Author |
Message |
grante
|
Posted: Thu Oct 18, 2007 8:25 pm |
|
Joined: Mon Jun 27, 2005 4:42 pm
Posts: 321
Location:
Minneapolis, Minnesota, USA
|
Is there any way to adjust audio sync other than in 10ms steps?
It's really painful when you have to adjust it by 4-5 minutes
and it has to be done in 10ms steps. What really sucks is that
after spending several minutes leaning on the remote button --
when I finally get it almost synced up -- the playback speed
often jumps to about 5X fast-forward mode (both audio and
video). Then I have to exit the player, start the program,
jump to where I left off, and start the long process of
adjusting the audio sync again.
Has the internal player always been this flakey or have I
broken it somehow? I just recently switched from the PVR-350
HW decoder to the internal player, and I've had a constant
string of problems with it...
_________________ Grant
|
|
Top |
|
 |
grante
|
Posted: Thu Oct 18, 2007 8:40 pm |
|
Joined: Mon Jun 27, 2005 4:42 pm
Posts: 321
Location:
Minneapolis, Minnesota, USA
|
I meant 4-5 seconds rather than 4-5 minutes. Every time I get
the audio sync up to about 3200ms, the playback goes into this
weird fast-forward mode where both video and audio skip forward
at about 5X normal speed. If I lower the audio sync back below
3200ms, then normal spped (but out of sync) playback resumes.
How do I get audio sync set to a value larger than 3200ms?
_________________ Grant
|
|
Top |
|
 |
datobin1
|
Posted: Fri Oct 19, 2007 6:36 am |
|
Joined: Wed Dec 21, 2005 1:14 pm
Posts: 145
Location:
Charlotte NC USA
|
Sounds like somthing else is broken, usually the audio delay is used to delay the audio to allow your tvs video proccessor extra time (especially true with digital displays).
Sounds like your seek tables may be corupt. If they are, ussing the arrow keys to skip forward 30 secounds should have a delay and commercial skip will be erratic.
If you see those simptoms also do a search for "seek Table" and you will find instruction on how to fix them. Its a simple command but, will take a little while depending on how many shows you have.
|
|
Top |
|
 |
grante
|
Posted: Fri Oct 19, 2007 10:54 am |
|
Joined: Mon Jun 27, 2005 4:42 pm
Posts: 321
Location:
Minneapolis, Minnesota, USA
|
datobin1 wrote: Sounds like somthing else is broken, usually the audio delay is used to delay the audio to allow your tvs video proccessor extra time (especially true with digital displays).
The audio/video in the recording is out of sync. At the end of the recording it's off by about 4 seconds. It's due to problems when transcoding an ATSC recording that has drop-outs. Quote: Sounds like your seek tables may be corupt. If they are, ussing the arrow keys to skip forward 30 secounds should have a delay and commercial skip will be erratic. I know they are. That's because you can't run commflagging on a transcoded MPEG2 recording to regenerate them. I write my own program to regenerate them, but it's not quite right, and I haven't been able to find any documentation on what's supposed to be _in_ the seek tables. I've been putting int the file offsets and timestamps of GOP frames, but that doesn't seem to be correct. Quote: If you see those simptoms also do a search for "seek Table" and you will find instruction on how to fix them. Its a simple command but, will take a little while depending on how many shows you have.
I've already tried that and it doesn't work for MPEG2
recordings due to bugs in the commflagging program [this seems
to be a known problem]. The seek tables generated are just
garbage. Since I can't generate proper seek tables, I usually
just use the audio sync adjust to watch the recordings. But
that only works up to 3200ms, after which the player goes nuts.
_________________ Grant
|
|
Top |
|
 |
mogator88
|
Posted: Fri Oct 19, 2007 3:59 pm |
|
Joined: Tue Jan 30, 2007 1:27 am
Posts: 299
|
can you pause the recording before sync'ing the audio?
Also, does the delay change from the beginning to the end of the recording?
|
|
Top |
|
 |
grante
|
Posted: Sat Oct 20, 2007 7:42 pm |
|
Joined: Mon Jun 27, 2005 4:42 pm
Posts: 321
Location:
Minneapolis, Minnesota, USA
|
mogator88 wrote: can you pause the recording before sync'ing the audio? I've never tried. I'll see if that works. Quote: Also, does the delay change from the beginning to the end of the recording?
Yes, the sound is initially sync'ed and goes progressively out
of sync each time the ATSC signal drops out. SD recordings
that haven't been transcoded don't have this problem, so the
sync issue is introduced when a recording with signal dropouts
is transcoded. When I watch the transcoded recordings, I use
the KM player's adjust audio sync feature to compensate, but it
only works up to 3200ms.
Hopefully the audio sync issue will soon be moot. I'm trying to
figure out both how to fix the transcoding and to build a
frontend that can playback HD recrodings without transcoding
them first. My current FE with a 2.4GHz P4 + FX5200 can't do
it: playback with XVMC works beautifully until the OSD comes up
at which point mythfrontend collapses in a heap using 100% CPU
and playing only a few frames per second.
