ffmpeg | branch: release/7.0 | Lynne | Wed Jan 1 18:03:33 2025
+0900| [d639a90f543c0cca9572d5dffeae54dfc934e931] | committer: Michael
Niedermayer
configure: update copyright year
On 01/01/2025 19:05, Peter Ross wrote:
> FFmpeg turns 25 this year.
(cherry picked from com
ffmpeg | branch: release/7.1 | Lynne | Sat Feb 8 04:35:31 2025
+0100| [c2184b65d214d60f2d3df86a11ca502567a3d134] | committer: Lynne
aacenc_tns: clamp filter direction energy measurement
The issue is that:
float en[2];
...
tns->n_filt[w] = is8 ? 1 : order != TNS_MAX_ORDER ? 2 : 3;
for (g
ffmpeg | branch: release/7.0 | Lynne | Sat Feb 8 04:35:31 2025
+0100| [f98f142da571653436596ccad2d09c7e39bfd4fb] | committer: Lynne
aacenc_tns: clamp filter direction energy measurement
The issue is that:
float en[2];
...
tns->n_filt[w] = is8 ? 1 : order != TNS_MAX_ORDER ? 2 : 3;
for (g
ffmpeg | branch: master | Lynne | Sat Feb 8 04:35:31 2025
+0100| [bedfb6eca402037f5cbb115fa767d106b8c14f1c] | committer: Lynne
aacenc_tns: clamp filter direction energy measurement
The issue is that:
float en[2];
...
tns->n_filt[w] = is8 ? 1 : order != TNS_MAX_ORDER ? 2 : 3;
for (g = 0
ffmpeg | branch: master | Lynne | Fri Jun 27 03:05:13 2025
+0900| [bd75fad85f37da92f83a4b36126d18112fce5b72] | committer: Lynne
hwcontext_vulkan: fix issues with importing a device
The previous fix just used a local variable for the memory properties,
which did not fix this issue.
> h
ffmpeg | branch: master | Lynne | Tue Jun 24 16:41:48 2025
+0900| [60fc0288ae39c3294091c2a4e06816836be95740] | committer: Lynne
hwcontext_vulkan: fix 85c0364b703
The issue is that by moving the mprops loading later, we inadvertently
broke some earlier dependencies.
> http://git.videolan.
ffmpeg | branch: master | Lynne | Mon Jun 16 19:11:20 2025
+| [11428896dba40814edd880272b3bcb9b778fd4a8] | committer: Lynne
hwcontext_vulkan: add workaround for broken Nvidia drivers
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=11428896dba40814edd880272b3bcb9b778fd
ffmpeg | branch: master | Lynne | Tue Jun 24 02:11:59 2025
+0900| [85c0364b703f4c5f581da6dd8cd6446769ed3439] | committer: Lynne
hwcontext_vulkan: move ReBAR check into a new function and use it later
The issue is that vulkan_device_create_internal() is only called for
devices that lavu creates
ffmpeg | branch: master | Lynne | Tue Jun 17 13:18:23 2025
+0900| [e5bb448543a60591b10a7070ec17d4799f621c9d] | committer: Lynne
vulkan: maintain compatibility with old headers
Previous patch to fix these issues was incomplete.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Thu Jun 12 00:17:01 2025
+0900| [922a1ca98955487df983a8ebe3e2b545f9f18de8] | committer: Lynne
vulkan: maintain compatibility with older headers
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=922a1ca98955487df983a8ebe3e2b545f9f18de8
---
libavu
ffmpeg | branch: master | Lynne | Wed Jun 11 13:41:59 2025
+0900| [3ac7d70291838d71a28cd31a5774bb69745d4216] | committer: Lynne
hwcontext_vulkan: fix image copy
The patch was applied by mistake in an unfinished form.
This fixes the build and lets the code run.
