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
squeaktoy has joined #ffmpeg
hightower3 has quit [Remote host closed the connection]
hightower3 has joined #ffmpeg
lexano has quit [Ping timeout: 268 seconds]
minimal has quit [Quit: Leaving]
finsternis has quit [Read error: Connection reset by peer]
vampi__ has joined #ffmpeg
dallemon4 has joined #ffmpeg
luva8 has joined #ffmpeg
HarshK23_ has joined #ffmpeg
acovrig601 has joined #ffmpeg
signalhunter7 has joined #ffmpeg
cworley_ has joined #ffmpeg
kepstin_ has joined #ffmpeg
NotFaely has joined #ffmpeg
DvdKhl has joined #ffmpeg
Su1c1deShow has joined #ffmpeg
q66_ has joined #ffmpeg
odrling has quit [Read error: Connection reset by peer]
HarshK23 has quit [Ping timeout: 268 seconds]
sonicrules1234 has quit [Ping timeout: 268 seconds]
kepstin has quit [Ping timeout: 268 seconds]
Arokh has quit [Quit: //ThisShouldNeverHappen]
vampirefrog has quit [Read error: Connection reset by peer]
Faely has quit [Remote host closed the connection]
signalhunter has quit [Read error: Connection reset by peer]
luva has quit [Read error: Connection reset by peer]
lec_thege8042726 has quit [Read error: Connection reset by peer]
SuicideShow has quit [Ping timeout: 268 seconds]
cworley has quit [Ping timeout: 268 seconds]
acovrig60 has quit [Read error: Connection reset by peer]
<contrapunctus>
bpmedley: After playing around a bit more, it does indeed seem to be the video height. The SVGs I used for the video were all 1500×1000 in size, so why is 1000 considered not divisible by 2 but 998 is? 🤔
<bpmedley>
contrapunctus: Did you run ffprobe on your output mp4?
<contrapunctus>
bpmedley: Thanks, just tried that...1500×999? Interesting. 😄
<bpmedley>
Feels like a fencepost error
another is now known as another|
noobaroo has quit [Quit: Konversation terminated!]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg
Muimi__ has joined #ffmpeg
AbleBacon has quit [Read error: Connection reset by peer]
rvalue- has joined #ffmpeg
rvalue has quit [Ping timeout: 268 seconds]
rvalue- is now known as rvalue
Tinos has quit [Remote host closed the connection]
Muimi__ has quit [Quit: Going offline, see ya! (www.adiirc.com)]
KDDLB has quit [Read error: Connection reset by peer]
TheSilentLink has quit [Ping timeout: 252 seconds]
TheSilentLink has joined #ffmpeg
YuGiOhJCJ has joined #ffmpeg
noobaroo has joined #ffmpeg
e^pi-1 has quit [Quit: WeeChat 4.3.0]
e^pi-1 has joined #ffmpeg
xx has joined #ffmpeg
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg
lavaball has joined #ffmpeg
evilscreww has joined #ffmpeg
celmor has joined #ffmpeg
<celmor>
I've tried converting a gif to an mp4 format with h264 codec but the target website I want to share it to doesn't support the output file. Using an online converter however worked fine. comparing `file` output between the 2 generated files is identical
<galad>
probably that website doesn't support the h.264 High 4:4:4 Predictive profile
<celmor>
then how do I get the same kind of output which the online converter produced for me?
<celmor>
guess I can't rely on ffmpeg (cli) and have to use a wrapper like handbrake
<galad>
set the pixel format to something like yuv420p
<galad>
-vf format=yuv420p
<celmor>
ah, I did find something like that but now that I read the numbers it makes sense: -pix_fmt yuv444p
<celmor>
was trying to get the acceptable parameters for -profile:v
<celmor>
does ffmpeg always default to 4:4:4 even if the input wouldn't have that?
<galad>
it defaults to the pixel format of the input
emanuele6 has quit [Read error: Connection reset by peer]
<galad>
then tries to match it with the list of supported pixel formats of the encoder
<celmor>
I'm not seeing the pixel format of my input. I guess I was assuming gif format would never have that
<galad>
if your input is rgb, and libx264 can support 4:4:4, it will use 4:4:4
emanuele6 has joined #ffmpeg
<celmor>
might need a smarter tool then, something that can auto select more appropriate options when transcoding
<celmor>
it's great and all that ffmpeg produced the highest quality/effecient output possible by itself but if then becomes incompatible that's not much of use
<galad>
that's up to you, converting to 420 will lose some info, but 444 is less compatibile
<celmor>
lose info from a gif container? is that even possible?
