ffmpeg keeps a trailing zero byte after the hvcC PPS that the in-band PPS lacks, so the byte-for-byte comparison read it as a parameter-set change and sent the stream to the software decoder from its first sample. A NAL unit never ends with 0x00 (H.264 7.4.1, H.265 7.4.2): those bytes are trailing_zero_8bits, so both sides are compared without them. The platform fallback test used a trailing zero byte as its changed SPS; it now changes nal_ref_idc.
With NVENC registered on Linux, `filmcraft-cli --test cli` aborts in 2 of 45 runs on an RTX 3090
(`malloc(): mismatching next->prev_size`, `malloc(): unaligned fastbin chunk detected`) right
after an export finishes; 0 of 30 runs fail with this reverted.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The NVENC HEVC encoder (Main 8-bit SDR and Main 10 HDR PQ / HLG) was gated to
Windows. It now also runs on 64-bit Linux, on the CUDA device the H.264 encoder
already uses: the H.265 format and HDR probes are registered there and the
HEVC tests run on both systems.
NVENC registration on Linux now happens after the VA-API probe. Started beside
it, the background HEVC probe's first cuInit could deadlock with
libva-nvidia-driver, which calls cuInit from its library constructor under the
loader lock.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflict in crates/engine/src/export_tools.rs: both sides added an
import (BitrateMode for CRF, GpuRendering from main); both kept.
main's new Chinese catalog requires every tl! string: added zh-CN for
"Quality (CRF)" and "CRF (constant quality)".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Session kept the parser's user data in a Box and later wrote through that Box,
which invalidates the raw pointer the callbacks use (Stacked/Tree Borrows).
Rust code now reaches the state only through the same into_raw pointer, never
across the parse call, and Drop frees it after the parser is destroyed.
The setting test accepts either Linux backend's decoder when NVDEC registered,
and the README notes perf.stats names NVDEC for streams VA-API decodes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds BitrateMode::Crf with ExportSettings::crf (0-51, default 23) and
maps it to the built-in encoder's RateControl::Crf, which export never
used. The Export panel offers "CRF (constant quality)" with a Quality
(CRF) slider in place of the bitrate fields; file.exportMedia takes
`crf` and `bitrateMode: "crf"`. H.265 (hardware only) rejects it, and
VideoToolbox / NVENC decline it so the built-in encoder runs. Settings
saved before this read as CRF 23.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The i18n CRLF normalisation already landed with #488, so main's version is kept.
abi_tests now asserts the HEVC VUI field offsets (inside the VUI and within
NV_ENC_CONFIG_HEVC) to rule out a binding mismatch behind the reversed timing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
tests/setting.rs: keep the check against the backend register() reported
(main's fix expects VA-API on Linux, which NVDEC now sits in front of).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds a second Linux decoder backend, crates/platform/src/nvdec, that talks to
NVIDIA's driver directly (libnvcuvid + the CUDA driver, both loaded at run
time). It is registered in front of VA-API: what NVDEC declines still reaches
VA-API, then the software decoders. H.264 8-bit and HEVC 8/10-bit, 4:2:0
progressive.
NVIDIA's parser takes Annex B access units, so there is no host-side DPB.
Pictures are read back as NV12 / P016 through the existing biplanar path and
are bit-exact with the software decoders. The CUDA context is the one
nvenc::device already retains; it gains push/pop and copy-to-host. A seek
recreates the parser and keeps the decoder.
Also fixes tests/setting.rs, which expected the Media Foundation backend name
on Linux, and updates the Hardware decoding tooltip and its Spanish row.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8-bit NVENC sessions (H.265 Main, hardware H.264) now take the export's RGBA
frames as ABGR input buffers and NVENC converts them to BT.709 limited 4:2:0
on the GPU, with the same codes as rgba_to_yuv420_8. The CPU conversion ran
on the encode thread, where it competed with the render of the next frames
for the thread pool (about 13 ms per 1080p frame instead of 1.5 ms) and held
NVENC back. A driver that refuses RGB input falls back to the CPU conversion
into NV12. HEVC Main 10 is unchanged.
