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
minimal has quit [Quit: Leaving]
<cone-332>
ffmpeg James Almer master:11032d819db2: swscale/swscale_unscaled: don't add offsets to more NULL pointers
BradleyS has quit [Read error: Connection reset by peer]
BradleyS_ has joined #ffmpeg-devel
BradleyS_ is now known as BradleyS
iive has quit [Quit: They came for me...]
nooticer has joined #ffmpeg-devel
nooticer has quit [Changing host]
nooticer has joined #ffmpeg-devel
mark4o has joined #ffmpeg-devel
markh has quit [Ping timeout: 276 seconds]
mark4o is now known as markh
Mirarora has quit [Quit: Mirarora encountered a fatal error and needs to close]
Mirarora has joined #ffmpeg-devel
_whitelogger has joined #ffmpeg-devel
<cone-332>
ffmpeg James Almer master:722a2170e832: avfilter/vf_curves: don't add offsets to NULL pointers
<BtbN>
Try again. Apparently mailman3 docs are flat out wrong that the two can coexist on the same postfix.
rvalue has quit [Read error: Connection reset by peer]
MisterMinister has quit [Ping timeout: 252 seconds]
<microchip_>
fo sho!
TheVibeCoder has joined #ffmpeg-devel
rvalue has quit [Ping timeout: 248 seconds]
rvalue has joined #ffmpeg-devel
bwu25 has quit [Ping timeout: 260 seconds]
mkver has joined #ffmpeg-devel
j45_ has joined #ffmpeg-devel
j45 has quit [Ping timeout: 248 seconds]
j45_ is now known as j45
j45 has quit [Changing host]
j45 has joined #ffmpeg-devel
<Lynne>
TheVibeCoder: those steps you said not to worry about? yeah, they're real, that's why the stairsteps happen
<Lynne>
a*x + b, with both a and b changing every 512 steps
<Lynne>
the change is slight but waveform was able to pick up on it, and so did my test program, though the noise made me think it was a colorspace issue
<Lynne>
dc-only, ac quant never changes with respect to the magnitude
<Lynne>
also dc == 0 -> 0, so that explains the jump
<TheVibeCoder>
Lynne: but output from videotoolbox, does not have those bands/steps in waveform
<Lynne>
yup, it has proper dequant
<Lynne>
I'm talking about that, its not a simple linear curve, but a piecewise-linear curve, where every 512 steps, the slope slightly increases
mkver has quit [Ping timeout: 252 seconds]
kurosu has quit [Quit: Connection closed for inactivity]
<kurosu>
haasn: re blackdetect, is movmask + popcnt + add faster than psabd+paddd (and a computation at the end of the loops)? I haven't looked closely at loops etc, mind you
<kurosu>
psadbw
<jamrial>
movmask is slow, so i doubt it
<haasn>
interesting idea, I'll give it a try sometime later
Everything has joined #ffmpeg-devel
System_Error has quit [Ping timeout: 244 seconds]
System_Error has joined #ffmpeg-devel
Mirarora has quit [Remote host closed the connection]
Mirarora has joined #ffmpeg-devel
<jamrial>
wbs: can you check what's different in the output of the mp4 apv test? those 18 extra bytes looks suspiciously like one extra frame_info entry is being written
<wbs>
jamrial: I'm a little short on time today and the coming week (I'm vacationing) unfortunately
<jamrial>
ok, no worries
<jamrial>
enjoy your time
<wbs>
thanks!
kurosu has quit [Quit: Connection closed for inactivity]
<TheVibeCoder>
ban vacations
<Lynne>
^^
minimal has joined #ffmpeg-devel
elvis_a_presley has quit [Quit: smoke-bomb ; grapple-hook]
elvis_a_presley has joined #ffmpeg-devel
Everything has quit [Quit: leaving]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
sepro3 has joined #ffmpeg-devel
sepro has quit [Ping timeout: 268 seconds]
sepro3 is now known as sepro
<BtbN>
new attempt at getting mailman2 and 3 to co-exist
<BtbN>
please tell me immediately if you notice anything wrong with the MLs
<BBB>
there's a lot wrong with our MLs, but not more or less than before ;)
<BBB>
"There are currently no mailing lists." is a bit problematic, practically speaking :-p
<Lynne>
hopefully we won't need it for much longer
<BtbN>
I doubt the MLs will go anywhere
<BtbN>
they'll stay if not for patches, for discussion
<BtbN>
And throwing away the archives would also suck
<BtbN>
So we need to get rid of mailman2 either way
<BtbN>
And there being no mailing lists is intended. Cause they would win over the mailman2 lists
<BtbN>
it now needs to be manually migrated, list by list
<kasper93>
discussions/rfcs can also be handled thorugh issues system on forgejo-like patforms. But for sure keeping ML is right thing to do, but I expect a lot less trafic
<kasper93>
BtbN: I reported this ffprobe_xsd failure on your fate workers upstream. new libxml2 doesn't support piped input correctly on Windows, hopefully will be fixed for next release.
