[FFmpeg-cvslog] avutil/vf_scdet_vulkan: add new filter

ffmpeg | branch: master | Niklas Haas | Thu Feb 20 21:23:28 2025 +0100| [9568e40aebee535ea3df97b8b8387b5399652d57] | committer: Niklas Haas avutil/vf_scdet_vulkan: add new filter Carbon copy of vf_scdet. Signed-off-by: Niklas Haas Sponsored-by: nxtedition > http://git.videolan.

[FFmpeg-cvslog] avutil/vulkan: add YUVA pixel formats support

ffmpeg | branch: master | Niklas Haas | Thu Feb 27 19:55:53 2025 +0100| [c2521c0cd2d3b590db8998e62c14edb4df059942] | committer: Niklas Haas avutil/vulkan: add YUVA pixel formats support Signed-off-by: Niklas Haas Sponsored-by: nxtedition > http://git.videolan.org/gitweb.cgi/ffmpeg.git

[FFmpeg-cvslog] avfilter/blackdetect_vulkan: add hw accelerated blackdetect filter

ffmpeg | branch: master | Niklas Haas | Tue Feb 18 16:54:40 2025 +0100| [0894cfb1683015347acaac204b4f18a2470e7736] | committer: Niklas Haas avfilter/blackdetect_vulkan: add hw accelerated blackdetect filter Like vf_blackdetect but better, faster, stronger, harder. Signed-off-by: Niklas Haas

[FFmpeg-cvslog] avfilter/vf_gblur_vulkan: omit unnecessary buffer usage flag

ffmpeg | branch: master | Niklas Haas | Sun May 18 13:57:36 2025 +0200| [02f45a7f3f013bcf2934b624083997d70d708946] | committer: Niklas Haas avfilter/vf_gblur_vulkan: omit unnecessary buffer usage flag Implied internally now when needed. > http://git.videolan.org/gitweb.cgi/ffmpeg.git

[FFmpeg-cvslog] avfilter/vf_blackdetect: add alpha option

ffmpeg | branch: master | Niklas Haas | Thu Feb 27 19:21:20 2025 +0100| [7e783fac5b4d0024640f0a09ab7bca9e24a39ab8] | committer: Niklas Haas avfilter/vf_blackdetect: add alpha option Check the alpha plane for (almost) transparent frames, instead of checking the luma channel for almost black

[FFmpeg-cvslog] avutil/vulkan: automatically enable shader device address usage bit

ffmpeg | branch: master | Niklas Haas | Sun May 18 13:52:46 2025 +0200| [0a4cdf8822e7388ca0930d71ac3cf90c9f2f9446] | committer: Niklas Haas avutil/vulkan: automatically enable shader device address usage bit We require this internally when using descriptor buffers, so it makes sense to enable

[FFmpeg-cvslog] avfilter/vf_libplacebo: add shader_cache option

ffmpeg | branch: master | Niklas Haas | Wed May 14 15:16:50 2025 +0200| [559317cc13577cebffc544c5b221819baaf957c4] | committer: Niklas Haas avfilter/vf_libplacebo: add shader_cache option Useful to speed up shader compilation. May significantly lower startup times, in particular with large or

[FFmpeg-cvslog] avfilter/vf_libplacebo: implement rotation option

ffmpeg | branch: master | Niklas Haas | Fri May 16 13:05:21 2025 +0200| [4f623b4c59c3c838f588e9def7b59dcc26dee7b9] | committer: Niklas Haas avfilter/vf_libplacebo: implement rotation option Flipping can already be accomplished by setting the crop_w/h expressions to their negative values, so

[FFmpeg-cvslog] avfilter/vf_interlace_vulkan: fix FPS and PTS calculation

ffmpeg | branch: master | Niklas Haas | Wed May 21 16:04:48 2025 +0200| [4099d53759ce5d9190a339c02b6bb486e2880f66] | committer: Niklas Haas avfilter/vf_interlace_vulkan: fix FPS and PTS calculation ol->frame_rate is 0/0, so we need to calcalute the correct value based on the il->fram

[FFmpeg-cvslog] doc: add swscale rewrite design document

ffmpeg | branch: master | Niklas Haas | Fri Mar 7 22:08:45 2025 +0100| [06cee0c681b32059bc7ea8bad9c5b4962a96e34a] | committer: Niklas Haas doc: add swscale rewrite design document This should hopefully serve as a better introduction to my new swscale redesign than hunting down random commit

[FFmpeg-cvslog] swscale/graph: prefer bools to ints

ffmpeg | branch: master | Niklas Haas | Thu Mar 6 10:37:40 2025 +0100| [6072e27e9aa809d64a16e89676763e67f081b448] | committer: Niklas Haas swscale/graph: prefer bools to ints This is more consistent with the rest of the newly added code, which universally switched to using bools for boolean

[FFmpeg-cvslog] swscale/graph: move vshift() and shift_img() to shared header

ffmpeg | branch: master | Niklas Haas | Thu Mar 6 10:25:13 2025 +0100| [d95944786e135a6e632fdf71b7d810e755a6c342] | committer: Niklas Haas swscale/graph: move vshift() and shift_img() to shared header I need to reuse these inside `ops.c`. > http://git.videolan.org/gitweb.cgi/ffmpeg.git

[FFmpeg-cvslog] swscale/graph: make noop loop more robust

ffmpeg | branch: master | Niklas Haas | Thu Mar 6 10:47:21 2025 +0100| [bc9696bff80d996b5d526c02b596caa5d64e2db2] | committer: Niklas Haas swscale/graph: make noop loop more robust The current loop only works if the input and output have the same number of planes. However, with the new

[FFmpeg-cvslog] swscale/graph: expose ff_sws_graph_add_pass

ffmpeg | branch: master | Niklas Haas | Fri Feb 21 16:09:38 2025 +0100| [51e912466f860f1ba7bcb2580b1a7383f07ab296] | committer: Niklas Haas swscale/graph: expose ff_sws_graph_add_pass So we can move pass-adding business logic outside of graph.c. > http://git.videolan.org/gitweb.cgi/ffmpeg.

