Marth64 changed the topic of #ffmpeg to: Welcome to the FFmpeg USER support channel | Development channel: #ffmpeg-devel | Bug reports: https://ffmpeg.org/bugreports.html | Wiki: https://trac.ffmpeg.org/ | This channel is publically logged | FFmpeg 7.1 is released
lusciouslover has joined #ffmpeg
HarshK23 has quit [Quit: Connection closed for inactivity]
five618480339176 has quit [Remote host closed the connection]
five618480339176 has joined #ffmpeg
Muimi__ has quit [Quit: Going offline, see ya! (www.adiirc.com)]
gchound has joined #ffmpeg
gchound has quit [Changing host]
gchound has joined #ffmpeg
<aaabbb>
is "SHA160" in framehash sha1? i assume it is
Muimi__ has joined #ffmpeg
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg
trenixjetix has joined #ffmpeg
System_Error has quit [Remote host closed the connection]
trenixjetix has quit [Quit: Leaving]
gchound has quit [Quit: WeeChat 4.1.1]
System_Error has joined #ffmpeg
^Neo_ has quit [Ping timeout: 248 seconds]
Muimi__ has quit [Quit: Going offline, see ya! (www.adiirc.com)]
StephenLynx has quit [Quit: Leaving]
<Traneptora>
aaabbb: it's SHA-1 which iirc is locked to 160 bits
<Traneptora>
calling av_hash_init with the enum (SHA160) calls av_sha_init, which has the following text:
<Traneptora>
Initialize SHA-1 or SHA-2 hashing.
<Traneptora>
and also: number of bits in digest (SHA-1 - 160 bits, SHA-2 224 or 256 bits)
<aaabbb>
ok thought so, just i never heard of sha1 called sha160 so i wasn't sure if it was a truncated sha2
<Traneptora>
so I didn't dig super deep but it looks like that is the case
<aaabbb>
wondering because it would be nice to have one of those use the cpu sha instructions for better performanace
<Traneptora>
as far as I understand it's all done in softwar
<Traneptora>
but some compilers might use them
<aaabbb>
a compiler isn't going to be able to detect a sha1 calculation and replace it with sha1 instructions
<aaabbb>
it would have to use the sha intrinsics
<aaabbb>
could do the same with -hash crc since most cpus have crc instructions too
<aaabbb>
shouldn't be hard to make a patch for but it probably hasn't been done just because framehash is not the most important thing to optimize
<Traneptora>
aaabbb: afaik michaelni knows quite a lot about crc32s and could probably explain more in detail why we don't use them
<Traneptora>
I don't really know
<aaabbb>
oh interesting, so there's a reason? i was just thinking about the crc32 framehash. but i guess crc32s was considered for other places where crc is used?
<aaabbb>
maybe it has to do with the instruction being for crc32c and the use case is not the c polynomial? idk
vlm_ has quit [Quit: vlm_]
vlm has joined #ffmpeg
System_Error has quit [Ping timeout: 264 seconds]
System_Error has joined #ffmpeg
System_Error has quit [Remote host closed the connection]
Some_Person has quit [Ping timeout: 245 seconds]
cptaffe has quit [Ping timeout: 248 seconds]
System_Error has joined #ffmpeg
parakovsky has quit [Read error: Connection reset by peer]
parakovsky_ has joined #ffmpeg
cptaffe has joined #ffmpeg
Some_Person has joined #ffmpeg
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
<Traneptora>
aaabbb: I have no idea if there's a reason
<Traneptora>
but there's lots of asm code for stuff. the reason may have just been "patches welcome"
<aaabbb>
do you know where crc32 is used, other than ffv1 slices and the framehash format?
