- July 19, 2011
- 152
- 4
- Home Country
- United States of America
Here is the tv server log
[2015-07-10 21:11:24,339] [Log ] [SchedulesDirect EPG Client] [INFO ] - Sucessfully processed 970 of 886 channels
[2015-07-10 21:11:24,349] [Log ] [SchedulesDirect EPG Client] [INFO ] - Starting processing of 19577 schedule entries
Hello and welcome mitchellmedia
First, the files/streams that the Colossus and TV Server are producing are 100% compliant with the relevant technical standards. I'm very confident there's nothing "hostile" or dodgy about them. I'd expect any reasonable hardware or software to be able to play them.
1. My recollection is that there was only trouble with client-only installations. If MediaPortal and TV Server were installed on the same PC then there was no problem.
2. I further recall that using UNC paths instead of the default RTSP streaming (previously a "debug" setting, now an advanced setting) as the "time-shifting protocol" eliminated the problem for at least some people. At the time using UNC paths introduced another problem (something related to the caching behaviour of the Windows implementation of the SMB file sharing protocol). However, my understanding is that the UNC paths problem has subsequently been eliminated, making it a viable solution.
3. As indicated in previous discussions about this problem, I'm very reluctant to introduce a delay in the TV Server tuning process for various reasons. A fixed delay is an effective solution, but rather crude. A variable delay is more desirable, but very complex (if not impossible) to implement within the current code framework.
4. A month or two ago I updated the library that MediaPortal and TV Server use for RTSP streaming. This is the first update of that particular component in more than 5 years. The update hasn't yet been integrated into a release, but I am able to provide patches.
Until today I had assumed the problem had been resolved. After all, I haven't seen any reports or references for years. Of course problems don't just solve themselves. My standing assumption would be that the problem was eliminated by the steady stream of TsReader improvements that have come in since the posts you referred to.
That brings me to another point: I had always thought that this problem was a MediaPortal (TsReader) problem rather than a TV Server problem. I couldn't and still can't prove that; it's just an informed opinion. However, your successful test with VLC seems to support that opinion. I mean: if the problem was with TV Server or the files/streams it produces, surely VLC wouldn't work.
As a logical consequence of the above opinions, I'm inclined to think that the problem you're experiencing is due to limitations or issues in OpenELEC rather than an issue in TV Server. Again, I have no way to prove that except the circumstantial evidence which is outlined above.
What can I do to help?
Unfortunately I have zero experience with OpenELEC. That means I'm unable to help debug the problem properly.
The only thing I feel able to do for you that might help is point you to the streaming server patch that I mentioned in (4) above. It's here:
https://forum.team-mediaportal.com/t...tively-refused-it.129178/page-10#post-1148645
To install it, stop the TV Server, download and extract the replacement DLL into your TV Server install directory (don't forget to take a backup first!) then start the TV Server again. I have no idea if it will help or not.
You mentioned that you think the solution would be to add a delay in the TV Server tuning process. Considering that TV Server currently doesn't appear to be at fault, I'm currently not willing to do that. Sorry. It isn't my intention to frustrate. Note that if you're really desperate, you may be able to hack the code for yourself. I already provided directions here:
https://forum.team-mediaportal.com/t...vr-colossus-support.94850/page-30#post-747830
The relevant code is here:
https://github.com/MediaPortal/Medi...ions/Analog/Graphs/HDPVR/HDPVRChannel.cs#L149
Ha - that'll teach me to be careful with the words I use!My use of the word hostile was stolen from a conversation you had back in 2011
Hmmm, very interesting. I think you've understood perfectly. Honestly, I don't recall the issue affecting recordings in the past, but that's a detail that I could be wrong about. I accept your report.This point confused me a bit...
I'd expect KODI is using common code across platforms, so if one platform fails they'll all fail. It's surprising really because in DVB TV streams (that KODI is presumably expected to be able to handle) it is possible for the stream configuration to change at any time. For example, some channels in the UK change audio stream channel count during the advertisements, or randomly change between progressive and interlaced video stream encoding. Anyhow, I'd appreciate it if you could report your situation and results to the KODI team [if you haven't done it already] so they can decide whether they consider the behaviour to be an issue worth fixing.However, I would counter that it's not just one client/software - in my testing the client/software that can't handle the change in encoding is any of the Kodi installations across mac, openelec, ipad and droid, goodplayer on droid.
Wow, that's a bit of a surprise too. Again I'd encourage you to report this. The VR guys really seem to know their stuff. My guess is that their code can handle changing audio configurations... but not a change in encoding on a single PID (stream). Really, Hauppauge would have been better to use one stream/PID for AAC audio and another for AC3/DD audio.In fact, video redo crashes when I try and skip across the change in encoding format too.
Yes, that's a fair point.So while MediaPortal single seat setup my handle it perfectly fine, given that we're in a world where media is enjoyed in many locations across many clients - I hope that we can work through a solution that gives works for the majority (?) that can't handle mid-stream changes in encoding.
C# is quite different to Basic... so please let me know if the attached patch works.My hope is that my additional reasoning here will lead to your working your magic on a crude delay patch. I'm going to look at your instructions, but to be honest, I'm almost certainly going to be out of my depth. I'm an economist by day and haven't coded a line in my life (unless your code resembles Commodore 64 basic from about 1985?)