libavcodec/libvpx: Add VPx alpha decode support

VPx (VP8/VP9) alpha encoding has been part of FFmpeg. Now, add the
ability to decode such files with alpha channel.

Signed-off-by: Vignesh Venkatasubramanian <vigneshv@google.com>

1 files changed, 87 insertions, 20 deletions

diff --git a/libavcodec/libvpxdec.c b/libavcodec/libvpxdec.c
index adbc6d0991..f7acc39258 100644
--- a/libavcodec/libvpxdec.c
+++ b/libavcodec/libvpxdec.c
@@ -29,6 +29,7 @@
 
 #include "libavutil/common.h"
 #include "libavutil/imgutils.h"
+#include "libavutil/intreadwrite.h"
 #include "avcodec.h"
 #include "internal.h"
 #include "libvpx.h"
@@ -36,10 +37,13 @@
 
 typedef struct VP8DecoderContext {
     struct vpx_codec_ctx decoder;
+    struct vpx_codec_ctx decoder_alpha;
+    int has_alpha_channel;
 } VP8Context;
 
 static av_cold int vpx_init(AVCodecContext *avctx,
-                            const struct vpx_codec_iface *iface)
+                            const struct vpx_codec_iface *iface,
+                            int is_alpha_decoder)
 {
     VP8Context *ctx = avctx->priv_data;
     struct vpx_codec_dec_cfg deccfg = {
@@ -50,7 +54,9 @@ static av_cold int vpx_init(AVCodecContext *avctx,
     av_log(avctx, AV_LOG_INFO, "%s\n", vpx_codec_version_str());
     av_log(avctx, AV_LOG_VERBOSE, "%s\n", vpx_codec_build_config());
 
-    if (vpx_codec_dec_init(&ctx->decoder, iface, &deccfg, 0) != VPX_CODEC_OK) {
+    if (vpx_codec_dec_init(
+            is_alpha_decoder ? &ctx->decoder_alpha : &ctx->decoder,
+            iface, &deccfg, 0) != VPX_CODEC_OK) {
         const char *error = vpx_codec_error(&ctx->decoder);
         av_log(avctx, AV_LOG_ERROR, "Failed to initialize decoder: %s\n",
                error);
@@ -61,7 +67,8 @@ static av_cold int vpx_init(AVCodecContext *avctx,
 }
 
 // returns 0 on success, AVERROR_INVALIDDATA otherwise
-static int set_pix_fmt(AVCodecContext *avctx, struct vpx_image *img)
+static int set_pix_fmt(AVCodecContext *avctx, struct vpx_image *img,
+                       int has_alpha_channel)
 {
 #if VPX_IMAGE_ABI_VERSION >= 3
     static const enum AVColorSpace colorspaces[8] = {
@@ -82,7 +89,8 @@ static int set_pix_fmt(AVCodecContext *avctx, struct vpx_image *img)
     case VPX_IMG_FMT_I420:
         if (avctx->codec_id == AV_CODEC_ID_VP9)
             avctx->profile = FF_PROFILE_VP9_0;
-        avctx->pix_fmt = AV_PIX_FMT_YUV420P;
+        avctx->pix_fmt =
+            has_alpha_channel ? AV_PIX_FMT_YUVA420P : AV_PIX_FMT_YUV420P;
         return 0;
 #if CONFIG_LIBVPX_VP9_DECODER
     case VPX_IMG_FMT_I422:
@@ -168,29 +176,75 @@ static int set_pix_fmt(AVCodecContext *avctx, struct vpx_image *img)
     }
 }
 
+static int decode_frame(AVCodecContext *avctx, vpx_codec_ctx_t *decoder,
+                        uint8_t *data, uint32_t data_sz)
+{
+    if (vpx_codec_decode(decoder, data, data_sz, NULL, 0) != VPX_CODEC_OK) {
+        const char *error  = vpx_codec_error(decoder);
+        const char *detail = vpx_codec_error_detail(decoder);
+
+        av_log(avctx, AV_LOG_ERROR, "Failed to decode frame: %s\n", error);
+        if (detail) {
+            av_log(avctx, AV_LOG_ERROR, "  Additional information: %s\n",
+                   detail);
+        }
+        return AVERROR_INVALIDDATA;
+    }
+    return 0;
+}
+
 static int vp8_decode(AVCodecContext *avctx,
                       void *data, int *got_frame, AVPacket *avpkt)
 {
     VP8Context *ctx = avctx->priv_data;
     AVFrame *picture = data;
     const void *iter = NULL;
-    struct vpx_image *img;
+    const void *iter_alpha = NULL;
+    struct vpx_image *img, *img_alpha;
     int ret;
+    uint8_t *side_data = NULL;
+    int side_data_size = 0;
 
-    if (vpx_codec_decode(&ctx->decoder, avpkt->data, avpkt->size, NULL, 0) !=
-        VPX_CODEC_OK) {
-        const char *error  = vpx_codec_error(&ctx->decoder);
-        const char *detail = vpx_codec_error_detail(&ctx->decoder);
+    ret = decode_frame(avctx, &ctx->decoder, avpkt->data, avpkt->size);
+    if (ret)
+        return ret;
 