<celmor>
I take the crappierst lowest quality input and ffmpeg ugesses an incompible high profile?
<celmor>
which I then can't share
<celmor>
Anyway. I'm just not able to deal with ffmpeg's cli semantic so I'll just not and use a wrapper. thanks for your help
<galad>
4:2:0 means you have one color sample for 4 luminance samples
yawkat has joined #ffmpeg
<celmor>
I know what those numbers mean but the input would maybe have 16 colors or so
<celmor>
or 265
<galad>
so yes, it will lose color info, and you will lose more info because you are recompressing it
<galad>
the best solution would be to keep the gif without recompressing it
<celmor>
if I could it as is I would've done so from the start but most services simply show a static image instead of the animation in that case
<galad>
then just add -vf format=yuv420p to your command
<celmor>
that works but I'd rather be interested in finding out the acceptable values cause no way I'm gonna remember this specific one. a wraper with a drop-down and perhaps recommended selection would still be more convenient in my use case
kus has joined #ffmpeg
zsoltiv_ has joined #ffmpeg
xx has quit [Ping timeout: 260 seconds]
rsx has joined #ffmpeg
zsoltiv_ has quit [Ping timeout: 268 seconds]
evilscreww has quit [Read error: Connection reset by peer]
finsternis has joined #ffmpeg
<Cracki>
chatgpt will quickly whip up any GUI you like. command line utilities do have the downside of not giving you a menu, so you have to remember the magic spell
<Cracki>
autocompletion in a terminal is still witchcraft
<Cracki>
even for simple tools, which ffmpeg is not
minimal has joined #ffmpeg
contrapunctus has left #ffmpeg [#ffmpeg]
e^pi-1 has quit [Quit: WeeChat 4.3.0]
iconoclasthero has quit [Killed (NickServ (GHOST command used by iconoclast_hero))]
emanuele6 has quit [Read error: Connection reset by peer]
Dagger has joined #ffmpeg
emanuele6 has joined #ffmpeg
rv1sr has quit []
rv1sr has joined #ffmpeg
evilscreww has joined #ffmpeg
kts has quit [Ping timeout: 268 seconds]
ttys000 has joined #ffmpeg
<celmor>
yeah, I still haven't gotten used to the semantic. there should be more examples in its tldr page
HerbY_NL has joined #ffmpeg
HerbY_NL has quit [Ping timeout: 252 seconds]
billchenchina has quit [Quit: Leaving]
evilscreww has quit [Quit: Leaving]
AbleBacon has joined #ffmpeg
elvis_a_presley has quit [Quit: smoke-bomb ; grapple-hook]
elvis_a_presley has joined #ffmpeg
<Cracki>
chatgpt is probably fluent in ffmpeg
<Cracki>
the greatest present to humanity, a machine that gives us some peace and quiet from other humans
<BtbN>
Given how many times people come here or flood issue trackers with absolute nonsense from some LLM, I highly doubt that.
<Cracki>
colleague in the open plan office had the nerve to complain that I use headphones to drown out everyone's jabbering. only SLIGHTLY less than half the people out there are fucking insane
<Cracki>
at least they asked the machine oracle before posing the same questions to humans
<BtbN>
They got absolute nonsense commandlines from an LLM, and then openend an issue on FFmpeg or related projects, claiming they're broken.
<Cracki>
lol ok
<BtbN>
That's just wasting everyones time
<Cracki>
rule 1: disclose LLM use
<Cracki>
open access issue trackers are a luxury. add some "i'm not a moron" test in front of it, kinda like the "i'm not a robot" tests
<BtbN>
If the two options are coming here and asking "How do i do Y?" or pasting a humongous convoluted commandline and saying "doesn't work"....
<BtbN>
Well, one of those seems largely superior
<Cracki>
in fact, an LLM could grade those test results :D
<BtbN>
Or we could just not waste peoples time with LLMs.
