Go to the documentation of this file.
24 #include <EbSvtAv1ErrorCodes.h>
25 #include <EbSvtAv1Enc.h>
65 #if FF_API_SVTAV1_OPTS
85 { EB_ErrorNone, 0,
"success" },
86 { EB_ErrorInsufficientResources,
AVERROR(ENOMEM),
"insufficient resources" },
87 { EB_ErrorUndefined,
AVERROR(EINVAL),
"undefined error" },
88 { EB_ErrorInvalidComponent,
AVERROR(EINVAL),
"invalid component" },
89 { EB_ErrorBadParameter,
AVERROR(EINVAL),
"bad parameter" },
90 { EB_ErrorDestroyThreadFailed,
AVERROR_EXTERNAL,
"failed to destroy thread" },
91 { EB_ErrorSemaphoreUnresponsive,
AVERROR_EXTERNAL,
"semaphore unresponsive" },
92 { EB_ErrorDestroySemaphoreFailed,
AVERROR_EXTERNAL,
"failed to destroy semaphore"},
96 { EB_NoErrorEmptyQueue,
AVERROR(EAGAIN),
"empty queue" },
110 *
desc =
"unknown error";
115 const char *error_string)
127 const int pack_mode_10bit =
128 (
config->encoder_bit_depth > 8) && (
config->compressed_ten_bit_format == 0) ? 1 : 0;
129 const size_t luma_size_8bit =
130 config->source_width *
config->source_height * (1 << pack_mode_10bit);
131 const size_t luma_size_10bit =
132 (
config->encoder_bit_depth > 8 && pack_mode_10bit == 0) ? luma_size_8bit : 0;
134 EbSvtIOFormat *in_data;
136 svt_enc->
raw_size = (luma_size_8bit + luma_size_10bit) * 3 / 2;
144 if (!svt_enc->
in_buf->p_buffer)
161 #if FF_API_SVTAV1_OPTS
163 param->tier = svt_enc->
tier;
164 param->scene_change_detection = svt_enc->
scd;
169 param->look_ahead_distance = svt_enc->
la_depth;
173 param->enc_mode = svt_enc->
enc_mode;
176 param->target_bit_rate = avctx->
bit_rate;
178 param->rate_control_mode = 1;
180 param->rate_control_mode = 2;
182 param->max_qp_allowed = avctx->
qmax;
183 param->min_qp_allowed = avctx->
qmin;
188 if (svt_enc->
crf > 0) {
189 param->qp = svt_enc->
crf;
190 param->rate_control_mode = 0;
191 }
else if (svt_enc->
qp > 0) {
192 param->qp = svt_enc->
qp;
193 param->rate_control_mode = 0;
194 param->enable_adaptive_quantization = 0;
201 param->transfer_characteristics = avctx->
color_trc;
208 #if SVT_AV1_CHECK_VERSION(1, 0, 0)
215 param->chroma_sample_position = EB_CSP_VERTICAL;
218 param->chroma_sample_position = EB_CSP_COLOCATED;
225 "Specified chroma sample location %s is unsupported "
226 "on the AV1 bit stream level. Usage of a container that "
227 "allows passing this information - such as Matroska - "
236 param->profile = avctx->
profile;
239 param->level = avctx->
level;
242 param->intra_period_length = avctx->
gop_size - 1;
255 #if SVT_AV1_CHECK_VERSION(0, 9, 1)
257 EbErrorType
ret = svt_av1_enc_parse_parameter(param, en->
key, en->
value);
258 if (
ret != EB_ErrorNone) {
268 av_log(avctx,
level,
"svt-params needs libavcodec to be compiled with SVT-AV1 "
269 "headers >= 0.9.1.\n");
275 param->source_width = avctx->
width;
276 param->source_height = avctx->
height;
278 param->encoder_bit_depth =
desc->comp[0].depth;
280 if (
desc->log2_chroma_w == 1 &&
desc->log2_chroma_h == 1)
281 param->encoder_color_format = EB_YUV420;
282 else if (
desc->log2_chroma_w == 1 &&
desc->log2_chroma_h == 0)
283 param->encoder_color_format = EB_YUV422;
284 else if (!