> http://git.videolan.
ffmpeg | branch: master | Lynne | Mon May 12 18:42:09 2025
+0200| [a9b2c10eee9cf28ecbce2f1972564f4aa826a855] | committer: Lynne
hwcontext_vulkan: use host image copy
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=a9b2c10eee9cf28ecbce2f1972564f4aa826a855
---
libavu
ffmpeg | branch: master | Lynne | Wed Jun 4 02:43:41 2025
+0900| [26d17709e7dca49d5959ca31f9626a603f0bbaee] | committer: Lynne
hwcontext_vulkan: minimize queue allocation on NVIDIA
On NVIDIA, there's a global maximum limit of approximately 112 queues,
which means it takes ONLY 7
ffmpeg | branch: master | Lynne | Thu Jun 5 00:52:23 2025
+0900| [f531c91170773032c2c8ab7d390e7a6a1382c77b] | committer: Lynne
hwcontext_vulkan: add a setting to limit queues
If its a problem, you'll likely want to set it to 1 than more fine-grained
control, which you can already do vi
ffmpeg | branch: master | Lynne | Wed Jun 4 02:34:40 2025
+0900| [b5262bccdba0a86b29c6cc58f099c869df752b61] | committer: Lynne
hwcontext_vulkan: do not use optical flow queueus by default
We don't use them, and on NVIDIA, each queue takes around 30ms
to allocate, and the driver has a g
ffmpeg | branch: master | Lynne | Fri May 30 01:19:06 2025
+0900| [ab37c7e49ff98a8c32c90e6df49dd11145bc64bb] | committer: Lynne
ffv1enc: do not hardcode 1024 slices
Instead use a field where possible and the defined constant when not.
Tested by using 4096 slices.
> http://git.videolan.
ffmpeg | branch: master | Lynne | Tue May 27 15:20:29 2025
+0900| [bf6d3dc339b6be1b07ba0812ab0d96d45f1dd53d] | committer: Lynne
ffv1enc_vulkan: allow slicecrc=2
For parity with the software encoder.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Tue May 27 06:38:47 2025
+0900| [3cbe3418b2b7b60299e54321b78922a3c3a9527b] | committer: Lynne
vulkan_ffv1: fix golomb coding for non-RGB streams
The run_index is reset on each plane, unlike with RGB, where
its reset once per slice.
> http://git.videolan.
ffmpeg | branch: master | Lynne | Tue May 27 06:32:55 2025
+0900| [c395ad7c2cdc18dea2b604e7471643b20fb2716e] | committer: Lynne
vulkan_ffv1: small cleanup for golomb
Split up computation of the offset in the same way that
the range coder version does it.
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Lynne | Fri May 23 06:42:38 2025
+0900| [56b85b689dcfcf76792860fdf75ae65e2b739e30] | committer: Lynne
aacdec_ac: fix signed overflow in ff_aac_ac_update_context()
The issue is that state->cur[] is 8-bits, but a+b+1 can overflow
before being clipped to 0xF in
ffmpeg | branch: master | Lynne | Fri May 23 05:11:07 2025
+0900| [977d1a24bc01ad2a9674f7bc749c6dc4adde0e13] | committer: Lynne
vulkan/ffv1: fix sync issue in cached bitstream reader/writer
The issue is that there is an explicit lack of synchronization as only the very
first invocation writes
ffmpeg | branch: master | Lynne | Wed May 21 03:10:46 2025
+0900| [842fa198e97937f546fd5001c8889aa15c9cfd34] | committer: Lynne
hwcontext_vulkan: fix build with old Vulkan header versions
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=842fa198e97937f546fd5001c8889aa15c9cf
ffmpeg | branch: master | Lynne | Thu May 1 19:45:08 2025
+0200| [bd41838b60660f2313f4822fa8899b657949d52e] | committer: Lynne
ffv1enc_vulkan: switch to 2-line cache, unify prediction code
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=bd41838b60660f2313f4822fa8899b657949d
ffmpeg | branch: master | Lynne | Mon May 12 17:48:58 2025
+0200| [7c3c5c8052700c6b7a95ad50c865090800918c59] | committer: Lynne
hwcontext_vulkan: correct image transfer usage flags
By pure coincidence, BUFFER and IMAGE flags were equal for those
two usage types.