[FFmpeg-cvslog] tests/swscale: improve colorization of speedup

ffmpeg | branch: master | Niklas Haas | Wed Apr 23 17:20:57 2025 +0200| [f297ebf97abbe300abfb7ca074fd56b9f3cc3f4a] | committer: Niklas Haas tests/swscale: improve colorization of speedup The old limits were a bit too tightly clustered around 1.0. Make the value range much more generous, and

[FFmpeg-cvslog] doc/filters: fix "ewa_lanczos" filter description

ffmpeg | branch: master | Niklas Haas | Sun Apr 20 14:32:18 2025 +0200| [853e66a0726b0a9d6d6269a22f6f9b5be7763738] | committer: Niklas Haas doc/filters: fix "ewa_lanczos" filter description > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=853e66a0726b0a9d6d6269a22f6

[FFmpeg-cvslog] tests/swscale: print performance stats on exit

ffmpeg | branch: master | Niklas Haas | Tue Feb 25 18:17:25 2025 +0100| [e1736d0d0bc50d32ae85e76695bd53b25f26726d] | committer: Niklas Haas tests/swscale: print performance stats on exit > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=e1736d0d0bc50d32ae85e76695bd53b25f267

[FFmpeg-cvslog] tests/swscale: allow testing only unscaled convertors

ffmpeg | branch: master | Niklas Haas | Thu Mar 6 12:22:55 2025 +0100| [995986e512d3f18a71aa12e139ba450d1b7205f3] | committer: Niklas Haas tests/swscale: allow testing only unscaled convertors I need this to be able to test the new unscaled conversion code more quickly. We re-order the flags

[FFmpeg-cvslog] tests/swscale: switch from MSE to SSIM

ffmpeg | branch: master | Niklas Haas | Tue Feb 25 15:47:43 2025 +0100| [6c12b1535a86011bedcab7cff5143a01b118bd8b] | committer: Niklas Haas tests/swscale: switch from MSE to SSIM And bias it towards Y. This is much better at ignoring errors due to differing dither patterns, and rewards

[FFmpeg-cvslog] tests/checkasm: fix wrong summation of bench time

ffmpeg | branch: master | Niklas Haas | Mon Mar 31 14:40:16 2025 +0200| [256a38101fd1bc26b020b870a202fd638616a0a6] | committer: Niklas Haas tests/checkasm: fix wrong summation of bench time This was changed 8 years ago with the introduction of the linux-perf path, with seemingly no

[FFmpeg-cvslog] tests/swscale: use hex format for flags values

ffmpeg | branch: master | Niklas Haas | Mon Mar 31 12:18:11 2025 +0200| [d467ceaa9b147b114edacf5808c234d0a476ec55] | committer: Niklas Haas tests/swscale: use hex format for flags values > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d467ceaa9b147b114edacf5808c234d0a476e

[FFmpeg-cvslog] tests/swscale: check supported inputs for legacy swscale separately

ffmpeg | branch: master | Niklas Haas | Tue Feb 25 21:12:56 2025 +0100| [a22faeb992b2ac1d4b0e5222c02a65ac77aeb15b] | committer: Niklas Haas tests/swscale: check supported inputs for legacy swscale separately The new code path supports more formats, so we can't test them all against the l

[FFmpeg-cvslog] tests/swscale: allow choosing specific flags and dither mode

ffmpeg | branch: master | Niklas Haas | Tue Mar 4 14:12:10 2025 +0100| [0e2742a693d5ce453a1f3e8d52cda057dd238786] | committer: Niklas Haas tests/swscale: allow choosing specific flags and dither mode So I can quickly iterate on the new swscale code. > http://git.videolan.org/gitweb.

[FFmpeg-cvslog] tests/swscale: remove stray whitespace in scanf format

ffmpeg | branch: master | Niklas Haas | Tue Feb 25 21:13:47 2025 +0100| [9549daa9963974f485cf07bd0b0ed9e7aaab31b6] | committer: Niklas Haas tests/swscale: remove stray whitespace in scanf format > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi

[FFmpeg-cvslog] tests/swscale: constrain reference SSIM for low bit depth formats

ffmpeg | branch: master | Niklas Haas | Tue Mar 4 15:49:50 2025 +0100| [92a57f1cfdfeb67eab3f0fac5448b494e668ea66] | committer: Niklas Haas tests/swscale: constrain reference SSIM for low bit depth formats Sometimes, the reference SSIM is significantly higher than the SSIM level expected for

