg-devel
th3synth4x has joined #ffmpeg-devel
<th3synth4x> Hi everyone! My name is thesynthax. I just wanted to know when tentatively the GSoC 2025 project ideas list will be released. I know that the list is released before the official announcement of the mentoring organizations by Google.
<th3synth4x> Or if someone can just write down the tentative projects FFmpeg is planning this year
th3synth4x has quit [Client Quit]
thesynthax has quit [Quit: WeeChat 4.5.1]
thesynthax has joined #ffmpeg-devel
thesynthax is now known as th3synth4x
ngaullie has quit [Remote host closed the connection]
ngaullie has joined #ffmpeg-devel
stupidoo has joined #ffmpeg-devel
^Neo has joined #ffmpeg-devel
^Neo has quit [Changing host]
^Neo has joined #ffmpeg-devel
^Neo has quit [Ping timeout: 252 seconds]
DauntlessOne4 has quit [Quit: Ping timeout (120 seconds)]
DauntlessOne4 has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
ngaullie has quit [Ping timeout: 245 seconds]
stupidoo has quit [Quit: Client closed]
rvalue has joined #ffmpeg-devel
paulk-bis has quit [Quit: WeeChat 3.0]
paulk has joined #ffmpeg-devel
^Neo has joined #ffmpeg-devel
^Neo has quit [Changing host]
^Neo has joined #ffmpeg-devel
GoldenLeader has joined #ffmpeg-devel
ccawley2011__ has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
ccawley2011_ has joined #ffmpeg-devel
av500 has joined #ffmpeg-devel
ccawley2011__ has quit [Ping timeout: 248 seconds]
ccawley2011 has quit [Ping timeout: 260 seconds]
ngaullie has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
ccawley2011__ has joined #ffmpeg-devel
ccawley2011_ has quit [Ping timeout: 248 seconds]
ccawley2011 has quit [Ping timeout: 276 seconds]
jamrial has joined #ffmpeg-devel
stupidoo has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
ccawley2011_ has joined #ffmpeg-devel
ccawley2011__ has quit [Ping timeout: 245 seconds]
ccawley2011 has quit [Ping timeout: 244 seconds]
ccawley2011 has joined #ffmpeg-devel
ccawley2011__ has joined #ffmpeg-devel
ccawley2011_ has quit [Ping timeout: 246 seconds]
ccawley2011 has quit [Ping timeout: 265 seconds]
th3synth4x has quit [Ping timeout: 260 seconds]
jamrial_ has joined #ffmpeg-devel
thesynthax has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
jamrial has quit [Read error: Connection reset by peer]
ccawley2011_ has joined #ffmpeg-devel
ccawley2011__ has quit [Ping timeout: 248 seconds]
ccawley2011 has quit [Ping timeout: 245 seconds]
abdu has joined #ffmpeg-devel
ngaullie has quit [Ping timeout: 252 seconds]
thesynthax has quit [Quit: WeeChat 4.5.1]
ngaullier has joined #ffmpeg-devel
abdu has quit [Quit: Client closed]
ngaullie has joined #ffmpeg-devel
ngaullier has quit [Ping timeout: 252 seconds]
System_Error has quit [Ping timeout: 264 seconds]
ccawley2011 has joined #ffmpeg-devel
ccawley2011__ has joined #ffmpeg-devel
ccawley2011_ has quit [Ping timeout: 246 seconds]
ccawley2011 has quit [Ping timeout: 248 seconds]
ccawley2011__ has quit [Ping timeout: 264 seconds]
<Lynne> its already nice having a proper repo, even if its just for testing
<Lynne> BtbN: thanks for working on compile/fate automation too
<BtbN> Well, it failed yesterday cause it needs bloody nodejs in the image :D
^Neo has quit [*.net *.split]
Mirarora has quit [*.net *.split]
microchip_ has quit [*.net *.split]
bilboed has quit [*.net *.split]
DodoGTA has quit [*.net *.split]
wyatt8740 has quit [*.net *.split]
APic has quit [*.net *.split]
^Neo has joined #ffmpeg-devel
Mirarora has joined #ffmpeg-devel
microchip_ has joined #ffmpeg-devel
bilboed has joined #ffmpeg-devel
DodoGTA has joined #ffmpeg-devel
wyatt8740 has joined #ffmpeg-devel
APic has joined #ffmpeg-devel
microchip_ has quit [Read error: Connection reset by peer]
microchip_ has joined #ffmpeg-devel
ngaullie has quit [Remote host closed the connection]
ngaullier has joined #ffmpeg-devel
ngaullier has quit [Ping timeout: 265 seconds]
<GoldenLeader> haasn: added 8bit/16bit support, adding on top of this also generic support for bitstream flagged pixdesc formats will be very long task
<haasn> Agreed, though with my design it hopefully shouldn’t be too painful
<haasn> I finished all of the RGB variant wrappers except for the oddly packed ones
<haasn> But I’m not yet sure what will be a good design for combining them together
<haasn> I don’t want to create all N^2 pixfmt conversion possibilities
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
<GoldenLeader> heh, i'will try do to it anyway. just to mess with fun
<haasn> One advantage of the current sws unscaled special converters is that it can handle multiple pixfmts with a single function
<haasn> Whereas my approach currently requires the function to be hard coded to an exact combination of pixfmts
<haasn> Otherwise it will have to go through the special path
<haasn> General*
<GoldenLeader> haasn: https://pastebin.com/YAvA2x9N
<GoldenLeader> i noticed that sws does some additional stuff when doing upconversion to higher bitdepth...