> http://git.videolan.
ffmpeg | branch: master | Lynne | Tue May 6 19:31:11 2025
+0200| [435db9bb49e48aae0ada537fe9ce9fe60d87a4f6] | committer: Lynne
vulkan: enable VK_KHR_shader_subgroup_rotate
Yet another thing that should've been always present.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=
ffmpeg | branch: master | Lynne | Wed Nov 20 08:27:02 2024
+0100| [7576410af776131eead51431993ee3a3c0bb01d8] | committer: Lynne
ffv1enc_vulkan: implement RCT search for level >= 4
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7576410af776131eead51431993ee3a3c0
ffmpeg | branch: master | Lynne | Tue May 6 11:53:12 2025
+0200| [7b45d9c5fdce010bbd00be1fb506685f6885a35b] | committer: Lynne
vulkan_ffv1: pipe through slice decoding status
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7b45d9c5fdce010bbd00be1fb506685f6885a35b
---
libavco
ffmpeg | branch: master | Lynne | Fri May 2 11:52:17 2025
+0200| [69f83bafd1faa69f3b05c87615c098eb0173d35f] | committer: Lynne
ffv1enc_vulkan: get rid of temporary data for the setup shader
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=69f83bafd1faa69f3b05c87615c098eb0173d
ffmpeg | branch: master | Lynne | Mon May 12 19:14:30 2025
+0200| [eabb62813e74fd71781159ac7dc38042a8ef3352] | committer: Lynne
hwcontext_vulkan: only try exporting DMABUF memory on !WIN32 and only for
DMABUF tiling
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Tue May 6 11:20:19 2025
+0200| [cb8f4b675d1e1074ddb6fe2bbb965933fb9e3571] | committer: Lynne
vulkan/ffv1: unify encode and decode get/put primitives
This simply makes a get_rac/put_rac_internal variant that can be
reused.
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Lynne | Thu May 1 12:47:16 2025
+0200| [ebbc7ff650676e59861ef49a5ffa3e32b5c7c784] | committer: Lynne
ffv1enc_vulkan: merge all encoder variants into one file
Makes it easier to work with, despite the heavy ifdeffery.
> http://git.videolan.org/gitweb.cgi/ffmpeg.
ffmpeg | branch: master | Lynne | Sat May 3 14:30:16 2025
+0200| [0156680f094f68de6d5e1d8e39994786b11f4c1a] | committer: Lynne
ffv1enc_vulkan: implement the cached EC writer from the decoder
This gives a 35% speedup on AMD and 50% on Nvidia.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git
ffmpeg | branch: master | Lynne | Sat May 3 14:15:04 2025
+0200| [a24ea37228f4a01d1adcf366d98ad112becc3e32] | committer: Lynne
vulkan_ffv1: fix PCM + cached symbol reader
writeout_rgb requires that all subgroups are active.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Sat May 3 13:44:04 2025
+0200| [f69db914cece68d92c0580ef4312eb6946c5a168] | committer: Lynne
ffv1enc_vulkan: use ff_get_encode_buffer
We used to create our own buffer, but still used the DR1 flag,
which is not how it's supposed to work.
Instead
ffmpeg | branch: master | Lynne | Fri May 2 15:39:55 2025
+0200| [8a2d9216275d7c32d1eb87c03c1a944b4a05bea2] | committer: Lynne
ffv1_common: minor RGB optimization
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=8a2d9216275d7c32d1eb87c03c1a944b4a05bea2
---
libavcodec/vul
ffmpeg | branch: master | Lynne | Fri May 2 13:34:49 2025
+0200| [7c0a8c07ce75143bcaf1d2996fadcf4bd4630a39] | committer: Lynne
ffv1enc_vulkan: unify EC code between setup and encode
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7c0a8c07ce75143bcaf1d2996fadcf4bd4630
ffmpeg | branch: master | Lynne | Thu May 1 14:36:23 2025
+0200| [a4078abd739a79390fdb8aefa1d5c3aa84d843e6] | committer: Lynne
vulkan/ffv1: synchronize get_pred implementations between encoder and decoder
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Fri May 2 15:27:52 2025
+0200| [52595025c5aab5986b39ea9276c959ea6f801b78] | committer: Lynne
ffv1enc_vulkan: minor EC optimizations
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=52595025c5aab5986b39ea9276c959ea6f801b78
---
libavcodec/vul
ffmpeg | branch: release/4.2 | Lynne | Wed Jan 1 18:03:33 2025
+0900| [4b61be93b8183fcfc1f02058829b99c4efc7b5b0] | committer: Michael
Niedermayer
configure: update copyright year
On 01/01/2025 19:05, Peter Ross wrote:
> FFmpeg turns 25 this year.
