Experimental TsReader development (3 Viewers)

[Owlsroost]

Re: AW: Experimental TsReader development

V43 was OK to me using UNC. But with V44 channel change takes up to 20s and more. Log attached.

Can you post all the logs (Client and Server) please ?

I can see this in the log:

25-01-2012 20:33:06.889 [818]CTsReaderFilter::Pause() - IsTimeShifting = 1 - state = 2
25-01-2012 20:33:06.961 [940]Aud/Ref : 12.046, Compensated = 9.787 ( 0.399 A/V buffers=02/56), Clk : 9.387562, SampCnt 60, Sleep 20 ms, stallPt 0.400
25-01-2012 20:33:07.625 [818]CTsReaderFilter::Pause() - END - state = 1
25-01-2012 20:33:07.626 [c64]CTsReaderFilter:: ForcePosnUpdate: 9.306 s
25-01-2012 20:33:17.979 [c64]CTsReaderFilter, Live Tv is paused for more than 10 secs => m_bLiveTv=false.
25-01-2012 20:33:39.284 [818]OnZapping 1
25-01-2012 20:33:39.304 [818]vidPin: UpdateFromSeek, m_rtStart 31.589000, m_dRateSeeking 1.000000
25-01-2012 20:33:39.304 [818]CTsReaderFilter::--SeekPreStart()-- LiveTv : 0, TimeShifting: 1 31.589 ( Abs 31.589000 / 31.589634 )- isSamePosn: 0, OnZap: 0, Force 0, ForceRC 0, Media changing 0
25-01-2012 20:33:39.304 [818]Zap to File Seek : 17770 mS ( 31.589000 / 31.589000 ) LiveTv : 1, Seek : 1

but there's no clue why there's 32s between the Pause() and the 'OnZapping 1' lines....

As far as I can tell from the TV plugin code, TVHome.cs is waiting for this method to return during that time:

       succeeded = server.StartTimeShifting(ref user, channel.IdChannel, out card, out cardChanged);

Also is this the standard MP 1.2.2 TV client/plugin ?

Tony

 

[disaster123]

AW: Experimental TsReader development

Owlsroost
no it is my custom git branch. But there are no changes regarding this stuff. Also it works fine with mod V43, V42, V41, ....

 

[Sebastiii]

disaster123, hi maybe you can send him your log in PM ?
Surely Tony can find the culprit
Thanks

 

[Owlsroost]

@disaster123, hi maybe you can send him your log in PM ?
Surely Tony can find the culprit
Thanks

You can't attach files to PM's (AFAIK)

I'm assuming the problem is being caused by my changes to the duration measuring code, but I'd like to know why it's breaking something else....

Tony

 

[disaster123]

AW: Experimental TsReader development

Will send you logs this evening (GMT+1).

 

[mrmojo666]

noStopMod44 added ->here<-

Includes a fix for random 'black screen at start of play' problem introduced in v41, and mods to duration update code to (hopefully) fix the 'live' recording stop problem.

Tony

seems that is working very well this v44. no issue to report.

thank you Tony

 

[FreakyJ]

AW: Experimental TsReader development

the first impression is good (24H use), but I need some more time to tell you more^^
- no black picture
- smaller jumps if there is corruption (now it seems to be only a small glitch before it was like a 5Sec jump...)

 

[disaster123]

AW: Experimental TsReader development

Strange today after rebooting the client i cannot reproduce it anymore ?!?! No idea why. Will now check for a few hours...

 

[tompa]

Hi!

I had some heavy stuttering after Swedish local news with v43 and Argus. With v44 (and 42d) I have not seen that! Maybe random, I will report if so.
Running UNC path.

thx!
/tompa

 

[Owlsroost]

Hi!

I had some heavy stuttering after Swedish local news with v43 and Argus. With v44 (and 42d) I have not seen that! Maybe random, I will report if so.
Running UNC path.

thx!
/tompa

Someone from Sweden (maybe yourself) has mentioned a similar problem before - it was caused by the broadcaster switching not being 'clean', so there were big jumps in the stream timestamps at the switching points - can't do much about that....

If you look in the TsReader log, you will probably see continuity error and PTS jump messages around the switching times.

Tony