VSE: Fix audaspace not reading ffmpeg files with start offset correctly

The duration and start time for audio strips were not correctly read in
audaspace.

Some video files have a "lead in" section of audio that plays before the
video starts playing back. Before this patch, we would play this lead in
audio at the same time as the video started and thus the audio would not
be in sync anymore.

Now the lead in audio is cut off and the duration should be correctly
calculated with this in mind.

If the audio starts after the video, the audio strip is shifted to
account for this, but it will also lead to cut off audio which might not
be wanted. However we don't have a simple way to solve this at this
point.

Differential Revision: http://developer.blender.org/D11917

2 files changed, 22 insertions, 6 deletions

diff --git a/source/blender/imbuf/intern/IMB_anim.h b/source/blender/imbuf/intern/IMB_anim.h
index cfeffcca0ea..c4e2ad9da7f 100644
--- a/source/blender/imbuf/intern/IMB_anim.h
+++ b/source/blender/imbuf/intern/IMB_anim.h
@@ -91,6 +91,7 @@ struct anim {
   int duration_in_frames;
   int frs_sec;
   double frs_sec_base;
+  double start_offset;
   int x, y;
 
   /* for number */
diff --git a/source/blender/imbuf/intern/anim_movie.c b/source/blender/imbuf/intern/anim_movie.c
index 6562bd7be5f..3585f585b91 100644
--- a/source/blender/imbuf/intern/anim_movie.c
+++ b/source/blender/imbuf/intern/anim_movie.c
@@ -425,6 +425,7 @@ static int startavi(struct anim *anim)
   }
 
   anim->duration_in_frames = anim->avi->header->TotalFrames;
+  anim->start_offset = 0.0f;
   anim->params = NULL;
 
   anim->x = anim->avi->header->Width;
@@ -597,6 +598,13 @@ static int startffmpeg(struct anim *anim)
     return -1;
   }
 
+  double video_start = 0;
+  double pts_time_base = av_q2d(video_stream->time_base);
+
+  if (video_stream->start_time != AV_NOPTS_VALUE) {
+    video_start = video_stream->start_time * pts_time_base;
+  }
+
   frame_rate = av_guess_frame_rate(pFormatCtx, video_stream, NULL);
   anim->duration_in_frames = 0;
 
@@ -620,20 +628,14 @@ static int startffmpeg(struct anim *anim)
    * This is because the video stream duration can be shorter than the pFormatCtx->duration.
    */
   if (anim->duration_in_frames == 0) {
-    double pts_time_base = av_q2d(video_stream->time_base);
     double stream_dur;
 
     if (video_stream->duration != AV_NOPTS_VALUE) {
       stream_dur = video_stream->duration * pts_time_base;
     }
     else {
-      double video_start = 0;
       double audio_start = 0;
 
-      if (video_stream->start_time != AV_NOPTS_VALUE) {
-        video_start = video_stream->start_time * pts_time_base;
-      }
-
       /* Find audio stream to guess the duration of the video.
        * Sometimes the audio AND the video stream have a start offset.
        * The difference between these is the offset we want to use to
@@ -662,6 +664,11 @@ static int startffmpeg(struct anim *anim)
     anim->duration_in_frames = (int)(stream_dur * av_q2d(frame_rate) + 0.5f);
   }
 
+  double ctx_start = 0;
+  if (pFormatCtx->start_time != AV_NOPTS_VALUE) {
+    ctx_start = (double)pFormatCtx->start_time / AV_TIME_BASE;
+  }
+
   frs_num = frame_rate.num;
   frs_den = frame_rate.den;
 
@@ -674,6 +681,9 @@ static int startffmpeg(struct anim *anim)
 
   anim->frs_sec = frs_num;
   anim->frs_sec_base = frs_den;
+  /* Save the relative start time for the video. IE the start time in relation to where playback
+   * starts. */
+  anim->start_offset = video_start - ctx_start;
 
   anim->params = 0;
 
@@ -1672,6 +1682,11 @@ int IMB_anim_get_duration(struct anim *anim, IMB_Timecode_Type tc)
   return IMB_indexer_get_duration(idx);
 }
 
+double IMD_anim_get_offset(struct anim *anim)
+{
+  return anim->start_offset;
+}
+
 bool IMB_anim_get_fps(struct anim *anim, short *frs_sec, float *frs_sec_base, bool no_av_base)
 {
   double frs_sec_base_double;