diff mbox series

[FFmpeg-devel,v4] avcodec/libvpxenc: add a way to set VP9E_SET_SVC_REF_FRAME_CONFIG

Message ID 20210108233329.363277-1-wonkap@google.com
State Accepted
Headers show
Series [FFmpeg-devel,v4] avcodec/libvpxenc: add a way to set VP9E_SET_SVC_REF_FRAME_CONFIG | expand

Checks

Context Check Description
andriy/x86_make success Make finished
andriy/x86_make_fate success Make fate finished
andriy/PPC64_make success Make finished
andriy/PPC64_make_fate success Make fate finished

Commit Message

Wonkap Jang Jan. 8, 2021, 11:33 p.m. UTC 
In order to fine-control referencing schemes in VP9 encoding, there
is a need to use VP9E_SET_SVC_REF_FRAME_CONFIG method. This commit
provides a way to use the API through frame metadata.
---
 doc/encoders.texi      | 32 ++++++++++++++++
 libavcodec/libvpxenc.c | 84 ++++++++++++++++++++++++++++++++++++++++++
 libavcodec/version.h   |  2 +-
 3 files changed, 117 insertions(+), 1 deletion(-)

Comments

Wonkap Jang Jan. 8, 2021, 11:35 p.m. UTC | #1 
On Fri, Jan 8, 2021 at 3:33 PM Wonkap Jang <wonkap@google.com> wrote:

