[FFmpeg-cvslog] avcodec/mpegvideoenc: Move tmp bframes to MPVMainEncContext
Andreas Rheinhardt
git at videolan.org
Wed Mar 26 06:08:36 EET 2025
ffmpeg | branch: master | Andreas Rheinhardt <andreas.rheinhardt at outlook.com> | Fri Jan 28 20:24:18 2022 +0100| [bb0a72b10087e1fa5c75a868514c507cebeff3d9] | committer: Andreas Rheinhardt
avcodec/mpegvideoenc: Move tmp bframes to MPVMainEncContext
Also move b_frame_strategy and b_sensitivity and brd_scale;
they are all only used by an encoder's main thread when
b_frame_strategy == 2.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt at outlook.com>
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=bb0a72b10087e1fa5c75a868514c507cebeff3d9
---
libavcodec/mpegvideo.h | 8 ----
libavcodec/mpegvideo_enc.c | 114 ++++++++++++++++++++++++---------------------
libavcodec/mpegvideoenc.h | 15 ++++--
3 files changed, 73 insertions(+), 64 deletions(-)
diff --git a/libavcodec/mpegvideo.h b/libavcodec/mpegvideo.h
index afd18591c4..1091c93d95 100644
--- a/libavcodec/mpegvideo.h
+++ b/libavcodec/mpegvideo.h
@@ -48,8 +48,6 @@
#define MAX_THREADS 32
-#define MAX_B_FRAMES 16
-
/**
* Scantable.
*/
@@ -446,7 +444,6 @@ typedef struct MpegEncContext {
int top_field_first;
int concealment_motion_vectors;
int q_scale_type;
- int brd_scale;
int intra_vlc_format;
int alternate_scan;
int repeat_first_field;
@@ -518,11 +515,6 @@ typedef struct MpegEncContext {
int error_rate;
- /* temporary frames used by b_frame_strategy = 2 */
- AVFrame *tmp_frames[MAX_B_FRAMES + 2];
- int b_frame_strategy;
- int b_sensitivity;
-
/* frame skip options for encoding */
int frame_skip_threshold;
int frame_skip_factor;
diff --git a/libavcodec/mpegvideo_enc.c b/libavcodec/mpegvideo_enc.c
index 720ded49c2..ffa9484669 100644
--- a/libavcodec/mpegvideo_enc.c
+++ b/libavcodec/mpegvideo_enc.c
@@ -419,7 +419,8 @@ static av_cold int init_matrices(MpegEncContext *s, AVCodecContext *avctx)
/* init video encoder */
av_cold int ff_mpv_encode_init(AVCodecContext *avctx)
{
- MpegEncContext *s = avctx->priv_data;
+ MPVMainEncContext *const m = avctx->priv_data;
+ MpegEncContext *const s = &m->s;
AVCPBProperties *cpb_props;
int i, ret;
int mb_array_size, mv_table_size;
@@ -456,10 +457,10 @@ av_cold int ff_mpv_encode_init(AVCodecContext *avctx)
}
s->gop_size = avctx->gop_size;
s->avctx = avctx;
- if (avctx->max_b_frames > MAX_B_FRAMES) {
+ if (avctx->max_b_frames > MPVENC_MAX_B_FRAMES) {
av_log(avctx, AV_LOG_ERROR, "Too many B-frames requested, maximum "
- "is %d.\n", MAX_B_FRAMES);
- avctx->max_b_frames = MAX_B_FRAMES;
+ "is " AV_STRINGIFY(MPVENC_MAX_B_FRAMES) ".\n");
+ avctx->max_b_frames = MPVENC_MAX_B_FRAMES;
} else if (avctx->max_b_frames < 0) {
av_log(avctx, AV_LOG_ERROR,
"max b frames must be 0 or positive for mpegvideo based encoders\n");