(cherry picked from com
ffmpeg | branch: master | Lynne | Mon Apr 28 07:57:13 2025
+0200| [9c0349cca1e47794cb0cd243020f211c8e932183] | committer: Lynne
aacdec_usac: correct Mps212 parsing location
It gets parsed after SBR, even if there is no SBR.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Tue Apr 29 07:41:36 2025
+0200| [33a4d3610162758ddb0390a93ea10c4e8fea7e9d] | committer: Lynne
hwcontext_vulkan: support AV_PIX_FMT_GBRP
Support was partially added previously in vulkan.c, but now it's fully
supported.
> http://git.videolan.org/gi
ffmpeg | branch: master | Lynne | Tue Apr 29 10:01:21 2025
+0200| [707c04fe0613ec9daa91361b889d5f9eedf93c5c] | committer: Lynne
ffv1enc_vulkan: support 8 and 16-bit 2-plane YUV formats
This adds support for all 8-bit and 16-bit 2-plane formats.
P010 and others require more work as the data&#
ffmpeg | branch: master | Lynne | Wed Apr 23 01:44:59 2025
+0200| [37bd91504250d21003a0397c2d1587a0d94bf0b8] | committer: Lynne
vulkan: use _KHR suffix for push descriptor properties struct ID
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=37bd91504250d21003a0397c2d1587a0d94bf
ffmpeg | branch: master | Lynne | Tue Apr 22 13:37:35 2025
+0200| [3bb2b8aff4f051f4f90b0161715afb2ef2f787e1] | committer: Lynne
hwcontext_vulkan: enable subgroupSizeControl
We already use this feature for setting the subgroup size,
but this feature was not enabled.
Fixes a validation warning
ffmpeg | branch: master | Lynne | Sun Apr 20 10:57:42 2025
+0200| [96ddce1b3c11a998ab249569b51a027cd5d74c45] | committer: Lynne
vulkan: move OPT_CHAIN out of hwcontext_vulkan
This allows for it to be shared.
Technically, implementations should not give drivers structs
that the drivers are not
ffmpeg | branch: master | Lynne | Sun Apr 20 10:35:15 2025
+0200| [5098b1a34545e8c16787c0a8d2823b1fb0194221] | committer: Lynne
vulkan: move feature<->usage mapping code outside of hwcontext_vulkan.c
Allows for it to be reused. In particular, for a future patch to make
vulkan hwaccels
ffmpeg | branch: master | Lynne | Sun Apr 20 10:29:17 2025
+0200| [cee34e0a550eddccdd4984926d519f5cd46a9d55] | committer: Lynne
vulkan: check that the max number of push descriptors is not exceeded
Just correctness. We don't exceed this on any known hardware, but
its better to check.
If
ffmpeg | branch: master | Lynne | Sat Apr 19 18:38:16 2025
+0200| [7cd1edeaa410d977a9f1ff8436f480cb45b80178] | committer: Lynne
vulkan: drop bgr_workaround
Vulkan's main issue around using BGR is simple.
The letters in the shader don't match up (rgba in shader, bgra in format).