_________________ Grant
|
|
Top |
|
 |
mogator88
|
Posted: Sat Oct 20, 2007 9:32 pm |
|
Joined: Tue Jan 30, 2007 1:27 am
Posts: 299
|
Something doesn't seem right about this. I record OTA ATSC only, and I have dropouts, but the audio never goes out of sync. I don't transcode now, but even when I did I don't recall having that problem. Might be a symptom of another deeper problem, or maybe not.
For my remote frontend I also have a p4 2.4 and 5200 video card, but I'm using the s-video output. No problems with that. But what are you driving from your card?
|
|
Top |
|
 |
grante
|
Posted: Wed Oct 24, 2007 6:05 pm |
|
Joined: Mon Jun 27, 2005 4:42 pm
Posts: 321
Location:
Minneapolis, Minnesota, USA
|
mogator88 wrote: Something doesn't seem right about this. I record OTA ATSC only, and I have dropouts, but the audio never goes out of sync. I don't transcode now, but even when I did I don't recall having that problem. Might be a symptom of another deeper problem, or maybe not. I was transcoding using mencoder, so I suspect it's a problem with mencoder. I've switched to using mythtranscode, so we'll see if that fixes the problem. Quote: For my remote frontend I also have a p4 2.4 and 5200 video card, but I'm using the s-video output. No problems with that. But what are you driving from your card?
X11 is configured to 640x480 60Hz and I'm driving a 15-year old
JVC TV using the svideo out on the NVidia 5200 card. It can
play back HD using xvmc, but anytime the OSD comes up (w/ xvmc)
mythfrontend sucks the CPU usage up to 100% and playback almost
stops. Even when playing back SD with xvmc, the OSD will cause
CPU usage to go to 100%. AFIACT, it almost has to be a bug in
the OSD code, but I'm not waiting for it to get fixed.
_________________ Grant
|
|
Top |
|
 |
mogator88
|
Posted: Thu Oct 25, 2007 7:09 am |
|
Joined: Tue Jan 30, 2007 1:27 am
Posts: 299
|
I can't check it now, but I remember the de-interlacing method did affect the kind of playback I got. Its on the same screen where you select XvMC.
|
|
Top |
|
 |
grante
|
Posted: Thu Oct 25, 2007 12:30 pm |
|
Joined: Mon Jun 27, 2005 4:42 pm
Posts: 321
Location:
Minneapolis, Minnesota, USA
|
mogator88 wrote: I can't check it now, but I remember the de-interlacing method did affect the kind of playback I got. Its on the same screen where you select XvMC.
I tried both with and without de-interlacing: the behavior was
the same. Anytime the OSD was up with xvmc during HD playback,
CPU usage jumps to 100% and playback drops to a few frames per
second. When the OSD goes away, normal playback usually
resumes, but sometimes the sound doesn't come back on.
During xvmc SD playback, the OSD causes the CPU to jump from
20% to to 90-100%, but the stuttering is very mild. There's another
thread where I hashed through all of the different setting combinations
I tried. The consensus was that that my system should "just work".
My new AM2/6150 box should be ready to go in a week or two. It
doesn't seem to be affected by the OSD+xvmc problem. It'll
also be smaller and have 3.5X the disk storage.
Unfortunately the case is going to look like something built by
an electrical engineer.
_________________ Grant
|
|
Top |
|
 |
mogator88
|
Posted: Thu Oct 25, 2007 2:51 pm |
|
Joined: Tue Jan 30, 2007 1:27 am
Posts: 299
|
Well, don't chuck the P4 box yet, it could possibly work as a remote frontend unit. That's how I'm using mine. Good luck with the new build!
|
|
Top |
|
 |
grante
|
Posted: Sun Oct 28, 2007 3:26 pm |
|
Joined: Mon Jun 27, 2005 4:42 pm
Posts: 321
Location:
Minneapolis, Minnesota, USA
|
mogator88 wrote: Well, don't chuck the P4 box yet, it could possibly work as a remote frontend unit.
I'm building a new box because my current one can't playback
recordings without transcoding them first: it chokes on the OSD
when playing back HD recordings. It won't be much good as a
remote frontend.
_________________ Grant
|
|
Top |
|
 |
mogator88
|
Posted: Sun Oct 28, 2007 5:18 pm |
|
Joined: Tue Jan 30, 2007 1:27 am
Posts: 299
|
A frontend doesn't have to do as much, it could make the difference. BTW, How much RAM do you have in the P4? If its less than 512mb that would be a problem.
|
|
Top |
|
 |
grante
|
Posted: Sun Oct 28, 2007 6:04 pm |
|
Joined: Mon Jun 27, 2005 4:42 pm
Posts: 321
Location:
Minneapolis, Minnesota, USA
|
mogator88 wrote: A frontend doesn't have to do as much, it could make the difference. Not that much of a difference. With the OSD displayed during xvmc HD playback, the video about 2 fps with 100% CPU usage. With xvmc enabled playback, the OSD still takes 100% of the CPU, but the frame rate is _almost_ real-time Quote: BTW, How much RAM do you have in the P4? If its less than 512mb that would be a problem.
It has 512MB, and applications only use about 1/3 of it. The
other 2/3 is caching disk blocks.
_________________ Grant
|
|
Top |
|
 |