Revisions of libplacebo
- Update to version 6.292.0+368~git20230914.808f1af: * demos/plplay: switch to new ICC profile API * renderer: remove pl_render_params.icc_params * options: refactor option macros slightly * renderer: add pl_frame.icc * renderer: re-use pl_icc_update internally * shaders/icc: fix empty profile check * shaders/icc: add pl_icc_update helper * shaders/icc: add note to size_r/g/b comment * shaders/icc: fix out-of-date / misleading comment * colorspace: add missing pl_icc_profile helper macro * colorspace: improve pl_icc_profile_compute_signature * demos/plplay: fix target ICC signature calculation * shaders/colorspace: document ICC signature requirements * shaders/colorspace: fix theoretical white shift * shaders/colorspace: fix gamut visualization * meson: fix typo (cosmetic) * shaders/custom: correctly update hook params * renderer: add pl_render_params.correct_subpixel_offsets * gamut_mapping: sanity clamp highlight() output * opengl/formats: update SNORM format documentation * shaders/colorspace: switch to UNORM 3DLUT texture * shaders/icc: handle format lookup failure gracefully * shaders/colorspace: fallback when no rgba16s format * filters: fix polar downscale antiringing
- Update to version 6.292.0+344~git20230913.a12b733: * renderer: fix pl_render_params.ignore_icc_profiles * demos/plplay: add option to control target zoom/crop * demos/utils: fix build on macOS * renderer: allow non-complex scalers without FBOs * gpu: suppress scan-build warning * utils/dav1d: remove use of deprecated type name * renderer: remove outdated and wrong comment * renderer: pass overlays through ICC profile * cache: sanitize/expose configured max object/total size * shaders/colorspace: use rgba16s for gamut 3DLUT * cache: add missing header * shaders/lut: fix texture LUT size generation * cache: make pl_cache_load() more robust * cache: defend against array overflow * cache: fix for 32-bit systems * demos/plplay: increase default cache size * cache: add verbose PL_TRACE logging * hash: improve golden ratio constant * global: remove MSAN hacks * cache: add sanity assertion (+ whitespace) * vulkan/context: log init times * demos/plplay: drop max object size * xxhash: add optional libxxhash implementation of pl_mem_hash * hash: inline pl_mem_hash * hash: move prototype from pl_string to new header * cache: fix deadlock on loading corrupt files
- Update to version 6.292.0+318~git20230908.07fa0f0: * demos/plplay: add cache statistics UI * demos/plplay: add basic shader caching support * demos/utils: add get_cache_dir helper * demos: add missing license headers * demos: remove old/unused assets * dispatch: deprecate pl_dispatch_save/load * gpu: remove pl_pass_params.cached_program * d3d11/pass: pl_str_hash(pl_str0()) -> pl_str0_hash * d3d11/pass: switch to pl_cache based caching * opengl/pass: switch to pl_cache based caching * vulkan/pass: switch to pl_cache based caching * shaders/colorspace: enable gamut mapping 3DLUT caching * shaders/colorspace: switch from update to signature * shaders/film_grain_h274: cache grain LUT * shaders/dithering: cache blue noise dither matrix * shaders/icc: refactor cache mechanism * shaders/lut: add ability to cache shader LUTs directly * gpu: add pl_gpu_set_cache() * tests/cache: add tests for pl_cache mechanism * cache: add common/core caching mechanism * pl_string: add pl_var_hash * d3d11/gpu: add missing pl_spirv_destroy call * glsl/spirv: API bikeshedding * convert: make fast_float optional dependency if not required * demos/plplay: work around exp10f non-portability * demos/plplay: adjust polar cutoff slider * filters: update ewa_lanczossharp blur value * meson: add missing variable * filters: fix ewa_lanczossharp(est) legacy description
- Update to version 6.292.0+269~git20230903.7737933: * tests: increase default test timeout to 120s * renderer: default to ewa_lanczossharp * filters: add pl_filter_ewa_lanczossharp(est) * options: add antiring options to custom scalers * filters: add pl_filter_config.antiring * filters: remove pl_filter_config.filter_scale * shaders/sampling: use pl_filter_config.blur instead of filter_scale * filters: apply blur to window as well * options: parse floats expressed as fractions * shaders/colorspace: bias PQ deltas for mostly black frames * shaders/colorspace: don't bypass smoothing for black frames * shaders/colorspace: distinguish between active/total wg count * shaders/colorspace: properly handle pl_buf_read failure * shaders/colorspace: don't warn on solid black frame