So
ffmpeg | branch: master | Lynne | Fri Apr 18 16:27:19 2025
+0200| [ec3f3457fdfdf49b64e57c1e6c8cefd0603bc309] | committer: Lynne
vulkan_decode: add STORAGE flag to output images
In filtering, and SDR encoding, we use storage images.
This fixes using Vulkan filters on Intel.
Tested not to break
ffmpeg | branch: master | Lynne | Fri Apr 18 13:47:32 2025
+0200| [ca6392e0a7072f8bd4871dcc2af2219581aff8de] | committer: Lynne
vulkan: always enable ReadWithoutFormat/WriteWithoutFormat
This implements support for reading and writing storage images with
no format.
The issue is that we define
ffmpeg | branch: master | Lynne | Wed Apr 16 13:03:28 2025
+| [bb3ce284d7cc99ddef9b2ceb00ec2db8b2c51338] | committer: Lynne
vulkan: use a single command buffer per command buffer pool
We violated the spec, which, despite the actual command buffer pool
*not* being involved in any functions
ffmpeg | branch: master | Lynne | Wed Apr 16 16:09:57 2025
+| [36c6c66deb56dcad0ca50778069cd7ef4ded5f08] | committer: Lynne
vulkan/rangecoder: minor cleanup
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=36c6c66deb56dcad0ca50778069cd7ef4ded5f08
---
libavcodec/vul
ffmpeg | branch: master | Lynne | Sat Apr 5 05:13:02 2025
+| [b72ada0a9652c24ac0bd7133e51839e727bdae8a] | committer: Lynne
vulkan_ffv1: fallback to upload if mapping packet fails, fix fallback
The commit which added support for host mapping accidentally broke the
original, upload route
ffmpeg | branch: master | Lynne | Fri Apr 4 05:49:31 2025
+| [defebd74c094f72af253b776882fc07b71a7db3c] | committer: Lynne
vulkan_ffv1: slightly optimize the range decoder
GPUs have cmovs as standard.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Sat Apr 5 18:17:29 2025
+| [6111aef5336fac475963891bcc5afc6ee7a5c908] | committer: Lynne
vulkan_ffv1: fix reset shader dependencies
Without a barrier upfront, the reset shader may read data fields not
yet set by the setup shader.
> h
ffmpeg | branch: master | Lynne | Sat Mar 29 22:44:11 2025
+| [79ff1f21c4bcd2b917973357138829d44c3836be] | committer: Lynne
vulkan_decode: only create sequence params in end_frame
We tried to create sequence params in both start_frame and end_frame.
This was redundant.
Just always create
ffmpeg | branch: master | Lynne | Sun Apr 13 03:39:01 2025
+| [7b0156201bf4c2543225ec3fb7f64b57f901d1dd] | committer: Lynne
vulkan: fix logging level when erroring upon creating shader module
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Tue Apr 8 05:45:16 2025
+| [985a26be28f838e47cf2208e0254a7edf731ab41] | committer: Lynne
vulkan_ffv1: shortcut +-1 coeffs in symbol reading
Slightly faster, and allows for further optimizations.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Sat Apr 5 00:58:21 2025
+| [45d7abf6d9eb004411e1088f3fe4699744c7025b] | committer: Lynne
vulkan_ffv1: init overread/corrupt fields
Forgotten.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=45d7abf6d9eb004411e1088f3fe4699744c70
ffmpeg | branch: master | Lynne | Sun Apr 6 07:45:44 2025
+| [8ceabb677c2d7796c36ec5318be8feb5847ffd1a] | committer: Lynne
vulkan_ffv1: externalize extended lookup check
8% speedup on nvidia on 4k.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Thu Apr 10 08:44:31 2025
+| [29b85cd4b8f2b3f9806332e15ef488748e895476] | committer: Lynne
vulkan_ffv1: add cached symbol reader for AMD
Speeds up everything on AMD by 3x.
This uses 32 local invocations to load state into cache, as well
as to do the RCT
ffmpeg | branch: master | Lynne | Fri Apr 11 23:44:26 2025
+| [e040c087c708c3c4a4c747d7042f1825a454ca75] | committer: Lynne
vulkan: add support for expect/assume
This commit adds support for compiler hints.