desc->log2_chroma_w && !
desc->log2_chroma_h)
285 param->encoder_color_format = EB_YUV444;
291 if ((param->encoder_color_format == EB_YUV422 || param->encoder_bit_depth > 10)
295 }
else if (param->encoder_color_format == EB_YUV444 && param->profile !=
FF_PROFILE_AV1_HIGH) {
300 avctx->
bit_rate = param->rate_control_mode > 0 ?
301 param->target_bit_rate : 0;
319 EbBufferHeaderType *header_ptr)
321 EbSvtIOFormat *in_data = (EbSvtIOFormat *)header_ptr->p_buffer;
322 ptrdiff_t linesizes[4];
324 int bytes_shift = param->encoder_bit_depth > 8 ? 1 : 0;
327 for (
int i = 0;
i < 4;
i++)
328 linesizes[
i] =
frame->linesize[
i];
336 for (
int i = 0;
i < 4;
i++) {
342 in_data->luma =
frame->data[0];
343 in_data->cb =
frame->data[1];
344 in_data->cr =
frame->data[2];
364 if (svt_ret != EB_ErrorNone) {
365 return svt_print_error(avctx, svt_ret,
"Error initializing encoder handle");
375 if (svt_ret != EB_ErrorNone) {
376 return svt_print_error(avctx, svt_ret,
"Error setting encoder parameters");
379 svt_ret = svt_av1_enc_init(svt_enc->
svt_handle);
380 if (svt_ret != EB_ErrorNone) {
385 EbBufferHeaderType *headerPtr =
NULL;
387 svt_ret = svt_av1_enc_stream_header(svt_enc->
svt_handle, &headerPtr);
388 if (svt_ret != EB_ErrorNone) {
389 return svt_print_error(avctx, svt_ret,
"Error building stream header");
396 "Cannot allocate AV1 header of size %d.\n", avctx->
extradata_size);
402 svt_ret = svt_av1_enc_stream_header_release(headerPtr);
403 if (svt_ret != EB_ErrorNone) {
418 EbBufferHeaderType *headerPtr = svt_enc->
in_buf;
422 EbBufferHeaderType headerPtrLast;
427 memset(&headerPtrLast, 0,
sizeof(headerPtrLast));
428 headerPtrLast.pic_type = EB_AV1_INVALID_PICTURE;
429 headerPtrLast.flags = EB_BUFFERFLAG_EOS;
431 svt_av1_enc_send_picture(svt_enc->
svt_handle, &headerPtrLast);
440 headerPtr->flags = 0;
441 headerPtr->p_app_private =
NULL;
442 headerPtr->pts =
frame->pts;
444 switch (
frame->pict_type) {
446 headerPtr->pic_type = EB_AV1_KEY_PICTURE;
450 headerPtr->pic_type = EB_AV1_INVALID_PICTURE;
454 svt_av1_enc_send_picture(svt_enc->
svt_handle, headerPtr);
462 const int max_frames = 8;
465 if (filled_len > svt_enc->
raw_size * max_frames) {
486 EbBufferHeaderType *headerPtr;
490 int ret = 0, pict_type;
507 if (svt_ret == EB_NoErrorEmptyQueue)
513 svt_av1_enc_release_out_buffer(&headerPtr);
519 memcpy(
pkt->
data, headerPtr->p_buffer, headerPtr->n_filled_len);
522 pkt->
size = headerPtr->n_filled_len;
523 pkt->
pts = headerPtr->pts;
524 pkt->
dts = headerPtr->dts;
526 switch (headerPtr->pic_type) {
527 case EB_AV1_KEY_PICTURE:
530 case EB_AV1_INTRA_ONLY_PICTURE:
533 case EB_AV1_INVALID_PICTURE:
541 if (headerPtr->pic_type == EB_AV1_NON_REF_PICTURE)
544 if (headerPtr->flags & EB_BUFFERFLAG_EOS)
549 svt_av1_enc_release_out_buffer(&headerPtr);
560 svt_av1_enc_deinit_handle(svt_enc->
svt_handle);
573 #define OFFSET(x) offsetof(SvtContext, x)