> In order to fine-control referencing schemes in VP9 encoding, there
> is a need to use VP9E_SET_SVC_REF_FRAME_CONFIG method. This commit
> provides a way to use the API through frame metadata.
> ---
>  doc/encoders.texi      | 32 ++++++++++++++++
>  libavcodec/libvpxenc.c | 84 ++++++++++++++++++++++++++++++++++++++++++
>  libavcodec/version.h   |  2 +-
>  3 files changed, 117 insertions(+), 1 deletion(-)
>
> diff --git a/doc/encoders.texi b/doc/encoders.texi
> index 0b1c69e982..ef85fbb062 100644
> --- a/doc/encoders.texi
> +++ b/doc/encoders.texi
> @@ -2129,6 +2129,38 @@ midpoint is passed in rather than calculated for a
> specific clip or chunk.
>  The valid range is [0, 10000]. 0 (default) uses standard VBR.
>  @item enable-tpl @var{boolean}
>  Enable temporal dependency model.
> +@item ref-frame-config
> +Using per-frame metadata, set members of the structure
> @code{vpx_svc_ref_frame_config_t} in @code{vpx/vp8cx.h} to fine-control
> referencing schemes and frame buffer management.
> +@*Use a :-separated list of key=value pairs.
> +For example,
> +@example
> +av_dict_set(&av_frame->metadata, "ref-frame-config", \
>
> +"rfc_update_buffer_slot=7:rfc_lst_fb_idx=0:rfc_gld_fb_idx=1:rfc_alt_fb_idx=2:rfc_reference_last=0:rfc_reference_golden=0:rfc_reference_alt_ref=0");
> +@end example
> +@table @option
> +@item rfc_update_buffer_slot
> +Indicates the buffer slot number to update
> +@item rfc_update_last
> +Indicates whether to update the LAST frame
> +@item rfc_update_golden
> +Indicates whether to update GOLDEN frame
> +@item rfc_update_alt_ref
> +Indicates whether to update ALT_REF frame
> +@item rfc_lst_fb_idx
> +LAST frame buffer index
> +@item rfc_gld_fb_idx
> +GOLDEN frame buffer index
> +@item rfc_alt_fb_idx
> +ALT_REF frame buffer index
> +@item rfc_reference_last
> +Indicates whether to reference LAST frame
> +@item rfc_reference_golden
> +Indicates whether to reference GOLDEN frame
> +@item rfc_reference_alt_ref
> +Indicates whether to reference ALT_REF frame
> +@item rfc_reference_duration
> +Indicates frame duration
> +@end table
>  @end table
>
>  @end table
> diff --git a/libavcodec/libvpxenc.c b/libavcodec/libvpxenc.c
> index a7c76eb835..80e677dbbc 100644
> --- a/libavcodec/libvpxenc.c
> +++ b/libavcodec/libvpxenc.c
> @@ -125,6 +125,10 @@ typedef struct VPxEncoderContext {
>       * encounter a frame with ROI side data.
>       */
>      int roi_warned;
> +#if CONFIG_LIBVPX_VP9_ENCODER &&
> defined(VPX_CTRL_VP9E_SET_MAX_INTER_BITRATE_PCT)
> +    vpx_svc_ref_frame_config_t ref_frame_config;
> +    AVDictionary *vpx_ref_frame_config;
> +#endif
>  } VPxContext;
>
>  /** String mappings for enum vp8e_enc_control_id */
> @@ -152,6 +156,7 @@ static const char *const ctlidstr[] = {
>      [VP9E_SET_SVC_LAYER_ID]            = "VP9E_SET_SVC_LAYER_ID",
>  #if VPX_ENCODER_ABI_VERSION >= 12
>      [VP9E_SET_SVC_PARAMETERS]          = "VP9E_SET_SVC_PARAMETERS",
> +    [VP9E_SET_SVC_REF_FRAME_CONFIG]    = "VP9E_SET_SVC_REF_FRAME_CONFIG",
>  #endif
>      [VP9E_SET_SVC]                     = "VP9E_SET_SVC",
>  #if VPX_ENCODER_ABI_VERSION >= 11
> @@ -394,6 +399,21 @@ static void vp8_ts_parse_int_array(int *dest, char
> *value, size_t value_len, int
>      }
>  }
>
> +#if CONFIG_LIBVPX_VP9_ENCODER &&
> defined(VPX_CTRL_VP9E_SET_MAX_INTER_BITRATE_PCT)
> +static void vp8_ts_parse_int64_array(int64_t *dest, char *value, size_t
> value_len, int max_entries)
> +{
> +    int dest_idx = 0;
> +    char *saveptr = NULL;
> +    char *token = av_strtok(value, ",", &saveptr);
> +
> +    while (token && dest_idx < max_entries) {
> +        dest[dest_idx++] = strtoull(token, NULL, 10);
> +        token = av_strtok(NULL, ",", &saveptr);
> +    }
> +}
> +
> +#endif
> +
>  static void set_temporal_layer_pattern(int layering_mode,
> vpx_codec_enc_cfg_t *cfg,
>                                         int *layer_flags, int
> *flag_periodicity)
>  {
> @@ -541,6 +561,48 @@ static int vpx_ts_param_parse(VPxContext *ctx, struct
> vpx_codec_enc_cfg *enccfg,
>      return 0;
>  }
>
> +#if CONFIG_LIBVPX_VP9_ENCODER &&
> defined(VPX_CTRL_VP9E_SET_MAX_INTER_BITRATE_PCT)
> +static int vpx_ref_frame_config_parse(VPxContext *ctx, const struct
> vpx_codec_enc_cfg *enccfg,
> +                                      char *key, char *value, enum
> AVCodecID codec_id)
> +{
> +    size_t value_len = strlen(value);
> +    int ss_number_layers = enccfg->ss_number_layers;
> +    vpx_svc_ref_frame_config_t *ref_frame_config = &ctx->ref_frame_config;
> +
> +    if (!value_len)
> +        return -1;
> +
> +    if (codec_id != AV_CODEC_ID_VP9)
> +        return -1;
> +
> +    if (!strcmp(key, "rfc_update_buffer_slot")) {
> +        vp8_ts_parse_int_array(ref_frame_config->update_buffer_slot,
> value, value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_update_last")) {
> +        vp8_ts_parse_int_array(ref_frame_config->update_last, value,
> value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_update_golden")) {
> +        vp8_ts_parse_int_array(ref_frame_config->update_golden, value,
> value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_update_alt_ref")) {
> +        vp8_ts_parse_int_array(ref_frame_config->update_alt_ref, value,
> value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_lst_fb_idx")) {
> +        vp8_ts_parse_int_array(ref_frame_config->lst_fb_idx, value,
> value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_gld_fb_idx")) {
> +        vp8_ts_parse_int_array(ref_frame_config->gld_fb_idx, value,
> value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_alt_fb_idx")) {
> +        vp8_ts_parse_int_array(ref_frame_config->alt_fb_idx, value,
> value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_reference_last")) {
> +        vp8_ts_parse_int_array(ref_frame_config->reference_last, value,
> value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_reference_golden")) {
> +        vp8_ts_parse_int_array(ref_frame_config->reference_golden, value,
> value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_reference_alt_ref")) {
> +        vp8_ts_parse_int_array(ref_frame_config->reference_alt_ref,
> value, value_len, ss_number_layers);
> +    } else if (!strcmp(key, "rfc_reference_duration")) {
> +        vp8_ts_parse_int64_array(ref_frame_config->duration, value,
> value_len, ss_number_layers);
> +    }
> +
> +    return 0;
> +}
> +#endif
> +
>  #if CONFIG_LIBVPX_VP9_ENCODER
>  static int set_pix_fmt(AVCodecContext *avctx, vpx_codec_caps_t codec_caps,
>                         struct vpx_codec_enc_cfg *enccfg,
> vpx_codec_flags_t *flags,
> @@ -1528,6 +1590,28 @@ static int vpx_encode(AVCodecContext *avctx,
> AVPacket *pkt,
>  #endif
>                  layer_id_valid = 1;
>              }
> +#if CONFIG_LIBVPX_VP9_ENCODER &&
> defined(VPX_CTRL_VP9E_SET_MAX_INTER_BITRATE_PCT)
> +            en = av_dict_get(frame->metadata, "ref-frame-config", NULL,
> 0);
> +
> +            if (en) {
> +                if (avctx->codec_id == AV_CODEC_ID_VP9) {
> +                    AVDictionaryEntry* en2 = NULL;
> +                    av_dict_parse_string(&ctx->vpx_ref_frame_config,
> en->value, "=", ":", 0);
> +
> +                    while ((en2 = av_dict_get(ctx->vpx_ref_frame_config,
> "", en2, AV_DICT_IGNORE_SUFFIX))) {
> +                        if (vpx_ref_frame_config_parse(ctx, enccfg,
> en2->key, en2->value, avctx->codec_id) < 0)
> +                            av_log(avctx, AV_LOG_WARNING,
> +                                   "Error parsing option '%s = %s'.\n",
> +                                   en2->key, en2->value);
> +                    }
> +
> +                    codecctl_intp(avctx, VP9E_SET_SVC_REF_FRAME_CONFIG,
> (int *)&ctx->ref_frame_config);
> +                } else {
> +                    av_log(avctx, AV_LOG_WARNING,
> +                           "Error using option ref-frame-config for a
> non-VP9 codec\n");
> +                }
> +            }
> +#endif
>          }
>
>          if (sd) {
> diff --git a/libavcodec/version.h b/libavcodec/version.h
> index 5b92afe60a..abfbd63ac8 100644
> --- a/libavcodec/version.h
> +++ b/libavcodec/version.h
> @@ -29,7 +29,7 @@
>
>  #define LIBAVCODEC_VERSION_MAJOR  58
>  #define LIBAVCODEC_VERSION_MINOR 115
> -#define LIBAVCODEC_VERSION_MICRO 102
> +#define LIBAVCODEC_VERSION_MICRO 103
>
>  #define LIBAVCODEC_VERSION_INT  AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
>                                                 LIBAVCODEC_VERSION_MINOR, \
> --
> 2.30.0.284.gd98b1dd5eaa7-goog
>
>
Jame,