[FFmpeg-cvslog] avfilter/vf_libplacebo: don't explode on zero FPS information

ffmpeg | branch: master | Niklas Haas | Mon Mar 17 15:22:16 2025 +0100| [46762c8b82b8c1f1ccf12b84f10c7f36b519566e] | committer: Niklas Haas avfilter/vf_libplacebo: don't explode on zero FPS information While technically not specifiad as valid by the AVFilterLink documentation, it is curr

[FFmpeg-cvslog] tests/swscale: use yuva444p as reference

ffmpeg | branch: master | Niklas Haas | Tue Feb 25 14:14:39 2025 +0100| [1707e81073cebd419333b8e115a02d1a9d9588c2] | committer: Niklas Haas tests/swscale: use yuva444p as reference Instead of the lossy yuva420p. This does change the results compared to the status quo, but is more reflective of

[FFmpeg-cvslog] tests/swscale: calculate theoretical expected SSIM

ffmpeg | branch: master | Niklas Haas | Tue Mar 4 13:59:13 2025 +0100| [8fc9808f18a18e7f0bfcc5de62bf7d9e94f7c455] | committer: Niklas Haas tests/swscale: calculate theoretical expected SSIM We can calculate with some confidence the theoretical expected SSIM from an "ideal" conv

[FFmpeg-cvslog] tests/swscale: allow setting log verbosity

ffmpeg | branch: master | Niklas Haas | Thu Feb 27 22:06:20 2025 +0100| [3e32dc8b0834b8a3f7487eb31c4aa41488846384] | committer: Niklas Haas tests/swscale: allow setting log verbosity Helpful for debugging the new swscale code, since it dumps the operations list in verbose logging mode. > h

[FFmpeg-cvslog] tests/swscale: print speedup numbers in color

ffmpeg | branch: master | Niklas Haas | Tue Mar 4 14:11:07 2025 +0100| [f438f3f8cd5265ef14895d7996e0d1fd9b36628d] | committer: Niklas Haas tests/swscale: print speedup numbers in color > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=f438f3f8cd5265ef14895d7996e0d1fd9b366

[FFmpeg-cvslog] swscale/swscale_unscaled: avoid nv12 <-> nv21 bug

ffmpeg | branch: master | Niklas Haas | Tue Mar 4 17:40:58 2025 +0100| [5b9356f18eaad4e156f90e1691fa4a321764a713] | committer: Niklas Haas swscale/swscale_unscaled: avoid nv12 <-> nv21 bug This is not handled by the planar copy wrapper, so exclude it. > http://git.videolan.org/g

[FFmpeg-cvslog] swscale: fix gray -> grayf32 SIGFPE

ffmpeg | branch: master | Niklas Haas | Fri Mar 7 19:36:29 2025 +0100| [8ab40ca984fa123e5aa8c45ec5d9fdadb1b12cb6] | committer: Niklas Haas swscale: fix gray -> grayf32 SIGFPE swscale internals don't distinguish between 16-bit and higher bit depth output formats internally when it come

[FFmpeg-cvslog] swscale/utils: split off format code into new file

ffmpeg | branch: master | Niklas Haas | Sun Feb 23 15:25:15 2025 +0100| [ae84aa775fa789c68e5e3a01b1b313acd901fad8] | committer: Niklas Haas swscale/utils: split off format code into new file utils.c is getting quite crowded, and I need a new place to dump a lot of format handling code for the

[FFmpeg-cvslog] avfilter/vf_libplacebo: set correct alpha mode before blending

ffmpeg | branch: master | Niklas Haas | Wed Feb 19 22:08:30 2025 +0100| [651bfd0dad07b9eb6de6bc810f665f32c4d01865] | committer: Niklas Haas avfilter/vf_libplacebo: set correct alpha mode before blending &pl_alpha_overlay expects straight alpha, but the alpha output may be premultiplied

[FFmpeg-cvslog] avfilter/vf_libplacebo: set colorspace properties on all inputs

ffmpeg | branch: master | Niklas Haas | Wed Feb 19 18:26:05 2025 +0100| [c85879c6012232e94c8be937ac90f96d0c66176d] | committer: Niklas Haas avfilter/vf_libplacebo: set colorspace properties on all inputs There is no reason to only do this on the first input. It doesn't actually matter fo

[FFmpeg-cvslog] avfilter/libplacebo: add option for outputting separate fields

ffmpeg | branch: master | Niklas Haas | Sun Feb 16 16:56:18 2025 +0100| [27814dc3b639917e38dd6913acebbdfbb25d056f] | committer: Niklas Haas avfilter/libplacebo: add option for outputting separate fields cf. the other deinterlacing-aware filters. > http://git.videolan.org/gitweb.cgi/ffmpeg.