While on AMD these are not used/needed, Nvidia benefits from them, and gives
a
ffmpeg | branch: master | Lynne | Sun Apr 6 20:11:32 2025
+| [4d561e6a1ed4ffe3bac6fdc2839c5adcf3a44699] | committer: Lynne
vulkan_ffv1: remove need for scratch data during setup
This saves on some VRAM, but mainly allows for a more unified path.
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Lynne | Sun Apr 13 03:35:56 2025
+| [11911aef4631971e3233f086b52caf06c8a7f55c] | committer: Lynne
vulkan_shaderc/glslang: print full shaders on TRACE rather than VERBOSE
Way too spammy.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Sat Apr 5 05:04:29 2025
+| [77f777d9258e2de84e743d77fcd4f1678910e81d] | committer: Lynne
ffv1/vulkan: redo context count tracking and quant_table_idx management
This commit also makes it possible for the encoder to choose a different
quantization table on
ffmpeg | branch: master | Lynne | Sun Apr 6 00:47:29 2025
+| [72953477a415d816e1152256f9cb3f74d00f245c] | committer: Lynne
vulkan_ffv1: fix left-2 sample addressing
Typo.
Not enough to fix context=1, but its a start.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Tue Apr 1 22:36:46 2025
+| [66b8c92df2b6644af96b25d6716664a2c23e5337] | committer: Lynne
vulkan_ffv1: cache only 2 lines when decoding RGB
This reduces the intermediate VRAM used for RGB decoding by a
factor of 100x for 6k video.
This also speeds the
ffmpeg | branch: master | Lynne | Sun Apr 6 03:49:28 2025
+| [694ebe890ce4cda41780006487fd351fbdebdec8] | committer: Lynne
vulkan_ffv1: improve buffer barrier correctness for slice state
This is likely a nanooptimization, but its more correct.
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Lynne | Wed Apr 2 00:03:11 2025
+| [1f09b55c943a63c85a92761ccf63a2a99b9d16ea] | committer: Lynne
vulkan_ffv1: allocate just as much memory for slice state as needed
Rather than always using the maximum allowed slices, just use the number
of slices present in this
ffmpeg | branch: master | Lynne | Wed Apr 2 20:46:12 2025
+| [fc960dafef9f1b30433bbf6ead84fbdc8d492539] | committer: Lynne
vulkan_ffv1: optimize symbol reader
This was the fastest variant tested.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Sat Apr 12 06:24:03 2025
+| [d077e00f3eb4d094cb16e099e5bb138458118d54] | committer: Lynne
vulkan_ffv1: enable acceleration on Intel
Fixed by previous commit.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Sat Apr 5 05:06:30 2025
+| [d7772da728bcb333ea49be98b84439a19e9381d2] | committer: Lynne
vulkan_ffv1: remove unused define
Leftover debug macro.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d7772da728bcb333ea49be98b84439a19e938
ffmpeg | branch: master | Lynne | Sat Apr 12 06:22:24 2025
+| [a1137f9214e140c79e5148d1aebee5a80e65d2b7] | committer: Lynne
hwcontext_vulkan: disable descriptor buffer extension on Intel
Temporary workaround. Will be replaced with a version check once a fix is
in the works and a known next
ffmpeg | branch: master | Lynne | Sun Apr 13 03:37:14 2025
+| [4f64df2928d7e54c89a5e1263c6768d3dcfe61a2] | committer: Lynne
vulkan: remove unused field from exec pools
This used to be involved in a mechanism to switch between queue indices,
but since the rewrite of the rewrite of the
ffmpeg | branch: master | Lynne | Thu Mar 27 12:44:04 2025
+| [4dadf71dceb6d211834fdfa6c9f2686b37132b5a] | committer: Lynne
lavu/vulkan: skip adding NULL buffers as deps in ff_vk_exec_add_dep_buf
Allows for cleaner code.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Thu Mar 27 12:47:11 2025
+| [5fc4acae9c764cf8b13d085d2d335bddd12230b5] | committer: Lynne
vulkan_decode: allow using NULL sequence_params when decoding
The function had some checks to allow for this, but as it always tried
to dereference a bufferref, it
ffmpeg | branch: master | Lynne | Thu Mar 27 17:21:16 2025
+0100| [8804ae158ba51d2b9d34d104a93e58b45ba07815] | committer: Lynne
vulkan_functions: add note to update ff_vk_extensions_to_mask
Otherwise, the extension looks like its enabled, but isn't
actually used by anything.