Thanks for checking. Sorry for the mistake in the last patch.

Please take a look,

Thank you,

Wonkap
James Zern Jan. 12, 2021, 11:28 p.m. UTC | #2 
On Fri, Jan 8, 2021 at 3:33 PM Wonkap Jang
<wonkap-at-google.com@ffmpeg.org> wrote:
>
> In order to fine-control referencing schemes in VP9 encoding, there
> is a need to use VP9E_SET_SVC_REF_FRAME_CONFIG method. This commit
> provides a way to use the API through frame metadata.
> ---
>  doc/encoders.texi      | 32 ++++++++++++++++
>  libavcodec/libvpxenc.c | 84 ++++++++++++++++++++++++++++++++++++++++++
>  libavcodec/version.h   |  2 +-
>  3 files changed, 117 insertions(+), 1 deletion(-)
>

lgtm. I'll push this soon if there aren't any other comments.

> [...]
> +}
> +

I removed this extra whitespace locally.

> +#endif
> +
James Zern Jan. 14, 2021, 7:13 p.m. UTC | #3 
On Tue, Jan 12, 2021 at 3:28 PM James Zern <jzern@google.com> wrote:
>
> On Fri, Jan 8, 2021 at 3:33 PM Wonkap Jang
> <wonkap-at-google.com@ffmpeg.org> wrote:
> >
> > In order to fine-control referencing schemes in VP9 encoding, there
> > is a need to use VP9E_SET_SVC_REF_FRAME_CONFIG method. This commit
> > provides a way to use the API through frame metadata.
> > ---
> >  doc/encoders.texi      | 32 ++++++++++++++++
> >  libavcodec/libvpxenc.c | 84 ++++++++++++++++++++++++++++++++++++++++++
> >  libavcodec/version.h   |  2 +-
> >  3 files changed, 117 insertions(+), 1 deletion(-)
> >
>
> lgtm. I'll push this soon if there aren't any other comments.
>

pushed. thanks for the patch.