[FFmpeg-cvslog] configure: align libplacebo version check with reality

ffmpeg | branch: master | Niklas Haas | Sun Feb 16 16:17:21 2025 +0100| [58875417afc63dd8ce801a05bcbd36b4fbc4c1ad] | committer: Niklas Haas configure: align libplacebo version check with reality This filter already failed to compile on older versions, because of an unconditional use of an API

[FFmpeg-cvslog] avfilter/vf_liplacebo: properly initialize variable

ffmpeg | branch: master | Niklas Haas | Wed Feb 19 19:30:29 2025 +0100| [05302767ac2b14f560f70d0a3e99de151b3e9539] | committer: Niklas Haas avfilter/vf_liplacebo: properly initialize variable Signed-off-by: Niklas Haas > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi

[FFmpeg-cvslog] avfilter/vf_libplacebo: allow differing formats per input

ffmpeg | branch: master | Niklas Haas | Wed Feb 19 18:21:53 2025 +0100| [d48745ba1cab3dbaf029e24aeef3c866b51cf6f9] | committer: Niklas Haas avfilter/vf_libplacebo: allow differing formats per input Each input is entirely independent and can have varying pixel formats, color spaces, etc. To

[FFmpeg-cvslog] avfilter/vf_libplacebo: switch to new background API

ffmpeg | branch: master | Niklas Haas | Wed Feb 19 18:20:11 2025 +0100| [6602b88df688a0697f93517028cdbec3e7c4d7c9] | committer: Niklas Haas avfilter/vf_libplacebo: switch to new background API Fixes a deprecation warning, and also fixes a bug where the depredated skip_target_clearing option

[FFmpeg-cvslog] avfilter/vf_libplacebo: switch to AV_OPT_TYPE_COLOR

ffmpeg | branch: master | Niklas Haas | Wed Feb 19 18:39:43 2025 +0100| [e23c79ebfcda9bfe75dd03b4798268857af93377] | committer: Niklas Haas avfilter/vf_libplacebo: switch to AV_OPT_TYPE_COLOR > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=e23c79ebfcda9bfe75dd03b4798268857af93

[FFmpeg-cvslog] avfilter/libplacebo: only mark frame fields when deinterlacing

ffmpeg | branch: master | Niklas Haas | Sun Feb 16 16:22:07 2025 +0100| [123cccb62104e3cdbe992aa6c55643fcfbdf7993] | committer: Niklas Haas avfilter/libplacebo: only mark frame fields when deinterlacing Avoids an unnecessary copy of the frame in the frame queue when not deinterlacing anyways

[FFmpeg-cvslog] avfilter/libplacebo: strip interlaced flag when deinterlacing

ffmpeg | branch: master | Niklas Haas | Sun Feb 16 16:53:11 2025 +0100| [83c1b349648b7cabff7ea2f7ff77e08c16d5fa20] | committer: Niklas Haas avfilter/libplacebo: strip interlaced flag when deinterlacing > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi

[FFmpeg-cvslog] avfilter/vf_libplacebo: remove deprecated options

ffmpeg | branch: master | Niklas Haas | Wed Feb 19 19:28:59 2025 +0100| [5944c4b20a32b4a127ca91f87ab13ca2dbfb5611] | committer: Niklas Haas avfilter/vf_libplacebo: remove deprecated options These were removed upstream and no longer have had any effect for a while now. Signed-off-by: Niklas

[FFmpeg-cvslog] avfilter/libplacebo: add deinterlacing options

ffmpeg | branch: master | Niklas Haas | Sun Feb 16 16:13:22 2025 +0100| [794cba812e74173b1598c7680a42e0f056d3d204] | committer: Niklas Haas avfilter/libplacebo: add deinterlacing options These were introduced in libplacebo API version 220. We actually already map the field by default, but

[FFmpeg-cvslog] avfilter/libplacebo: use a transparent default fillcolor

ffmpeg | branch: master | Niklas Haas | Mon Feb 17 15:17:20 2025 +0100| [cd81f084917dacaffda3112fee6876501d399898] | committer: Niklas Haas avfilter/libplacebo: use a transparent default fillcolor This will leave the background transparent when the target format contains an alpha channel

[FFmpeg-cvslog] avfilter/interlace_vulkan: add interlace_vulkan filter

ffmpeg | branch: master | Niklas Haas | Mon Feb 17 15:46:09 2025 +0100| [4dc2ae69e72b57ba828bc57a1d0d1e09c65de5e3] | committer: Niklas Haas avfilter/interlace_vulkan: add interlace_vulkan filter This is a Vulkan-accelerated version of the existing interlace filter. > http://git.videolan.

[FFmpeg-cvslog] swscale/utils: lazily allocate XYZ tables on CONFIG_SMALL

ffmpeg | branch: master | Niklas Haas | Mon Dec 23 13:21:21 2024 +0100| [59258fb90ee0cff10c3668374dbbb13edbdac86c] | committer: Niklas Haas swscale/utils: lazily allocate XYZ tables on CONFIG_SMALL This has the downside of requiring these tables to be recomputed on every init, but saves ~270

[FFmpeg-cvslog] swscale: use 16-bit intermediate precision for RGB/XYZ conversion

ffmpeg | branch: master | Niklas Haas | Mon Dec 16 14:49:39 2024 +0100| [af6d52eec66961f6a502b0f2f390c12226d087cd] | committer: Niklas Haas swscale: use 16-bit intermediate precision for RGB/XYZ conversion The current logic uses 12-bit linear light math, which is woefully insufficient and

[FFmpeg-cvslog] swscale/utils: add return code to fill_xyztables()

ffmpeg | branch: master | Niklas Haas | Mon Dec 23 13:13:00 2024 +0100| [60bfafcd8a943f4ec03a7e80460f8e22abbb4781] | committer: Niklas Haas swscale/utils: add return code to fill_xyztables() Needed for lazy allocation of XYZ tables. Signed-off-by: Niklas Haas > http://git.videolan.