&
ffmpeg | branch: master | Lynne | Thu Mar 27 15:47:18 2025
+| [5595068be562b2a975d4f60ad691307dbb1bf079] | committer: Lynne
vulkan_loader: remap video_maintenance2 in ff_vk_extensions_to_mask
This was broken.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Thu Mar 27 12:49:06 2025
+| [193610d9bac1fc00f99975f13c1370aec816f930] | committer: Lynne
vulkan_decode: allow using NULL offsets/nb_slices in ff_vk_decode_add_slice()
For codecs like VP9 which use a single slice.
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Lynne | Mon Mar 10 03:04:39 2025
+| [6bad55eb17d03e7e54106bd195d1afd8ebf36197] | committer: Lynne
ffv1: add a Vulkan-based decoder
This patch adds a fully-featured level 3 and 4 decoder for FFv1,
supporting Golomb and all Range coding variants, all pixel formats
ffmpeg | branch: master | Lynne | Fri Mar 14 17:53:02 2025
+| [3c09aadcf47a1cbb990c48cac1c4bea9bf0593cf] | committer: Lynne
hwcontext_vulkan: allow using concurrent access images with more than 5 queues
This limit was not bumped when the queue family rewrite happened.
Fixes validation
ffmpeg | branch: master | Lynne | Mon Mar 10 16:44:00 2025
+| [53ff9005b5bd4bedad7080121d4220b3bd193561] | committer: Lynne
vulkan: add support for AV_PIX_FMT_GBRAP32
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=53ff9005b5bd4bedad7080121d4220b3bd193561
---
libavu
ffmpeg | branch: master | Lynne | Mon Mar 10 16:40:04 2025
+| [49af9746e8f6b65f0a3981c749af5a97194d6849] | committer: Lynne
pixfmt: add AV_PIX_FMT_GBRAP32
This commit adds a 32-bit *integer* planar RGBA format.
Vulkan FFv1 decoding is best performed on separate planes, rather than
packed
ffmpeg | branch: master | Lynne | Fri Mar 14 15:07:45 2025
+0100| [82864c21112157951ce91b4430a9018edd02f5ab] | committer: Lynne
vulkan_hevc: use VK_KHR_video_maintenance2 if available
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=82864c21112157951ce91b4430a9018edd02f
ffmpeg | branch: master | Lynne | Fri Mar 14 12:44:24 2025
+0100| [d3d2e254ebc2a9bee849605336467f6b2a60f36d] | committer: Lynne
av1dec: update hwaccel decode_params on AV1_OBU_SEQUENCE_HEADER
Previously, the callback was only called on init. This makes it
get called on every frame.
We should
ffmpeg | branch: master | Lynne | Fri Mar 14 12:53:47 2025
+0100| [67eb01b458371b0cd885e99a8943cb95c01bf4f5] | committer: Lynne
hwcontext_vulkan: enable VK_KHR_video_maintenance2
Enables some fixes for the video specifications, and a new feature.
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Lynne | Mon Mar 10 03:03:41 2025
+| [4de023f95362b0a13567e81329efd22cbb8c6a34] | committer: Lynne
vulkan: add ff_vk_exec_add_dep_wait_sem()
This adds a function which adds a regular timeline semaphore
as a wait-only dependency.