60 s of a 1080p30 project as H.265 in the app: 18.7 s -> 15.0 s (96 -> 120
fps); headless 30.3 s -> 16.8 s.
Co-Authored-By: Claude <noreply@anthropic.com>
Every VA-API session opened its own display and terminated it when dropped. With NVIDIA's driver
that is a CUDA context each: 0.2-0.3 s to open, 60-70 ms to terminate and about 100 MB of video
memory, on the thread creating or dropping the decoder (a frame worker at every clip change, the
UI thread when media is removed: a ~0.6 s freeze deleting five played clips, and dropped frames at
clip changes).
Sessions now take the process's shared display (the startup probe's, kept for the life of the
process like the libraries) and make only their own configuration, context and surfaces. Display
is Sync on the strength of libva's Multithreading Guide (va.h): its functions and drivers are
thread-safe, and every session uses only its own VA objects. A session that cannot be made on the
shared display is tried once on a new display, which is shared from then on if it works (a GPU
reset).
On NVIDIA (RTX 3060): deleting five played clips 650-690 ms -> 132-198 ms, dropped frames in 20 s
of playback 2-3 -> 0, peak video memory 1281 -> 1010 MiB. Intel and software unchanged.
Test: tests/vaapi_displays.rs (decoders one after another and six at once on six threads open no
display; pictures bit-exact). The VA-API hardware suite passes on NVIDIA and Intel, and the
concurrency tests 20 times each.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
After a hardware decoder failed, the replay of the run through the software decoder stopped at the
first sample the software decoder rejected (`?` in the replay loop): the pictures it had already
output were dropped, the rest of the run was never fed, the error was returned for the failed
sample even when the software decoder accepted it, and the replay log was not restored. That
happens when the GPU tolerated a damaged sample our decoder rejects and then fails later in the
same run; one picture went missing on a damaged HEVC file through VA-API (Intel).
The replay now behaves like software decoding from the restart point: a rejected sample is skipped
(logged when the hardware had already decoded it); if the failed sample itself is rejected its
error is this call's, and the pictures replayed before it are kept in `carry` for the next output.
Regression test (every platform, no GPU): a stand-in hardware decoder that decodes one damaged
sample the software decoder rejects and fails at that sample or later; every software picture
comes out once and in order, errors are reported where the software decoder reports them, and the
pictures from the next restart point are the software decoder's.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Delegate HEVC VUI timing to NVENC's configured frame rate instead of the
explicit timing pair reversed by the affected driver. Preserve existing
stream quality/timing assertions and cover integer and fractional rates.
Normalize line endings in the translation test's source scanner so it
excludes test modules in Windows checkouts as well as LF checkouts.
Assisted-by: Codex
main gained NVENC H.265 Main / Main 10 (#328, #360), export stage timings
(#370) and VA-API decoding (#358). Resolution: main's NVENC session, FFI and
export factory, with this PR's changes ported onto them:
- ffi: the portable Guid (now also for the HEVC GUIDs), NV_ENC_DEVICE_TYPE_CUDA;
- session: device.rs (Direct3D 11 / retained CUDA primary context), the
Linux libnvidia-encode.so.1 loader, Device::KIND;
- export: factory_with and its tests, the even-size check, the saturating
max bitrate; H.265 through NVENC stays Windows-only (the factory declines
it on Linux, where nothing probed or tested it);
- Cargo.toml: main's Linux libloading dependency covers both users;
- docs: README / ROADMAP / architecture now say NVENC H.264 runs on Windows
and Linux; ADR and platform README keep both addenda / test rows;
- abi_tests: note why the MSVC-generated expectations hold on LP64 Linux.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
encode / encode_10 compute the plane sizes they check with saturating_mul;
fill_p010, rgbf_to_yuv420_10 and the HDR export test use as_chunks instead of
chunks_exact with a constant size (clippy::chunks_exact_to_as_chunks).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Runtime-load the NVIDIA encode and CUDA driver libraries, retaining the primary context through session destruction. Preserve the Windows Direct3D path and default-off selection, and test driver rejection and deterministic software fallback.
