BtbN 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 7.0 is released
minimal has quit [Quit: Leaving]
Warcop has joined #ffmpeg
Mister_Magister has quit [Quit: bye]
Mister_Magister has joined #ffmpeg
\\Mr_C\\ has quit [Remote host closed the connection]
<Marth64>
nemo_magneet: have you ffproved it yet
<Marth64>
ffprobe*
lucasta has joined #ffmpeg
vincejv_ has joined #ffmpeg
vincejv has quit [Ping timeout: 260 seconds]
vincejv_ is now known as vincejv
xx has quit [Ping timeout: 260 seconds]
lucasta has quit [Quit: Leaving]
five6184803391 has quit [Remote host closed the connection]
five6184803391 has joined #ffmpeg
NotWarcop has joined #ffmpeg
Warcop has quit [Ping timeout: 272 seconds]
bip0pular has quit [Ping timeout: 256 seconds]
<Marth64>
nemo_magneet: or even run plain old `file` on it. maybe the magic number hints something
robobub has joined #ffmpeg
Suchiman has quit [Quit: Connection closed for inactivity]
Vonter has quit [Quit: WeeChat 4.4.2]
Vonter has joined #ffmpeg
billchenchina- has joined #ffmpeg
foul_owl has quit [Ping timeout: 276 seconds]
foul_owl has joined #ffmpeg
FH_thecat has quit [Quit: Leaving]
billchenchina- has quit [Ping timeout: 244 seconds]
billchenchina- has joined #ffmpeg
FH_thecat has joined #ffmpeg
ewomer has quit [Ping timeout: 255 seconds]
five61848033915 has joined #ffmpeg
five6184803391 has quit [Ping timeout: 255 seconds]
five61848033915 is now known as five6184803391
Juest has quit [Ping timeout: 246 seconds]
billchenchina- has quit [Ping timeout: 246 seconds]
billchenchina- has joined #ffmpeg
FH_thecat has quit [Quit: Leaving]
Juest has joined #ffmpeg
mven97 has quit [Ping timeout: 252 seconds]
five61848033910 has joined #ffmpeg
five6184803391 has quit [Ping timeout: 245 seconds]
five61848033910 is now known as five6184803391
rvalue has quit [Ping timeout: 260 seconds]
coldfeet has joined #ffmpeg
luc4 has joined #ffmpeg
FH_thecat has joined #ffmpeg
<nemo_magneet>
?
<nemo_magneet>
?2
manwithluck has quit [Ping timeout: 252 seconds]
manwithluck has joined #ffmpeg
darkapex has quit [Ping timeout: 260 seconds]
rvalue has joined #ffmpeg
darkapex has joined #ffmpeg
manwithluck has quit [Ping timeout: 245 seconds]
manwithluck has joined #ffmpeg
luc4 has quit [Quit: Konversation terminated!]
Blacker47 has joined #ffmpeg
nigetilly has joined #ffmpeg
lavaball has joined #ffmpeg
emmanuelux has quit [Quit: au revoir]
Suchiman has joined #ffmpeg
billchenchina- has quit [Remote host closed the connection]
billchenchina- has joined #ffmpeg
tranzistor has quit [Read error: Connection reset by peer]
tranzistor has joined #ffmpeg
billchenchina- has quit [Remote host closed the connection]
billchenchina- has joined #ffmpeg
billchenchina- has quit [Remote host closed the connection]
HerbY_NL has joined #ffmpeg
fling has joined #ffmpeg
fling_ has joined #ffmpeg
<another|>
nemo_magneet: what exactly do you not understand here?
billchenchina- has joined #ffmpeg
fling has quit [Ping timeout: 260 seconds]
billchenchina- has quit [Max SendQ exceeded]
billchenchina- has joined #ffmpeg
turlando has quit [Quit: No Ping reply in 180 seconds.]
billchenchina- has quit [Max SendQ exceeded]
billchenchina- has joined #ffmpeg
turlando has joined #ffmpeg
billchenchina- has quit [Max SendQ exceeded]
billchenchina- has joined #ffmpeg
fling_ is now known as fling
billchenchina- has quit [Max SendQ exceeded]
billchenchina- has joined #ffmpeg
xx has joined #ffmpeg
billchenchina- has quit [Ping timeout: 260 seconds]
billchenchina- has joined #ffmpeg
billchenchina- has quit [Max SendQ exceeded]
billchenchina- has joined #ffmpeg
billchenchina- has quit [Max SendQ exceeded]
billchenchina- has joined #ffmpeg
billchenchina- has quit [Max SendQ exceeded]
lavaball has quit [Remote host closed the connection]
lavaball has joined #ffmpeg
fling has quit [Ping timeout: 260 seconds]
nigetilly has quit [Quit: Konversation terminated!]