-        av_log(avctx, AV_LOG_ERROR, "Failed to decode frame: %s\n", error);
-        if (detail)
-            av_log(avctx, AV_LOG_ERROR, "  Additional information: %s\n",
-                   detail);
-        return AVERROR_INVALIDDATA;
+    side_data = av_packet_get_side_data(avpkt,
+                                        AV_PKT_DATA_MATROSKA_BLOCKADDITIONAL,
+                                        &side_data_size);
+    if (side_data_size > 1) {
+        const uint64_t additional_id = AV_RB64(side_data);
+        side_data += 8;
+        side_data_size -= 8;
+        if (additional_id == 1) {  // 1 stands for alpha channel data.
+            if (!ctx->has_alpha_channel) {
+                ctx->has_alpha_channel = 1;
+                ret = vpx_init(avctx,
+#if CONFIG_LIBVPX_VP8_DECODER && CONFIG_LIBVPX_VP9_DECODER
+                               (avctx->codec_id == AV_CODEC_ID_VP8) ?
+                               &vpx_codec_vp8_dx_algo : &vpx_codec_vp9_dx_algo,
+#elif CONFIG_LIBVPX_VP8_DECODER
+                               &vpx_codec_vp8_dx_algo,
+#else
+                               &vpx_codec_vp9_dx_algo,
+#endif
+                               1);
+                if (ret)
+                    return ret;
+            }
+            ret = decode_frame(avctx, &ctx->decoder_alpha, side_data,
+                               side_data_size);
+            if (ret)
+                return ret;
+        }
     }
 
-    if ((img = vpx_codec_get_frame(&ctx->decoder, &iter))) {
-        if ((ret = set_pix_fmt(avctx, img)) < 0) {
+    if ((img = vpx_codec_get_frame(&ctx->decoder, &iter)) &&
+        (!ctx->has_alpha_channel ||
+         (img_alpha = vpx_codec_get_frame(&ctx->decoder_alpha, &iter_alpha)))) {
+        uint8_t *planes[4];
+        int linesizes[4];
+        if ((ret = set_pix_fmt(avctx, img, ctx->has_alpha_channel)) < 0) {
 #ifdef VPX_IMG_FMT_HIGHBITDEPTH
             av_log(avctx, AV_LOG_ERROR, "Unsupported output colorspace (%d) / bit_depth (%d)\n",
                    img->fmt, img->bit_depth);
@@ -210,8 +264,19 @@ static int vp8_decode(AVCodecContext *avctx,
         }
         if ((ret = ff_get_buffer(avctx, picture, 0)) < 0)
             return ret;
-        av_image_copy(picture->data, picture->linesize, (const uint8_t **)img->planes,
-                      img->stride, avctx->pix_fmt, img->d_w, img->d_h);
+
+        planes[0] = img->planes[VPX_PLANE_Y];
+        planes[1] = img->planes[VPX_PLANE_U];
+        planes[2] = img->planes[VPX_PLANE_V];
+        planes[3] =
+            ctx->has_alpha_channel ? img_alpha->planes[VPX_PLANE_Y] : NULL;
+        linesizes[0] = img->stride[VPX_PLANE_Y];
+        linesizes[1] = img->stride[VPX_PLANE_U];
+        linesizes[2] = img->stride[VPX_PLANE_V];
+        linesizes[3] =
+            ctx->has_alpha_channel ? img_alpha->stride[VPX_PLANE_Y] : 0;
+        av_image_copy(picture->data, picture->linesize, (const uint8_t**)planes,
+                      linesizes, avctx->pix_fmt, img->d_w, img->d_h);
         *got_frame           = 1;
     }
     return avpkt->size;
@@ -221,13 +286,15 @@ static av_cold int vp8_free(AVCodecContext *avctx)
 {
     VP8Context *ctx = avctx->priv_data;
     vpx_codec_destroy(&ctx->decoder);
+    if (ctx->has_alpha_channel)
+        vpx_codec_destroy(&ctx->decoder_alpha);
     return 0;
 }
 
 #if CONFIG_LIBVPX_VP8_DECODER
 static av_cold int vp8_init(AVCodecContext *avctx)
 {
-    return vpx_init(avctx, &vpx_codec_vp8_dx_algo);
+    return vpx_init(avctx, &vpx_codec_vp8_dx_algo, 0);
 }
 
 AVCodec ff_libvpx_vp8_decoder = {
@@ -246,7 +313,7 @@ AVCodec ff_libvpx_vp8_decoder = {
 #if CONFIG_LIBVPX_VP9_DECODER
 static av_cold int vp9_init(AVCodecContext *avctx)
 {
-    return vpx_init(avctx, &vpx_codec_vp9_dx_algo);
+    return vpx_init(avctx, &vpx_codec_vp9_dx_algo, 0);
 }
 
 AVCodec ff_libvpx_vp9_decoder = {