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
Guest44 has joined #ffmpeg-devel
Guest44 has quit [Client Quit]
HarshK23 has quit [Quit: Connection closed for inactivity]
Guest9 has joined #ffmpeg-devel
Guest9 is now known as finix
finix has quit [Client Quit]
jamrial has quit []
thilo has quit [Ping timeout: 246 seconds]
thilo has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
Mirarora has joined #ffmpeg-devel
abdu has joined #ffmpeg-devel
minimal has quit [Quit: Leaving]
MisterMinister has quit [Remote host closed the connection]
MisterMinister has joined #ffmpeg-devel
^Neo has quit [Ping timeout: 260 seconds]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 252 seconds]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
kasper93 has quit [Remote host closed the connection]
cone-296 has quit [Quit: transmission timeout]
abdu has quit [Ping timeout: 240 seconds]
jamrial has quit []
Guest92 has joined #ffmpeg-devel
Guest92 has quit [Client Quit]
cone-626 has joined #ffmpeg-devel
<cone-626>
ffmpeg Andreas Rheinhardt master:ad4f6a3d6e5f: avcodec/mpegvideo(_enc)?: Mark init, close functions as av_cold
<cone-626>
ffmpeg Andreas Rheinhardt master:bb0a72b10087: avcodec/mpegvideoenc: Move tmp bframes to MPVMainEncContext
<cone-626>
ffmpeg Andreas Rheinhardt master:f507f26e4763: avcodec/mpegvideoenc: Add MPVMainEncContext
<cone-626>
ffmpeg Andreas Rheinhardt master:3275d788ec7d: avcodec/mpegvideo: Move frame_skip_* fields to MPVMainEncContext
<cone-626>
ffmpeg Andreas Rheinhardt master:d8c7925e3291: avcodec/mpegvideoenc: Move ratecontrol to MPVMainEncContext
<cone-626>
ffmpeg Andreas Rheinhardt master:4d79175b9a2a: avcodec/mpegvideo_enc: Precalculate which frames to reconstruct
<cone-626>
ffmpeg Andreas Rheinhardt master:b7ee7994cc76: avcodec/h261enc: Make h261_encode_init() call ff_mpv_encode_init()
<cone-626>
ffmpeg Andreas Rheinhardt master:c757f948d1d7: avcodec/mpegvideo: Don't initialize [yc]_dc_scale_table by default
<cone-626>
ffmpeg Andreas Rheinhardt master:565e57ea87dc: avcodec/speedhqenc: Make speedhq_encode_init() call ff_mpv_encode_init()
<cone-626>
ffmpeg Andreas Rheinhardt master:9928b9e9faef: avcodec/speedhqenc: Avoid indirection
<cone-626>
ffmpeg Andreas Rheinhardt master:dcd4daee8c2a: avcodec/mjpegenc: Make mjpeg_encode_init() call ff_mpv_encode_init()
<cone-626>
ffmpeg Andreas Rheinhardt master:24dbc4c2e824: avcodec/mjpegenc: Simplify allocating huffman table
<cone-626>
ffmpeg Andreas Rheinhardt master:edebd1d9a465: avcodec/mpegvideo_enc: Only set fcode_tab if unset
<cone-626>
ffmpeg Andreas Rheinhardt master:34557b1d30ad: avcodec/mpeg12enc: Move encode_init() to the end of the file
<cone-626>
ffmpeg Andreas Rheinhardt master:6ecdbcc4540a: avcodec/mpeg12enc: Merge ff_mpeg1_encode_init() into ordinary init
<haasn>
ramiro: I found a way to integrate the new sws ops into checkasm
<haasn>
and it already uncovered bugs in my x86 impl, nice
<wbs>
haasn: nice
<haasn>
it's a slight bit of a hack
<haasn>
but nothing too egregious
witchymary has quit [Remote host closed the connection]
<haasn>
(problem mainly being that the function being tested also needs its associated private data that was generated at compile time, but the private data differs between the ref and the test func)
<haasn>
(so we just generate the private data twice, once for the backend chosen by the cpu flags and once for the C backend, and then assert(func_ref == func_c)
pross has quit [Ping timeout: 252 seconds]
witchymary has joined #ffmpeg-devel
IndecisiveTurtle has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
quietvoid has quit [Remote host closed the connection]
<ramiro>
haasn: cool. it is becoming a bit tedious to check things manually without checkasm
<ramiro>
haasn: I just had an issue with ssim. I accidentally interleaved the output pixels per vector, so instead of y1y2y3y4y5y6y7y8 I had y1y5y2y6y3y7y4y8. ssim did point out a difference, but a very small difference.