<haasn> going from one src->fmt function call to two function calls with a temporary variable as the intermediate is about 20% slower
<haasn> I think we can survive this slowdown in the general path, not to need all NxM combinations of possible conversions
<haasn> GoldenLeader: afaict swscale does not really have special convertors for most bit depth upconversions
<haasn> so it goes through the general scaling path
ngaullier has joined #ffmpeg-devel
abdu has joined #ffmpeg-devel
stupidoo has quit [Ping timeout: 240 seconds]
<haasn> GoldenLeader: https://0x1.st/GdIq.txt example assembly output of rgb24 -> rgba routine
<haasn> looks pretty dense but somehow still slightly slower than swscale status quo
<haasn> that is comparing against rgb24tobgr32_mmxext
<GoldenLeader> how much is slightly? overall?
<haasn> 900 us vs 600 us roughly
<haasn> I mean, it's quite a lot slower in relative terms
<haasn> though most swscale functions are closer to 2000 us - 3000 us
<haasn> and they got 3x ish speedup from the new approach
<GoldenLeader> so its generally faster just current special path is sometimes faster
<haasn> yes
<haasn> well, this is testing with -march=native
<haasn> I haven't yet figured out which minimum level we would need to target for the code to be efficient
<haasn> since the scalar equivalent is very slow compared to swscale SIMD
<haasn> I may need to force the compiler to assume memory alignment
<GoldenLeader> if x86 simd special code is very good written than do not expect that compiler will be faster or at similar speed, even for old mmxext
<haasn> seems like -march=haswell is the minimum level to get it to produce SIMD
<haasn> sure
<haasn> my goal is not to remove all SIMD, just as much as possible
stupidoo has joined #ffmpeg-devel
<GoldenLeader> yea, but note, that for gcc, ffmpeg explicitly disables auto-vectorization...
<haasn> or rather, my goal is to not need to write so much SIMD :)
<GoldenLeader> yes...
<haasn> since the new approach is incompatible with the old
<haasn> interestingly clang produces better loops than gcc for me almost across the board
<GoldenLeader> you can play also with this: https://godbolt.org
<haasn> I do, though it's a pain to import my code there
<haasn> since it uses libavutil dependencies heavily
ngaullier has quit [Ping timeout: 252 seconds]
witchymary has quit [Ping timeout: 245 seconds]
<GoldenLeader> i wonder how to automate picking functions that are generated from templates, more templates  - more functions to take care of and properly map
<GoldenLeader> some sort of auto generating functions at runtime depending on pixdesc parameters
<GoldenLeader> so that every N^2 conversion can have SIMD
<GoldenLeader> i guess it would be much easier to write script which would write all conversion functions at once
ngaullier has joined #ffmpeg-devel
<haasn> seems like I can set loop vectorization pragmas on all loops to get it to produce faster code on older x86
<haasn> but still not as fast as with AXV
<haasn> AXV
<haasn> AVX even
<haasn> well, something we definitely could do is compile our template once with AVX enabled and once without and pick the right symbol table at runtime the "manual" way
<haasn> GoldenLeader: I'm also not sure yet, I was going to go with a big array approach
<haasn> we can have an array per platform / optimization level and pick the right one at runtime
<GoldenLeader> 2 endians * 16bit_depth * shift variations * depth variations - this grows very fast and there are many path gaps that remain unused, ....
<GoldenLeader> if i pick another approach, that would reduce from N^2 to 2*N functions
<GoldenLeader> but couldn't I use self-modifying C functions and still get N^2 C optimized functions at ease
ngaullier has quit [Ping timeout: 248 seconds]
<haasn> what we really need is a custom ABI
<GoldenLeader> why?