Co-Authored-By: OpenAI Codex <noreply@openai.com>
HDR sequences export as real 10-bit HEVC Main 10 through NVENC where the GPU
reports 10-bit HEVC: float BT.2020 R'G'B' -> rgbf_to_yuv420_10 -> P010 input;
VUI BT.2020 + PQ/HLG; HDR10 mastering-display / content-light as SEI on every
IDR (PQ); colr/mdcv/clli in the sample entry. New register_hdr_probe /
hdr_available in the export crate; macOS and GPUs without 10-bit keep
tone-mapping HDR sequences to 8-bit SDR HEVC. ABI asserts extended and
re-verified against nvEncodeAPI.h 12.1 (0 drift). 8-bit SDR HEVC unchanged;
the two #328 tests that asserted HDR is declined now expect Main 10.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
HEVC through VA-API the same way as H.264: the software HEVC decoder's
parsers and DPB (now generic over the picture type in filmcraft-hevc:
POC, reference picture sets, reference lists, output order, RASL
handling) drive the stateless hardware. Main, Main 10 and Main Still
Picture, 4:2:0 8/10-bit (P010 read-back); scaling lists, tiles, WPP,
dependent slices, weighted prediction. Missing references and range
extensions go to the software decoder.
Tests: hardware parity on HEVC Main / Main 10 incl. open GOP, scaling
lists, slices+WPP, weighted prediction, tskip+AMP, 1080p/2160p (streamed
comparison to stay within memory); HEVC front end against a recording
stand-in on every OS; ABI checks extended to va_dec_hevc.h. Benchmark on
Iris Xe: 4K HEVC 203 -> 10 ms CPU per frame, 4K Main 10 212 -> 18 ms.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Stateless VA-API decoding with FilmCraft's own H.264 host side: the
software decoder's SPS/PPS/slice-header parsers and its DPB (now generic
over the picture type in filmcraft-h264) drive the GPU, so reference
marking, reference lists and output order are shared and the pictures are
bit-exact with the software decoder. libva is loaded at run time; no libva
or driver means software decoding as before. frame_num gaps (open-GOP
seeks) get the software decoder's stand-in frames. HybridDecoder keeps the
transparent software fallback.
Tests: hardware parity on 7 H.264 streams incl. 1080p/2160p, seeks,
flushes, forced failures, fuzzing, concurrency; front end against a
recording stand-in on every OS; ABI checks against va.h. Docs: platform
README, ADR 0001 addendum, architecture, performance (H.264 2160p CPU per
frame 417 -> 11 ms on Iris Xe), roadmap.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Left to choose, NVENC answered a peak bitrate above its level's Main tier
limit with the High tier: 1080p30 at 12/18 Mbit/s was written as level 4 High
tier, 2160p30/60 at 40/60 Mbit/s as 5.0/5.1 High tier, which many hardware
decoders refuse. The level is now the lowest whose Main tier limits hold the
picture size, sample rate and peak bitrate (H.265 Table A.8), e.g. 1080p30 at
18 Mbit/s is level 4.1 Main tier. Unit test of the table; hardware test that
the hvcC says Main tier at the expected level.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- register() asks the HEVC probe on a thread of its own (the first probe is
about half a second: D3D11 device + NVENC session), so the first draw of
Export's format list does not wait for it; tests/nvenc_hevc_warm.rs.
- A machine with no HEVC encoder is not a declined hardware attempt: the
factory returns None without counting `declined`.
- HEVC tests skip only when the probe says there is no HEVC encoder; on a
machine that has one, a configuration NVENC refuses fails the test.
- Wording: NVENC is FilmCraft's only HEVC encoder on Windows (Windows itself
ships others).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
One NVENC session, ring and NV12 path now serve H.264 and HEVC (Codec enum,
Profile::HevcMain). HEVC Main 8-bit 4:2:0 SDR BT.709, MP4/MOV hvc1, parameter
sets only in an hvcC built from the SPS the encoder wrote. The H.265 format is
available through a cached NVENC probe; choosing it is the opt-in, declines are
counted errors. ABI asserts regenerated from nvEncodeAPI.h 12.1 (no drift) and
extended for HEVC; tests, docs and roadmap updated.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
`Format::Hevc` is available only where a hardware HEVC session can be created, and
`hardware_encode::hevc_available` finds out by creating a small one on the first question. That
question comes from the first draw of Export mode's format list, on the UI thread, and the session
takes about 0.1 s (88-100 ms in seven of eight fresh processes on an Apple M1, 222 ms in the first):
the app stalls for that long the first time Export is opened.