coldfeet has quit [Remote host closed the connection]
nigetilly has joined #ffmpeg
HerbY_NL has quit [Ping timeout: 252 seconds]
bitbinge has quit [Quit: bitbinge]
earthwormjim has joined #ffmpeg
earthw0rmjim has joined #ffmpeg
<nemo_magneet>
how you probe files
earthw0rmjim has quit [Client Quit]
earthwormjim has quit [Ping timeout: 244 seconds]
<another|>
ffprobe <file>
<another|>
file <file>
<another|>
mediainfo <file>
coldfeet has joined #ffmpeg
nigetilly has quit [Remote host closed the connection]
manwithluck has quit [Ping timeout: 252 seconds]
manwithluck has joined #ffmpeg
nigetilly has joined #ffmpeg
microlappy has joined #ffmpeg
microlappy has quit [Remote host closed the connection]
talismanick has quit [Ping timeout: 260 seconds]
evilscreww has joined #ffmpeg
bitbinge has joined #ffmpeg
nigetilly has quit [Ping timeout: 245 seconds]
lavaball has quit [Remote host closed the connection]
DennisBagley has joined #ffmpeg
lavaball has joined #ffmpeg
microlappy has joined #ffmpeg
microlappy has quit [Remote host closed the connection]
<DennisBagley>
TIA : I am converting an mp4 (h264/aac) to a wav, the mp4 has a 4 second glitch in it - which is skipped in the wav (so wav is 4 seconds shorter ... how do I tell ffmpeg to include 4 seconds of nothing to maintain teh original timeline (so if mp4 and wav played together they stay in sync)
evilscreww has quit [Ping timeout: 248 seconds]
coldfeet has quit [Remote host closed the connection]
pa has quit [Ping timeout: 248 seconds]
xue has joined #ffmpeg
<xue>
Hi, how to re-encode mp3 to add album cover but also keep existing metadata
<xue>
id3v2 or smth, idk
pah has joined #ffmpeg
nigetilly has joined #ffmpeg
<another|>
metadata is copied by default
<xue>
yeah, but i want to add album cover
microlappy has joined #ffmpeg
nrg has quit [Ping timeout: 260 seconds]
microlappy has quit [Remote host closed the connection]
<enva2712>
if i'm transcoding with libavcodec how can i ensure hardware acceleration is used to decode? source is getting demuxed via libavformat
arbitercoin has quit [Ping timeout: 252 seconds]
<xue>
another|: thx
gronk has joined #ffmpeg
<gronk>
If I'm using libvpx-vp9, is there a reason to use 2-stage to get a very specific filesize, or can I use the -minrate x -maxrate x -b:v x recommendation to get a very similar result? What will the difference really be?
<gronk>
No, I won't test it, because I don't know how to properly test it in a way that informs me, as I'm too ignorant.
<xue>
another|: and if i wanted to do opposite thing, remove album cover? Car sound system doesn't like files with it
<xue>
album cover only
rsx has joined #ffmpeg
<enva2712>
\quit
enva2712 has quit [Quit: leaving]
xue has quit [Ping timeout: 245 seconds]
xue has joined #ffmpeg
<another|>
gronk: yes, because libvpx requires 2pass for some optimisations
<gronk>
Shit.
<gronk>
I really wish, since 2-pass is literally made so you can optimize for a GIVEN bitrate, that I could just pass a final filesize and tell it to go.
<another|>
xue: -map 0:a -c copy out.mp3
Tano has joined #ffmpeg
ewomer has quit [Quit: WeeChat 4.4.2]
shmocking has joined #ffmpeg
ewomer has joined #ffmpeg
<xue>
another|: thx
<DennisBagley>
seems -af aresample=async=1:first_pts=0 was enough to fill my 4 second glich with silent audio (and maintain time alignment with input file)
nigetilly has quit [Quit: Konversation terminated!]
DennisBagley has quit [Quit: Konversation terminated!]