- Update to version 6.292.0+255~git20230901.ec60dd1: * shaders/colorspace: significantly lower default smoothing period * shaders/colorspace: significantly reduce scene threshold * tone_mapping: remove pl_tone_map_auto * shaders/colorspace: default to pl_tone_map_spline * shaders/colorspace: exclude black pixels from peak detection * shaders/colorspace: simplify sh_luma_coeffs signature * shaders/colorspace: use correct luma coeffs for peak detection * options: exclude special options from `preset=` loading
- Update to version 6.292.0+247~git20230828.defc776: * renderer: only disable peak detection if HDR10+ OOTF in use * tone_mapping: simplify auto-TM selection * demos/plplay: add detected peak metadata * renderer: add pl_renderer_get_hdr_metadata * shaders/colorspace: remove minimum_peak * shaders/colorspace: fix order of pl_tone_map_params_infer * tone_mapping: also constrain output max if no inverse TM * tone_mapping: properly constrain input peak * tone_mapping: used inferred peak in fix_params() * tone_mapping: lower default adaptation strength * tone_mapping: add dynamic knee point adaptation tuning * tone_mapping: reformulate st2094_pick_knee (cosmetic) * tone_mapping: rename duplicate expression (cosmetic) * gamut_mapping: slightly increase perceptual strength default * demos/plplay: simplify option value ranges * options: simplify value ranges
- Update to version 6.292.0+231~git20230826.612071d: * renderer: fix nominal luma comparison * gamut_mapping: slightly boost default perceptual deadzone * gamut_mapping: drop perceptual strength to 0.6 * tests/tone_mapping: force max strength for saturation test * gamut_mapping: add tunable perceptual_strength * renderer: use nominal max luminance for peak detection check
- Update to version 6.292.0+225~git20230826.b19e441: * tools/glsl_preproc: fix vecType on 1-component mask * shaders: make sh_float_type accept comp mask * vulkan/formats: drop format probing spam verbosity * dispatch: reduce verbosity of cache load/save spam * options: lower verbosity of option parsing * Revert "tools/glsl_preproc: add support for const char variables" * Revert "docs/glsl: expand GLSL variables documentation" * shaders/sampling: switch to new swizzle/vecType fmt * tools/glsl_preproc: add support for swizzle and vecType * tools/glsl_preproc: add ability to wrap bound references * tools/glsl_preproc: move FmtSpec to statement.py * demos/plplay: add advanced constants section * demos/plplay: remove deprecated tone mapping param UI * options: map pl_tone_map_constants * tone_mapping: refactor tone mapping parameters API * tone_mapping: change gamma default knee to 0.3 * tone_mapping: change st2094-10 knee point default to 0.7 * tests/options: explicitly test setting 0.0 * options: don't reject 0.0 floats * options: expose pl_color_map_constants options * options: fix gamut_mapping options order * gamut_mapping: make hard-coded constants tunable * common: protect pl_smoothstep against edge0 = edge1 * shaders: replace magic constants with values from default struct
- Update to version 6.292.0+201~git20230821.69e40e5: * utils/frame_queue: don't crash on empty pl_queue_update
- Update to version 6.292.0+200~git20230821.e9b39ab: * RELEASING: add note to create github release tags * vulkan/malloc: bump up maximum slab age * vulkan/malloc: profile malloc times * tests: limit array log size and start with an offset * Revert "d3d11: remove static texels count limit" * vulkan/tex: split texel transfers into slices * d3d11/tex: split texel transfers into slices * gpu/utils: add pl_tex_transfer_slices * gpu/utils: make tex offset dynamic in pl_tex_upload/download_texel * gpu/utils: fix pl_tex_upload/download_texel to respect params.buf_offset * gpu/utils: drop UBO support from pl_tex_upload_texel * demos/video-filtering: switch to new pl_thread helpers * pl_thread: properly document pl_thread_sleep * docs/glsl: expand GLSL variables documentation * shaders/deinterlacing: re-use swizzle helpers * shader/sampling: 80col align GLSL pragma (cosmetic) * shaders: simplify sh_tex_swiz helper * tools/glsl_preproc: add support for const char variables * tools/glsl_preproc: fix ident_t size * renderer: alignment (cosmetic) * renderer: set proper error attributes on hook * renderer: don't wipe `unique` attrib on custom hooks * options: allow : as separator in pl_options_load * renderer: correctly hook CHROMA/ALPHA_SCALED * renderer: simplify plane shader handling * win32: include manifest in libplacebo and plplay * shaders/colorspace: allow smoothing period of 0.0 * shaders/colorspace: simplify iir_coeff math * vulkan/malloc: fix allocation with large pages