`register()` now starts the probe on a thread of its own (`warm_hevc_probe`, once however many times
`register` is called). Whoever asks while it runs waits for the same `OnceLock` answer, no longer than
asking first; a thread that cannot start, or a probe that panics (caught), leaves the question to the
first caller as before. `hevc_probed()` says whether the answer exists, for diagnostics and tests.
Test: `tests/hevc_probe.rs` registers, never calls `hevc_available`, and waits for `hevc_probed()` (it
fails after 60 s without the call in `register`).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Merge origin/main into the NVENC branch. #183 already provides the
HardwareEncoding enum, ExportSettings.hardware_encoding, the
hardwareEncoding export param and the Export > Video > Hardware Encoding
checkbox, so the duplicates from this branch are dropped and NVENC uses
them. filmcraft-export is now an all-platform dependency of
filmcraft-platform (VideoToolbox and NVENC both register encoders).
Never-crash fixes in the NVENC backend:
- submit() rejects a driver input pitch narrower than a row (and unlocks
the buffer); fill_nv12 uses get_mut instead of slicing rows.
- composition offsets use saturating math and i32::try_from; frame-rate,
size and timestamp casts are checked; sizes above u16::MAX decline.
- Annex B scanning and NAL type checks use get()/first().
- A failed nvEncOpenEncodeSessionEx that returns a handle destroys it.
ffi.rs carries NVIDIA's MIT notice for nvEncodeAPI.h, listed in
ATTRIBUTION.md under a new "Third-party source code" section.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds `Format::Hevc` (Export > Format > H.265 (HEVC); `"format": "hevc"` in
`file.exportMedia`): HEVC Main, 8-bit 4:2:0, in MP4 (`hvc1`) or QuickTime, with AAC audio,
encoded by the macOS hardware encoder. There is no built-in H.265 encoder (one in pure Rust
is a project of its own, and x265 is GPL), so the format exists only where the OS has one.
- `videotoolbox_encode`: the session wrapper creates HEVC sessions too (`VtProfile::HevcMain`;
the profile also says which codec). Keyframes are the HEVC random access points (NAL types
16 to 21); the parameter sets are VPS / SPS / PPS; the `hvcC` record in the MP4 sample
entry is the one VideoToolbox writes (sample description extension atoms), so profile,
level and flags are the encoder's own. `VtEncoder::uses_hardware()` asks VideoToolbox
whether the session runs on a hardware encoder (a test checks H.264 and HEVC), so a
requirement dropped by accident could never become a silent software encode.
- `hardware_encode`: one adapter for both codecs. Choosing the format is the opt-in (the
`hardware_encoding` setting is for H.264). `hevc_available()` (one small hardware session,
created once) is registered with the new `filmcraft_export::register_format_probe`, so
`available(Format::Hevc)`, the format list of Export mode and `exportFormats` follow the
machine. `Format::has_builtin_encoder` says which formats need such a probe.
- `export`: `Format::is_h26x` where H.264 and H.265 behave alike (MP4 or QuickTime, AAC, size
estimate, summary line). Two-pass VBR is refused up front ("H.265 export has no two-pass
mode") and HDR sequences export as SDR, because there is no encoder to fall back to.
- Export mode: H.265 shows H.264's multiplexer, audio, metadata and bitrate controls, without
profile, level, two-pass or the hardware checkbox, and says "Encoder: Hardware".
Apple M1 (8 cores), 1080p25, a 9:29 timeline (camera clip, ProRes 4444 overlays, AAC,
loudness): hardware H.265 at 10 Mb/s 195 s, hardware H.264 at 20 Mb/s 204 s (CPU 1029 s and
1042 s; the same H.264 export took 193 s and 200 s on other runs). It is not faster: the codec
that compresses further is the more expensive one (raw encoders: 166 and 189 frames/s), and
the export is bound by the CPU compositor, not by the media engine.
