Skip Problem

[Graham]

I have read dozens of posts and tried numerous combinations of decoders and I cannot fix this.

There is a delay of anything from 10 seconds to 2 minutes when I hit skip forward (or back) when playing a recording. The picture skips immediately (and freezes) but there is a random and variable delay before the audio starts and the picture unfreezes. It does not seem to be anything to do with the recording because I can play the same recording and get different results each time.

I have imported some GBPVR recording in msdvr format into the npvr recording database and these skip nicely using the same decoders as the npvr ts files that do not skip nicely.

I have tried combinations of the Windows 7 MS decoders, PDVD 12, ffdshow and LAV and any permutation gives me variable delays.

I would welcome any suggestions that I could try.

Thanks

Graham

[sub]

Do you have your audio renderer set to something other than the default? This has caused problems like this for some users. Otherwise it's probably down to the decoders you're using.

[b00sfuk]

Exactly the same for me, I've also spent a long time going through the same process. No solution other than I find if it does freeze I can speed up the resume by doing another skip quickly, e.g. forward then back, and it generally comes back straightaway.

[Graham]

Do you have your audio renderer set to something other than the default? This has caused problems like this for some users. Otherwise it's probably down to the decoders you're using.

I tried changing the audio renderer for a couple of tests but it made no difference so I put it back to "Default". Almost all of my attempts had it on Default.

[Graham]

A ts file graph is ...

Playback graph running...
2012-02-18 17:47:23.625 [DEBUG][1] Graph filter list:
2012-02-18 17:47:23.625 [DEBUG][1] - NPVR CC Dump
2012-02-18 17:47:23.625 [DEBUG][1] - EVR Renderer
2012-02-18 17:47:23.625 [DEBUG][1] - NPVR PES Collector (DVBSubs)
2012-02-18 17:47:23.625 [DEBUG][1] - MPEG1 Audio Renderer
2012-02-18 17:47:23.625 [DEBUG][1] - Line 21 Decoder
2012-02-18 17:47:23.625 [DEBUG][1] - MPEG Audio Decoder
2012-02-18 17:47:23.625 [DEBUG][1] - Microsoft DTV-DVD Video Decoder
2012-02-18 17:47:23.625 [DEBUG][1] - MPEG-2 Demultiplexer
2012-02-18 17:47:23.625 [DEBUG][1] - NPVR TS Reader

And skips are variable.

A dvrms graph is ...

Graph filter list:
2012-02-18 17:49:20.746 [DEBUG][1] - Default DirectSound Device
2012-02-18 17:49:20.746 [DEBUG][1] - EVR Renderer
2012-02-18 17:49:20.746 [DEBUG][1] - Microsoft DTV-DVD Video Decoder
2012-02-18 17:49:20.746 [DEBUG][1] - MPEG Audio Decoder
2012-02-18 17:49:20.746 [DEBUG][1] - C:\$Media\TVvideo\Earthflight\Earthflight_20111229 _20002100.dvr-ms

and skips are instant.

The full logs are attached. I am at my wits end. Is there anything else that I can try?

Thank you.

[johnsonx42]

It's not just you Graham, this is a problem for many NPVR users (me included). NPVR has had multiple problems with skipping since day one yet since none of these problems seem to affect sub, so it's tough for him to find the problem. I really don't believe it has anything to do with decoders, having tried various decoders without any change to the problems.

[Graham]

... none of these problems seem to affect sub, so it's tough for him to find the problem.

A feeble attempt to help ... I can hit skip, video skips and freezes, I can open npvr.log to see the most recent log msg, close log and wait for the skip to complete. Every time that I tried this, the most recent log msg during the freeze was "Hiding OSD" as shown ...

2012-02-18 17:47:28.861 [DEBUG][1] OnKeyDown() called: {CTRL}ControlKey, Control
2012-02-18 17:47:28.861 [DEBUG][1] OnKeyDown() called: {CTRL}Right, Control
2012-02-18 17:47:28.881 [DEBUG][1] User is trying to skip.
2012-02-18 17:47:28.881 [DEBUG][1] Current position is: 12.0939998626709
2012-02-18 17:47:28.881 [DEBUG][1] Asking for position: 72.0940017700195
2012-02-18 17:47:28.885 [DEBUG][1] Using timing.info, jumping to time 72.094002 (offset 35635400)
2012-02-18 17:47:28.977 [DEBUG][1] After skipping position reports as: 83.6880035400391
2012-02-18 17:47:28.977 [DEBUG][1] OnKeyDown() done. Took 116.210899353027 ms
2012-02-18 17:47:29.005 [DEBUG][1] Updating OSD.
2012-02-18 17:47:32.985 [DEBUG][1] Hiding OSD.
2012-02-18 17:47:35.364 [DEBUG][1] GetEventCode() returned: 0x8104
2012-02-18 17:47:36.682 [DEBUG][1] GetEventCode() returned: 0x8104

I'll carry on wrestling with this if anyone has any suggestions !!!

[mikeh49]

Same skipping problem here, on one system ("lounge"), but not the other two.

[johnsonx42]

Code:

2012-02-18 17:47:28.881 [DEBUG][1] Current position is: 12.0939998626709
2012-02-18 17:47:28.881 [DEBUG][1] Asking for position: 72.0940017700195
2012-02-18 17:47:28.885 [DEBUG][1] Using timing.info, jumping to time 72.094002 (offset 35635400)
2012-02-18 17:47:28.977 [DEBUG][1] After skipping position reports as: 83.6880035400391

I see you have the same problem I do - forward skips are too long. You asked for a 60 second skip, and got a ~71.5 second skip. Mine are usually more like 3-4 seconds too long, while my comskip jumps are about 3-4 seconds too early. My backwards skips are similarly shorter than they should be.

[mvallevand]

For a contrasting view on my files I get (from the keyboard)

Code:

2012-02-19 01:03:58.285    [DEBUG][1]    Current position is: 1860.53002929688
2012-02-19 01:03:58.285    [DEBUG][1]    Asking for position: 1920.53002929688
2012-02-19 01:03:58.287    [DEBUG][1]    Using timing.info, jumping to time 1920.530029 (offset 893222780)
2012-02-19 01:03:58.335    [DEBUG][1]    After skipping position reports as: 1921.86999511719

2012-02-19 01:04:03.844    [DEBUG][1]    Current position is: 1926.36206054688
2012-02-19 01:04:03.844    [DEBUG][1]    Asking for position: 2226.36206054688
2012-02-19 01:04:03.845    [DEBUG][1]    Using timing.info, jumping to time 2226.362061 (offset 1028292508)
2012-02-19 01:04:03.921    [DEBUG][1]    After skipping position reports as: 2226.4150390625

Skips on the NMT are pretty good too, typically within a second with Timing.Info

Martin