[FFmpeg-cvslog] swscale/unscaled: add pal8 -> gbr(a)p special converter

ffmpeg | branch: master | Niklas Haas | Wed Dec 18 15:07:33 2024 +0100| [8cf2d97280400b2c94ceeb3ed8f3aa1c687aaf40] | committer: Niklas Haas swscale/unscaled: add pal8 -> gbr(a)p special converter Fixes: #9520 Signed-off-by: Niklas Haas Sponsored-by: Sovereign Tech Fund >

[FFmpeg-cvslog] avfilter/vf_lut*: strip color volume metadata

ffmpeg | branch: master | Niklas Haas | Fri Apr 26 13:19:26 2024 +0200| [d23f443dfcad65c7abff728f7835d5fb78654393] | committer: Niklas Haas avfilter/vf_lut*: strip color volume metadata These filters, in general, will apply some arbitrary color volume transformation. Strip corresponding

[FFmpeg-cvslog] avfilter/vf_libplacebo: update metadata stripping logic

ffmpeg | branch: master | Niklas Haas | Fri Apr 26 13:03:57 2024 +0200| [0d468730bc4ecc1a1b5870d10abc59fa2b18dfa6] | committer: Niklas Haas avfilter/vf_libplacebo: update metadata stripping logic Switches to av_frame_side_data_remove_by_props(), covering a number of cases that we previously

[FFmpeg-cvslog] avfilter/vf_scale*: strip metadata on size change

ffmpeg | branch: master | Niklas Haas | Fri Apr 26 12:56:35 2024 +0200| [f22965b8823c92b16c645e7ef5a0ef95b81c7821] | committer: Niklas Haas avfilter/vf_scale*: strip metadata on size change > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=f22965b8823c92b16c645e7ef5a0ef95b81c7

[FFmpeg-cvslog] avutil/frame: add av_frame_side_data_remove_by_props()

ffmpeg | branch: master | Niklas Haas | Wed Dec 4 12:37:15 2024 +0100| [b88944a8aa599be0f16603380e7bc51402d3caf8] | committer: Niklas Haas avutil/frame: add av_frame_side_data_remove_by_props() As discussed in the previous commit, we often need a convenient way of stripping all side data

[FFmpeg-cvslog] avfilter/vf_colorspace: strip color volume metadata

ffmpeg | branch: master | Niklas Haas | Fri Apr 26 13:07:43 2024 +0200| [eecdc25d4b7c6be368be38279880ed38fc3e1bd2] | committer: Niklas Haas avfilter/vf_colorspace: strip color volume metadata > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=eecdc25d4b7c6be368be38279880ed38fc3e1

[FFmpeg-cvslog] avfilter/vf_zscale: strip metadata on change

ffmpeg | branch: master | Niklas Haas | Fri Apr 26 12:59:37 2024 +0200| [d1eadc43b91a84ddf6ae5d130b4752e9a36d0a92] | committer: Niklas Haas avfilter/vf_zscale: strip metadata on change Required for both size changes and color volume changes (as a result of changing primaries/transfer). > h

[FFmpeg-cvslog] avutil/frame: add AV_SIDE_DATA_PROP_{SIZE,COLOR}_DEPENDENT

ffmpeg | branch: master | Niklas Haas | Wed Dec 4 12:20:22 2024 +0100| [3428a8d830357c7c3a30cabbf22f74a362541a5b] | committer: Niklas Haas avutil/frame: add AV_SIDE_DATA_PROP_{SIZE,COLOR}_DEPENDENT Many filters modify certain aspects of frame data, e.g. through resizing (vf_*scale* family

[FFmpeg-cvslog] avutil/frame: add AVSideDataDescriptor for AV_FRAME_DATA_VIDEO_HINT

ffmpeg | branch: master | Niklas Haas | Wed Dec 4 12:14:46 2024 +0100| [01084f3d16aeff2129fb1d2944921fc986cd4475] | committer: Niklas Haas avutil/frame: add AVSideDataDescriptor for AV_FRAME_DATA_VIDEO_HINT This entry was seemingly missing. > http://git.videolan.org/gitweb.cgi/ffmpeg.git

[FFmpeg-cvslog] swscale/options: add -sws_dither none alias

ffmpeg | branch: master | Niklas Haas | Thu Dec 19 12:41:22 2024 +0100| [5ca5bbd462bcd8040b35957ccc00d0392e6de57d] | committer: Niklas Haas swscale/options: add -sws_dither none alias While this one was technically supported on account of the generic options code allowing "none"

[FFmpeg-cvslog] swscale: remove primaries/trc change warning

ffmpeg | branch: master | Niklas Haas | Mon Dec 2 11:22:14 2024 +0100| [253b8977c02078faef057542ca647fc676e747e1] | committer: Niklas Haas swscale: remove primaries/trc change warning This is now supported when using the new API. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi

[FFmpeg-cvslog] swscale/csputils: add internal colorspace math helpers

ffmpeg | branch: master | Niklas Haas | Thu Nov 28 14:05:49 2024 +0100| [2e674780b7402d8b75dc222340fd51511f7856aa] | committer: Niklas Haas swscale/csputils: add internal colorspace math helpers Logic is, for the most part, a straight port of similar logic in liplacebo's colorspace.c,