HerbY_NL has joined #ffmpeg
DauntlessOne has quit [Remote host closed the connection]
DauntlessOne has joined #ffmpeg
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<BtbN>
2pass in libvpx is weird as hell. You need to turn it on also for single pass iirc, else its rate control sucks and the quality is bad or something
<gronk>
BtbN: that's so annoying
<BtbN>
libvpx is just terrible
<BtbN>
VP9 is a really nice codec. Which suffers from libvpx being the only widely available encoder
<BtbN>
stuff like eve vp9 does amazing things with VP9.
rv1sr has joined #ffmpeg
<gronk>
subscription based? yick
<another|>
eve is not for personal use, but for netflix style customers
<BtbN>
yes, but it shows what the codec is capable of
<BtbN>
A GPL or even AGPL version of it for non-commercial use would be amazing, but would likely get abused :/
<BtbN>
I wonder why Nvidia never implemented VP9 encoding support. If the hardware can do AV1, it surely also has the capabilities for VP9 already there
rsx has quit [Quit: rsx]
Unit640 has joined #ffmpeg
<Unit640>
So, I'm using ffmpeg to maximum-compress all my PNG images in order to save on storage space. For a few really big ones, it fails. From the output, it mentions "[png @ 0x5560dacae680] [IMGUTILS @ 0x7ffc699e5c50] Picture size 30000x20000 is invalid" and "[png @ 0x5560dacae680] Invalid image size". I think I asked about this here a while ago, but I can't remember if there was a solution. Why is there an upper limit at all? Surely my 32 GB of RAM can hold
<Unit640>
even such a big PNG image and all the data required to hold in RAM while it produces the better-compressed version?
<BtbN>
It means that av_image_check_size() failed
<BtbN>
basically, 30000x20000 pixels, time 4 byte per pixel cause RGBA, is 2400000000, which is larger than 2^31-1, so it's impossible to address all those pixels on 32 bit systems, which is what the check verifies
<gronk>
>3KB off a 54 second video
<gronk>
I'm gonna kill myself
<Unit640>
But it's a 64-bit system...?
<BtbN>
Well, it's according to the PNG spec. It wants all images to be decodable everywhere
lockywolf has quit [Ping timeout: 264 seconds]
pikapika_lunar has joined #ffmpeg
<BtbN>
It's also a bit of a security measure. Without this check, you could craft a png image that's tiny but would make ffmpeg try to allocate multiple exabytes
pikapika_lunar is now known as militantorc
BtbN has quit [Remote host closed the connection]
BtbN has joined #ffmpeg
<nasso>
if i want to process AVFrames with the gpu (using vulkan/opengl/whatever) what do i gain from using libplacebo vs just mapping the textures myself?
HerbY_NL has joined #ffmpeg
five6184803391 has quit [Remote host closed the connection]
five6184803391 has joined #ffmpeg
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
vincejv has quit [Remote host closed the connection]
hjckr has joined #ffmpeg
shmocking has quit [Quit: Client closed]
coldfeet has joined #ffmpeg
pah is now known as pa
HerbY_NL has joined #ffmpeg
vincejv has joined #ffmpeg
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<snoriman>
nasso: I'm looking into something similar ... so I'm interested to the answer too ^.^
wyatt8750 has quit [Remote host closed the connection]
<snoriman>
currently my goal is to leverage hardware decoding and interop with GL. JEEB pointed me to MVP which has done interop with their own abstraction layer (if I'm correct).
<snoriman>
my current take is that it's a lot less work to use Media Engine on Windows and CUVID where possible; of course this is for my own set of requirements which might be different from yours.
System_Error has quit [Remote host closed the connection]
rv1sr has quit [Remote host closed the connection]
System_Error has joined #ffmpeg
rv1sr has joined #ffmpeg
Vonter has quit [Ping timeout: 244 seconds]
iive has joined #ffmpeg
Vonter has joined #ffmpeg
fling has joined #ffmpeg
HerbY_NL has joined #ffmpeg
<nasso>
yeah in my case i only care about linux pretty much, maybe macos
coldfeet has quit [Remote host closed the connection]
mven97 has quit [Quit: Goodbye.]
mven97 has joined #ffmpeg
beastd has joined #ffmpeg
mven971 has joined #ffmpeg
xx has quit [Ping timeout: 260 seconds]
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
xx has joined #ffmpeg
mven97 has quit [Ping timeout: 260 seconds]
mven971 is now known as mven97
Blacker47 has quit [Quit: Life is short. Get a V.90 modem fast!]