@@ -700,10 +701,10 @@ av_cold int ff_mpv_encode_init(AVCodecContext *avctx)
return AVERROR(EINVAL);
}
- if (s->b_frame_strategy && (avctx->flags & AV_CODEC_FLAG_PASS2)) {
+ if (m->b_frame_strategy && (avctx->flags & AV_CODEC_FLAG_PASS2)) {
av_log(avctx, AV_LOG_INFO,
"notice: b_frame_strategy only affects the first pass\n");
- s->b_frame_strategy = 0;
+ m->b_frame_strategy = 0;
}
i = av_gcd(avctx->time_base.den, avctx->time_base.num);
@@ -914,8 +915,8 @@ av_cold int ff_mpv_encode_init(AVCodecContext *avctx)
return ret;
if (!(avctx->stats_out = av_mallocz(256)) ||
- !FF_ALLOCZ_TYPED_ARRAY(s->input_picture, MAX_B_FRAMES + 1) ||
- !FF_ALLOCZ_TYPED_ARRAY(s->reordered_input_picture, MAX_B_FRAMES + 1) ||
+ !FF_ALLOCZ_TYPED_ARRAY(s->input_picture, MPVENC_MAX_B_FRAMES + 1) ||
+ !FF_ALLOCZ_TYPED_ARRAY(s->reordered_input_picture, MPVENC_MAX_B_FRAMES + 1) ||
!(s->new_pic = av_frame_alloc()) ||
!(s->picture_pool = ff_mpv_alloc_pic_pool(0)))
return AVERROR(ENOMEM);
@@ -1010,17 +1011,17 @@ av_cold int ff_mpv_encode_init(AVCodecContext *avctx)
if ((ret = ff_rate_control_init(s)) < 0)
return ret;
- if (s->b_frame_strategy == 2) {
+ if (m->b_frame_strategy == 2) {
for (i = 0; i < s->max_b_frames + 2; i++) {
- s->tmp_frames[i] = av_frame_alloc();
- if (!s->tmp_frames[i])
+ m->tmp_frames[i] = av_frame_alloc();
+ if (!m->tmp_frames[i])
return AVERROR(ENOMEM);
- s->tmp_frames[i]->format = AV_PIX_FMT_YUV420P;
- s->tmp_frames[i]->width = s->width >> s->brd_scale;
- s->tmp_frames[i]->height = s->height >> s->brd_scale;
+ m->tmp_frames[i]->format = AV_PIX_FMT_YUV420P;
+ m->tmp_frames[i]->width = s->width >> m->brd_scale;
+ m->tmp_frames[i]->height = s->height >> m->brd_scale;
- ret = av_frame_get_buffer(s->tmp_frames[i], 0);
+ ret = av_frame_get_buffer(m->tmp_frames[i], 0);
if (ret < 0)
return ret;
}
@@ -1039,8 +1040,8 @@ av_cold int ff_mpv_encode_init(AVCodecContext *avctx)
av_cold int ff_mpv_encode_end(AVCodecContext *avctx)
{
- MpegEncContext *s = avctx->priv_data;
- int i;
+ MPVMainEncContext *const m = avctx->priv_data;
+ MpegEncContext *const s = &m->s;
ff_rate_control_uninit(&s->rc_context);
@@ -1048,13 +1049,13 @@ av_cold int ff_mpv_encode_end(AVCodecContext *avctx)
av_refstruct_pool_uninit(&s->picture_pool);
if (s->input_picture && s->reordered_input_picture) {
- for (int i = 0; i < MAX_B_FRAMES + 1; i++) {
+ for (int i = 0; i < MPVENC_MAX_B_FRAMES + 1; i++) {
av_refstruct_unref(&s->input_picture[i]);
av_refstruct_unref(&s->reordered_input_picture[i]);
}
}
- for (i = 0; i < FF_ARRAY_ELEMS(s->tmp_frames); i++)
- av_frame_free(&s->tmp_frames[i]);
+ for (int i = 0; i < FF_ARRAY_ELEMS(m->tmp_frames); i++)
+ av_frame_free(&m->tmp_frames[i]);
av_frame_free(&s->new_pic);
@@ -1249,8 +1250,9 @@ static int prepare_picture(MpegEncContext *s, AVFrame *f, const AVFrame *props_f
