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
<jamrial>
jkqxz: want to test the patch i just sent? fixed several issues in Dawid's encoder implementation
System_Error has quit [Ping timeout: 264 seconds]
realies has quit [Read error: Connection reset by peer]
realies has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
haihao_ is now known as haihao
System_Error has quit [Remote host closed the connection]
aaabbb has quit [Ping timeout: 260 seconds]
System_Error has joined #ffmpeg-devel
aaabbb has joined #ffmpeg-devel
mkver has quit [Ping timeout: 260 seconds]
novaphoenix has quit [Quit: i quit]
novaphoenix has joined #ffmpeg-devel
Gramner has quit [Remote host closed the connection]
novaphoenix has quit [Quit: i quit]
Gramner has joined #ffmpeg-devel
novaphoenix has joined #ffmpeg-devel
cone-064 has quit [Quit: transmission timeout]
Martchus_ has joined #ffmpeg-devel
novaphoenix has quit [Quit: i quit]
Martchus has quit [Ping timeout: 244 seconds]
novaphoenix has joined #ffmpeg-devel
jamrial has quit []
secondcreek has quit [Remote host closed the connection]
secondcreek has joined #ffmpeg-devel
secondcreek has quit [Remote host closed the connection]
secondcreek has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
j45_ has joined #ffmpeg-devel
secondcreek has quit [Remote host closed the connection]
vriska has joined #ffmpeg-devel
bencoh_ has joined #ffmpeg-devel
nto_ has joined #ffmpeg-devel
termos_ has joined #ffmpeg-devel
abdu has joined #ffmpeg-devel
j45 has quit [Ping timeout: 260 seconds]
leo60228 has quit [Ping timeout: 260 seconds]
termos has quit [Ping timeout: 260 seconds]
bencoh has quit [Ping timeout: 260 seconds]
nto has quit [Ping timeout: 260 seconds]
j45 has joined #ffmpeg-devel
j45_ is now known as j45
j45 has quit [Changing host]
termos_ is now known as termos
secondcreek has joined #ffmpeg-devel
abdu has quit [Ping timeout: 240 seconds]
secondcreek has quit [Remote host closed the connection]
secondcreek has joined #ffmpeg-devel
j-b has quit [Remote host closed the connection]
j-b has joined #ffmpeg-devel
LainIwakura has joined #ffmpeg-devel
DodoGTA has joined #ffmpeg-devel
abdu has joined #ffmpeg-devel
LainIwakura has quit [Ping timeout: 240 seconds]
System_Error has quit [Remote host closed the connection]
abdu has quit [Ping timeout: 240 seconds]
mkver has joined #ffmpeg-devel
<fflogger>
[newticket] lazarcf: Ticket #11564 ([undetermined] Chaining 4 video fade filters does not work) created https://trac.ffmpeg.org/ticket/11564
<Lynne>
so for the Mps212 extension in AAC-USAC (that's not MPEG-H, MPEG-H is different, but yet it's the same codec...), they have a very convoluted way of parsing all coefficients for the stereo mix
<Lynne>
using about a hundred-ish small Huffman tables
<Lynne>
it ocurred to me that this is just a pathetic way of emulating an entropy coding system by just stacking huffman tables on top of huffman tables (2d huffman tables is what they call them)
<Lynne>
all this to emulate like 20 lines of code
usagi_mimi has joined #ffmpeg-devel
lexano has joined #ffmpeg-devel
_av500_ is now known as av500
jamrial has joined #ffmpeg-devel
abdu has joined #ffmpeg-devel
Anthony_ZO1 has joined #ffmpeg-devel
<Lynne>
> will help you write a flappy bird clone
<Lynne>
> will not help you parse 100 tables from a spec document reliably or help you write a program to do that
Anthony_ZO has quit [Ping timeout: 276 seconds]
Anthony_ZO1 is now known as Anthony_ZO
<Lynne>
we'll all going to work in the AI Gulag in 5 years, aren't we?
<haasn>
I asked ChatGPT which distro to install and it suggested AlmaLinux and Rocky
<haasn>
my requirements were "RPM based, good package support, up to date packages, good ZFS support (i.e. offers LTS kernel), no strange out-of-place bloatware like YaST"
<haasn>
yeah.. I guess I'm stuck with Tumbleweed and its strange administration UX
<llyyr>
you can just install TW and then remove yast, which fulfills the criteria
<haasn>
well, the other think I don't like about TW is having to hunt down third party repos for something as basic as ffmpeg
<haasn>
I have 13 third party repos installed
<jamrial>
msys2 package manager already provides gcc 15.1, before even Arch
<jamrial>
a bit too eager maybe
System_Error has quit [Ping timeout: 264 seconds]
<jamrial>
doesn't seem to miscompile ffmpeg at least
System_Error has joined #ffmpeg-devel
bencoh_ is now known as bencoh
kode540 has joined #ffmpeg-devel
kode54 has quit [Ping timeout: 248 seconds]
kode540 is now known as kode54
minimal has quit [Quit: Leaving]
Traneptora has quit [Quit: Quit]
ccawley2011 has joined #ffmpeg-devel
<fflogger>
[editedticket] marto55555: Ticket #11201 ([avformat] Atypical AVIF stream handling in FFmpeg 7: Misaligning major implementations) updated https://trac.ffmpeg.org/ticket/11201#comment:12
Teukka has quit [Read error: Connection reset by peer]
<cone-870>
ffmpeg Andreas Rheinhardt master:8279d02cf199: avformat/apvdec: Check before access
<cone-870>
ffmpeg Andreas Rheinhardt master:0a12b84d3b40: avformat/apvdec: Use ffio_read_size()
<cone-870>
ffmpeg Andreas Rheinhardt master:f98c63e417d1: avformat/apvdec: Fix seeking
<cone-870>
ffmpeg Andreas Rheinhardt master:0b588bfb51ec: avcodec/apv_dsp: Fix left-shift of negative value
<cone-870>
ffmpeg Andreas Rheinhardt master:c681d8e8a818: avformat/apvdec: Remove inappropriate INIT_CLEANUP flag
<cone-870>
ffmpeg Mark Thompson master:585455f7b327: apv_decode: Replace division with shift
<cone-870>
ffmpeg Mark Thompson master:a8bae9b18d1c: fate: Add test for APV 422-10 profile
ccawley2011 has quit [Read error: Connection reset by peer]
<cone-870>
ffmpeg James Almer master:e80f32f3bdd5: avcodec/cbs_apv: don't return an error when reading empty buffers
<jkqxz>
jamrial: Why does the openapv encoder need a copy at all? The input frame should already be platform-simd-aligned (which I assume is what they want).
<jamrial>
jkqxz: i interpreted it as effective allocated buffer dimensions being multiple of 16 (as in coded_{width,height}), not buffer size aligned to 16 bytes
<jamrial>
mm, i guess i could add an APV case in avcodec_align_dimensions2()
<jamrial>
so the input buffer is already as required
<jkqxz>
Oh, so they probably avoid having any check and encode the extra pixels over the edges.
<jamrial>
yeah, Dawid's version also aligned dimensions to a multiple of 16, not the buffer size
<jkqxz>
Make sure to copy the edges to avoid bad encode / information leaks then I guess.
<jkqxz>
Same question for the output: can that be refcounted by the packet instead of copied?
<jamrial>
not sure. we need to call oapve_encode() with a buffer where the bitstream data is going to be written (Dawid made it 128mb, which i guess is worst case scenario?)
<jamrial>
and only after that i get a size value for the actual output
<jkqxz>
Which might use an unreasonable amount of memory if every packet is maximum size. Urgh.