> > [...]
> > +}
> > +
>
> I removed this extra whitespace locally.
>
> > +#endif
> > +
diff mbox series

Patch

diff --git a/doc/encoders.texi b/doc/encoders.texi
index 0b1c69e982..ef85fbb062 100644
--- a/doc/encoders.texi
+++ b/doc/encoders.texi
@@ -2129,6 +2129,38 @@  midpoint is passed in rather than calculated for a specific clip or chunk.
 The valid range is [0, 10000]. 0 (default) uses standard VBR.
 @item enable-tpl @var{boolean}
 Enable temporal dependency model.
+@item ref-frame-config
+Using per-frame metadata, set members of the structure @code{vpx_svc_ref_frame_config_t} in @code{vpx/vp8cx.h} to fine-control referencing schemes and frame buffer management.
+@*Use a :-separated list of key=value pairs.
+For example,
+@example
+av_dict_set(&av_frame->metadata, "ref-frame-config", \
+"rfc_update_buffer_slot=7:rfc_lst_fb_idx=0:rfc_gld_fb_idx=1:rfc_alt_fb_idx=2:rfc_reference_last=0:rfc_reference_golden=0:rfc_reference_alt_ref=0");
+@end example
+@table @option
+@item rfc_update_buffer_slot
+Indicates the buffer slot number to update
+@item rfc_update_last
+Indicates whether to update the LAST frame
+@item rfc_update_golden
+Indicates whether to update GOLDEN frame
+@item rfc_update_alt_ref
+Indicates whether to update ALT_REF frame
+@item rfc_lst_fb_idx
+LAST frame buffer index
+@item rfc_gld_fb_idx
+GOLDEN frame buffer index
+@item rfc_alt_fb_idx
+ALT_REF frame buffer index
+@item rfc_reference_last
+Indicates whether to reference LAST frame
+@item rfc_reference_golden
+Indicates whether to reference GOLDEN frame
+@item rfc_reference_alt_ref
+Indicates whether to reference ALT_REF frame
+@item rfc_reference_duration
+Indicates frame duration
+@end table
 @end table
 
 @end table
diff --git a/libavcodec/libvpxenc.c b/libavcodec/libvpxenc.c
index a7c76eb835..80e677dbbc 100644
--- a/libavcodec/libvpxenc.c
+++ b/libavcodec/libvpxenc.c
@@ -125,6 +125,10 @@  typedef struct VPxEncoderContext {
      * encounter a frame with ROI side data.
      */
     int roi_warned;
+#if CONFIG_LIBVPX_VP9_ENCODER && defined(VPX_CTRL_VP9E_SET_MAX_INTER_BITRATE_PCT)
+    vpx_svc_ref_frame_config_t ref_frame_config;
+    AVDictionary *vpx_ref_frame_config;
+#endif
 } VPxContext;
 
 /** String mappings for enum vp8e_enc_control_id */
@@ -152,6 +156,7 @@  static const char *const ctlidstr[] = {
     [VP9E_SET_SVC_LAYER_ID]            = "VP9E_SET_SVC_LAYER_ID",
 #if VPX_ENCODER_ABI_VERSION >= 12
     [VP9E_SET_SVC_PARAMETERS]          = "VP9E_SET_SVC_PARAMETERS",
+    [VP9E_SET_SVC_REF_FRAME_CONFIG]    = "VP9E_SET_SVC_REF_FRAME_CONFIG",
 #endif
     [VP9E_SET_SVC]                     = "VP9E_SET_SVC",
 #if VPX_ENCODER_ABI_VERSION >= 11
@@ -394,6 +399,21 @@  static void vp8_ts_parse_int_array(int *dest, char *value, size_t value_len, int
     }
 }
 
+#if CONFIG_LIBVPX_VP9_ENCODER && defined(VPX_CTRL_VP9E_SET_MAX_INTER_BITRATE_PCT)
+static void vp8_ts_parse_int64_array(int64_t *dest, char *value, size_t value_len, int max_entries)
+{
+    int dest_idx = 0;
+    char *saveptr = NULL;
+    char *token = av_strtok(value, ",", &saveptr);
+
+    while (token && dest_idx < max_entries) {
+        dest[dest_idx++] = strtoull(token, NULL, 10);
+        token = av_strtok(NULL, ",", &saveptr);
+    }
+}
+
+#endif
+
 static void set_temporal_layer_pattern(int layering_mode, vpx_codec_enc_cfg_t *cfg,
                                        int *layer_flags, int *flag_periodicity)
 {
@@ -541,6 +561,48 @@  static int vpx_ts_param_parse(VPxContext *ctx, struct vpx_codec_enc_cfg *enccfg,
     return 0;
 }
 