<Cracki>
don't blame the tool, blame the users
<JEEB>
LLMs seem most useful for menial tasks the results of which you can easily verify
<BtbN>
Everything that comes out of there could be absolute bullshit or worse. So it takes more time to cross-check it thouroughly, than to just do it yourself
<BtbN>
Combine that with loads of people who for some reason treat it like unfailable truth, and you just got a huge waste of time and energy, both of people and electricity.
<Cracki>
people are finally learning to stop trusting and to start verifying
xx has joined #ffmpeg
<BtbN>
yeah, no
<Cracki>
ok, people aren't learning. you'll sign that, right?
<BtbN>
The vast majority of people are morons and it's impressive they even manage to use a PC to navigate to any LLM
<Cracki>
they don't actually use a PC
<Cracki>
they use an app
<Cracki>
it's always on, reboots itself. you don't even need to know how to turn it on
<BtbN>
Well, if they ask it for ffmpeg commandline, I think PC usage is relatively safe to assume :D
<Cracki>
sadly that's true
<Cracki>
irc being irc is another filter that rejects technically useless people
<furq>
you'd think so wouldn't you
<Cracki>
there's an app for that
markizano has quit [Remote host closed the connection]
ivanich has joined #ffmpeg
<Cracki>
LLMs are trained to be inoffensive. you could build the image that real human support is positively offensive, so that keeps the LLM users away
<BtbN>
The actual main issue is, that an LLM will never say "I don't know", unless an active filter was put on top to remove "forbidden information"
<BtbN>
Instead of not knowing, it'll just make shit up
markizano has joined #ffmpeg
<Cracki>
an issue of training
<BtbN>
yeah no
<Cracki>
yeah yes
<BtbN>
unless you somehow feed it literally all information there is, which is impossible, that's not an issue of training, but an inherent conceptual issue
<Cracki>
it's a controversial topic. my mistake for getting that started.
<Cracki>
you don't know that they literally feed it the entire internet, including sites like libgen and whatnot
<Cracki>
it is an issue of training, not on data, but on its types of responses
<BtbN>
the internet is not allknowing.
<furq>
most of the internet is convincing-sounding seo-friendly garbage written by people who don't know what they're talking about
<furq>
so it's nice that we can automate that now
<Cracki>
if your concept of internet is "reddit", then sure, but tell me relevant (!) repositories of information that are not reachable by internet
<BtbN>
sounds exactly like an LLM
<furq>
if there's one job that's definitely at risk
<BtbN>
There simply is no magic library of all knowledge of the universe.
<Cracki>
that's a reductionist argument
<Cracki>
and a level of effort not required of humans
<BtbN>
Do you claim that knowledge is finite and we can cram it all into some magic computer model? That's just silly.
<Cracki>
I do not claim that, nor is that relevant to the issue
<furq>
i mean i think that might be true enough
<Cracki>
dude, please don't view me as an adversary
<BtbN>
There will always be unknowns. And LLMs are inherently a thing that are unable of "not knowing"
<furq>
i just don't think that the model is an llm
<Cracki>
if your premise is that these things are "inherently unable of 'not knowing'", I can see how you'd conclude all that you did. please challenge your premise. for your own sake.
<BtbN>
lol
<Cracki>
your response saddens me
<BtbN>
Have you seen the nonsense that those LLMs produce? Beleving that that's just solved by shoving even more dats into it is just denial.
<BtbN>
Or buying into the hype created by some of the "AI-Bros" it spawned
<Cracki>
distinguish between what is currently being done, and what the technology *could* do, if they actually cared to figure out how to fine-tune it for some measure of humility ("I don't know")
<Cracki>
I can't help you if you start assigning opinions to me I never professed
<BtbN>
Well, all I see is those things wasting my and other peoples time, and at the same time wasting INSANE amounts of electric energy.
<Cracki>
I'm very sorry if I somehow offended you but speaking of "denial" is wholly unfounded. I never denied anything. I spoke of capability.
<BtbN>
As far as I'm concerned, I'd rather see the hype end sooner than later.
<Cracki>
I think I'm wasting my time right now. bye.
<BtbN>
Sorry for bursting your dream bubble...