Haxxa has quit [Quit: Haxxa flies away.]
Haxxa has joined #ffmpeg
wyatt8740 has joined #ffmpeg
Juest has quit [Ping timeout: 246 seconds]
Juest has joined #ffmpeg
beastd has quit [Ping timeout: 260 seconds]
microlappy has joined #ffmpeg
microlappy has quit [Remote host closed the connection]
fling has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
fling has joined #ffmpeg
wyatt8740 has quit [Ping timeout: 246 seconds]
wyatt8740 has joined #ffmpeg
<CounterPillow>
You gain all the things libplacebo lists as features, which are not trivial to implement performantly and correctly.
beastd has joined #ffmpeg
ewomer has quit [Read error: Connection reset by peer]
rv1sr has quit []
Marth64 has quit [Quit: Leaving]
ewomer has joined #ffmpeg
<Unit640>
BtbN: "Without this check, you could craft a png image that's tiny but would make ffmpeg try to allocate multiple exabytes" <-- How so?
Keshl has quit [Ping timeout: 246 seconds]
<haasn>
nasso: all of the processing shaders, basically
<haasn>
Including stuff like conversion to rgb
<BtbN>
Unit640: what do you mean? Just make it stupid large
emmanuelux has joined #ffmpeg
beastd has quit [Ping timeout: 248 seconds]
Juest has quit [Ping timeout: 272 seconds]
Juest has joined #ffmpeg
Marth64 has joined #ffmpeg
<bpmedley>
nasso: You articulated interest in macOS and processing frames with a gpu. I don't see it mentioned, so I'll state that you might have good luck working with Metal Performance Shaders (e.g MPSScaleTransform) augmented with compute shaders (e.g. MTLComputeCommandEncoder).
<bpmedley>
One reason I mention is because I don't see macOS Metal support with libplacebo.
YuGiOhJCJ has joined #ffmpeg
lavaball has quit [Remote host closed the connection]
<DeHackEd>
Unit640: zlib can compress solid black at about a 1000:1 ratio, and better ratios are doable. so a 4 megabyte compressed all-black image easily turns into a 32 bit blow-out.
SuicideShow has quit [Ping timeout: 276 seconds]
SuicideShow has joined #ffmpeg
FlorianBad has quit [Quit: Konversation terminated!]
<Unit640>
BtbN: Well, he said the image was "tiny".
<Unit640>
DeHackEd: Interesting.
<BtbN>
A 4 MB png expanding to multiple petabytes seems comparatively tiny to me
<Unit640>
BTW, today I found a 10-hour YouTube video with literally nothing but a 100% orange screen 100% of the time. Out of curiosity, I downloaded it with yt-dlp in the highest quality and checked the resulting file size. It was like 25 MB. Not very efficient if you ask me. Should be like "orange, full image, repeated for entire video". Like a few bytes.
<BtbN>
25MB for a 10h video seems pretty dang efficient to me?
<BtbN>
I expected more than that
<Unit640>
Regardless of the reason, this limitation makes me unable to use ffmpeg to make those larger images take up less space. Feels frustrating. I wish the upper limit at least would be bigger than what it apparently is.
<Unit640>
BtbN: But it's a video that has nothing changing at all.
<BtbN>
It'll still have to contain the usual video structure and an I-Frame, i.e. one full picture of solid orange, every GOP
<BtbN>
So the I-Frame will be however tiny it cane make "solid orange", and all the P/B ones will be virtually nothing
<Unit640>
Can't say I understand that, but it sounds like you know what you're talking about.
<Unit640>
And I guess they don't exactly have too many of those "special" videos for them to ultra-optimize such bizarre cases.
<BtbN>
Video encoders and formats simply don't expect to deal with special cases like that
<BtbN>
it'd be a waste to design them for it
<BtbN>
So of course they'll be slightly inefficient in those cases
<BtbN>
But 25MB for a 10h long video is still next to nothing
FlorianBad has joined #ffmpeg
<Unit640>
Sure, but remember the "contents" still.
lusciouslover has quit [Read error: Connection reset by peer]
lusciouslover has joined #ffmpeg
Unit640 has quit [Quit: Leaving]
<BtbN>
If the content was anything else, it'd be multiple gigabytes