<haasn> to take and return value in %xmm registers
<haasn> without going through stack
<haasn> then you can just call multiple functions without performance overhead other than the jump/ret
System_Error has joined #ffmpeg-devel
ngaullier has joined #ffmpeg-devel
<GoldenLeader> see all 10 parameters passed in pf2pf_loop, the point is to unroll all such values into special functions, so they can get better compiler code generated.
ccawley2011__ has joined #ffmpeg-devel
ngaullier has quit [Remote host closed the connection]
ngaullier has joined #ffmpeg-devel
ngaullier has quit [Remote host closed the connection]
darkapex has quit [Remote host closed the connection]
darkapex has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
<Traneptora> BtbN: I logged into code.ffmpeg.org and it appears there's no repositories at all. is this intentional? or am I missing something?
<BtbN> There is exactly one so far (and peoples forks)
<Traneptora> ah, I had to click 'explore'
<Traneptora> it was showing personal ones
<BtbN> if you log in, you should see more I think
ccawley2011 has joined #ffmpeg-devel
ccawley2011__ has quit [Ping timeout: 252 seconds]
<Traneptora> ye, it was just that I didn't click "explore" to show ones I didn't own
<Traneptora> it was showing my repos, which was none
<Traneptora> just to be clear, the forgejo is a mirror, right?
<Traneptora> or can we actually send PRs to it and have it update the main repo
<GoldenLeader> demo, showcase..
<Traneptora> BtbN: for some reasons it's marking my gpg key as "suspicious" even though github never does. do you know why? e.g. https://code.ffmpeg.org/Traneptora/FFmpeg/commit/eb5fc63deb1e812c71d0520fce1dd13cb080e25d
<BtbN> no idea, I don't use gpg stuff
<BtbN> The error message seems to be relatively clear thouth?
<GoldenLeader> was it  compromised?
<BtbN> And it's not even a mirror
<BtbN> I just pushed stuff to it. It's purely a test so far
<Traneptora> well I just added the gpg key to my account and then pushed a commit to it with the key
<Traneptora> the ssh worked fine
<Traneptora> pushing the same commit to my github fork doesn't error out
<Traneptora> and running git verify-commit works locally
ccawley2011 has quit [Ping timeout: 272 seconds]
System_Error has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
GoldenLeader has quit [Quit: Client closed]
stupidoo has quit [Quit: Client closed]
witchymary has joined #ffmpeg-devel
abdu has quit [Quit: Client closed]
abdu has joined #ffmpeg-devel
<Traneptora> BtbN: update: apparently there's a bug where this happens ifyou add the gpg key before verifying your email. so I tried to click "send verification email" or whatever it's called in settings and never received one
<Traneptora> can you check logs? maybe there's some SMTP config issue
<BtbN> Check your spam folder
<BtbN> I get E-Mails fine, and bunch of other peoples did as well
<Traneptora> I can re-check, but I did
<BtbN> Just tested it, it sends mail to me just fine
<Traneptora> interesting
<Traneptora> wonder if it's an issue with the UI
<Traneptora> is pushing that button supposed to reload the page?
<BtbN> Which button?
<Traneptora> Send Activation https://0x0.st/8Xqo.png
abdu87 has joined #ffmpeg-devel
<BtbN> I doubt that'd refresh anything
<BtbN> do you have JS disabled?
<Traneptora> I do not
<BtbN> No clue then
<Traneptora> pushing the button causes the page to reload but nothing else happens
abdu has quit [Ping timeout: 240 seconds]
<BtbN> for Lynne it worked to just add another address
<BtbN> Maybe there is some bug about activating the initial E-Mail when in admin-approval mode
<Traneptora> it sends a post request to https://code.ffmpeg.org/user/settings/account/email and then it receives a 303 status with Location: /user/settings/account
<Traneptora> which causes the page to reload
GoldenLeader has joined #ffmpeg-devel
<Traneptora> BtbN: did some testing. added an alt email address and it activated it immediately (nothing required)
<Traneptora> made it primary, sent activation email to the original address and it worked as expected
<BtbN> Yeah, must be some confusion going on with the admin-approval-mode
ccawley2011 has quit [Ping timeout: 252 seconds]
odrling has quit [Remote host closed the connection]
odrling has joined #ffmpeg-devel
GoldenLeader has quit [Quit: Client closed]
stupidoo has joined #ffmpeg-devel
Traneptora has quit [Ping timeout: 260 seconds]
abdu has joined #ffmpeg-devel
LainExperiments has joined #ffmpeg-devel
abdu87 has quit [Ping timeout: 240 seconds]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
Everything has joined #ffmpeg-devel
ccawley2011 has quit [Client Quit]
Everything has quit [Ping timeout: 265 seconds]
Everything has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
DodoGTA has quit [Ping timeout: 252 seconds]
psykose has quit [Remote host closed the connection]
psykose has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
Everything has quit [Quit: Lost terminal]
abdu has quit [Ping timeout: 240 seconds]
abdu has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
abdu has quit [Quit: Client closed]
System_Error has joined #ffmpeg-devel