[FFmpeg-cvslog] swscale/graph: allow dynamically updating HDR metadata

ffmpeg | branch: master | Niklas Haas | Sun Dec 1 15:00:35 2024 +0100| [6da940e1189af566614bdd39ca6ac2ae54615248] | committer: Niklas Haas swscale/graph: allow dynamically updating HDR metadata Without triggering a full graph reinit. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=com

[FFmpeg-cvslog] swscale/utils: add helper function to infer colorspace metadata

ffmpeg | branch: master | Niklas Haas | Thu Nov 28 13:49:23 2024 +0100| [1f0c5007840932b30c162982219750f8c9ed89c7] | committer: Niklas Haas swscale/utils: add helper function to infer colorspace metadata Logic is loosely on equivalent decisions in libplacebo. The basic idea is to try and be a

[FFmpeg-cvslog] swscale/graph: add color mapping pass

ffmpeg | branch: master | Niklas Haas | Fri Nov 29 19:18:34 2024 +0100| [efff80c8f65ec3114811cc8261174ca0a81b4c75] | committer: Niklas Haas swscale/graph: add color mapping pass This leverages the previously introduced color management subsystem in order to adapt between transfer functions and

[FFmpeg-cvslog] swscale/lut3d: add 3DLUT dispatch system

ffmpeg | branch: master | Niklas Haas | Fri Nov 29 15:51:09 2024 +0100| [a57fe519b60a23ead01f24c20638160d93a6cb72] | committer: Niklas Haas swscale/lut3d: add 3DLUT dispatch system This is a lightweight wrapper around the underlying color management system, whose job it is merely to manage the

[FFmpeg-cvslog] avfilter/vf_scale: strip metadata when changing colorspace

ffmpeg | branch: master | Niklas Haas | Sun Dec 1 12:19:27 2024 +0100| [01ce9d796300be815351579fe6cb343c732e36f4] | committer: Niklas Haas avfilter/vf_scale: strip metadata when changing colorspace This is no longer relevant after a change in color space. > http://git.videolan.org/gitweb.

[FFmpeg-cvslog] avfilter/vf_scale: make options alignment consistent (cosmetic)

ffmpeg | branch: master | Niklas Haas | Mon Dec 16 12:10:28 2024 +0100| [4315cc943aa2c82df640ae2ba04ca1242e760c45] | committer: Niklas Haas avfilter/vf_scale: make options alignment consistent (cosmetic) > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi

[FFmpeg-cvslog] swscale/cms: add color management subsystem

ffmpeg | branch: master | Niklas Haas | Fri Nov 29 15:05:26 2024 +0100| [dddf536d3d575a8f7cf7eb85bdd2b2d20fc36369] | committer: Niklas Haas swscale/cms: add color management subsystem The underlying color mapping logic was ported as straightforwardly as possible from libplacebo, although the

[FFmpeg-cvslog] avfilter/vf_scale: add colorspace and transfer property options

ffmpeg | branch: master | Niklas Haas | Thu Nov 28 13:51:06 2024 +0100| [b7a25719f5ec9792af38bfb66da2d10308e8] | committer: Niklas Haas avfilter/vf_scale: add colorspace and transfer property options In the long run, it would be ideal if we could add these to the avfilter negotiation as

[FFmpeg-cvslog] swscale: add ICC intent enum and option

ffmpeg | branch: master | Niklas Haas | Thu Nov 28 13:59:02 2024 +0100| [45f0a7ad338cef639b32c8bc5330d3ff5415bdb5] | committer: Niklas Haas swscale: add ICC intent enum and option This setting can be used to infuence the type of tone and gamut mapping used internally when color space

[FFmpeg-cvslog] swscale/utils: read dynamic HDR10+ metadata from AVFrame

ffmpeg | branch: master | Niklas Haas | Sat Nov 30 14:50:36 2024 +0100| [9084d581e8e5196b661d2841b45a9e773a1378a4] | committer: Niklas Haas swscale/utils: read dynamic HDR10+ metadata from AVFrame Logic ported from libplacebo's AVFrame helpers. The basic idea is to use the provided M

[FFmpeg-cvslog] swscale/utils: read HDR mastering metadata from AVFrame

ffmpeg | branch: master | Niklas Haas | Sat Nov 30 14:50:07 2024 +0100| [7432fa19cdd014bcc120eb8e403ceb301bb5d0d2] | committer: Niklas Haas swscale/utils: read HDR mastering metadata from AVFrame > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi

[FFmpeg-cvslog] swscale/utils: fix XYZ primaries tagging

ffmpeg | branch: master | Niklas Haas | Fri Dec 6 15:29:12 2024 +0100| [7b7c32322d1f754b3cd78c4dab02ac4c3992114c] | committer: Niklas Haas swscale/utils: fix XYZ primaries tagging Swscale currently handles XYZ by embedding a forced conversion to BT.709 RGB with a hardcoded matrix. This is not

[FFmpeg-cvslog] swscale/utils: add HDR metadata to SwsFormat

ffmpeg | branch: master | Niklas Haas | Thu Nov 28 13:43:48 2024 +0100| [a8d01dff9a24520f80b1d4243a9787925d57539f] | committer: Niklas Haas swscale/utils: add HDR metadata to SwsFormat Only add the condensed values that we actually care about. Group them into a new struct to make it easier to

[FFmpeg-cvslog] swscale/utils: set static/implied HDR metadata

ffmpeg | branch: master | Niklas Haas | Sat Nov 30 14:48:57 2024 +0100| [5b21b7f52c7a306e4a087f4822089e29d3b8a709] | committer: Niklas Haas swscale/utils: set static/implied HDR metadata Provide default values for the fields added in the previous commit. > http://git.videolan.org/gitweb.