return 0;
}
-static int load_input_picture(MpegEncContext *s, const AVFrame *pic_arg)
+static int load_input_picture(MPVMainEncContext *const m, const AVFrame *pic_arg)
{
+ MpegEncContext *const s = &m->s;
MPVPicture *pic = NULL;
int64_t pts;
int display_picture_number = 0, ret;
@@ -1376,9 +1378,9 @@ static int load_input_picture(MpegEncContext *s, const AVFrame *pic_arg)
}
/* shift buffer entries */
- for (int i = flush_offset; i <= MAX_B_FRAMES; i++)
+ for (int i = flush_offset; i <= MPVENC_MAX_B_FRAMES; i++)
s->input_picture[i - flush_offset] = s->input_picture[i];
- for (int i = MAX_B_FRAMES + 1 - flush_offset; i <= MAX_B_FRAMES; i++)
+ for (int i = MPVENC_MAX_B_FRAMES + 1 - flush_offset; i <= MPVENC_MAX_B_FRAMES; i++)
s->input_picture[i] = NULL;
s->input_picture[encoding_delay] = pic;
@@ -1451,10 +1453,11 @@ static int encode_frame(AVCodecContext *c, const AVFrame *frame, AVPacket *pkt)
return size;
}
-static int estimate_best_b_count(MpegEncContext *s)
+static int estimate_best_b_count(MPVMainEncContext *const m)
{
+ MpegEncContext *const s = &m->s;
AVPacket *pkt;
- const int scale = s->brd_scale;
+ const int scale = m->brd_scale;
int width = s->width >> scale;
int height = s->height >> scale;
int i, j, out_size, p_lambda, b_lambda, lambda2;
@@ -1491,18 +1494,18 @@ static int estimate_best_b_count(MpegEncContext *s)
data[2] += INPLACE_OFFSET;
}
- s->mpvencdsp.shrink[scale](s->tmp_frames[i]->data[0],
- s->tmp_frames[i]->linesize[0],
+ s->mpvencdsp.shrink[scale](m->tmp_frames[i]->data[0],
+ m->tmp_frames[i]->linesize[0],
data[0],
pre_input_ptr->f->linesize[0],
width, height);
- s->mpvencdsp.shrink[scale](s->tmp_frames[i]->data[1],
- s->tmp_frames[i]->linesize[1],
+ s->mpvencdsp.shrink[scale](m->tmp_frames[i]->data[1],
+ m->tmp_frames[i]->linesize[1],
data[1],
pre_input_ptr->f->linesize[1],
width >> 1, height >> 1);
- s->mpvencdsp.shrink[scale](s->tmp_frames[i]->data[2],
- s->tmp_frames[i]->linesize[2],
+ s->mpvencdsp.shrink[scale](m->tmp_frames[i]->data[2],
+ m->tmp_frames[i]->linesize[2],
data[2],
pre_input_ptr->f->linesize[2],
width >> 1, height >> 1);
@@ -1539,10 +1542,10 @@ static int estimate_best_b_count(MpegEncContext *s)
goto fail;
- s->tmp_frames[0]->pict_type = AV_PICTURE_TYPE_I;
- s->tmp_frames[0]->quality = 1 * FF_QP2LAMBDA;
+ m->tmp_frames[0]->pict_type = AV_PICTURE_TYPE_I;
+ m->tmp_frames[0]->quality = 1 * FF_QP2LAMBDA;
- out_size = encode_frame(c, s->tmp_frames[0], pkt);
+ out_size = encode_frame(c, m->tmp_frames[0], pkt);
if (out_size < 0) {
ret = out_size;
goto fail;
@@ -1553,11 +1556,11 @@ static int estimate_best_b_count(MpegEncContext *s)
for (i = 0; i < s->max_b_frames + 1; i++) {
int is_p = i % (j + 1) == j || i == s->max_b_frames;
- s->tmp_frames[i + 1]->pict_type = is_p ?
+ m->tmp_frames[i + 1]->pict_type = is_p ?