574 #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
576 #if FF_API_SVTAV1_OPTS
577 {
"hielevel",
"Hierarchical prediction levels setting (Deprecated, use svtav1-params)",
OFFSET(hierarchical_level),
582 {
"la_depth",
"Look ahead distance [0, 120] (Deprecated, use svtav1-params)",
OFFSET(la_depth),
585 {
"tier",
"Set operating point tier (Deprecated, use svtav1-params)",
OFFSET(tier),
590 {
"preset",
"Encoding preset",
595 #define LEVEL(name, value) name, NULL, 0, AV_OPT_TYPE_CONST, \
596 { .i64 = value }, 0, 0, VE, "avctx.level"
597 {
LEVEL(
"2.0", 20) },
598 {
LEVEL(
"2.1", 21) },
599 {
LEVEL(
"2.2", 22) },
600 {
LEVEL(
"2.3", 23) },
601 {
LEVEL(
"3.0", 30) },
602 {
LEVEL(
"3.1", 31) },
603 {
LEVEL(
"3.2", 32) },
604 {
LEVEL(
"3.3", 33) },
605 {
LEVEL(
"4.0", 40) },
606 {
LEVEL(
"4.1", 41) },
607 {
LEVEL(
"4.2", 42) },
608 {
LEVEL(
"4.3", 43) },
609 {
LEVEL(
"5.0", 50) },
610 {
LEVEL(
"5.1", 51) },
611 {
LEVEL(
"5.2", 52) },
612 {
LEVEL(
"5.3", 53) },
613 {
LEVEL(
"6.0", 60) },
614 {
LEVEL(
"6.1", 61) },
615 {
LEVEL(
"6.2", 62) },
616 {
LEVEL(
"6.3", 63) },
617 {
LEVEL(
"7.0", 70) },
618 {
LEVEL(
"7.1", 71) },
619 {
LEVEL(
"7.2", 72) },
620 {
LEVEL(
"7.3", 73) },
623 {
"crf",
"Constant Rate Factor value",
OFFSET(crf),
625 {
"qp",
"Initial Quantizer level value",
OFFSET(qp),
627 #if FF_API_SVTAV1_OPTS
628 {
"sc_detection",
"Scene change detection (Deprecated, use svtav1-params)",
OFFSET(scd),
635 {
"svtav1-params",
"Set the SVT-AV1 configuration using a :-separated list of key=value parameters",
OFFSET(svtav1_opts),
AV_OPT_TYPE_DICT, { 0 }, 0, 0,
VE },
649 {
"flags",
"+cgop" },
657 .
p.
name =
"libsvtav1",
670 .p.priv_class = &
class,
672 .p.wrapper_name =
"libsvtav1",
AVBufferPool * av_buffer_pool_init(size_t size, AVBufferRef *(*alloc)(size_t size))
Allocate and initialize a buffer pool.
#define AV_LOG_WARNING
Something somehow does not look correct.
AVPixelFormat
Pixel format.
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this just let it vf default minimum maximum flags name is the option name
#define FF_CODEC_CAP_INIT_CLEANUP
The codec allows calling the close function for deallocation even if the init function returned a fai...
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later. That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another. Frame references ownership and permissions
enum AVColorSpace colorspace
YUV colorspace type.
static const FFCodecDefault eb_enc_defaults[]
static AVBufferRef * get_output_ref(AVCodecContext *avctx, SvtContext *svt_enc, int filled_len)
static const struct @88 svt_errors[]
const AVPixFmtDescriptor * av_pix_fmt_desc_get(enum AVPixelFormat pix_fmt)
#define AVERROR_EOF
End of file.
int ff_side_data_set_encoder_stats(AVPacket *pkt, int quality, int64_t *error, int error_count, int pict_type)
int err_recognition
Error recognition; may misdetect some more or less valid parts as errors.
