diff mbox series

[FFmpeg-devel] Allow enabling SVC in libaomenc

Message ID 20240624185059.16797-1-chun.m.chang@gmail.com
State New
Headers show
Series [FFmpeg-devel] Allow enabling SVC in libaomenc | expand

Checks

Context Check Description
yinshiyou/commit_msg_loongarch64 warning The first line of the commit message must start with a context terminated by a colon and a space, for example "lavu/opt: " or "doc: ".
andriy/commit_msg_x86 warning The first line of the commit message must start with a context terminated by a colon and a space, for example "lavu/opt: " or "doc: ".
andriy/make_x86 success Make finished
andriy/make_fate_x86 success Make fate finished

Commit Message

Chun-Min Chang June 24, 2024, 6:50 p.m. UTC
This patch updates libaomenc.c to accept parameters for SVC (Scalable
Video Coding) settings via the FFmpeg API `av_opt_set`. The SVC
configuration is applied based on the provided parameters. As libaom's
SVC functionality only operates with constant bitrate encoding [1],
these parameters will only take effect when the bitrate is set to
constant.

[1] https://aomedia.googlesource.com/aom/+/a7ef80c44bfb34b08254194b1ab72d4e93ff4b07/av1/encoder/svc_layercontext.h#115
---
 libavcodec/libaomenc.c | 75 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 75 insertions(+)

Comments

Anton Khirnov June 25, 2024, 8:45 a.m. UTC | #1
Quoting Chun-Min Chang (2024-06-24 20:50:59)
> +        const AVDictionaryEntry *en = NULL;
> +        while ((en = av_dict_iterate(ctx->svc_parameters, en))) {
> +            if (!strlen(en->value))
> +                return AVERROR(EINVAL);
> +
> +            if (!strcmp(en->key, "number_spatial_layers"))
> +                svc_params.number_spatial_layers = strtoul(en->value, NULL, 10);
> +            else if (!strcmp(en->key, "number_temporal_layers"))
> +                svc_params.number_temporal_layers = strtoul(en->value, NULL, 10);
> +            else if (!strcmp(en->key, "max_quantizers"))
> +                aom_svc_parse_int_array(svc_params.max_quantizers, en->value, AOM_MAX_LAYERS);
> +            else if (!strcmp(en->key, "min_quantizers"))
> +                aom_svc_parse_int_array(svc_params.min_quantizers, en->value, AOM_MAX_LAYERS);
> +            else if (!strcmp(en->key, "scaling_factor_num"))
> +                aom_svc_parse_int_array(svc_params.scaling_factor_num, en->value, AOM_MAX_SS_LAYERS);
> +            else if (!strcmp(en->key, "scaling_factor_den"))
> +                aom_svc_parse_int_array(svc_params.scaling_factor_den, en->value, AOM_MAX_SS_LAYERS);
> +            else if (!strcmp(en->key, "layer_target_bitrate"))
> +                aom_svc_parse_int_array(svc_params.layer_target_bitrate, en->value, AOM_MAX_LAYERS);
> +            else if (!strcmp(en->key, "framerate_factor"))
> +                aom_svc_parse_int_array(svc_params.framerate_factor, en->value, AOM_MAX_TS_LAYERS);

So you declare a new dict-type option, then parse this dict manually
instead of AVOptions doing this work for you? Why?
diff mbox series

Patch

diff --git a/libavcodec/libaomenc.c b/libavcodec/libaomenc.c
index dec74ebecd..a8602a6b56 100644
--- a/libavcodec/libaomenc.c
+++ b/libavcodec/libaomenc.c
@@ -30,6 +30,7 @@ 
 #include <aom/aomcx.h>
 
 #include "libavutil/avassert.h"
+#include "libavutil/avstring.h"
 #include "libavutil/base64.h"
 #include "libavutil/common.h"
 #include "libavutil/cpu.h"
@@ -137,6 +138,7 @@  typedef struct AOMEncoderContext {
     int enable_diff_wtd_comp;
     int enable_dist_wtd_comp;
     int enable_dual_filter;
+    AVDictionary *svc_parameters;
     AVDictionary *aom_params;
 } AOMContext;
 
@@ -201,6 +203,7 @@  static const char *const ctlidstr[] = {
     [AV1E_GET_TARGET_SEQ_LEVEL_IDX]     = "AV1E_GET_TARGET_SEQ_LEVEL_IDX",
 #endif
     [AV1_GET_NEW_FRAME_IMAGE]           = "AV1_GET_NEW_FRAME_IMAGE",
+    [AV1E_SET_SVC_PARAMS]               = "AV1E_SET_SVC_PARAMS",
 };
 
 static av_cold void log_encoder_error(AVCodecContext *avctx, const char *desc)
@@ -382,6 +385,31 @@  static av_cold int codecctl_imgp(AVCodecContext *avctx,
     return 0;
 }
 
