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 7.0.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
IndecisiveTurtle has quit [Ping timeout: 256 seconds]
thilo has quit [Ping timeout: 256 seconds]
thilo has joined #ffmpeg-devel
thilo has quit [Changing host]
thilo has joined #ffmpeg-devel
iive has quit [Quit: They came for me...]
mkver has quit [Ping timeout: 264 seconds]
arch1t3cht2 has joined #ffmpeg-devel
arch1t3cht has quit [Ping timeout: 260 seconds]
arch1t3cht2 is now known as arch1t3cht
lemourin has quit [Quit: The Lounge - https://thelounge.chat]
cone-088 has joined #ffmpeg-devel
<cone-088> ffmpeg Shiyou Yin master:6b35fcacdb74: swscale: [loongarch] Fix undeclared functions prob.
lemourin has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
natto17 has quit [Quit: a.]
natto has joined #ffmpeg-devel
zsoltiv_ has quit [Ping timeout: 256 seconds]
Warcop has joined #ffmpeg-devel
jamrial has quit []
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 255 seconds]
System_Error has quit [Ping timeout: 260 seconds]
AbleBacon has quit [Read error: Connection reset by peer]
System_Error has joined #ffmpeg-devel
pal_ is now known as pal
<pal> ssh is being super slow for me... is anyone else experiencing this?
<cone-088> ffmpeg Pierre-Anthony Lemieux master:249c66bb225b: avcodec/jpeg2000dec: fix HT block decoder
<courmisch> Lynne: it's not really useful if you need to support free(). so really it's only for embedded
kurosu has joined #ffmpeg-devel
<Lynne> eh, they say it can be useful for optimizing
<courmisch> well yes, you can make a simpler/safer/faster heap if you can assume that you'll be given the length and alignment on free
<courmisch> but that's just never going to happen in existing libc
DHE has quit [Remote host closed the connection]
DHE has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
cone-088 has quit [Quit: transmission timeout]
ngaullier has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
Krowl has joined #ffmpeg-devel
j45_ has joined #ffmpeg-devel
j45 has quit [Ping timeout: 260 seconds]
j45_ is now known as j45
j45 has joined #ffmpeg-devel
j45 has quit [Changing host]
j45_ has joined #ffmpeg-devel
j45 has quit [Ping timeout: 268 seconds]
j45_ is now known as j45
j45 has quit [Changing host]
j45 has joined #ffmpeg-devel
<tr4nq_> this is probably a bit far-fetched. do any of you build/debug/run ffmpeg in Clion under Windows?
<tr4nq_> i got a license for clion and i'm trying to figure out how to build ffmpeg
<tr4nq_> this should probably be in the other channel though i guess
<tr4nq_> i can't get cmake to properly recognize the project structure
<JEEB> not surprising, it's not cmake :)
<JEEB> back in ye olden days I think I did once integrate another project that had configure/make based flow in the Qt Creator thing, by defining explicitly the commands
<JEEB> but for vscode I only add required flags generated by whatever the tool was
<JEEB> builds I execute by myself anyways
<tr4nq_> yeah i tried to make a CMakeLists.txt file but i can't seem to properly load it
<tr4nq_> i'll figure it out eventually
<tr4nq_> can't wait to have it set up
<tr4nq_> apparently next to mingw and cmake windows clion uses something called ninja which i've never encountered before
<tr4nq_> i was hoping something like this would do the trick https://pastebin.com/raw/vBmMu2r7
<tr4nq_> oh looks like i'm missing some things
Krowl has quit [Read error: Connection reset by peer]
kasper93_ has joined #ffmpeg-devel
kasper93 has quit [Ping timeout: 264 seconds]
<tr4nq_> maybe i'm better off trying to get clion to play nice with autotools
<elenril> we are not using autotools
<tr4nq_> huh
<tr4nq_> now i'm confused o_O
<elenril> and you're VASTLY underestimating the complexity of our build system
<tr4nq_> i had a sense that might be the case
