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
Marth64 has joined #ffmpeg
alexherbo2 has quit [Remote host closed the connection]
bitbinge has joined #ffmpeg
nitrix is now known as nitrix-or-treat
relue has joined #ffmpeg
relue has quit [Client Quit]
relue has joined #ffmpeg
Unit640 has quit [Quit: Leaving]
Marth64 has quit [Quit: Leaving]
ewomer has quit [Quit: WeeChat 4.4.2]
bitbinge has quit [Quit: bitbinge]
ewomer has joined #ffmpeg
bitbinge has joined #ffmpeg
ewomer has quit [Read error: Connection reset by peer]
bertieb has quit [Read error: Connection reset by peer]
bertieb has joined #ffmpeg
FlorianBad has quit [Remote host closed the connection]
FlorianBad has joined #ffmpeg
bitbinge has quit [Quit: bitbinge]
emmanuelux_ has joined #ffmpeg
emmanuelux has quit [Read error: Connection reset by peer]
emmanuelux_ has quit [Read error: Connection reset by peer]
emmanuelux_ has joined #ffmpeg
bitbinge has joined #ffmpeg
bitbinge has quit [Quit: bitbinge]
bitbinge has joined #ffmpeg
stolen has joined #ffmpeg
bitbinge has quit [Quit: bitbinge]
bitbinge has joined #ffmpeg
Suchiman has quit [Quit: Connection closed for inactivity]
bitbinge has quit [Quit: bitbinge]
Sketch has quit [Ping timeout: 248 seconds]
emmanuelux_ has quit [Read error: Connection reset by peer]
emmanuelux_ has joined #ffmpeg
StephenLynx has quit [Remote host closed the connection]
Sketch has joined #ffmpeg
relue has quit [Ping timeout: 248 seconds]
Sl4yer has joined #ffmpeg
l4yer has quit [Ping timeout: 265 seconds]
Nact has joined #ffmpeg
billchenchina- has joined #ffmpeg
stolen has quit [Quit: Connection closed for inactivity]
rsx has joined #ffmpeg
rpthms has quit [Remote host closed the connection]
rpthms has joined #ffmpeg
rv1sr has joined #ffmpeg
x_x has joined #ffmpeg
lolok has joined #ffmpeg
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg
arbitercoin has joined #ffmpeg
Hackerpcs has quit [Quit: Hackerpcs]
Hackerpcs has joined #ffmpeg
System_Error has quit [Ping timeout: 260 seconds]
crossby1004 has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
coldfeet has joined #ffmpeg
billchenchina- has quit [Quit: Leaving]
System_Error has joined #ffmpeg
FH_thecat has quit [Quit: Leaving]
FH_thecat has joined #ffmpeg
hightower2 has joined #ffmpeg
Hackerpcs has joined #ffmpeg
cosimone has joined #ffmpeg
Hackerpcs has quit [Remote host closed the connection]
Hackerpcs has joined #ffmpeg
rv1sr has quit [Ping timeout: 244 seconds]
KombuchaKip has quit [Ping timeout: 245 seconds]
noobaroo has quit [Quit: noobaroo]
rv1sr has joined #ffmpeg
kepstin has quit [Remote host closed the connection]
kepstin has joined #ffmpeg
chiselfuse has quit [Remote host closed the connection]
chiselfuse has joined #ffmpeg
kepstin has quit [Remote host closed the connection]
kepstin has joined #ffmpeg
lavaball has joined #ffmpeg
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
Suchiman has joined #ffmpeg
ewomer has joined #ffmpeg
hightower2 has quit [Remote host closed the connection]
ewomer has quit [Read error: Connection reset by peer]
cosimone has quit [Remote host closed the connection]
cosimone has joined #ffmpeg
cosimone has quit [Remote host closed the connection]
cosimone has joined #ffmpeg
ewomer has joined #ffmpeg
ewomer has quit [Ping timeout: 246 seconds]
ewomer has joined #ffmpeg
Hobbyboy has joined #ffmpeg
Sketch has quit [Ping timeout: 260 seconds]
Sketch has joined #ffmpeg
coldfeet has quit [Remote host closed the connection]
FH_thecat has quit [Quit: Leaving]
Nact has quit [Quit: Konversation terminated!]