[FFmpeg-cvslog] swscale/utils: check for supported color transfers

ffmpeg | branch: master | Niklas Haas | Fri Nov 29 21:39:15 2024 +0100| [b9dfe8138eb2df49c85a085b73c20d652ecf579a] | committer: Niklas Haas swscale/utils: check for supported color transfers We will use the av_csp_itu_eotf() functions to decode these internally, so check this function to see

[FFmpeg-cvslog] checkasm/sw_scale: add assertion for hscale assumption

ffmpeg | branch: master | Niklas Haas | Wed Dec 11 10:13:46 2024 +0100| [fe9bf7cd5235bbdbe95eb610540e6e8851e58b7d] | committer: Niklas Haas checkasm/sw_scale: add assertion for hscale assumption This code only checks hcScale. In practice this is not an issue because the function pointers

[FFmpeg-cvslog] swscale/unscaled: allow semiplanar copies

ffmpeg | branch: master | Niklas Haas | Thu Dec 19 15:05:19 2024 +0100| [6c9218d748e06da019b6d15c37e5849d8a69bb07] | committer: Niklas Haas swscale/unscaled: allow semiplanar copies As fixed in the previous commit, this enables semipacked range and bit depth conversions. Previously these would

[FFmpeg-cvslog] swscale/unscaled: correctly round yuv2yuv when not dithering

ffmpeg | branch: master | Niklas Haas | Tue Dec 17 15:20:14 2024 +0100| [c6bf7f664527842fa3b89e49dc2a43f639430c33] | committer: Niklas Haas swscale/unscaled: correctly round yuv2yuv when not dithering We should at least bias towards the nearest integer, instead of always rounding down, when

[FFmpeg-cvslog] swscale/unscaled: correctly copy semiplanar formats

ffmpeg | branch: master | Niklas Haas | Thu Dec 19 14:41:34 2024 +0100| [77db7f9b87811aa588e6ab6a7256f7cd8267265d] | committer: Niklas Haas swscale/unscaled: correctly copy semiplanar formats This fixes multiple bugs with semiplanar formats like NV12. Not only do these false positive the

[FFmpeg-cvslog] checkasm/sw_scale: add test for yuv2nv12cX

ffmpeg | branch: master | Niklas Haas | Mon Dec 16 12:44:10 2024 +0100| [57bbdb4fb1e1dde4c8c447ba42c941c1422abd11] | committer: Niklas Haas checkasm/sw_scale: add test for yuv2nv12cX Mirroring yuv2yuvX. Signed-off-by: Niklas Haas Sponsored-by: Sovereign Tech Fund > http://git.videolan.

[FFmpeg-cvslog] checkasm/sw_rgb: add tests for yuv2packed{1,2,X}

ffmpeg | branch: master | Niklas Haas | Thu Dec 12 16:52:18 2024 +0100| [c601bb8df5ae896061c42dd0e23cea8fba938ecb] | committer: Niklas Haas checkasm/sw_rgb: add tests for yuv2packed{1,2,X} Signed-off-by: Niklas Haas Sponsored-by: Sovereign Tech Fund > http://git.videolan.org/gitweb.

[FFmpeg-cvslog] checkasm/sw_rgb: add alpToYV12 check

ffmpeg | branch: master | Niklas Haas | Fri Dec 13 15:27:11 2024 +0100| [a9ae2cc14d43ea35b0f90dc3ac3ee35fed586446] | committer: Niklas Haas checkasm/sw_rgb: add alpToYV12 check Mirroring lumToYV12 and chrToYV12. Signed-off-by: Niklas Haas Sponsored-by: Sovereign Tech Fund > h

[FFmpeg-cvslog] swscale/output: fix bilinear yuv2rgb chroma interpolation

ffmpeg | branch: master | Niklas Haas | Tue Dec 17 13:23:14 2024 +0100| [095f8038fa9180842cd38d4d61c7c47a02aad9ed] | committer: Niklas Haas swscale/output: fix bilinear yuv2rgb chroma interpolation These functions were divided into two special cases; one assuming that uvalpha == 0, and the

[FFmpeg-cvslog] tests/swscale: allow nonzero positive return codes from sws_scale_frame()

ffmpeg | branch: master | Niklas Haas | Tue Dec 17 13:59:54 2024 +0100| [b38f6f9990cb2445a718f290cbef772c99c54232] | committer: Niklas Haas tests/swscale: allow nonzero positive return codes from sws_scale_frame() See previous commit. Signed-off-by: Niklas Haas Sponsored-by: Sovereign Tech

[FFmpeg-cvslog] swscale: fix documentation of sws_scale_frame()

ffmpeg | branch: master | Niklas Haas | Tue Dec 17 13:57:08 2024 +0100| [e05a1bb87962f4ade57a2a9664356978a08485ce] | committer: Niklas Haas swscale: fix documentation of sws_scale_frame() Since its introduction, this function has claimed to return 0 on success, yet never actually did so (until

