On Tue, 2005-11-29 at 09:25 -0500, Daniel Kristjansson wrote:
> On Sat, 2005-11-26 at 21:59 +0000, MythTV wrote:
> > #692: Odd seek behavior in DVB/HDTV recordings
> > Changes (by danielk):
> > * status: new => assigned
> > It looks like recordings with [7300] are ok.
> > But recordings made with [8042] are not.
> > I'll try isolate the revision where things break and report back on what I
> > find.
>
> It looks like the problem first appears in [7675].
>
> SVN/trac appears to be down at the moment, but I'll check out the
> changeset and fix this shortly after it comes back online...
Can someone experiencing the seek problem give the attached patch a try.
I'm still having problems accessing mythtv.org, so this patch is against
a several day old checkout I have here. But I don't think there have
been any commits that break the patch since then...
I think the mythtv.org access problem is with my ISP, I also lost access
to cuymedia.net for a few hours today... oy wey es RCN.
-- Daniel
--- ../mythtv-2/mythtv/libs/libmythtv/dtvrecorder.cpp 2005-11-25 22:40:09.696454312 -0500
+++ libs/libmythtv/dtvrecorder.cpp 2005-11-29 14:47:19.482134080 -0500
@@ -157,6 +157,13 @@
* streams, and if all else fails we just look for the picture
* start codes and call every 16th frame a keyframe.
*
+ * NOTE: This does not only find keyframes but also tracks the
+ * total frames as well. At least a couple times seeking
+ * has been broken by short-circuiting the search once
+ * a keyframe stream id has been found. This may work on
+ * some channels, but will break on others so algorithmic
+ * optimizations should be done with great care.
+ *
* \startcode
* PES header format
* byte 0 byte 1 byte 2 byte 3 [byte 4 byte 5]
@@ -182,7 +189,6 @@
const unsigned char *buffer = tspacket->data();
bool hasKeyFrame = false;
- int stream_id = 0xffff; // sentinel value (0xffff is always invalid)
// Scan for PES header codes; specifically picture_start
// and group_start (of_pictures). These should be within
@@ -192,7 +198,7 @@
// 00 00 01 B3: seq_start_code
// (there are others that we don't care about)
uint i = tspacket->AFCOffset();
- for (; (i+1) < TSPacket::SIZE && !hasKeyFrame; i++)
+ for (; (i+1) < TSPacket::SIZE; i++)
{
const unsigned char k = buffer[i];
if (0 == _position_within_gop_header)
@@ -210,7 +216,7 @@
// At this point we have seen the start code 0 0 1
// the next byte will be the PES packet stream id.
- stream_id = buffer[i+1];
+ const int stream_id = buffer[i+1];
if (PESStreamID::PictureStartCode == stream_id)
{
_frames_written_count += (_first_keyframe > 0) ? 1 : 0;
@@ -220,19 +226,19 @@
// last GOP or SEQ stream_id, then pretend this picture
// is a keyframe. We may get artifacts but at least
// we will be able to skip frames.
- hasKeyFrame = (0 == (_frames_seen_count & 0xf)) &&
+ hasKeyFrame |= (0 == (_frames_seen_count & 0xf)) &&
(_last_gop_seen + maxKFD < frameSeenNum) &&
(_last_seq_seen + maxKFD < frameSeenNum);
}
else if (PESStreamID::GOPStartCode == stream_id)
{
- _last_gop_seen = frameSeenNum;
- hasKeyFrame = true;
+ _last_gop_seen = frameSeenNum;
+ hasKeyFrame |= true;
}
else if (PESStreamID::SequenceStartCode == stream_id)
{
- _last_seq_seen = frameSeenNum;
- hasKeyFrame = _last_gop_seen + maxKFD < frameSeenNum;
+ _last_seq_seen = frameSeenNum;
+ hasKeyFrame |= _last_gop_seen + maxKFD < frameSeenNum;
}
_position_within_gop_header = 0;
}
_______________________________________________
mythtv-dev mailing list
[email protected]
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
