resume playback position problem

[sub]

See if this helps

[johnsonx42]

yep, that works just fine. thanks sub.

[johnsonx42]

can I assume that this patch to Unmanaged.dll also includes the HDHR fixes from earlier today? http://forums.gbpvr.com//showthread.php?...g&p=388287

edit: actually I have already assumed it; please let me know if I should un-assume this.

[johnsonx42]

sub,

that patch did resolve the resume problem, but it fouled up the skipping. now it's back to the skipping problem that existed when 1.5.28 first came out, as described and patched here: http://forums.gbpvr.com//showthread.php?...recordings

I can't say that it's exactly the same as before, but the behavior is certainly very similar. When I try to skip backwards in a recording, the small skips (I guess what is called rewind) either stay in the same spot or go forward a little, while the big skips go back only about 10 seconds when they should go back 30. Conversely, the forward skips go further than they should - 10 second skips go 20+, and 60 seconds skips go over 2 minutes ahead. Near the beginning of a recording the skips seem closer to normal, but from about half-way through on they get worse and worse.

As a sanity check, I put the original 1.5.31 Unmanaged.dll back on, and the skipping is back to normal from the beginning of the recording to the end. I'd rather have the resume problem than be unable to skip properly.

[johnsonx42]

for anyone following along with this issue, troubleshooting transferred over to here: http://forums.gbpvr.com//showthread.php?...post388528
there is a patch in post 35 of that thread that puts proper skipping back in place, at the expense of breaking the resume position fix again. presumably either that thread or this one here will eventually have a patch that fixes the resume position problem without breaking skip timing.

[sub]

Here is a patch your can try for the problem where resuming playback showed the position as 12:48.

[johnsonx42]

sub, I took the liberty of moving that patch from my "paradise lost" thread to here; it made no sense to discuss it there, since nothing else in that thread pertained to this problem, but made no sense to discuss it here in absence of the patch either.

The bad news is that patch has horribly broken skipping again, but in a new way. When I skip in to a show to about 20 minutes, it jumps back to the beginning of the show and the beginning timeline position goes to -35791394; at that point any further skips go back to the beginning again, and after a bit playback just ends. This happens on every single recording I tried.

Resume does seem to work correctly for positions prior to 20 minutes in.

I did just notice though that I am using the Unmanaged.dll dated 11/8, so it could be the problem was introduced in the later version; I will revert to the 11/3 version posted here and see if the problems are the same.

[johnsonx42]

Yes, I just confirmed it does the same thing with the 11/3 version of Unmanaged.dll posted here.

I've attached my NPVR.log. I played back tonight's Family Guy. First I skipped in to around 20 minutes, and it jumped back to the beginning, then I tried to skip again and it exited back to the Recordings list. On the next playback attempt I skipped in only about 14 minutes, then exited and resumed playback. It did resume correctly.

[johnsonx42]

I realized that Family Guy show has a comskip break at the key point in the timeline, which may muddy your interpretation of the log. Here's the log showing playback of a 1-hour show (House) that does not have a comskip break at the magic point (which actually appears to be just after 18 minutes). The problem does affect shows without comskip files as well - at the end of this log I played Sesame Street, a show on which I do not run comskip, and the problem occurs there at some point after 30 minutes.

Note: look at the playback of House that begins at 22:42:40.041; the one before it did have a commercial at the 18 minute mark. The playback of Sesame Street is immediately after that. They're the last two events in the log.

[johnsonx42]

two final notes:

1) the timeline still corrupts to -35791934 if I just let it play to the problem point, but it appears to keep playing correctly as long as I don't skip.

2) forward skips go farther than they should, while backwards skips either don't skip at all or go forward a little. so this version of the patch has the same trouble in that area as the last one did.