[Approved] Fix for stuttering/corrupted streaming of live tv (3 Viewers)

[davidf]

Cheers very much

Just to confirm again - I've suffered from stuttering since the demise of my RAID 0 timeshift array failed (I blamed MP Audio Renderer, so sorry Tourettes ). I've been watching TV most of the evening and haven't noticed any stuttering on SD or HD with Audio Renderer also running. Having 20/20 hindsight it fairly obvious that the RAID failure and stuttering were around the same time, but now it's gone again , should have put 2+2 together myself.

disaster123 - try uping the size of each timeshift file and see if that reduces the problem (they are about 15-20 minutes each for an HD file on the defaults which is about 3 times per hour).

 

[disaster123]

AW: Fix for stuttering/corrupted streaming of live tv

davidf
i get this now on SD channels

 

[tourettes]

can someone compile this for main trunk as well ?

/gibman

Here you are.

Simone

Does "main trunk" mean MP 1.1.2?

Main trunk is always the latest SVN content, so now it is the post 1.2.0 alpha or pre 1.2.0 beta if you want to have a name for it

 

[pilehave]

I think this patch cured the pixelation on our only HD channel, "DR HD", but I still see pixelation on all my channels if I record them to my normal disk (WD Green 1 TB disk) or if I watch Live TV while I record. If I only watch live TV, I never see any pixellation due to bad data.

I timeshift to a 4 GB ramdrive and record to a normal disk. When only timeshifting, Windows Resource Monitor reports a response time to this disk ~ 1 ms (obviously a ramdisk is superfast accessed). When I record, the response time to my normal disk (the WD Green) is up at ~ 350-550 ms.

My theory is that writing the recordings to disk, is a massive load for some reason (many writecycles?). It is better with this patch, but I'm pretty sure it didn't cure anything.

My signal (DVB-T) is just about as good as it gets, and even then TV-Server sometimes reports that my signal craps out totally, like down to 33% signal and 0% quality or so. And this happens while I happily watches TV on my Sony DVB-T 26", and even while I watch the built-in signal status which has never left "Very high" signal and always have 0 errors and pre-viterbi is ~ 1.3 E-04 which is actually very good signal.

So far I have tried 3 different DVB-T cards and the result have been pretty much the same, files with what looks like bad signal (bad data) but what I believe is a result of stress on the disk.

I find it hard to believe that PC-based TV-cards are so much crappier than an average Sony TV which today can be had for ~ £150. I mean, I have seen TV-cards that alone are more expensive...

What's your take on my conspiracy theory?

 

[mm1352000]

I find it hard to believe that PC-based TV-cards are so much crappier than an average Sony TV which today can be had for ~ £150. I mean, I have seen TV-cards that alone are more expensive...

What's your take on my conspiracy theory?

My take is that PC TV cards often have to work in a hotter or [electrically] noiser environment. This can compromise signal quality. It doesn't mean the components are different or more expensive. Just means the PCB designs have to be more careful. Could be completely wrong of course

There are all manner of reasons why your HDD might be so slow:
- I'd be interested to know if your HDD is an "EARS" green drive. If it is then you may have the partitions mis-aligned.
- Have you tried a defrag on it? Fragmented drive space can slow down access times dramatically.
- Is the drive interface running in AHCI, SCSI or legacy IDE mode, and is the access with UDMA or PIO?

 

[pilehave]

I find it hard to believe that PC-based TV-cards are so much crappier than an average Sony TV which today can be had for ~ £150. I mean, I have seen TV-cards that alone are more expensive...

What's your take on my conspiracy theory?

My take is that PC TV cards often have to work in a hotter or [electrically] noiser environment. This can compromise signal quality. It doesn't mean the components are different or more expensive. Just means the PCB designs have to be more careful. Could be completely wrong of course

There are all manner of reasons why your HDD might be so slow:
- I'd be interested to know if your HDD is an "EARS" green drive. If it is then you may have the partitions mis-aligned.
- Have you tried a defrag on it? Fragmented drive space can slow down access times dramatically.
- Is the drive interface running in AHCI, SCSI or legacy IDE mode, and is the access with UDMA or PIO?

