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 6.1.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
<Lynne>
I'll update my second patch during the night on top of that branch then, thanks
<mkver>
Lynne: I think I could ExtChannelElement from aacsbr.h (and move it into aacsbr_template.c). Do you prefer this?
cone-068 has joined #ffmpeg-devel
<cone-068>
ffmpeg Andreas Rheinhardt master:bbeee3339fec: avcodec/libxevd: Remove FF_CODEC_CAP_SETS_PKT_DTS cap
<cone-068>
ffmpeg Andreas Rheinhardt master:4e14a8e6be99: avcodec/libxevd: Set AV_CODEC_CAP_DR1
<stevenliu>
5. In the development stage, in order to implement #4, I am modifying the local code. The muxer for mov mkv is written to side_data to produce a flv file.
<stevenliu>
Via the command line ./ffmpeg -i *** -metadata:s:v **.mkv
<stevenliu>
6. When implementing the fate use case, I found that if mov demuxer wants to write side_data, it reads and writes through mkv_parse_video_color; I temporarily did not find the data source with MasteringMeta in the sample on the official website, so the data of #5 was not constructed.
<stevenliu>
The idea is to construct a data similar to #5, but it is not constructed with existing data.
<stevenliu>
~
<mkver>
The mov demuxer surely does not use mkv_parse_video_color.
<stevenliu>
:qa!
<stevenliu>
s/mov/mkv/g
<Lynne>
"libavcodec/aac/%.o: CPPFLAGS += -I$(SRC_PATH)/libavcodec/", that's a nice trick
<mkver>
It unfortunately does not work for headers.
<mkver>
Lynne: It does not work for file in libavcodec/ including headers from libavcodec/aac (because the -I in use is only for the files in the subfolder). If all files including aacdec_tab.h were to be moved into the subfolder, it would work (even when checking the header).
rvalue has joined #ffmpeg-devel
lexano has joined #ffmpeg-devel
kurosu has quit [Quit: Connection closed for inactivity]
<jamrial>
mkver: if it removes that wrapper, then sure
<mkver>
No, I meant because inserting an argument in the middle of other arguments is more work than inserting it at the end.
<mkver>
(HEVC and VVC use different values for the dst_stride parameter.)
cone-747 has quit [Quit: transmission timeout]
<Lynne>
mkver: yeah, I figured, not sure if I like needing to address 2 include namespaces individuall than just splitting them up
<Lynne>
the fact that channel element mapping is fixed/float specific now makes deduplication difficult, as everything tries to adjust element mapping at some point
BetweenUs has quit [Quit: Leaving]
Guest48 has joined #ffmpeg-devel
Guest48 is now known as zhupengfei
zhupengfei is now known as zhupengfei08
kurosu has joined #ffmpeg-devel
Traneptora_ has quit [Ping timeout: 240 seconds]
briano has joined #ffmpeg-devel
briano has quit [Remote host closed the connection]
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
AbleBacon has joined #ffmpeg-devel
<zhupengfei08>
Hi mkver, stevenliu: Please ignore #5 #6, there is something wrong with my description.
<zhupengfei08>
I need a video file containing type AV_PKT_DATA_MASTERING_DISPLAY_METADATA(AVPacketSideData) to complete the lavf-container test case, but I did not find a suitable file in the fate suite, so I based on the "multimedia/ffmpeg/fate-suite/hevc/hdr10_plus_h265_sample.hevc" through mov_write_mdcv_tag fake ***.mp4 are used for testing.