<tr4nq_> i'm trying to move away from vim/gdb
<tr4nq_> clion seems really useful
mkver has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
psykose has quit [Remote host closed the connection]
<tr4nq_> i like being in over my head though
<tr4nq_> means i always have something to do
psykose has joined #ffmpeg-devel
<tr4nq_> and smarter people to annoy :)
<tr4nq_> MPEG-5 Essential Video Coding (EVC), standardized as ISO/IEC 23094-1, is a video compression standard that has been completed in April 2020 by decision of MPEG Working Group 11 at its 130th meeting.[1][2][3] [4]
<tr4nq_> sorry that was a right click paste fail
lexano has joined #ffmpeg-devel
psykose_ has joined #ffmpeg-devel
kasper93_ is now known as kasper93
psykose has quit [Ping timeout: 268 seconds]
psykose_ is now known as psykose
Krowl has joined #ffmpeg-devel
cone-212 has joined #ffmpeg-devel
<cone-212> ffmpeg Wu Jianhua master:09d3370c2884: avcodec/x86/vvc/vvc_alf: fix integer overflow
<cone-212> ffmpeg Wu Jianhua master:9950f1486493: avcodec/x86/vvc/vvc_alf: use xq to match ptrdiff_t
<cone-212> ffmpeg Wu Jianhua master:442e94e5e476: tests/checkasm/vvc_alf: change alf step size to 8
<cone-212> ffmpeg James Almer master:b70289f35446: tests/checkasm/vvc_mc: fix indentation
<cone-212> ffmpeg James Almer master:4008a80c1bed: tests/checkasm/vvc_mc: don't zero the SAD buffers
<cone-212> ffmpeg Zhao Zhili release/7.0:edc7b57e2637: avcodec/mediacodecenc: workaround the alignment requirement only for H.264
<cone-212> ffmpeg Zhao Zhili release/7.0:f0b747ef1a9b: avcodec/mediacodecenc: workaround the alignment requirement for H.265
<cone-212> ffmpeg Andreas Rheinhardt master:dfd0320e23e3: avfilter/vf_signalstats: Use 64bit for processing histogram
<cone-212> ffmpeg Andreas Rheinhardt master:996e0a57cad3: avformat/matroskadec: Add assert to silence Coverity false positive
<cone-212> ffmpeg Andreas Rheinhardt master:37cc648fbd9e: avfilter/af_aiir: Remove dead code
<cone-212> ffmpeg Andreas Rheinhardt master:21bfc62642cb: avfilter/af_aiir: Avoid unchecked allocation
<cone-212> ffmpeg Andreas Rheinhardt master:2d818644a3f1: avfilter/af_afir: Merge header into af_afir.c
<cone-212> ffmpeg Andreas Rheinhardt master:6d86146fce70: avcodec/vc2enc: Avoid void* where possible
<cone-212> ffmpeg Andreas Rheinhardt master:b1702afdfdfa: avcodec/vc2enc: Constify slices->main context pointers
<cone-212> ffmpeg Andreas Rheinhardt master:eac8dcb1871c: avcodec/vc2enc: Remove superfluous error message
<cone-212> ffmpeg Andreas Rheinhardt master:02ecf8d7f3fe: avcodec/vc2enc: Fix slice length
<cone-212> ffmpeg Andreas Rheinhardt master:c30928566623: avcodec/vc2enc: Avoid relocations for short strings
<cone-212> ffmpeg Andreas Rheinhardt master:d1d40a7c9b59: avcodec/vc2enc: Move transient PutBitContext from ctx to stack
<cone-212> ffmpeg Andreas Rheinhardt master:d9bd5baf9d29: avcodec/vc2enc: Use already available AVPixFmtDescriptor
<cone-212> ffmpeg Andreas Rheinhardt master:0f3090cbd125: avcodec/diracdec: Use FF_CODEC_CAP_INIT_CLEANUP
<cone-212> ffmpeg Andreas Rheinhardt master:8cbf7e840846: avcodec/diracdec: Mark flush as av_cold
Krowl has quit [Read error: Connection reset by peer]
ccawley2011 has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
Livio has joined #ffmpeg-devel
kurufu has quit [Ping timeout: 256 seconds]
Traneptora has joined #ffmpeg-devel
kurufu has joined #ffmpeg-devel
<cone-212> ffmpeg James Almer master:8a1c491354bc: avcodec/packet: remove reference to old AV_SIDE_DATA_PARAM_CHANGE_ values
Livio has quit [Ping timeout: 240 seconds]
* Sean_McG peeks in
Livio has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
Livio has quit [Ping timeout: 268 seconds]
Krowl has joined #ffmpeg-devel
<courmisch> something is very wrong in tests/checkasm/motion.c
<courmisch> very weird side effect going on here
<courmisch> oooh, got it.