<kasper93>
hmm, next zlib won't be compatible with our `-lz) echo zlib.lib ;;` subtitution for msvc
<frankplow>
kasper93: Sorry, I intended to check it this week before it was pushed but have been rather busy.
<kasper93>
which target is that?
<kasper93>
make target I mean
<frankplow>
tools/target_dec_%_fuzzer, for vvc specifically
<kasper93>
this should have added $(LIBFUZZER_PATH)
<kasper93>
which should by default contain `-fsanitize=fuzzer`
<kasper93>
which should add main()
<frankplow>
According to make --dry-run it hasn't been added
<frankplow>
I'm on MacOS btw, I've tried both Apple Clang 16.0.0 and LLVM Clang 18.1.8 and both give the same result.
<kasper93>
make distclean && ../configure --toolchain=clang-fuzz-asan-ubsan --enable-ossfuzz && make -j`nproc` tools/target_dec_vvc_fuzzer
<kasper93>
works for me
<kasper93>
unfortunatelly we still need --enable-ossfuzz for `codec_list` that is altered in fuzzer build, if we didn't do it, our fuzzing binaries could be transparent to other tools
<frankplow>
Sorry I didn't realise that --enable-ossfuzz was no longer sufficient and you now also need to set --toolchain=clang-fuzz-*.
<frankplow>
I still have an error, but now it's about not being able to find a libclang_rt.fuzzer_osx.a. It might be my environment
<kasper93>
I was just about to explain that. Yes, fuzzer is handled like any other sanitizer in --toolchain. I find this easier to manage different build types, istead of having set of hardcoded flags for enable-ossfuzz
<frankplow>
Ok, Apple Clang still has some complaint but it seems LLVM Clang is happy. Thanks for your help!
<frankplow>
Unrelated: does anyone know if there's been a change made which might affect availability of debug symbols recently? I only seem to be getting symbols for ~3/4 of objects today and I'm not sure why. It was working fine a few days ago and I don't think I've changed anything in my environment.
<kasper93>
np, sorry for not waiting longer on the patches, but they got ok'ed so I figured should be fine for now
termos has quit [Ping timeout: 252 seconds]
termos has joined #ffmpeg-devel
<BtbN>
Do we have any low-traffic mailing lists that can relatively safely serve as a test for migrations?
<carinobra>
Hi. I'm trying to open an encrypted stream, with KID and KEY. But I'm receiving "Option decryption_key not found" error. Here is the command I'm doing: ffmpeg -decryption_key KID:KEY -i URL -c copy -f hls pipe:1
<jamrial>
the stream is probably not mov. in any case, this kind of question goes to #ffmpeg, not here
TheVibeCoder has quit [Ping timeout: 276 seconds]
<kasper93>
jamrial: memcmp is used to compare APVDecoderFrameInfo
<kasper93>
but APVDecoderFrameInfo has padding
<kasper93>
so it fails
<kasper93>
and adds another frame info
<jamrial>
kasper93: but it's zero initialized
<kasper93>
zero initialization doesn't initialize padding in C
<kasper93>
doing memset(&frame_info, 0, sizeof(frame_info)); instead fixed the issue
<kasper93>
#pragma pack on the struct also fixes the issue
<jamrial>
alright, will use memset then
<jamrial>
thanks
<kasper93>
np
cone-012 has joined #ffmpeg-devel
<cone-012>
ffmpeg James Almer master:85ac5892116e: avformat/apv: use memset to zero initialize frame_info
mkver has quit [Ping timeout: 276 seconds]
<iive>
jamrial, doesn't ffmpeg have some av_mallocz ?
indecisiveturtle has quit [Ping timeout: 268 seconds]
mkver has joined #ffmpeg-devel
kurosu has quit [Quit: Connection closed for inactivity]