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 7.0.2 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
<Lynne> wow, drivers are hopeless at writing SPS/PPS headers
<Lynne> it was common to joke about AMD's drivers writing VUIs with full_range_flag always set to 1, but at least they didn't get something wrong like golomb codes
<Lynne> nvidia chokes up coding scaling list deltas, but says they're written correctly, and says they totally weren't overwritten
<Lynne> even though they're very very disabled
thilo_ has quit [Ping timeout: 248 seconds]
thilo_ has joined #ffmpeg-devel
<Lynne> correction, its in radv, we can fix this!
auri has quit []
auri has joined #ffmpeg-devel
feiw has quit [Remote host closed the connection]
feiw has joined #ffmpeg-devel
cone-283 has quit [Quit: transmission timeout]
jamrial has quit []
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 246 seconds]
arch1t3cht7 has joined #ffmpeg-devel
arch1t3cht has quit [Ping timeout: 248 seconds]
arch1t3cht7 is now known as arch1t3cht
tufei has quit [Remote host closed the connection]
tufei has joined #ffmpeg-devel
cone-355 has joined #ffmpeg-devel
<cone-355> ffmpeg Lynne master:034e25e1a468: hw_base_encode: move VAAPI SPS/PPS constructors to a shared file
<cone-355> ffmpeg Lynne master:9db68ed042a9: hw_base_encode: refactor picture allocation/freeing
<cone-355> ffmpeg Lynne master:e6019ed075d6: hw_base_encode: allocate DPB image upfront
<cone-355> ffmpeg Lynne master:0a1f724bf4a2: vulkan_video: add utilities for H264 level/profile mapping
<cone-355> ffmpeg Lynne master:fd86c373c281: vulkan: use correct return codes for query errors
<cone-355> ffmpeg Lynne master:3de73f126273: hw_base_encode: make recon_frames_ref optional
<cone-355> ffmpeg Lynne master:06483d039a8e: vulkan: error out if query is called without being initialized
<cone-355> ffmpeg Lynne master:66e950fcacf2: vulkan_video: move imageview creation and DPB fields to common context
<cone-355> ffmpeg Lynne master:c41ef7f2ff80: hwcontext_vulkan: add PREP_MODE_GENERAL for non-transfer_dst images
<cone-355> ffmpeg Lynne master:eb5088d28cdf: vulkan_filter: require storage images properly, set usage flags explicitly
<cone-355> ffmpeg Lynne master:0ffa96717035: hwcontext_vulkan: ask for storage images by default
<cone-355> ffmpeg Lynne master:620822c0c5c0: vulkan: add a ff_vk_init function
<cone-355> ffmpeg Lynne master:9c6532581986: vulkan_decode: use ff_vk_init
<cone-355> ffmpeg Lynne master:3415e0533f97: hwcontext_vulkan: disable more false positive validation checks
<cone-355> ffmpeg Lynne master:5e9845f11edd: vulkan(_decode): fix, simplify and improve queries
<Lynne> really tested the hell out of this
tufei_ has joined #ffmpeg-devel
tufei has quit [Ping timeout: 260 seconds]
Martchus_ has joined #ffmpeg-devel
microchip_ has quit [Ping timeout: 265 seconds]
Martchus has quit [Ping timeout: 246 seconds]
microchip_ has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 246 seconds]
microchip__ has joined #ffmpeg-devel
microchip_ has quit [Ping timeout: 276 seconds]
feiw has quit [Remote host closed the connection]
feiw has joined #ffmpeg-devel
feiw has quit [Remote host closed the connection]
feiw has joined #ffmpeg-devel
Martchus has joined #ffmpeg-devel
microchip__ is now known as microchip_
HarshK23 has joined #ffmpeg-devel
tufei_ has quit [Remote host closed the connection]
tufei_ has joined #ffmpeg-devel
mkver has quit [Remote host closed the connection]
mkver has joined #ffmpeg-devel
MetaNova has quit [Ping timeout: 276 seconds]
MetaNova has joined #ffmpeg-devel