<Traneptora>
PNG is one example of where it is used
<Traneptora>
it does not use zlib for that, it uses the inhouse version
<aaabbb>
i imagine that's the crc32c polynomial
MisterMinister has quit [Ping timeout: 245 seconds]
<aaabbb>
oh no it's not crc32c
<Traneptora>
it's the one specified in V.42
<aaabbb>
it makes sense that crc32 instructions can't be used for that because they are polynomial 0x11EDC6F41, png (and zlib) use 0xEDB88320
<Traneptora>
av_crc32 is a bit more flexible but yea
<aaabbb>
would have been nice if intel's crc32 instruction let you load a 32bit polynomial into some register to initialize it and then support any 32bit crc not just crc32c
Marth64 has joined #ffmpeg
MisterMinister has joined #ffmpeg
MisterMinister has quit [Remote host closed the connection]
MisterMinister has joined #ffmpeg
vtorri_ has joined #ffmpeg
signalhunter has quit [Ping timeout: 272 seconds]
HarshK23 has joined #ffmpeg
signalhunter has joined #ffmpeg
antranigv_ is now known as antranigv
derpydoo has joined #ffmpeg
emanuele6 has joined #ffmpeg
vtorri_ has quit [Ping timeout: 268 seconds]
MisterMinister has quit [Ping timeout: 268 seconds]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg
Shuriko has joined #ffmpeg
tokyovigilante has quit [Ping timeout: 260 seconds]
vtorri_ has joined #ffmpeg
lavaball has joined #ffmpeg
makidoll has quit [Remote host closed the connection]
tokyovigilante has joined #ffmpeg
SuicideShow has quit [Ping timeout: 276 seconds]
SuicideShow has joined #ffmpeg
Shine_ has joined #ffmpeg
tokyovigilante has quit [Ping timeout: 268 seconds]
System_Error has quit [Remote host closed the connection]
tokyovigilante has joined #ffmpeg
System_Error has joined #ffmpeg
vtorri_ has quit [Remote host closed the connection]
rv1sr has joined #ffmpeg
Hackerpcs has quit [Ping timeout: 268 seconds]
Sketch has quit [Remote host closed the connection]
Sketch has joined #ffmpeg
zenmov_ has joined #ffmpeg
zenmov_ has quit [Client Quit]
justaguy has joined #ffmpeg
Hackerpcs has joined #ffmpeg
zenmov_ has joined #ffmpeg
zenmov_ has quit [Client Quit]
StephenLynx has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Blacker47 has joined #ffmpeg
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Tano has quit [Quit: WeeChat 4.4.3]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg
Guest97 has joined #ffmpeg
Guest97 has quit [Client Quit]
Shine_ has quit [Read error: Connection reset by peer]
lavaball has quit [Remote host closed the connection]
rappy has joined #ffmpeg
<rappy>
Hey guys, I coudln't get 7.1 to run on my system
Shine_ has joined #ffmpeg
emanuele6 has quit [Read error: Connection reset by peer]
rv1sr has quit []
damian101__ has joined #ffmpeg
TheSilentLink has quit [Ping timeout: 260 seconds]
damian101_ has quit [Ping timeout: 248 seconds]
<rappy>
7.0 runs fine
^Neo has joined #ffmpeg
^Neo has joined #ffmpeg
rappy has quit [Quit: Client closed]
damian101 has joined #ffmpeg
damian101__ has quit [Ping timeout: 248 seconds]
derpydoo has quit [Quit: derpydoo]
l4yer has quit [Ping timeout: 272 seconds]
DauntlessOne4 has quit [Ping timeout: 252 seconds]
l4yer has joined #ffmpeg
DauntlessOne4 has joined #ffmpeg
l4yer has quit [Ping timeout: 268 seconds]
cptaffe has quit [Ping timeout: 245 seconds]
HerculesHD has quit [Ping timeout: 245 seconds]
cptaffe` has joined #ffmpeg
Krusher has joined #ffmpeg
HerculesHD has joined #ffmpeg
cptaffe` is now known as cptaffe
makidoll has joined #ffmpeg
HerculesHD has quit [Ping timeout: 268 seconds]
HerculesHD has joined #ffmpeg
HerculesHD has quit [Changing host]
HerculesHD has joined #ffmpeg
Some_Person has quit [Ping timeout: 265 seconds]
Some_Person has joined #ffmpeg
coldfeet has joined #ffmpeg
coldfeet has quit [Quit: leaving]
coldfeet has joined #ffmpeg
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
nd has quit [Ping timeout: 260 seconds]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
coldfeet has quit [Remote host closed the connection]
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
TheSilentLink has joined #ffmpeg
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
Tano has joined #ffmpeg
Hackerpcs has joined #ffmpeg
nd has joined #ffmpeg
pyre has quit [Remote host closed the connection]
pyre has joined #ffmpeg
vtorri_ has joined #ffmpeg
Guest24 has joined #ffmpeg
Guest24 has quit [Client Quit]
zenmov is now known as m5
Shine_ has quit [Ping timeout: 248 seconds]
makidoll2 has joined #ffmpeg
makidoll has quit [Quit: Ping timeout (120 seconds)]
makidoll2 is now known as makidoll
Muimi__ has joined #ffmpeg
putacho has joined #ffmpeg
putacho has quit [Client Quit]
elChippo has joined #ffmpeg
microchip_ has quit [Ping timeout: 248 seconds]
elChippo has quit [Client Quit]
elChippo has joined #ffmpeg
HarshK23 has quit [Quit: Connection closed for inactivity]
devinheitmueller has joined #ffmpeg
coldfeet has joined #ffmpeg
minimal has joined #ffmpeg
lavaball has joined #ffmpeg
EmleyMoor has quit [Ping timeout: 244 seconds]
kepstin has quit [Remote host closed the connection]
[R]x has quit [Quit: One of God's own prototypes : a high powered mutant of some kind never even considered for mass production. Too weird to live and too rare to die.]
justaguy has quit [Remote host closed the connection]
[R]x has joined #ffmpeg
justaguy has joined #ffmpeg
NotFaely is now known as Faely
HarshK23 has joined #ffmpeg
gebra has joined #ffmpeg
iive has joined #ffmpeg
gchound has joined #ffmpeg
coldfeet has quit [Quit: Lost terminal]
FlorianBad has joined #ffmpeg
FlorianBad has quit [Remote host closed the connection]
ryoskzypu has quit [Ping timeout: 264 seconds]
ryoskzypu has joined #ffmpeg
System_Error has quit [Ping timeout: 264 seconds]
lavaball has quit [Remote host closed the connection]
<Guest58>
I'm testing on the buck bunny video from blender: https://peach.blender.org/download/ What am I misunderstanding about how ffmpeg in terms of `rate` or `fps`?