<Cracki>
I'll be another two years until I show up here again. last time it was some "durandal" or whatever type that really showed unnecessary character traits.
Cracki has left #ffmpeg [Leaving]
<furq>
speaking of which, how's librempeg going
<CounterPillow>
BtbN can simply not see the future of cooking spaghetti in gasoline
gothicserpent has quit [Quit: gothicserpent]
rsx has quit [Quit: rsx]
gothicserpent has joined #ffmpeg
ossifrage has quit [Quit: Leaving]
GingerGeek has quit [Quit: Follow ma Twitter: @GingerGeek]
Tinos has joined #ffmpeg
xx has quit [Ping timeout: 260 seconds]
xx has joined #ffmpeg
<llyyr>
>add some "i'm not a moron" test in front of it, kinda like the "i'm not a robot" tests
<llyyr>
>last time it was some "durandal" or whatever type that really showed unnecessary character traits
<llyyr>
failed the test...
hightower3 has quit [Remote host closed the connection]
hightower3 has joined #ffmpeg
bitbinge has joined #ffmpeg
jagannatharjun has quit [Quit: Connection closed for inactivity]
Narrat has joined #ffmpeg
mrelcee has quit [Quit: I want Waffles!]
HerbY_NL has joined #ffmpeg
jagannatharjun has joined #ffmpeg
markizano has quit [Quit: Poweroff]
markizano has joined #ffmpeg
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
HerbY_NL has joined #ffmpeg
celmor has quit [Quit: Client closed]
HerbY_NL has quit [Client Quit]
deus0ww has quit [Ping timeout: 240 seconds]
deus0ww has joined #ffmpeg
darkapex has quit [Remote host closed the connection]
darkapex has joined #ffmpeg
beaver has quit [Remote host closed the connection]
beaver has joined #ffmpeg
HarshK23 has quit [Quit: Connection closed for inactivity]
yans has joined #ffmpeg
bitbinge has quit [Ping timeout: 260 seconds]
ivanich has quit [Ping timeout: 252 seconds]
iive has joined #ffmpeg
beaver has quit [Remote host closed the connection]
beaver has joined #ffmpeg
minimal has quit [Quit: Leaving]
travisghansen has quit [Ping timeout: 255 seconds]
yans has quit [Remote host closed the connection]
yans has joined #ffmpeg
yans has quit [Read error: Connection reset by peer]
rv1sr has quit []
bitbinge has joined #ffmpeg
Livio has joined #ffmpeg
e^pi-1 has joined #ffmpeg
ossifrage has joined #ffmpeg
Livio has quit [Ping timeout: 268 seconds]
five61848033 has quit [Remote host closed the connection]
five61848033 has joined #ffmpeg
poorboy_ is now known as poorboy
yeahitsme has quit [Ping timeout: 246 seconds]
System_Error has quit [Remote host closed the connection]
<noobaroo>
Everytime I cut with -ss before the -i, timestamps don't start at 0. I use -avoid_negative_ts 2 -reset_timestamps 1
<noobaroo>
Not doing stream copy
ossifrage has quit [Remote host closed the connection]
System_Error has joined #ffmpeg
hightower4 has joined #ffmpeg
Narrat has quit [Quit: They say a little knowledge is a dangerous thing, but it's not one half so bad as a lot of ignorance.]
hightower3 has quit [Ping timeout: 268 seconds]
ossifrage has joined #ffmpeg
System_Error has quit [Remote host closed the connection]
xx has quit [Ping timeout: 260 seconds]
System_Error has joined #ffmpeg
Su1c1deShow has quit [Ping timeout: 260 seconds]
SuicideShow has joined #ffmpeg
lavaball has quit [Remote host closed the connection]
jagannatharjun has quit [Quit: Connection closed for inactivity]
lexano has quit [Ping timeout: 268 seconds]
ossifrage has quit [Quit: Leaving]
ferreum has quit [Quit: bye]
ferreum has joined #ffmpeg
ossifrage has joined #ffmpeg
<kepstin>
noobaroo: if you've told it to seek to, say 30 seconds, but the first frame after that time isn't until, say 30.1s, then the timestamp of the first frame after seeking and resetting timestamps will be 0.1s