<elenril>
zhupengfei08: mkv/hdr10tags-both.mkv has it
<JEEB>
right, that was the name of the sample I wanted to refer to :)
<elenril>
very long names are a hallmark of java, in my impression at least
<JEEB>
just sync up with jamrial so that I won't end up going here & back, and then just tell me what you want. at this point this patch set has been around for way too long :V
<JEEB>
and thanks for taking a look at it
<elenril>
I'll send something more specific to the ML once I've read through it
<JEEB>
roger
<elenril>
this is just something that annoys me in most of the patches
<elenril>
I generally prefer <object>_<action> when possible/applicable
<JEEB>
yea the last changes to names happened when I finally got the brain strength to rebase & rewrite it at all from the set struct to an array and length arguments which at point of v6 jamrial wanted
<JEEB>
I am happy enough that I got to that point
<JEEB>
but most of the naming style was from the previous iteration where IIRC you also voiced in
<jamrial>
i don't find the names wrong
<jamrial>
av_frame_side_data_ is the namespace. then from_{sd,buf} is as short as it gets
<elenril>
I'm not talking about these
<elenril>
but things like add_side_data_to_set_from_buf
<jamrial>
ah
<jamrial>
the "to_set" part is a remnant from previous versions it seems
<JEEB>
yea, it'd be to_array now
<JEEB>
(or if the context is clear, just without to_xyz
<elenril>
side_data_add_from_buf()
<elenril>
little endian naming is godly
<Sean_McG>
javaUsuallyUsesCamelCaseMrElenril()
<cone-857>
ffmpeg Andreas Rheinhardt master:2d63379cae59: avutil/hwcontext_vulkan: Allocate public and priv device hwctx together
<cone-857>
ffmpeg Andreas Rheinhardt master:e70e9b655426: avutil/hwcontext_vulkan: Allocate pub and priv frames hwctx together
<elenril>
mkver: do you remember why does mpegts.c need avcodec.h? (2b41463b8706638b9aaf967655f95ddc195c32eb)
<elenril>
it seems to compile without it
<mkver>
avcodec_is_open, removed in 0d54ae40123ef8dbf77210faf1edbeacc121854e
<elenril>
ah
MisterMinister has joined #ffmpeg-devel
facefamousee has joined #ffmpeg-devel
microchip_ is now known as OfficerPendejo
marcj has quit [Ping timeout: 255 seconds]
hamzah has joined #ffmpeg-devel
microchip_ has joined #ffmpeg-devel
OfficerPendejo has quit [Ping timeout: 260 seconds]
microchip_ is now known as OfficerPendejo
OfficerPendejo has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
hamzah has quit [Quit: Client closed]
dykai has joined #ffmpeg-devel
dionisis has quit [Ping timeout: 256 seconds]
dykai has quit [Ping timeout: 264 seconds]
hamzah has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
loltweyz has joined #ffmpeg-devel
loltweyz has quit [Remote host closed the connection]
hamzah has quit [Quit: Client closed]
jamrial has left #ffmpeg-devel [#ffmpeg-devel]
jamrial has joined #ffmpeg-devel
<jamrial>
JEEB: i like elenril's suggestion. _from_sd -> _clone(), and then you can do _from_buf() -> _add()
<JEEB>
ok. I will soon take a look at the comments and attempt to apply them
<jdek>
elenril: if c11 is a req now can we use _Generic
iive has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
<Lynne>
yes, just not in public headers
cone-857 has quit [Quit: transmission timeout]
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
darkapex has quit [Remote host closed the connection]
darkapex has joined #ffmpeg-devel
user23 has joined #ffmpeg-devel
elastic_dog has quit [Ping timeout: 268 seconds]
elastic_dog has joined #ffmpeg-devel
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
microchip_ has quit [Client Quit]
microchip_ has joined #ffmpeg-devel
<JEEB>
went through patch #5, I hope there are no comments going the opposite way for the same thing later in the set :P
<JEEB>
elenril: > These extra blocks are ugly.
<JEEB>
but you cannot do AVBufferRef *buf = av_buffer_ref(src->buf); after a sanity check without them
<JEEB>
and defining a variable and initializing it to the final value should generally be the best way :P
<JEEB>
also wtf @ removing by the pointer is ugly
<JEEB>
this fucking went through how many revisions of the patch set >:|
<JEEB>
ok, not reworking this at least tonight.
<JEEB>
did the stuff noted in patch #5 at least
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
<JEEB>
(and don't get me wrong, I'm thankful for reviews but at the point of doing `git checkout -b avcodec_cll_mdcv_side_data_v8` I was kind of hoping that maybe some docstring updates and renames
<JEEB>
but noope
<JEEB>
because my next step would have been making the supported-by-module side data stuff listable and then visible via `ffmpeg -h encoder=libx264`
<jamrial>
i only suggested renames
<JEEB>
yea
<jamrial>
the "Removing by the pointer is ugly" stuff is internal, so it could be changed later if you want to not deal with it now
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
microchip_ has quit [Client Quit]
microchip_ has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
jarthur has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
mkver has quit [Ping timeout: 256 seconds]
System_Error has quit [Remote host closed the connection]