Nact has joined #ffmpeg
FH_thecat has joined #ffmpeg
cosimone has quit [Remote host closed the connection]
StephenLynx has joined #ffmpeg
coldfeet has joined #ffmpeg
JanC has quit [Ping timeout: 245 seconds]
JanC has joined #ffmpeg
Unit640 has joined #ffmpeg
bitbinge has joined #ffmpeg
cosimone has joined #ffmpeg
bitbinge has quit [Quit: bitbinge]
microlappy has joined #ffmpeg
microlappy has quit [Ping timeout: 248 seconds]
Dagger has quit [Ping timeout: 260 seconds]
minimal has joined #ffmpeg
Dagger has joined #ffmpeg
cosimone has quit [Remote host closed the connection]
Narrat has joined #ffmpeg
System_Error has quit [Remote host closed the connection]
haritz has quit [Ping timeout: 252 seconds]
System_Error has joined #ffmpeg
bitbinge has joined #ffmpeg
haritz has joined #ffmpeg
arbitercoin has quit [Ping timeout: 265 seconds]
KombuchaKip has joined #ffmpeg
rsx has quit [Quit: rsx]
lavaball has quit [Remote host closed the connection]
StephenLynx has quit [Ping timeout: 265 seconds]
Sketch has quit [Ping timeout: 265 seconds]
StephenLynx has joined #ffmpeg
Sketch has joined #ffmpeg
coldfeet has quit [Remote host closed the connection]
Sketch has quit [Ping timeout: 252 seconds]
Dagger has quit [Ping timeout: 265 seconds]
Dagger has joined #ffmpeg
Sketch has joined #ffmpeg
zsoltiv_ has joined #ffmpeg
sjs has quit [Quit: sjs]
luva888 has quit [Ping timeout: 252 seconds]
ewomer has quit [Read error: Connection reset by peer]
Sketch has quit [Ping timeout: 252 seconds]
sfan5 has joined #ffmpeg
<sfan5>
how do I strip the "Mastering Display Metadata" side data?
<sfan5>
because it appears during encoding ffmpeg copies it from the source stream and I don't want that
ewomer has joined #ffmpeg
JanC has joined #ffmpeg
JanC is now known as Guest3510
Guest3510 has quit [Killed (silver.libera.chat (Nickname regained by services))]
StephenLynx has quit [Ping timeout: 246 seconds]
coldfeet has joined #ffmpeg
Sketch has joined #ffmpeg
sjs has joined #ffmpeg
sjs has quit [Remote host closed the connection]
StephenLynx has joined #ffmpeg
sjs has joined #ffmpeg
<sfan5>
sidedata=delete in the filter chain unfortunately doesn't do anything
<JEEB>
sfan5: fun if the (a)sidedata filters don't work
<JEEB>
since I think I added that stuff so that it passes the first filtered AVFrame's side data to the encoder
<JEEB>
in case the loop across all side data is broken, sidedata=type=MASTERING_DISPLAY_METADATA,mode=delete would be worth a check. gotta love how this filter has another set of names for these things :D
ewomer has quit [Read error: Connection reset by peer]
<sfan5>
sidedata=delete works with ffmpeg 7.1
<sfan5>
but not 7.0
<sfan5>
it'll be related to that change I linked
Macer has joined #ffmpeg
<JEEB>
that change I think is mostly related to container stuff, so not encoding.
yans has quit [Quit: Let us play... Hide and Slay!]