ryoskzypu has quit [Remote host closed the connection]
ryoskzypu has joined #ffmpeg
<kepstin>
Guest58: the fps filter will use the last frame where rounding the frame timestamp to the output timebase produces the same value. (i.e. with a 30fps input and 1/60fps output, frames 0 through 15 all round to frame 0, and frames 16 through 45 all round to frame 1)
<kepstin>
Guest58: this behaviour is tweakable with the "round" option to the fps filter
<kepstin>
Guest58: for the behaviour you want, i think you should use "fps=fps=1/60:round=up"
EmleyMoor has joined #ffmpeg
<Guest58>
Ah ok, I played with that a little but couldn't get it right. Both round up and down sort of work, round up makes the frame at 60 seconds the first 2 frames, and round down seems to work but doesn't give me the frame at 0.
<Guest58>
is the fps=fps also necessary? Having fps=1/60:round=down seems to have the same effect
<kepstin>
if the timestamp of the first frame is > 0 that would happen, yeah :/
<kepstin>
i prefer to always explicitly write option names when using multiple options like that.
<kepstin>
the first value passed to fps filter will always be used as the value of the fps option, but round is the .. third option i think?
EmleyMoor has quit [Ping timeout: 244 seconds]
<Guest58>
Above when you said frames 16 through 45, did you mean frames 16 through 30? Or am I confused on where the 45 came from.
<furq>
frames between 0.5 and 1.5 seconds round to 1
<kepstin>
er, my numbers in my example were wrong.
<kepstin>
with fps=1, yes frames between 0.5 and 1.5 seconds would round to frame 1
<kepstin>
and it would use whichever the _last_ of the frames in that timespan is, which would be the one closes to 1.5s
<kepstin>
anyways, you could consider doing something like "setpts=PTS+30/TB,fps=1/60" i guess
<Guest58>
I see, hence the 30s frame being considered the 0s frame for an fps of 1/60...
<Guest58>
Ok let me poke around
<Guest58>
Yeah basically I just want a way to get frames at a specified fps but it's important for my used case that they be... from zero I guess? An fps of 1/10 should be frame 0s, 10s, 20s, etc., and fps of 1/60 should be 0s, 60s, etc.
<Guest58>
I figured with a video with a higher frame rate than that this would be pretty trivial
<Guest58>
(but I'm sure I'm missing something about the complexity here)
<kepstin>
so i guess what you want is actually the "first frame to occur after each multiple of N seconds"?
<kepstin>
note that frames don't necessarily line up nicely with even numbers of seconds; you might not have a frame at 10s but rather some slightly before or slightly after
<Guest58>
Yeah, sub-second errors in the frames are totally fine... but yeah the N second multiple... which in my mind is just say I want frames at a specified FPS
<Guest58>
But maybe the terminology ffmpeg uses is different
<kepstin>
might be an interesting thing to improve the fps filter to use the "nearest" frame to the desired output timestamp rather than the last frame which rounds to the desired output timestamp.
<kepstin>
figuring out the timestamp calculations to do that would make my head hurt tho
<kepstin>
the current calculation works in integer pts values with only one conversion step, nice and simple ;)
<kepstin>
(not actually nice and simple, but at least not hard to reason about)
<Guest58>
haha I gotcha
<Guest58>
The PTS suggestion doesn't appear to have done anything
<kepstin>
oh, you should also set the "start_time=0" option on the fps filter
<Guest58>
Sorry, is that `ffmpeg -i ~/Downloads/bbb_sunflower_1080p_30fps_normal.mp4 -vf "start_time=0,fps=1/60" img%03d.jpg`?
<kepstin>
huh, i guess they must have been using a web irc client and clicking links didn't open in a new tab? weird.
Guest58 has joined #ffmpeg
<Guest58>
ha, that gets us back to seeing the frame at 30s as the first frame
<kepstin>
you'd use that in addition to the setpfs filter to apply the offset
<kepstin>
setpts*
<kepstin>
basically, without the "start_time" option, the fps filter automatically aligns itself with the timestamp of the first frame that it sees, cancelling out the effect of the offset.
<Guest58>
Huzzah! I think that did it. I can't say I totally understand but I'll read the docs related to these options to see if I can figure it out.
<kepstin>
i do kinda like the idea of the fps filter outputting the nearest frame, but i doubt i'll ever find time to work on that :(
<Guest58>
Haha that's how it goes isn't it! Do you have a buy-me-a-coffee link or GitHub sponsorship I could contribute to to express my gratitude?
<Guest58>
Or whatever?
<kepstin>
nah, i don't need it. the little work i've done on ffmpeg has been stuff for the benefit of my employer so i'm paid fine for it :)