avcodec: Add picture structure information to AVCodecParserContext.

Signed-off-by: Anton Khirnov <anton@khirnov.net>

1 files changed, 17 insertions, 0 deletions

diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index b48ba0cce8..21e049136a 100644
--- a/libavcodec/avcodec.h
+++ b/libavcodec/avcodec.h
@@ -3407,6 +3407,13 @@ int avcodec_decode_subtitle2(AVCodecContext *avctx, AVSubtitle *sub,
  * @{
  */
 
+enum AVPictureStructure {
+    AV_PICTURE_STRUCTURE_UNKNOWN,      //< unknown
+    AV_PICTURE_STRUCTURE_TOP_FIELD,    //< coded as top field
+    AV_PICTURE_STRUCTURE_BOTTOM_FIELD, //< coded as bottom field
+    AV_PICTURE_STRUCTURE_FRAME,        //< coded as frame
+};
+
 typedef struct AVCodecParserContext {
     void *priv_data;
     struct AVCodecParser *parser;
@@ -3542,6 +3549,16 @@ typedef struct AVCodecParserContext {
     int duration;
 
     enum AVFieldOrder field_order;
+
+    /**
+     * Indicate whether a picture is coded as a frame, top field or bottom field.
+     *
+     * For example, H.264 field_pic_flag equal to 0 corresponds to
+     * AV_PICTURE_STRUCTURE_FRAME. An H.264 picture with field_pic_flag
+     * equal to 1 and bottom_field_flag equal to 0 corresponds to
+     * AV_PICTURE_STRUCTURE_TOP_FIELD.
+     */
+    enum AVPictureStructure picture_structure;
 } AVCodecParserContext;
 
 typedef struct AVCodecParser {