<JEEB>
although yes, if the problem is stuff that you could filter out from encoding but it still gets pushed to the muxer - yup
<JEEB>
since it will thus still show up in the output, even if not as part of the bit stream
Sketch has quit [Ping timeout: 264 seconds]
rv1sr has quit []
Sketch has joined #ffmpeg
<sfan5>
now for the next problem the logging about implicit filters is woefully incomplete
<sfan5>
Impossible to convert between the formats supported by the filter 'Parsed_sidedata_1' and the filter 'auto_scale_0'
<sfan5>
ok cool I did not insert a scale filter, but the preceding line tells me it was automatically inserted "between the filter 'Parsed_sidedata_1' and the filter 'format'"
<JEEB>
format is basically a meta-filter which sets settings for the previous filter
<sfan5>
however, I did not specify a filter called 'format' either. I can guess that this mirrors the requirement for the encoder.
<JEEB>
so that you can do `scale,format=pix_fmts=yuv420p10` for example
<sfan5>
but what *are* those? and what *is* the format it supposedly cannot convert? I have no clue as it doesn't tell me
<JEEB>
yeh, it does add some extra logging with verbose or higher log level (verbose is what I utilize), but indeed that error message doesn't tell (usually the problem is that you're dealing with lists of formats)
<JEEB>
still doesn't mean that the message couldn't be improved, which it totally could and should be
<furq>
just after the not found using pkg-config message gets improved
<sfan5>
conceptually there should be no problem. since i'm trying to pass frames from vf_libplacebo to the hevc_vulkan encoder. which AFAICT are both AV_PIX_FMT_VULKAN
<JEEB>
ahh, hw pix_fmts.
<JEEB>
for hw related stuff I generally recommend for now that one disables any auto-conversions. the error spots might become better
<JEEB>
-noauto_conversion_filters
<sfan5>
thanks for the hint
<sfan5>
however: The filters 'Parsed_sidedata_1' and 'format' do not have a common format and automatic conversion is disabled.
<sfan5>
this is just the same not-useful message with different flavoring
<JEEB>
yeh
<JEEB>
it's just the first level where it fails now
<JEEB>
which is the output/encoder at hand?
<JEEB>
because format is supposed to be added to the last filter in the chain as far as I can quickly see from this code in ffmpeg_filter.c
<sfan5>
let me check if I can find a minimal reproducer first
<sfan5>
aha, I should have read the warning it printed earlier
<sfan5>
now the vaapi hwaccell is totally out of place there but it made sense in the original arrangement
<sfan5>
and ffmpeg says "There are 2 hardware devices. device vaapi0 of type vaapi is picked for filters by default. Set hardware device explicitly with the filter_hw_device option if device vaapi0 is not usable for filters."
<sfan5>
if I honor that advice and use: ffmpeg -f lavfi -hwaccel vaapi -i testsrc -init_hw_device vulkan=vk -filter_hw_device vk -vf hwupload,libplacebo -c:v hevc_vulkan -f matroska -y /dev/null
<sfan5>
...it works again
<JEEB>
fun
<sfan5>
I'm not sure what exactly is happening because there is no format conversion needed in the first place, but it seems to assume that the output should be AV_PIX_FMT_VAAPI
<sfan5>
this almost makes me tolerate dealing with ML just to improve the logging
<JEEB>
that should be checkable through logging the contents of the bprint string in ffmpeg_filter.c::configure_output_video_filter
<JEEB>
which sets the arguments to format
<sfan5>
now the sad part in this entire story is the following:
<sfan5>
[hevc_vulkan @ 0x5a63b0015ac0] Device does not support the VK_KHR_video_encode_queue extension!
<JEEB>
:)
<JEEB>
so back to vaapi I guess
<JEEB>
(or an update to latest git main branch mesa)
<sfan5>
can I map hw frames from vulkan to vaapi?
<sfan5>
vaapi to vulkan clearly works, I've used that to decode on GPU and then pass to vf_libplacebo