> http://git.videolan.
ffmpeg | branch: master | Lynne | Fri Mar 14 13:23:55 2025
+0100| [9a78ebc11a8845a9e9bdc2dd1fa5c474263833bf] | committer: Lynne
vulkan_av1: use VK_KHR_video_maintenance2 if available
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=9a78ebc11a8845a9e9bdc2dd1fa5c47426383
ffmpeg | branch: master | Lynne | Sat Feb 22 19:55:53 2025
+| [bbb06f1ffc804ed55c56632e4ccfcc7704333bf2] | committer: Lynne
vulkan: workaround BGR storage image undefined behaviour
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=bbb06f1ffc804ed55c56632e4ccfcc7704333
ffmpeg | branch: master | Lynne | Thu Mar 13 16:26:40 2025
+| [caff29dbb18feeb87cb00fc4c33d20cf01667be0] | committer: Lynne
FFHWAccel: add buffer_ref argument to start_frame
This commit adds a reference to the buffer as an argument to
start_frame, and adapts all existing code.
This allows
ffmpeg | branch: master | Lynne | Wed Feb 19 16:09:50 2025
+| [a5b90caacdbf0c5c9e1f88ae903fe16bc194d371] | committer: Lynne
vulkan: copy host-mapping buffer code from hwcontext
This is useful elsewhere.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Fri Mar 14 22:09:11 2025
+| [dd7cc557a5040d3e8a9a072f01155b591607] | committer: Lynne
ffv1enc_vulkan: clip micro_version to 3 for level 4
This unbreaks level 4 encoding.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Lynne | Sun Jan 5 15:27:30 2025
+0900| [23eb499b282ab1952461872ddea008adc7f95413] | committer: Lynne
ffv1dec: add support for hwaccels
This commit adds support for hardware accelerated decoding to
the decoder.
The previous commits already refactored the decoder, this
ffmpeg | branch: master | Lynne | Tue Dec 24 15:57:22 2024
+0900| [b2ebe9884ec156f1550aa7c677b66156db77ea9d] | committer: Lynne
ffv1enc_vulkan: refactor code to support sharing with decoder
The shaders were written to support sharing, but needed slight
tweaking.
> http://git.videolan.
ffmpeg | branch: master | Lynne | Sun Feb 23 14:09:13 2025
+| [f2a0bdd6b15523f062ac782727696effd449d347] | committer: Lynne
vulkan: unify handling of BGR and simplify ffv1_rct
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=f2a0bdd6b15523f062ac782727696effd449d
ffmpeg | branch: master | Lynne | Fri Mar 14 13:41:10 2025
+0100| [53f156bccc4886b687f6433ed5e69cf827e9f666] | committer: Lynne
vulkan_h264: use VK_KHR_video_maintenance2 if available
This avoids having to copy and update the entire SPS/PPS stack.
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Lynne | Fri Mar 14 13:23:36 2025
+0100| [31176b16accd26b0a9cbeb0552f5a8ef7a8d2c0c] | committer: Lynne
vulkan_decode: use VK_KHR_video_maintenance2 if available
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=31176b16accd26b0a9cbeb0552f5a8ef7a8d2
ffmpeg | branch: master | Lynne | Fri Feb 21 01:33:54 2025
+| [4495802bdb4d2bd4b4eabbce3936ceb89236ffd3] | committer: Lynne
vulkan_decode: support multiple image views
Enables non-monochrome video decoding using all our existing functions
in the context of an SDR decoder.
> h
ffmpeg | branch: master | Lynne | Fri Mar 7 21:07:41 2025
+| [e15e85b869c086fab9e926a7fbc7b6a2ca759997] | committer: Lynne
vulkan_decode: adjust number of async contexts created
This caps the number of contexts we create based on thread count.
This saves VRAM and filters out cases where
ffmpeg | branch: master | Lynne | Tue Dec 3 18:21:57 2024
+0900| [491b65e3437cb5f4e8647eb5d780bc9aa62841a1] | committer: Lynne
vulkan_decode: support software-defined decoders
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=491b65e3437cb5f4e8647eb5d780bc9aa62841a1
---
libavco
1 - 100 of 817 matches
Mail list logo