<cone-212> ffmpeg Gyan Doshi master:2f184aa89be8: configure: correct libopenjpeg description
Livio has joined #ffmpeg-devel
DHE has quit [Remote host closed the connection]
<elenril> how nice, mips/hevcpred_msa.c duplicates the entirety of intra_pred() at least twice, with random intrinsics mixed in
<jamrial> yeah, the mips code is something else
<elenril> I wonder who decodes HEVC in software on MIPS
<elenril> can we kill it?
<nevcairiel> that sounds like an adventure in waiting
<jamrial> i'll not oppose it
<nevcairiel> didnt even mips kill mips
<jamrial> afaik, yeah
<Lynne> yup, they're purveyors of quality risc-v cores now, and loongson is so extended its almost proprietary at this point
<elenril> now that I think of it, is the compiler even allowed to optimize away multiple accesses to foo->bar?
<courmisch> didn't Intel kill 386?
<courmisch> Lynne: isn't Loongson also dead in favour of glorious harmonious people's LoongArch?
<kurosu> elenril: kill what? HEVC? MIPS? MIPS HEVC? All?
<kurosu> :D
<elenril> kurosu: bink2 is the only relevant codec today, nothing else is needed
<kurosu> 3 > 2, hence bink3 > bink2
<cone-212> ffmpeg James Almer master:e0db1f51d6dd: avcodec/lpc: account for odd len values
<elenril> bink3 exists?
<kurosu> That was RAD tools, some of them going to specifically work on these LZ codecs, with great success (afaik, HW implementation on the PS5 console)
<jamrial> courmisch: your compute_autocorr test works now on x86 with the existing assembly
<elenril> for what purpose?
<Lynne> jamrial: would extending the loop to be <= also fix it?
<courmisch> sorry busy watching latest Tensura episode
<jamrial> >isekai
ngaullier has quit [Ping timeout: 260 seconds]
<elenril> >watching anime in 2024
<Lynne> watch variety shows, they're better
<Lynne> I got an OOM somehow and the only emacs instance to not get deleted was the lpc assembly, so I guess that's what I'm working on tonight probably
<elenril> i only watch buffy the vampire slayer and other quality 90s shows
<kurosu> elenril: well, it's only a matter of for f in snow*; do mv $f ${f//snow/bink3_}; done
<kurosu> https://tsdr.uspto.gov/#caseNumber=75587217&caseSearchType=US_APPLICATION&caseType=DEFAULT&searchType=statusSearch
<kurosu> The RAD Bink trademark ran out
kasper93 has quit [Remote host closed the connection]
<kurosu> Ah, they reregistered it
<courmisch> reverse isekai is already caught up to so that leaved the straight one
Krowl has quit [Read error: Connection reset by peer]
<jamrial> Lynne: not sure what you mean
<Lynne> for(i=j-1; i<=len; i+=2){ ?
j45_ has joined #ffmpeg-devel
<jamrial> no, it fails
j45 has quit [Ping timeout: 240 seconds]
j45_ is now known as j45
j45 has quit [Changing host]
j45 has joined #ffmpeg-devel
<jamrial> the inline sse2 asm still gives different results for odd len values, but the checaksm will not bother with those for now
<jamrial> still would be ideal to adapt it to handle them (while being ported to nasm preferably)
<Lynne> is the inline asm correct, and this just makes the C version correct?
deus0ww has quit [Ping timeout: 252 seconds]
<jamrial> the inline asm is correct for even len values only. and yes, the c version was wrong for those it seems
deus0ww has joined #ffmpeg-devel
<Lynne> okay, I'll fix my asm to be correct for odd values too and add that check to checkasm
<jamrial> Lynne: enabling odd values in checkasm depends on if courmisch's riscv implementation also supports them
<Lynne> I think they get odd-length handling for free with rvv, right?