#define FF_AV1_PROFILE_OPTS
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
This structure describes decoded (raw) audio or video data.
enum AVColorTransferCharacteristic color_trc
Color Transfer Characteristic.
@ AVCOL_RANGE_JPEG
Full range content.
#define AV_PIX_FMT_YUV420P10
static int eb_receive_packet(AVCodecContext *avctx, AVPacket *pkt)
#define AV_DICT_IGNORE_SUFFIX
Return first entry in a dictionary whose first part corresponds to the search key,...
static av_cold int eb_enc_init(AVCodecContext *avctx)
@ AVCOL_SPC_RGB
order of coefficients is actually GBR, also IEC 61966-2-1 (sRGB), YZX and ST 428-1
static av_cold int eb_enc_close(AVCodecContext *avctx)
#define AV_PKT_FLAG_DISPOSABLE
Flag is used to indicate packets that contain frames that can be discarded by the decoder.
#define AVERROR_UNKNOWN
Unknown error, typically from an external library.
AVCPBProperties * ff_add_cpb_side_data(AVCodecContext *avctx)
Add a CPB properties side data to an encoding context.
int qmax
maximum quantizer
#define AV_PKT_FLAG_KEY
The packet contains a keyframe.
const char * av_chroma_location_name(enum AVChromaLocation location)
#define AV_CODEC_FLAG_GLOBAL_HEADER
Place global headers in extradata instead of every keyframe.
AVCodec p
The public AVCodec.
static int eb_send_frame(AVCodecContext *avctx, const AVFrame *frame)
int flags
AV_CODEC_FLAG_*.
AVFrame * av_frame_alloc(void)
Allocate an AVFrame and set its fields to default values.
enum AVColorPrimaries color_primaries
Chromaticity coordinates of the source primaries.
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
#define FF_ARRAY_ELEMS(a)
AVDictionaryEntry * av_dict_get(const AVDictionary *m, const char *key, const AVDictionaryEntry *prev, int flags)
Get a dictionary entry with matching key.
AVBufferRef * av_buffer_pool_get(AVBufferPool *pool)
Allocate a new AVBuffer, reusing an old buffer from the pool when available.
#define AV_CEIL_RSHIFT(a, b)
int ticks_per_frame
For some codecs, the time base is closer to the field rate than the frame rate.
#define AV_CODEC_CAP_OTHER_THREADS
Codec supports multithreading through a method other than slice- or frame-level multithreading.
#define FF_PROFILE_UNKNOWN
#define av_assert0(cond)
assert() equivalent, that is always enabled.
@ AV_PIX_FMT_YUV420P
planar YUV 4:2:0, 12bpp, (1 Cr & Cb sample per 2x2 Y samples)
int64_t rc_max_rate
maximum bitrate
This structure describes the bitrate properties of an encoded bitstream.
int rc_buffer_size
decoder bitstream buffer size
AVBufferRef * buf
A reference to the reference-counted buffer where the packet data is stored.
#define LIBAVUTIL_VERSION_INT
Describe the class of an AVClass context structure.
static const int sizes[][2]
enum AVColorRange color_range
MPEG vs JPEG YUV range.
@ AVCHROMA_LOC_LEFT
MPEG-2/4 4:2:0, H.264 default for 4:2:0.
int av_image_fill_plane_sizes(size_t sizes[4], enum AVPixelFormat pix_fmt, int height, const ptrdiff_t linesizes[4])
Fill plane sizes for an image with pixel format pix_fmt and height height.
@ AVCHROMA_LOC_TOPLEFT
ITU-R 601, SMPTE 274M 296M S314M(DV 4:1:1), mpeg2 4:2:2.
#define FF_CODEC_RECEIVE_PACKET_CB(func)
int64_t bit_rate
the average bitrate
const char * av_default_item_name(void *ptr)
Return the context name.