Quality per bit, 30 s of the same timeline against a ProRes 422 HQ export (average PSNR over
the three planes, SSIM): 20 Mb/s 46.88 dB / 0.9902 (H.264) and 46.91 dB / 0.9903 (H.265);
10 Mb/s 44.96 / 0.9863 and 45.09 / 0.9863; 6 Mb/s 43.72 / 0.9833 and 44.31 / 0.9843; 4 Mb/s
42.91 / 0.9807 and 43.64 / 0.9825; 3 Mb/s 42.37 / 0.9786 and 43.05 / 0.9807. The same picture
at 10 Mb/s and above; at 3 to 6 Mb/s H.265 needs 22 to 31 % less bitrate for the same PSNR.
Both formats therefore keep the same default bitrate. Details in docs/performance.md (HW3).
Tests: tests/hardware_encode.rs gets the HEVC counterparts of the H.264 ones (what is taken
and declined, a round trip through our HEVC decoder with luma PSNR above 30 dB, keyframes, MP4
and QuickTime exports with AAC, ffmpeg / ffprobe reading `hevc` / `Main` / `hvc1` /
`yuv420p` / BT.709, exact sizes, hostile configurations, encoders dropped at any point, the
format list agreeing with the probe, sessions on the hardware encoder); settings_tests.rs
checks names, ids, defaults, the summary, two-pass refusal and the probe; export_ui.rs the
controls H.265 shows and hides.
Stacked on the hardware H.264 encoder (the first commit of this branch).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Export encodes H.264 on the CPU with the built-in encoder, which is most of the time
of an export (about two thirds of the busy CPU in a profile of a 20 s export). On
macOS the media engine can do it: a VideoToolbox compression session behind
`filmcraft_export::VideoEncoder`, registered by `filmcraft_platform::register()`.
- `ExportSettings::hardware_encoding` (`HardwareEncoding::Off` | `Auto`; Export > Video >
Hardware Encoding, `"hardwareEncoding": "off|auto"` in `file.exportMedia`, a boolean is
understood too), off by default: the built-in encoder's output is byte-identical on
every machine, a hardware encoder's is not.
- `videotoolbox_encode` (FFI, same containment as the decoder: SAFETY comments, no
panic across the callback, safe API) and `hardware_encode` (safe adapter and factory).
- The factory declines, so the built-in encoder is used, for other formats, MXF, two-pass
VBR, HDR, odd sizes, non-square pixels and any configuration VideoToolbox cannot create a
hardware session for. An encoder that fails mid-export stops it with an error: unlike a
decoder, it cannot hand a half-written stream to another one.
- No B-frames: on real 1080p footage the quality is the same with and without them
(+-0.3 dB at equal bitrate) and the bitrate lands closer to the target, and without
them every frame is presented at its decode time (no composition offsets, no edit list).
- `export::register_encoder` ignores a factory registered twice; `encoder_registered`.
Apple M1, 8 cores, single runs: a 9:29 timeline (camera clip, ProRes 4444 overlays with
alpha, AAC, loudness) exports in 311 s instead of 793 s (2.5x faster); its densest 134 s
in 84 s instead of 395 s (4.7x). The outputs compare at SSIM 0.990 / PSNR 46.5 dB. The encoder
alone runs 200 fps at 1080p (docs/performance.md).
Tests (tests/hardware_encode.rs): what is taken and declined, a round trip through our
software decoder, an export that decodes in our decoder and in ffmpeg, the fallback for
everything hardware declines, exact output sizes, an empty range, hostile configurations
and encoders dropped at any point; the engine's `hardwareEncoding` parameter.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
NVENC through nvEncodeAPI64.dll (API 12.1, FFI layout checked against the C header by generated
tests), registered as an Export encoder in front of the software one; Export > Hardware encoding
(off by default). Declines two-pass VBR, HDR, MXF, unsupported sizes; B-frames, IDR/GOP, VUI, avcC.
Tests: decode with our decoder (PSNR), pipeline export vs software, decline paths.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MMorBHwct72udS2LiC3ZHp
Extends the Windows backend (#141) to VP9 (profiles 0 and 2) and AV1 (main, 8/10-bit), 4:2:0.