ngaullier has joined #ffmpeg-devel
tufei_ has quit [Remote host closed the connection]
tufei_ has joined #ffmpeg-devel
cone-355 has quit [Quit: transmission timeout]
psykose has quit [Remote host closed the connection]
psykose has joined #ffmpeg-devel
psykose has quit [Remote host closed the connection]
psykose has joined #ffmpeg-devel
rajivharlalka6 has joined #ffmpeg-devel
feiw has quit [Remote host closed the connection]
feiw has joined #ffmpeg-devel
mkver has quit [Ping timeout: 260 seconds]
acryo has joined #ffmpeg-devel
acryo has quit [Changing host]
acryo has joined #ffmpeg-devel
rajivharlalka6 has quit [Quit: The Lounge - https://thelounge.chat]
feiw has quit [Remote host closed the connection]
feiw has joined #ffmpeg-devel
rajivharlalka5 has joined #ffmpeg-devel
rajivharlalka has quit [Quit: The Lounge - https://thelounge.chat]
rajivharlalka5 is now known as rajivharlalka
arbitercoin has joined #ffmpeg-devel
<Daemon404> i saw the uncompressed video in isobmff spec eas finally published fully and went to look at it
<Daemon404> itbis comparable to a lovecraftian horror
<Daemon404> with no test vectors
<Daemon404> y4m is safe.
tufei__ has joined #ffmpeg-devel
<nevcairiel> it sounds like such a simple task, how can they mess it up
<nevcairiel> y4m manages the same with a single line of text headers
<Daemon404> super super overcomplicated
<Daemon404> and weird compat shit
<Daemon404> the spec is dozens of pages.
<Daemon404> overcomplicated while not solving all the issues with y4m mught i add
<Daemon404> it is peak design-by-committee
<Daemon404> probably why it took years.
<elenril> I still find it amazing how mxf, with all its insane complexity, can only handle CFR
<ramiro> wbs: regarding the getauxval patch, I noticed that some of its uses are under #ifdef __linux__ (such as in lavu/aarch64/cpu.c). it's best to avoid system ifdefs if we can properly detect the feature in configure. wouldn't #if HAVE_GETAUXVAL be enough? (and then "|| HAVE_ELF_AUX_INFO" later on)
<wbs> ramiro: yes, that's probably better in the long run
tufei_ has quit [Ping timeout: 260 seconds]
<Daemon404> elenril, mxf vfr idea: file per framerate change and xml sidecarriage to link thek
<Daemon404> imf for vfr
<elenril> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
* elenril stabs Daemon404
<Daemon404> youre welcome
<JEEB> Daemon404: yea they did the same for raw audio
<Daemon404> wut?
<Daemon404> oh mp4
<JEEB> yea
<Daemon404> not imf
<Daemon404> mixing my trolling
<Daemon404> :D
<JEEB> overcomplication galore
<Daemon404> yea
rossy has quit [Remote host closed the connection]
<ramiro> Daemon404: reminds me of that xml quote from attila that michael used as a signature :P
<ramiro> If you really think that XML is the answer, then you definitly missunderstood the question -- Attila Kinali
rossy has joined #ffmpeg-devel
<Daemon404> :D
<kurosu> Daemon404: now, only if you knew people in Netflix and Apple to whom to complain. Possibly even the authors of these specs.
cone-279 has joined #ffmpeg-devel
<cone-279> ffmpeg Brad Smith master:fe4b9ef69f10: avutil/cpu_internal: Provide ff_getauxval() wrapper for getauxvaul()
<Daemon404> kurosu, i tried during its development and failes :)
<Daemon404> failed*
<kurosu> I was mostly kidding. I can't blame you for not sinking a month of your life every year during MPEG meetings on these topics
<Daemon404> im not even a memher
<Daemon404> and have no way to be one
<kurosu> Indeed, first step would be to be with a company that has registered through ANSI or ITU.