[FFmpeg-cvslog] swscale/utils: fix sws_getCachedContext check

ffmpeg | branch: master | Niklas Haas | Mon Dec 16 12:25:44 2024 +0100| [2df655bc2cf8bc83e282a779b17192555860dc77] | committer: Niklas Haas swscale/utils: fix sws_getCachedContext check This logic was inverted, but || was not replaced by &&. Fixes: ed5dd675624c83d9c69b406ce30e4e09

[FFmpeg-cvslog] swscale/slice: fix init of 32 bpc planes

ffmpeg | branch: master | Niklas Haas | Wed Dec 11 09:14:13 2024 +0100| [ce457bfccdd2558dc0e508f2ddf9925e1b741209] | committer: Niklas Haas swscale/slice: fix init of 32 bpc planes In input.c and output.c and many other places, swscale follows the rule of using 15-bit intermediate if output

[FFmpeg-cvslog] tests/swscale: fix sscanf return value check

ffmpeg | branch: master | Niklas Haas | Mon Dec 2 14:55:22 2024 +0100| [ee903c4786ed13d2d9ca63cff02fcef3238f881b] | committer: Niklas Haas tests/swscale: fix sscanf return value check We only parse 12 values, so this check always failed. Regression caused by a change to the print format

[FFmpeg-cvslog] avutil/csp: add av_csp_trc_inv_from_id()

ffmpeg | branch: master | Niklas Haas | Mon Nov 25 13:35:34 2024 +0100| [bf0a6c484ffab9f157db7ee7c8a0f59df077] | committer: Niklas Haas avutil/csp: add av_csp_trc_inv_from_id() Mathematical inverse of av_csp_trc_from_id(), plus testing to make sure it roundtrips correctly with the

[FFmpeg-cvslog] avutil/csp: add EOTF function definitions

ffmpeg | branch: master | Niklas Haas | Mon Nov 25 13:52:30 2024 +0100| [06f084468e034ecb0bfbe522953c1efd77595454] | committer: Niklas Haas avutil/csp: add EOTF function definitions The existing av_csp_trc_func_from_id() mostly implements the OETF, except for PQ. As such, we are currently

[FFmpeg-cvslog] avutil/tests/color_utils: add tests for av_csp_itu_eotf

ffmpeg | branch: master | Niklas Haas | Mon Nov 25 14:10:46 2024 +0100| [7b73ea501dc738c52daf2a2fef0c41b4ce95928f] | committer: Niklas Haas avutil/tests/color_utils: add tests for av_csp_itu_eotf > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi

[FFmpeg-cvslog] avutil/csp: fix documentation of av_csp_trc_function

ffmpeg | branch: master | Niklas Haas | Mon Nov 25 11:45:13 2024 +0100| [ec0489e35c2b996ce70e13201bb44a545736d3b5] | committer: Niklas Haas avutil/csp: fix documentation of av_csp_trc_function This explanation was inaccurate and highly misleading. The new wording is taken more or less directly

[FFmpeg-cvslog] avutil/tests/color_utils: clean up slightly (cosmetic)

ffmpeg | branch: master | Niklas Haas | Mon Nov 25 13:26:54 2024 +0100| [28f217780bcdc3f0f5b8b1678ab7f68283f109ea] | committer: Niklas Haas avutil/tests/color_utils: clean up slightly (cosmetic) Rewrite this test slightly to fix the formatting, improve the readability of the output, and

[FFmpeg-cvslog] avutil/csp: eliminate redundant branch

ffmpeg | branch: master | Niklas Haas | Mon Nov 25 13:07:35 2024 +0100| [feb5982f435fe9d86775ba430a489120a8a90226] | committer: Niklas Haas avutil/csp: eliminate redundant branch > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=feb5982f435fe9d86775ba430a489120a8a90

[FFmpeg-cvslog] swscale/utils: disable full_chr_h_input optimization for odd width

ffmpeg | branch: master | Niklas Haas | Mon Dec 2 13:08:52 2024 +0100| [2f95bc3cb3c0d2e88e33b219409e5ca2719f0c73] | committer: Niklas Haas swscale/utils: disable full_chr_h_input optimization for odd width The basic problem here is that the rgb*ToUV_half_* functions hard-code a bilinear

[FFmpeg-cvslog] swscale/graph: fix memleak of cascaded graphs

ffmpeg | branch: master | Niklas Haas | Mon Dec 2 11:28:45 2024 +0100| [79452d382f7b55f258c17f8ed8a51feb5411713a] | committer: Niklas Haas swscale/graph: fix memleak of cascaded graphs Just free them directly and discard the parent context. Fixes: bf738412e849bcb8c63a330dfb814281b3d97f6b

[FFmpeg-cvslog] avfilter/vf_scale: fix off-by-one in loop bounds

ffmpeg | branch: master | Niklas Haas | Mon Nov 25 14:27:38 2024 +0100| [bcbf3a56303d043b1c74e500772ed73a5ebab759] | committer: Niklas Haas avfilter/vf_scale: fix off-by-one in loop bounds Results in over-read of the array. Fortunately, the excess element was never actually used, but it still

[FFmpeg-cvslog] tests/swscale: add a benchmarking mode

ffmpeg | branch: master | Niklas Haas | Sat Oct 5 16:21:37 2024 +0200| [3edd1e42b93e2f534cfb66954c36a2846d73bc5e] | committer: Niklas Haas tests/swscale: add a benchmarking mode With the ability to set the thread count as well. This benchmark includes the constant overhead of context

  1   2   3   4   >