+static av_cold int codecctl_svcp(AVCodecContext *avctx,
+#ifdef UENUM1BYTE
+                                 aome_enc_control_id id,
+#else
+                                 enum aome_enc_control_id id,
+#endif
+                                 aom_svc_params_t *svc_params)
+{
+    AOMContext *ctx = avctx->priv_data;
+    char buf[80];
+    int res;
+
+    snprintf(buf, sizeof(buf), "%s:", ctlidstr[id]);
+
+    res = aom_codec_control(&ctx->encoder, id, svc_params);
+    if (res != AOM_CODEC_OK) {
+        snprintf(buf, sizeof(buf), "Failed to get %s codec control",
+                 ctlidstr[id]);
+        log_encoder_error(avctx, buf);
+        return AVERROR(EINVAL);
+    }
+
+    return 0;
+}
+
 static av_cold int aom_free(AVCodecContext *avctx)
 {
     AOMContext *ctx = avctx->priv_data;
@@ -673,6 +701,18 @@  static int choose_tiling(AVCodecContext *avctx,
     return 0;
 }
 
+static void aom_svc_parse_int_array(int *dest, char *value, 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++] = strtoul(token, NULL, 10);
+        token = av_strtok(NULL, ",", &saveptr);
+    }
+}
+
 static av_cold int aom_init(AVCodecContext *avctx,
                             const struct aom_codec_iface *iface)
 {
@@ -968,6 +1008,40 @@  static av_cold int aom_init(AVCodecContext *avctx,
     if (ctx->enable_intrabc >= 0)
         codecctl_int(avctx, AV1E_SET_ENABLE_INTRABC, ctx->enable_intrabc);
 
+    if (enccfg.rc_end_usage == AOM_CBR) {
+        aom_svc_params_t svc_params = {};
+        svc_params.framerate_factor[0] = 1;
+        svc_params.number_spatial_layers = 1;
+        svc_params.number_temporal_layers = 1;
+
+        const AVDictionaryEntry *en = NULL;
+        while ((en = av_dict_iterate(ctx->svc_parameters, en))) {
+            if (!strlen(en->value))
+                return AVERROR(EINVAL);
+
+            if (!strcmp(en->key, "number_spatial_layers"))
+                svc_params.number_spatial_layers = strtoul(en->value, NULL, 10);
+            else if (!strcmp(en->key, "number_temporal_layers"))
+                svc_params.number_temporal_layers = strtoul(en->value, NULL, 10);
+            else if (!strcmp(en->key, "max_quantizers"))
+                aom_svc_parse_int_array(svc_params.max_quantizers, en->value, AOM_MAX_LAYERS);
+            else if (!strcmp(en->key, "min_quantizers"))
+                aom_svc_parse_int_array(svc_params.min_quantizers, en->value, AOM_MAX_LAYERS);
+            else if (!strcmp(en->key, "scaling_factor_num"))
+                aom_svc_parse_int_array(svc_params.scaling_factor_num, en->value, AOM_MAX_SS_LAYERS);
+            else if (!strcmp(en->key, "scaling_factor_den"))
+                aom_svc_parse_int_array(svc_params.scaling_factor_den, en->value, AOM_MAX_SS_LAYERS);
+            else if (!strcmp(en->key, "layer_target_bitrate"))
+                aom_svc_parse_int_array(svc_params.layer_target_bitrate, en->value, AOM_MAX_LAYERS);
+            else if (!strcmp(en->key, "framerate_factor"))
+                aom_svc_parse_int_array(svc_params.framerate_factor, en->value, AOM_MAX_TS_LAYERS);
+        }
+
+        res = codecctl_svcp(avctx, AV1E_SET_SVC_PARAMS, &svc_params);
+        if (res < 0)
+            return res;
+    }
+
 #if AOM_ENCODER_ABI_VERSION >= 23
     {
         const AVDictionaryEntry *en = NULL;
@@ -1517,6 +1591,7 @@  static const AVOption options[] = {
     { "enable-masked-comp",           "Enable masked compound",                            OFFSET(enable_masked_comp),           AV_OPT_TYPE_BOOL, {.i64 = -1}, -1, 1, VE},
     { "enable-interintra-comp",       "Enable interintra compound",                        OFFSET(enable_interintra_comp),       AV_OPT_TYPE_BOOL, {.i64 = -1}, -1, 1, VE},
     { "enable-smooth-interintra",     "Enable smooth interintra mode",                     OFFSET(enable_smooth_interintra),     AV_OPT_TYPE_BOOL, {.i64 = -1}, -1, 1, VE},
+    { "svc-parameters", "SVC configuration using a :-separated list of key=value parameters (only applied in CBR mode)", OFFSET(svc_parameters), AV_OPT_TYPE_DICT, { 0 }, 0, 0, VE},
 #if AOM_ENCODER_ABI_VERSION >= 23
     { "aom-params",                   "Set libaom options using a :-separated list of key=value pairs", OFFSET(aom_params), AV_OPT_TYPE_DICT, { 0 }, 0, 0, VE },
 #endif