@ AV_PICTURE_TYPE_I
Intra.
void av_buffer_pool_uninit(AVBufferPool **ppool)
Mark the pool as being available for freeing.
#define AV_EF_EXPLODE
abort decoding on minor error detection
const FFCodec ff_libsvtav1_encoder
@ AVCOL_RANGE_UNSPECIFIED
#define LEVEL(name, value)
AVRational time_base
This is the fundamental unit of time (in seconds) in terms of which frame timestamps are represented.
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification.
int gop_size
the number of pictures in a group of pictures, or 0 for intra_only
#define AV_PIX_FMT_FLAG_RGB
The pixel format contains RGB-like data (as opposed to YUV/grayscale).
static int config_enc_params(EbSvtAv1EncConfiguration *param, AVCodecContext *avctx)
EbComponentType * svt_handle
AVDictionary * svtav1_opts
@ AVCHROMA_LOC_UNSPECIFIED
@ AV_PICTURE_TYPE_NONE
Undefined.
static int alloc_buffer(EbSvtAv1EncConfiguration *config, SvtContext *svt_enc)
int64_t dts
Decompression timestamp in AVStream->time_base units; the time at which the packet is decompressed.
static const AVOption options[]
#define AVERROR_EXTERNAL
Generic error in an external library.
int flags
A combination of AV_PKT_FLAG values.
static int svt_map_error(EbErrorType eb_err, const char **desc)
int64_t avg_bitrate
Average bitrate of the stream, in bits per second.
#define i(width, name, range_min, range_max)
int64_t pts
Presentation timestamp in AVStream->time_base units; the time at which the decompressed packet will b...
uint8_t * extradata
some codecs need / can use extradata like Huffman tables.
int64_t max_bitrate
Maximum bitrate of the stream, in bits per second.
void av_frame_unref(AVFrame *frame)
Unreference all the buffers referenced by frame and reset the frame fields.
void * av_mallocz(size_t size)
Allocate a memory block with alignment suitable for all memory accesses (including vectors if availab...
const char * name
Name of the codec implementation.
enum AVChromaLocation chroma_sample_location
This defines the location of chroma samples.
enum AVPixelFormat pix_fmt
Pixel format, see AV_PIX_FMT_xxx.
EbSvtAv1EncConfiguration enc_params
#define FF_PROFILE_AV1_PROFESSIONAL
#define AV_CODEC_FLAG_CLOSED_GOP
const char * class_name
The name of the class; usually it is the same name as the context structure type to which the AVClass...
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return or at least make progress towards producing a frame
int64_t buffer_size
The size of the buffer to which the ratecontrol is applied, in bits.
#define AV_INPUT_BUFFER_PADDING_SIZE
static int svt_print_error(void *log_ctx, EbErrorType err, const char *error_string)
main external API structure.
int qmin
minimum quantizer
static int ref[MAX_W *MAX_W]
#define AV_CODEC_CAP_DELAY
Encoder or decoder requires flushing with NULL input at the end in order to give the complete and cor...
EbBufferHeaderType * in_buf
@ AV_PICTURE_TYPE_P
Predicted.
int ff_encode_get_frame(AVCodecContext *avctx, AVFrame *frame)
Called by encoders to get the next frame for encoding.
A reference to a data buffer.
#define FF_CODEC_CAP_AUTO_THREADS
Codec handles avctx->thread_count == 0 (auto) internally.
Descriptor that unambiguously describes how the bits of a pixel are stored in the up to 4 data planes...
This structure stores compressed data.
int width
picture width / height.
#define FF_PROFILE_AV1_HIGH
#define FF_QP2LAMBDA
factor to convert from H.263 QP to lambda
static int read_in_data(EbSvtAv1EncConfiguration *param, const AVFrame *frame, EbBufferHeaderType *header_ptr)
#define AV_OPT_FLAG_DEPRECATED
set if option is deprecated, users should refer to AVOption.help text for more information