- Update to version 6.292.0+158~git20230811.e9043c5: * shaders/film_grain_av1: fix name collision on 4:4:4
- Update to version 6.292.0+157~git20230809.2af3c56: * tools/glls_preproc: use correct shader asprintf * README: more minor fixes * tests/fuzz: add pl_options_load fuzzer * tests/fuzz: fix various issues * README: fix minimum vulkan version * README: add github sponsors link * README: rewrite marketing blurbs * README: historical revisionism * README: remove obsolete sponsorship section * README: document options.h * demos/plplay: allow pasting/copying settings from clipboard * demos/window: support clipboard copy/paste * demos/plplay: add raw settings dump * demos/plplay: switch to pl_options * options: add new options parsing API * pl_string: add pl_str_split_chars * pl_alloc: support negative indices in PL_ARRAY_* * tests: add REQUIRE_STREQ
- Update to version 6.292.0+139~git20230807.95d44fb: * vulkan/context: don't load synchronization2 layer * shaders/sampling: improve numerical accuracy of antiringing * shaders/custom: add support for NAMED parameters * shaders/sampling: redesign polar anti-ringing shader * renderer: fix outdated comment * shaders/sampling: enable antiringing for polar downscaling * shaders/sampling: fix stray parameter * shaders/sampling: optimize/improve polar anti-ringing
- Update to version 6.292.0+131~git20230806.a09b834: * common: fix PL_PRINTF nonnull annotation * shaders/custom: formatting (cosmetic) * demos/plplay: switch to new filters API * renderer: soft-deprecate old filter configuration API * filters: add "linear" pseudo-alias for bilinear * filters: add extra aliases to pl_filter_configs/functions * filters: avoid unnecessary double precision * filters: fix description on mitchell_clamp * filters: fix comment on spline16/36/64 * filters: fix box window * filters: remove ancient aliases * filters: remove slightly misleading comment * filters: add pl_filter_functions/configs * filters: move pl_filter_oversample to filters.h * filters: add pl_filter_config usage semantics * filters: error out on opaque kernels/windows * filters: refactor filter configuration * renderer: reprioritize default frame mixers * renderer: fix typo (cosmetic) * renderer: add safe params to PL_RENDER_DEFAULTS * shaders/icc: switch pl_icc_params.size_r/g/b to int * colorspace: make pl_color_adjustment consistent with other params * shaders/colorspace: properly default to pl_tone_map_clip * gamut_mapping: fix stray comment (cosmetic) * shaders/sampling: lower default deband threshold * shaders/colorspace: clamp CR output to legal range * shaders/colorspace: fix comment * filters: fix comment (cosmetic) * shaders/colorspace: only visualize legal part of 3DLUT
- Update to version 6.292.0+38~git20230729.fcba744: * opengl/swapchain: glFlush on submit_frame() * meson: disable UCRT math usage on 32-bit Windows target * demos/plplay: make antiringing slider slower to adjust * shaders/sampling: enable antiringing for EWA scalers * shaders/sampling: rename dmax to dmin * shaders/sampling: use #pragma GLSL in EWA footer * shaders/sampling: slight refactor to polar_sample * demos/plplay: add ability to override rendering FPS * demos/plplay: stall render loop when frame mixing disabled * demos/plplay: add frame/timing statistics * demos/plplay: implement more robust timing loop * pl_clock: add pl_clock_sleep() * utils/frame_queue: add queue state introspection API * utils/frame_queue: bump max FPS estimate hard limit * utils/frame_queue: use doubles for PTS * utils/frame_queue: handle FPS updates more quickly * utils/frame_queue: dynamically increase PREFETCH_FRAMES * vulkan: add pl_vulkan_swapchain_params.disable_10bit_sdr * vulkan: fix crash without synchronization2
- Update to version 6.292.0+19~git20230723.1d70923: * vulkan/context: don't require synchronization2 * vulkan/command: add compatibility for pre-QueueSubmit2 * vulkan/command: add compatibility for pre-sync2 barriers * vulkan/command: drop redundant vk_ctx param from vk_cmd_* * vulkan/command: drop unused param from vk_sem_barrier * vulkan/command: add struct vk_ctx *vk to vk_cmdpool * filters: fix polar downscaling
- Update to version 6.292.0+12~git20230717.e6a89f3: * shaders/colorspace: reverse LUT-less fallback logic * Revert "vulkan/pass: switch to dynamic rendering" * utils/libav: add pl_test_pixfmt_caps
Displaying revisions 21 - 40 of 165