<cone-279> ffmpeg Marvin Scholz master:9556379943b6: lavfi/avfiltergraph: fix leak on error
<vjaquez> c!
<Lynne> see, it wasn't that hopeless
<Lynne> vulkan encoding works fully with nvidia, and with radv with https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31086
<JEEB> nice
<elenril> it sure would be nice to have at least some testing for all this hwaccel code
<cone-279> ffmpeg Lynne master:1709f3830e14: fftools/opt_common: add missing newline after printing codecs
<JEEB> elenril: I think thankfully at least the hwaccel stuff in theory has a variable in FATE, but for the rest there isn't I think
<Lynne> hw_base_encode is the sort of stuff that would keep elenril up at night if he decided to give it a look
<elenril> jkqxz approved it, can't be that bad
<elenril> and it's based off his vaapi encode, which should be decent
<Lynne> it was sort of ripped out of there, internals and all
<JEEB> hw stuff reminds me of the whole X vs A RGB formats, which I recently hit with a DRM module
<elenril> though of course it should be called hw_encode_base
<Lynne> so it expects a very specific initialization sequence that isn't really documented
<elenril> documentation is for the weak
<Lynne> I was wondering why my B-frames had wrong PTS for 3 hours until I found out b_per_p was 0 because I called gop_init a few lines under
<elenril> go beyond documentation and kick apis to the curb
<elenril> row row
<Lynne> fight the mower
___nick___ has joined #ffmpeg-devel
<JEEB> :D
<cone-279> ffmpeg Brad Smith master:a3f79fd22a36: aarch64: Implement support for elf_aux_info(3) on FreeBSD and OpenBSD
acryo has quit [Quit: ZNC 1.8.2 - https://znc.in]
acryo has joined #ffmpeg-devel
acryo has quit [Changing host]
acryo has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
<elenril> jamrial: if you move lcevc bitstream filter to lavf, then does the codec id still need to exist?
<jamrial> yes if we want to export the enhancement stream for caller handling and remuxing purposes
<elenril> but do we?
blb has quit [Quit: blb]
___nick___ has quit [Ping timeout: 276 seconds]
___nick___ has joined #ffmpeg-devel
blb has joined #ffmpeg-devel
cone-279 has quit [Quit: transmission timeout]
mkver has joined #ffmpeg-devel
<jamrial> elenril: what's the alternative? we want muxing support for split enhancement metadata
<elenril> if you say so
<elenril> I just didn't see any muxing patches yet
<jamrial> haven't gotten to it yet
<jamrial> right now i'm making the decode.c code opaque (so we don't litter it with lcevc specifc stuff), as well as trying to do the merging in lavf's demux.c instead of ffmpeg_demux.c
<jamrial> which i assume you will appreciate :p
<elenril> eh, I'm not that sure
<elenril> lavf generic code could use a lot of untangling
<elenril> I'm not objecting to handling it in CLI in principle, it just shouldn't affect the main patch so much
<elenril> s/patch/path
cone-326 has joined #ffmpeg-devel
<cone-326> ffmpeg Anton Khirnov master:cfbf10412f9c: tests/fate/filter-audio: add a test for handling more than 64 channels
<cone-326> ffmpeg Anton Khirnov master:18d492ff41b2: lavfi/af_channelsplit: support arbitrary channel layouts
<cone-326> ffmpeg Anton Khirnov master:7dc81d33c241: lavfi/af_channelmap: remove an arbitrary limit on channel count
<cone-326> ffmpeg Anton Khirnov master:e37d1efe6575: lavfi/af_dialoguenhance: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:84fda7de7948: lavfi/af_channelsplit: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:7896b9543001: lavfi/af_earwax: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:ffbafbfdeddf: lavfi/af_crossfeed: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:408587c83673: lavfi/af_extrastereo: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:0ec382f4948e: lavfi/af_haas: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:c9c6f0743621: lavfi/af_hdcd: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:fdc6e3172576: