On 4/25/2025 7:55 PM, Andreas Rheinhardt wrote:
James Almer:On 4/25/2025 7:08 PM, Andreas Rheinhardt wrote:toq...@gmail.com:From: Wu Jianhua <toq...@outlook.com>Signed-off-by: Wu Jianhua <toq...@outlook.com> --- libavcodec/cbs_h2645.c | 6 ++++++ libavcodec/cbs_sei.h | 12 ++++++++++++ libavcodec/cbs_sei_syntax_template.c | 26 ++++++++++++++++++++++++++ 3 files changed, 44 insertions(+) diff --git a/libavcodec/cbs_h2645.c b/libavcodec/cbs_h2645.c index c868770209..6677134d77 100644 --- a/libavcodec/cbs_h2645.c +++ b/libavcodec/cbs_h2645.c @@ -2199,6 +2199,12 @@ static const SEIMessageTypeDescriptor cbs_sei_common_types[] = { sizeof(SEIRawDisplayOrientation), SEI_MESSAGE_RW(sei, display_orientation) }, + { + SEI_TYPE_FRAME_FIELD_INFO, + 1, 0, + sizeof(SEIRawFrameFieldInformation), + SEI_MESSAGE_RW(sei, frame_field_information) + }, SEI_MESSAGE_TYPE_END, }; diff --git a/libavcodec/cbs_sei.h b/libavcodec/cbs_sei.h index 7be12e33e0..81867b79a7 100644 --- a/libavcodec/cbs_sei.h +++ b/libavcodec/cbs_sei.h @@ -125,6 +125,18 @@ typedef struct SEIRawDisplayOrientation { uint8_t display_orientation_reserved_zero_3bits; } SEIRawDisplayOrientation; +typedef struct SEIRawFrameFieldInformation { + uint8_t ffi_field_pic_flag; + uint8_t ffi_bottom_field_flag; + uint8_t ffi_pairing_indicated_flag; + uint8_t ffi_paired_with_next_field_flag; + uint8_t ffi_display_fields_from_frame_flag; + uint8_t ffi_top_field_first_flag; + uint8_t ffi_display_elemental_periods_minus1; + uint8_t ffi_source_scan_type; + uint8_t ffi_duplicate_flag; +} SEIRawFrameFieldInformation; + typedef struct SEIRawMessage { uint32_t payload_type; uint32_t payload_size; diff --git a/libavcodec/cbs_sei_syntax_template.c b/libavcodec/ cbs_sei_syntax_template.c index fb19f6ee42..e6863a0fd7 100644 --- a/libavcodec/cbs_sei_syntax_template.c +++ b/libavcodec/cbs_sei_syntax_template.c @@ -295,6 +295,32 @@ SEI_FUNC(display_orientation, (CodedBitstreamContext *ctx, RWContext *rw, return 0; } +SEI_FUNC(frame_field_information, (CodedBitstreamContext *ctx, RWContext *rw, + SEIRawFrameFieldInformation *current, + SEIMessageState *state)) +{ + int err; + + HEADER("Frame-field information"); + + flag(ffi_field_pic_flag); + if (current->ffi_field_pic_flag) { + flag(ffi_bottom_field_flag); + flag(ffi_pairing_indicated_flag); + if (current->ffi_pairing_indicated_flag) + flag(ffi_paired_with_next_field_flag); + } else { + flag(ffi_display_fields_from_frame_flag); + if (current->ffi_display_fields_from_frame_flag) + flag(ffi_top_field_first_flag); + u(8, ffi_display_elemental_periods_minus1, 0, 0xff); + } + u(2, ffi_source_scan_type, 0, 3); + flag(ffi_duplicate_flag); + + return 0; +} + static int FUNC(message)(CodedBitstreamContext *ctx, RWContext *rw, SEIRawMessage *current) {This does really not seem like a common SEI, yet you add it to the common ones.It's in ITU-T H.274, which is what the common list contains. It's used as fallback if a SEI type is not a H.26x specific one.I know and this makes it wrong. It will e.g. be used for H.264 SEI type 168, although this SEI does not exist in H.264 (unless I am mistaken about the last point, but I don't think I am).
Then maybe ff_cbs_sei_find_type() should not look at cbs_sei_common_types for H.264 and H.265, given that those don't reference H.274, unlike H.266 which does (while not defining any SEI type at all).
OpenPGP_signature.asc
Description: OpenPGP digital signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