AV_PICTURE_TYPE_P : AV_PICTURE_TYPE_B;
- s->tmp_frames[i + 1]->quality = is_p ? p_lambda : b_lambda;
+ m->tmp_frames[i + 1]->quality = is_p ? p_lambda : b_lambda;
- out_size = encode_frame(c, s->tmp_frames[i + 1], pkt);
+ out_size = encode_frame(c, m->tmp_frames[i + 1], pkt);
if (out_size < 0) {
ret = out_size;
goto fail;
@@ -1603,8 +1606,10 @@ fail:
* input_picture[0] is always NULL when exiting this function, even on error;
* reordered_input_picture[0] is always NULL when exiting this function on error.
*/
-static int set_bframe_chain_length(MpegEncContext *s)
+static int set_bframe_chain_length(MPVMainEncContext *const m)
{
+ MpegEncContext *const s = &m->s;
+
/* Either nothing to do or can't do anything */
if (s->reordered_input_picture[0] || !s->input_picture[0])
return 0;
@@ -1649,11 +1654,11 @@ static int set_bframe_chain_length(MpegEncContext *s)
}
}
- if (s->b_frame_strategy == 0) {
+ if (m->b_frame_strategy == 0) {
b_frames = s->max_b_frames;
while (b_frames && !s->input_picture[b_frames])
b_frames--;
- } else if (s->b_frame_strategy == 1) {
+ } else if (m->b_frame_strategy == 1) {
int i;
for (i = 1; i < s->max_b_frames + 1; i++) {
if (s->input_picture[i] &&
@@ -1668,7 +1673,7 @@ static int set_bframe_chain_length(MpegEncContext *s)
for (i = 0; i < s->max_b_frames + 1; i++) {
if (!s->input_picture[i] ||
s->input_picture[i]->b_frame_score - 1 >
- s->mb_num / s->b_sensitivity)
+ s->mb_num / m->b_sensitivity)
break;
}
@@ -1678,8 +1683,8 @@ static int set_bframe_chain_length(MpegEncContext *s)
for (i = 0; i < b_frames + 1; i++) {
s->input_picture[i]->b_frame_score = 0;
}
- } else if (s->b_frame_strategy == 2) {
- b_frames = estimate_best_b_count(s);
+ } else if (m->b_frame_strategy == 2) {
+ b_frames = estimate_best_b_count(m);
if (b_frames < 0) {
av_refstruct_unref(&s->input_picture[0]);
return b_frames;
@@ -1733,17 +1738,18 @@ static int set_bframe_chain_length(MpegEncContext *s)
return 0;
}
-static int select_input_picture(MpegEncContext *s)
+static int select_input_picture(MPVMainEncContext *const m)
{
+ MpegEncContext *const s = &m->s;
int ret;
av_assert1(!s->reordered_input_picture[0]);
- for (int i = 1; i <= MAX_B_FRAMES; i++)
+ for (int i = 1; i <= MPVENC_MAX_B_FRAMES; i++)
s->reordered_input_picture[i - 1] = s->reordered_input_picture[i];
- s->reordered_input_picture[MAX_B_FRAMES] = NULL;
+ s->reordered_input_picture[MPVENC_MAX_B_FRAMES] = NULL;
- ret = set_bframe_chain_length(s);
+ ret = set_bframe_chain_length(m);
av_assert1(!s->input_picture[0]);
if (ret < 0)
return ret;
@@ -1867,7 +1873,8 @@ static void frame_start(MpegEncContext *s)
int ff_mpv_encode_picture(AVCodecContext *avctx, AVPacket *pkt,
const AVFrame *pic_arg, int *got_packet)
{
- MpegEncContext *s = avctx->priv_data;
+ MPVMainEncContext *const m = avctx->priv_data;
+ MpegEncContext *const s = &m->s;
int stuffing_count, ret;
int context_count = s->slice_context_count;