- codecs::hw: FrameStreamInfo (vpcC / av1C): profile, bit depth, chroma, picture size and colour as
the software decoders report them (vp9_color / av1_color are now shared with them), random-access
samples, in-band format changes; StreamInfo lets HybridDecoder handle NAL and frame-coded streams.
- platform: media_foundation::stream (what each codec needs: DXVA profile, MFT request, input
format), MfDecoder generalised; VP9 frames / AV1 temporal units go in as they are; the MFT's
declared picture size (display aperture) is checked against the bitstream's.
- Declined up front: VP9 profiles 1/3, AV1 profiles 1/2, 4:2:2/4:4:4, mono, 12-bit, no codec
extension, no DXVA profile. A key frame of another size or a different sequence header falls back
to the software decoder.
- Tests: bit-exact parity (360p / 1080p / 2160p, seeks, resets, flush, hidden frames), forced
failure, damaged samples and av1C, format change fallback, WebM / Matroska through the media stack.
- Bench: AV1 fixtures with libaom when ffmpeg lacks SVT-AV1.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MMorBHwct72udS2LiC3ZHp
New backend in crates/platform (media_foundation/): a Direct3D-aware decoder MFT is driven at the
level of single access units (Annex B), given the process's D3D11 video device through an
IMFDXGIDeviceManager so it decodes with DXVA, and its NV12 / P010 textures are read back through a
staging texture into planar Yuv8 / Yuv16. H.264 Baseline/Main/High and HEVC Main / Main 10, 4:2:0.
- Hardware is verified: non-D3D-aware or async MFTs, streams the GPU's DXVA decoder does not list,
and system-memory output (Microsoft's MFT software fallback) are declined or fail the stream so
HybridDecoder continues with our decoder.
- mfplat.dll is loaded at run time so Windows N editions still start.
- register() / registered() / hardware_decoder_for() cover Windows; perf.stats reports the backend.
- Tests (Windows): bit-exact parity with the software decoders at 640x360, 1080p and 2160p
(resets, seeks, flush), forced mid-stream failure, damaged input, concurrency, repeated init.
- Bench: runs on Windows (ffmpeg on PATH, process/thread CPU clocks), HEVC Main 10 2160p row,
hardware counters per row.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MMorBHwct72udS2LiC3ZHp
- New crates/platform (filmcraft-platform, L5): OS media FFI only. The one
crate allowed `unsafe` (own [lints] with unsafe_code = "deny",
#[allow(unsafe_code)] only on the videotoolbox module, SAFETY comment on
every block, safe Result-returning public API). AGENTS.md §0.3 names the
exception; docs/adr/0001-platform-ffi.md records why and the rules.
- VideoToolbox decoder for avcC / hvcC, 8/10-bit 4:2:0 and 4:2:2: session from
the sample entry's parameter sets with hardware required, asynchronous
decode (2 in flight), biplanar CVPixelBuffer copied to planar Yuv8/Yuv16,
reorder buffer of the stream's own depth, RASL skipped after a starting CRA.
Output callback under catch_unwind.
- HybridDecoder: transparent mid-stream fallback to the software decoder
(replays samples since the last restart point, drops already-returned
pictures). New codecs::software_video_decoder (built-in factories only) and
codecs::hw (switch, counters, NalStreamInfo shared with software decoders;
vui_color / sar_par helpers now used by both).
- Settings ▸ Playback ▸ Hardware decoding: Auto / Off (engine applies it to
the codecs switch); desktop app, CLI/MCP and bench call register().
- perf.stats decode.framesDecoded and decode.hardware counters.
- Tests: bit-exact parity with the software decoders (H.264 High, HEVC Main
open GOP, Main 10; resets, reseeks, flushes), forced-failure fallback,
mutation fuzz of samples and parameter sets, cross-platform fallback with a
stand-in decoder, setting Off, HEVC 4:2:2 vs ffmpeg.
- Bench --hw auto|off; results in docs/performance.md (4K CPU/frame
~86-119 ms -> ~4 ms).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Marco Arnulfo