My recording disc is this one: Western Digital's Caviar Green hard drive - The Tech Report - Page 1
I have Windows running automated defrag every week in the night, so all my drives are unfragmented.
Everything is running ACHI, just to be sure I disabled the onboard IDE controller, and everything still works

About the heat, the server is placed in my garage which at the moment is down around 10 degrees celsius, and the discs are showing ~ 18-20 degrees temperatures. Perfect condition IMHO

Happy XMAS!

 

[mm1352000]

My recording disc is this one: Western Digital's Caviar Green hard drive - The Tech Report - Page 1

There are at least 2 models of WD Green 1TB HDDs. Are you *certain* that you have the EADS model which was reviewed and *not* the 4kB sector EARS model? The difference is important.

Everything else you described seems perfectly reasonable to me, although I'd still recommend confirming that you're not running in PIO mode.

 

[gibman]

while CPU is idle try and do a simple copy-paste operation of a bigger file (on the server). Then watch CPU usage.
It should be very low when doing this file copy.

If u r running PIO then cpu usage would rise very quickly.

This is how I often test these things.

/gibman

 

[disaster123]

AW: Fix for stuttering/corrupted streaming of live tv

hi miroslav22 since your patch my biggest problem is this:

26-12-2010 22:25:49.298 [e78]FileReader::Read() no open file
26-12-2010 22:25:49.298 [e78]READ FAILED1
26-12-2010 22:25:49.319 [d08]SEEK FAILED
26-12-2010 22:25:49.319 [d08]FileReader::Read() no open file
26-12-2010 22:25:49.319 [d08]READ FAILED1
26-12-2010 22:25:49.339 [e78]SEEK FAILED
26-12-2010 22:25:49.339 [e78]FileReader::Read() no open file
26-12-2010 22:25:49.339 [e78]READ FAILED1
26-12-2010 22:25:49.357 [d08]SEEK FAILED
26-12-2010 22:25:49.358 [d08]FileReader::Read() no open file
26-12-2010 22:25:49.358 [d08]READ FAILED1
26-12-2010 22:25:49.377 [d08]SEEK FAILED
26-12-2010 22:25:49.378 [d08]FileReader::Read() no open file
26-12-2010 22:25:49.378 [d08]READ FAILED1
26-12-2010 22:25:49.401 [e78]SEEK FAILED
26-12-2010 22:25:49.402 [e78]FileReader::Read() no open file
26-12-2010 22:25:49.402 [e78]READ FAILED1
26-12-2010 22:25:49.419 [d08]SEEK FAILED
26-12-2010 22:25:49.419 [d08]FileReader::Read() no open file
26-12-2010 22:25:49.420 [d08]READ FAILED1
26-12-2010 22:25:49.441 [e78]SEEK FAILED
26-12-2010 22:25:49.442 [e78]FileReader::Read() no open file
26-12-2010 22:25:49.442 [e78]READ FAILED1
26-12-2010 22:25:49.460 [d08]SEEK FAILED
26-12-2010 22:25:49.460 [d08]FileReader::Read() no open file
26-12-2010 22:25:49.461 [d08]READ FAILED1

Full log file can be found here:
TsReader-2010-12-26.Log

Stefan

 

[pilehave]

My recording disc is this one: Western Digital's Caviar Green hard drive - The Tech Report - Page 1

There are at least 2 models of WD Green 1TB HDDs. Are you *certain* that you have the EADS model which was reviewed and *not* the 4kB sector EARS model? The difference is important.

Everything else you described seems perfectly reasonable to me, although I'd still recommend confirming that you're not running in PIO mode.

Yes, I'm sure