@@ -1877,12 +1884,13 @@ int ff_mpv_encode_picture(AVCodecContext *avctx, AVPacket *pkt,
s->picture_in_gop_number++;
- if (load_input_picture(s, pic_arg) < 0)
- return -1;
+ ret = load_input_picture(m, pic_arg);
+ if (ret < 0)
+ return ret;
- if (select_input_picture(s) < 0) {
- return -1;
- }
+ ret = select_input_picture(m);
+ if (ret < 0)
+ return ret;
/* output? */
if (s->new_pic->data[0]) {
diff --git a/libavcodec/mpegvideoenc.h b/libavcodec/mpegvideoenc.h
index f7aad17650..7b5635e50e 100644
--- a/libavcodec/mpegvideoenc.h
+++ b/libavcodec/mpegvideoenc.h
@@ -33,8 +33,16 @@
#include "libavutil/opt.h"
#include "mpegvideo.h"
+#define MPVENC_MAX_B_FRAMES 16
+
typedef struct MPVMainEncContext {
MpegEncContext s; ///< The main slicecontext
+
+ /// temporary frames used by b_frame_strategy = 2
+ AVFrame *tmp_frames[MPVENC_MAX_B_FRAMES + 2];
+ int b_frame_strategy;
+ int b_sensitivity;
+ int brd_scale;
} MPVMainEncContext;
#define MAX_FCODE 7
@@ -85,6 +93,7 @@ typedef struct MPVMainEncContext {
{ "msad", "Sum of absolute differences, median predicted", 0, AV_OPT_TYPE_CONST, {.i64 = FF_CMP_MEDIAN_SAD }, INT_MIN, INT_MAX, FF_MPV_OPT_FLAGS, .unit = "cmp_func" }
#define FF_MPV_OFFSET(x) offsetof(MpegEncContext, x)
+#define FF_MPV_MAIN_OFFSET(x) offsetof(MPVMainEncContext, x)
#define FF_RC_OFFSET(x) offsetof(MpegEncContext, rc_context.x)
#define FF_MPV_OPT_FLAGS (AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM)
#define FF_MPV_COMMON_OPTS \
@@ -126,9 +135,9 @@ FF_MPV_OPT_CMP_FUNC, \
{"ps", "RTP payload size in bytes", FF_MPV_OFFSET(rtp_payload_size), AV_OPT_TYPE_INT, {.i64 = 0 }, INT_MIN, INT_MAX, FF_MPV_OPT_FLAGS }, \
#define FF_MPV_COMMON_BFRAME_OPTS \
-{"b_strategy", "Strategy to choose between I/P/B-frames", FF_MPV_OFFSET(b_frame_strategy), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 2, FF_MPV_OPT_FLAGS }, \
-{"b_sensitivity", "Adjust sensitivity of b_frame_strategy 1", FF_MPV_OFFSET(b_sensitivity), AV_OPT_TYPE_INT, {.i64 = 40 }, 1, INT_MAX, FF_MPV_OPT_FLAGS }, \
-{"brd_scale", "Downscale frames for dynamic B-frame decision", FF_MPV_OFFSET(brd_scale), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 3, FF_MPV_OPT_FLAGS },
+{"b_strategy", "Strategy to choose between I/P/B-frames", FF_MPV_MAIN_OFFSET(b_frame_strategy), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 2, FF_MPV_OPT_FLAGS }, \
+{"b_sensitivity", "Adjust sensitivity of b_frame_strategy 1", FF_MPV_MAIN_OFFSET(b_sensitivity), AV_OPT_TYPE_INT, {.i64 = 40 }, 1, INT_MAX, FF_MPV_OPT_FLAGS }, \
+{"brd_scale", "Downscale frames for dynamic B-frame decision", FF_MPV_MAIN_OFFSET(brd_scale), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 3, FF_MPV_OPT_FLAGS },
#define FF_MPV_COMMON_MOTION_EST_OPTS \
{"motion_est", "motion estimation algorithm", FF_MPV_OFFSET(motion_est), AV_OPT_TYPE_INT, {.i64 = FF_ME_EPZS }, FF_ME_ZERO, FF_ME_XONE, FF_MPV_OPT_FLAGS, .unit = "motion_est" }, \
More information about the ffmpeg-cvslog
mailing list