lavfi/af_headphone: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:48d9a4d7ca75: lavfi/af_join: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:bfb139bb8d10: lavfi/af_ladspa: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:e13592c22d48: lavfi/af_loudnorm: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:7887d00ee6e2: lavfi/af_volume: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:2ee4625d4bac: lavfi/af_virtualbass: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:d90035c94137: lavfi/af_surround: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:c4d9282a5688: lavfi/af_stereowiden: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:449f670a14b1: lavfi/af_stereotools: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:4c4199758c9d: lavfi/af_sofalizer: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:75b2f5a7d316: lavfi/af_sidechaincompress: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:363d4a75593b: lavfi/af_replaygain: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:d0f4c08207e3: lavfi/af_lv2: convert to query_func2()
<cone-326> ffmpeg Anton Khirnov master:b38f4e7f865d: lavfi/af_pan: stop modifying private data in query_formats()
<cone-326> ffmpeg Anton Khirnov master:691e0a108247: lavfi/af_pan: convert to query_func2()
<frankplow> Does MSVC support C11 threads properly or is their implementation still non-standard?
<elenril> it doesn't even support atomics without special magic flags
<kasper93> I think threads support is ok. There is detach difference, that thread cannot detech itself using current handle
<kasper93> atomics are indeed still behind experimental flag
tufei_ has joined #ffmpeg-devel
tufei__ has quit [Remote host closed the connection]
Traneptora has quit [Quit: Quit]
ngaullier has quit [Ping timeout: 245 seconds]
___nick___ has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
___nick___ has joined #ffmpeg-devel
___nick___ has quit [Client Quit]
___nick___ has joined #ffmpeg-devel
<cone-326> ffmpeg Zhao Zhili master:071c4990265a: avfilter/unsharp: Call function directly rather than via function pointer
<jkqxz> elenril: I don't think I approved it; I did look at it a few times but have limited bandwidth.
<jkqxz> And yeah, the name is rather awful but what would actually be better?
<jkqxz> Lynne: "it expects a very specific initialization sequence that isn't really documented" - well what did you expect? The whole thing is "abstract away common parts of VAAPI and D3D12" not "make a coherent API".
<Lynne> just saying it can be improved, that's all
<jkqxz> I am shocked to hear that.
<elenril> jkqxz: you replied to the thread with something other than AAAAAAAAAAAAAAAAAA
<elenril> that counts as approval
<elenril> but yeah, I wish more people had "make a coherent API" as a personal goal
compnn has quit [Read error: Connection reset by peer]
compnn has joined #ffmpeg-devel
<another|> something something .... make an abhorrent API ... okay
<jkqxz> It must score at least one elenril point, since it doesn't have any global state!
<elenril> it could get more if it didn't have any local state either
<elenril> abandon state and be saved
rvalue- has joined #ffmpeg-devel
<jkqxz> Abandon state all ye who enter here.
rvalue has quit [Ping timeout: 260 seconds]
rvalue- is now known as rvalue
arbitercoin has quit [Ping timeout: 246 seconds]
<elenril> precisely
Krowl has joined #ffmpeg-devel
___nick___ has quit [Ping timeout: 260 seconds]
___nick___ has joined #ffmpeg-devel
cone-326 has quit [Quit: transmission timeout]
witchymary has quit [Ping timeout: 248 seconds]
iive has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
cone-631 has joined #ffmpeg-devel
<cone-631> ffmpeg Michael Riedl master:9bcb86b0fa58: avdevice/decklink_dec: extend available actions on signal loss
witchymary has joined #ffmpeg-devel
SystemError has quit [Remote host closed the connection]
SystemError has joined #ffmpeg-devel
iive has quit [Quit: They came for me...]