Krowl has joined #ffmpeg-devel
<courmisch> we even get zero-length for free
ccawley2011 has quit [Ping timeout: 260 seconds]
ccawley2011 has joined #ffmpeg-devel
Livio has quit [Ping timeout: 256 seconds]
<cone-212> ffmpeg Anton Khirnov master:a1471ec8ada0: lavc/hevcdec: rename HEVCContext.HEVClcList to local_ctx
<cone-212> ffmpeg Anton Khirnov master:25ce44efa5f1: lavc/hevcdec: track local context count separately from WPP thread count
<cone-212> ffmpeg Anton Khirnov master:f0aece90d99c: lavc/hevcdec: allocate local_ctx as array of structs rather than pointers
<cone-212> ffmpeg Anton Khirnov master:72bdbce00d4d: lavc/hevcdec: drop a useless execute() call with 1 job
<cone-212> ffmpeg Anton Khirnov master:bd1a06dc4394: lavc/hevc_ps: reduce the size of used_by_curr_pic_lt_sps_flag
<cone-212> ffmpeg Anton Khirnov master:6fed1841a1f5: lavc/hevc_ps/HEVCSPS: change flags into uint8_t
<cone-212> ffmpeg Anton Khirnov master:4264e4056c41: lavc/hevc_ps: fix variable signedness in ff_hevc_decode_short_term_rps()
<cone-212> ffmpeg Anton Khirnov master:d8936678673d: lavc/hevc_ps: do not store delta_poc_s[01] in ShortTermRPS
<cone-212> ffmpeg Anton Khirnov master:9127819d514a: lavc/hevc_ps: reduce the size of ShortTermRPS.used
<cone-212> ffmpeg Anton Khirnov master:63a96dbcced2: lavc/hevc_ps: compactify ShortTermRPS
Krowl has quit [Read error: Connection reset by peer]
Livio has joined #ffmpeg-devel
Warcop has quit [Remote host closed the connection]
kasper93 has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
IndecisiveTurtle has joined #ffmpeg-devel
<courmisch> anyone seeing Nintendo laywers on the Horizon?
<courmisch> lawyers even
<Lynne> I haven't seen anyone following me or spying on me, if that's what you mean
<Lynne> that's how I heard sony and the KGB operate, maybe nintendo are more devious and don't want you to know they've got someone watching you
<cone-212> ffmpeg Rémi Denis-Courmont master:06fc919aadd1: lavc/sbrdsp: add support for 256-bit vectors
<cone-212> ffmpeg Rémi Denis-Courmont master:6a7c4d60a149: lavu/float_dsp: add double-precision scalar product
<cone-212> ffmpeg Rémi Denis-Courmont master:98405d28fab1: checkasm/float_dsp: add double-precision scalar product
<cone-212> ffmpeg Rémi Denis-Courmont master:73c278d2702c: lavu/lls: use ff_scalarproduct_double_c()
<cone-212> ffmpeg Rémi Denis-Courmont master:b11495203058: lavfi: get rid of bespoke double scalar products
<cone-212> ffmpeg Rémi Denis-Courmont master:9238f6cb415d: lavu/float_dsp: R-V V scalarproduct_double
<courmisch> I heard rumours of a Mordor86 double-precision scalar product
<jamrial> Lynne: the benchmark is run with a fixed length, so randomized length is not a problem
<cone-212> ffmpeg Rémi Denis-Courmont master:16132a810d04: checkasm/lpc: test compute_autocorr
<jamrial> courmisch: chill
<Lynne> ah, okay, that's fine then
b50d has joined #ffmpeg-devel
<jamrial> courmisch: there, x86 asm for scalarproduct_double
<courmisch> ugh, the address operands that I can't unsee
<courmisch> jamrial: why no FMA? aren't mulpd and addpd vying to marry happy ever after
<jamrial> courmisch: because it was slower in my tests
<jamrial> it chokes somewhere
<jamrial> fmadd is too finicky it seems
<courmisch> on the flip side why evaluate_lls has no avx
<jamrial> there's a comment about alignment and loads in the sse2 version
<courmisch> now I'm happy RVV has unordered float vector sum
<courmisch> how it wouldn't be fun to slide add slide add slide add
<cone-212> ffmpeg James Almer master:0a949aacaedb: checkasm/lpc: use fixed length to bench apply_welch_window
b50d has quit [Remote host closed the connection]
<Lynne> a horizontal add?
<Lynne> how fast is it compared to slide + add, I wonder
<courmisch> I have not tried, but it can't be good. You'd have to loop over log2(length) since you don't know what it is.
Krowl has quit [Read error: Connection reset by peer]
AbleBacon has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
ccawley2011 has quit [Read error: Connection reset by peer]
Livio has quit [Ping timeout: 256 seconds]
cone-212 has quit [Quit: transmission timeout]
kurosu has quit [Quit: Connection closed for inactivity]
pross has quit [Remote host closed the connection]
pross has joined #ffmpeg-devel