<ramiro>
since I'm testing this over ssh, I'm not checking the output visually. it took me a while to pinpoint what the error was
<JEEB>
chromium seems to be limiting matrix coefficients to zero like with PNG? if (matrix_coefficients != 0) return nullptr
<haasn>
jamrial: I would argue that erroring out here is the correct result and if the user wants a wrong output just for debug purposes they can use vf_setparams to override the csp metadata
<haasn>
anyway, should hopefully soon be a non-issue
<jamrial>
alright
<JEEB>
but yea, ICC.2022 apparently contains the definition
<JEEB>
would be lulz if it was limited to no matrix since JPEG is most commonly YCbCr
grillo_0 has quit [Quit: Ping timeout (120 seconds)]
grillo_0 has joined #ffmpeg-devel
abdu93 has quit [Ping timeout: 240 seconds]
abdu93 has joined #ffmpeg-devel
abdu93 has quit [Client Quit]
abdu93 has joined #ffmpeg-devel
ccawley2011 has joined #ffmpeg-devel
rvalue- has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 268 seconds]
rvalue- is now known as rvalue
abdu93 has quit [Quit: Client closed]
abdu93 has joined #ffmpeg-devel
abdu93 has quit [Quit: Client closed]
abdu93 has joined #ffmpeg-devel
MetaNova has quit [Ping timeout: 276 seconds]
MetaNova has joined #ffmpeg-devel
IndecisiveTurtle has joined #ffmpeg-devel
cone-159 has joined #ffmpeg-devel
<cone-159>
ffmpeg Martin Storsjö master:d256118b7e62: videotoolbox: Fix building with older SDKs
<Traneptora>
what's the point of adding cicpTag to ICC profiles
<Traneptora>
since ICC profiles are already a superset of those tags
<Traneptora>
in terms of what can be expressed. or is it just for convenience
<Traneptora>
like I can attach a PQ profile and then just label it as PQ so readers can read the tag even if they don't fully understand ICC
<JEEB>
Traneptora: for the transfer functions is what they mention in the spec
<JEEB>
IIRC the PQ ICC was inexact or so
<Traneptora>
but are there any H.273 transfers that aren't representable with non-LUT ICC Profiles?
<Traneptora>
oh
<JEEB>
probably HLG as well since it's a mix of gamma and logarithmic
<Traneptora>
iirc though H.273 doesn't have any black-point data
<Traneptora>
wouldn't it make more sene to just make a more expressive version of curv or para tags that can encode the PQ curve directly?
Guest52 has joined #ffmpeg-devel
<JEEB>
I think the nextgen version of ICC profiles added HDR support
<JEEB>
although not that I know the versions of ICC profiles :D
Guest52 has quit [Client Quit]
<JEEB>
if v4 is the current-gen ICC profile, then they just extended the one that is actually in use
<JEEB>
I do like the random limitations chromium then put onto interpreting the values
<Lynne>
maybe it has something to do with icc profiles being defined as part of tiff, and tiff needing an hdr flag?
PAUL007 has quit [Quit: Client closed]
pross has joined #ffmpeg-devel
<JEEB>
&34
s55 has quit [Quit: Bye]
s55 has joined #ffmpeg-devel
martinr1 has joined #ffmpeg-devel
s55 has quit [Read error: Connection reset by peer]
s55 has joined #ffmpeg-devel
<martinr1>
Lynne: I've finished the ProRes qualification task for GSoC and have a gh repo with all my work. What's the best way to share this with you? I can make it public but maybe this isn't a good idea - email?