+#if CONFIG_LIBVPX_VP9_ENCODER && defined(VPX_CTRL_VP9E_SET_MAX_INTER_BITRATE_PCT)
+static int vpx_ref_frame_config_parse(VPxContext *ctx, const struct vpx_codec_enc_cfg *enccfg,
+                                      char *key, char *value, enum AVCodecID codec_id)
+{
+    size_t value_len = strlen(value);
+    int ss_number_layers = enccfg->ss_number_layers;
+    vpx_svc_ref_frame_config_t *ref_frame_config = &ctx->ref_frame_config;
+
+    if (!value_len)
+        return -1;
+
+    if (codec_id != AV_CODEC_ID_VP9)
+        return -1;
+
+    if (!strcmp(key, "rfc_update_buffer_slot")) {
+        vp8_ts_parse_int_array(ref_frame_config->update_buffer_slot, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_update_last")) {
+        vp8_ts_parse_int_array(ref_frame_config->update_last, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_update_golden")) {
+        vp8_ts_parse_int_array(ref_frame_config->update_golden, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_update_alt_ref")) {
+        vp8_ts_parse_int_array(ref_frame_config->update_alt_ref, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_lst_fb_idx")) {
+        vp8_ts_parse_int_array(ref_frame_config->lst_fb_idx, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_gld_fb_idx")) {
+        vp8_ts_parse_int_array(ref_frame_config->gld_fb_idx, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_alt_fb_idx")) {
+        vp8_ts_parse_int_array(ref_frame_config->alt_fb_idx, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_reference_last")) {
+        vp8_ts_parse_int_array(ref_frame_config->reference_last, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_reference_golden")) {
+        vp8_ts_parse_int_array(ref_frame_config->reference_golden, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_reference_alt_ref")) {
+        vp8_ts_parse_int_array(ref_frame_config->reference_alt_ref, value, value_len, ss_number_layers);
+    } else if (!strcmp(key, "rfc_reference_duration")) {
+        vp8_ts_parse_int64_array(ref_frame_config->duration, value, value_len, ss_number_layers);
+    }
+
+    return 0;
+}
+#endif
+
 #if CONFIG_LIBVPX_VP9_ENCODER
 static int set_pix_fmt(AVCodecContext *avctx, vpx_codec_caps_t codec_caps,
                        struct vpx_codec_enc_cfg *enccfg, vpx_codec_flags_t *flags,
@@ -1528,6 +1590,28 @@  static int vpx_encode(AVCodecContext *avctx, AVPacket *pkt,
 #endif
                 layer_id_valid = 1;
             }
+#if CONFIG_LIBVPX_VP9_ENCODER && defined(VPX_CTRL_VP9E_SET_MAX_INTER_BITRATE_PCT)
+            en = av_dict_get(frame->metadata, "ref-frame-config", NULL, 0);
+
+            if (en) {
+                if (avctx->codec_id == AV_CODEC_ID_VP9) {
+                    AVDictionaryEntry* en2 = NULL;
+                    av_dict_parse_string(&ctx->vpx_ref_frame_config, en->value, "=", ":", 0);
+
+                    while ((en2 = av_dict_get(ctx->vpx_ref_frame_config, "", en2, AV_DICT_IGNORE_SUFFIX))) {
+                        if (vpx_ref_frame_config_parse(ctx, enccfg, en2->key, en2->value, avctx->codec_id) < 0)
+                            av_log(avctx, AV_LOG_WARNING,
+                                   "Error parsing option '%s = %s'.\n",
+                                   en2->key, en2->value);
+                    }
+
+                    codecctl_intp(avctx, VP9E_SET_SVC_REF_FRAME_CONFIG, (int *)&ctx->ref_frame_config);
+                } else {
+                    av_log(avctx, AV_LOG_WARNING,
+                           "Error using option ref-frame-config for a non-VP9 codec\n");
+                }
+            }
+#endif
         }
 
         if (sd) {
diff --git a/libavcodec/version.h b/libavcodec/version.h
index 5b92afe60a..abfbd63ac8 100644
--- a/libavcodec/version.h
+++ b/libavcodec/version.h
@@ -29,7 +29,7 @@ 
 
 #define LIBAVCODEC_VERSION_MAJOR  58
 #define LIBAVCODEC_VERSION_MINOR 115
-#define LIBAVCODEC_VERSION_MICRO 102
+#define LIBAVCODEC_VERSION_MICRO 103
 
 #define LIBAVCODEC_VERSION_INT  AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
                                                LIBAVCODEC_VERSION_MINOR, \