Setting up Twinhan 102g with new release

[sub]

It WORKSSSSSSSSSSS. Thank you so much guys. some sanity after a long time. I was going crazy abt this. I switched thru all the channels and it has not crashed. I guess I should have asked Ashwin abt this longtime back. I bet he is using 1.4 version hence he didn't c any crashes.

Lets move on to some more testing and make gbpvr THE BEST software for DVB period.

Thanks Sub.

Great!

[Stevie77]

Aragwal's API only allows me to pass the one caid and one ecmid.

I'm so attached to the multidec plugins now though, that if we need this and he cant add it in the near future, I'll implement support for calling the multidec API directly.

Excuse me! It's just that I am soooo happy! I was waiting for md-plugin support in gppvr for years! It was the only very important feature to me that was missing within your great software.

Over the years I tested a lot of DVB-applications. However you decide to improve the api, I think you will have to define at least ecm and/or caid somewhere statically or let the plugins do the work via pmt-pid.

[sub]

Over the years I tested a lot of DVB-applications. However you decide to improve the api, I think you will have to define at least ecm and/or caid somewhere statically or let the plugins do the work via pmt-pid.

My PSI parser does get all this PMT info within a couple of hundred milliseconds of attempting to change channels, including multiple ECM etc, but unfortunately no where to pass them currently.

That said I may look at a cache of some type (registry?), that could be submitted immediately to minimize any delay, and when the real info is parsed a few hundred milliseconds later, if it doesnt match the cached copy, it could be sent then updated in the cache. I'd have to check whether it'd really speed things up before bothering to implement though... (the info is available in the transport stream very quickly)

[Stevie77]

I'm not concerned about the speed of your excellent parser (working great here!). In my opinion it should be possible to manually define (choose) ecm/ca values that get hand out to the plugin.

vplug is able to choose working values automatically by itself. So the decission about right values could also be left to this plugin. Though it's not the only one MD-plugin ....

At this point i may have to explain what it's all about:

I receive one programme here, which is scrambled in irdeto AND cryptoworks AND nagra2. gbpvr passes mdapifilter to use irdeto because it's the first transmitted ca-value within TS. I never get the chance to choose one of the others. This could become very disappointing if I have no keys for system 1 but keys for system 2 or 3. The screen will stay dark though I have good keys!
I receive some other programmes which stay dark ATM, because the "wrong" (first) ca/ecm is choosen by default.

[ashwincr]

some updates from me. I have got Diseqc working. Currently scanned in 4 satellites and working beautifully! if the problem of unknown channels during scan can be fixed, this will be a fantastic all in one media center.
I read through some questions in the forums, and I understand that ffdshow post processing will not work with Nvidia codecs. is this still true? I couldn't get it to work just by checking enable ffdshow post processing in GBPVR configuration

[sub]

I read through some questions in the forums, and I understand that ffdshow post processing will not work with Nvidia codecs.

I dont remember reading that. ffdshow can be quirky and cause problems for some users, but it should be able to at least do this wi the nvidia decoder.

I couldn't get it to work just by checking enable ffdshow post processing in GBPVR configuration

You probably need to go into the ffdshow configuration (filter property pages) and enable it first, or something similar. The easiest think to use to check its going is enable the OSD which puts green text in the top left of the screen.

[spyder40]

Sub/Ashwincr,

Excellent thread everyone! Don't want to hijack/off topic this thread but ashwincr's success brings a question. Does a dp44 work the same as a Diseqc switch from the driver standpoint? I briefly looked at one when installing my stuff (2 dishes 110/119 and 91 wired to 2 locations in anticipation of GBPVR functionality and a Viewsat Ultra) but went with 4x1 Chieta diseqc switches and multiple cables instead due to cost/requirement for power on port 1. I was also hoping to get a hardware based card but with the success of the 102g that looks like the way to go. How much resource (cpu/memory) is being chewed up? Time to search fleabay....

[sub]

I havnt had much exposure to different diseqc switches, and dont know what a dp44 is so cant really answer that question. I've only ever tried my simple four way diseqc switch.

[ashwincr]

i've a diseqc switch with 4 satellites coming into my home. I use a twinhan 102G and in gbpvr, set up four BDA recording sources each using a different diseqc setting with slightly different LOF settings for each satellite. and I can change channels seamlessly inside gbpvr and seems to be working fine. the only issue I am having is during scanning a lot of channels come in as unknown, and so i cannot manage my channel database very well.
sub, I know ffdshow is not being chained to the graph, because I added the registry setting so that i can connect to the remote graph, and I can see that ffdshow is not being used. In ffdshow I have enabled all filters including mpeg4 and raw video filters, but gbpvr doesn't see it. I even reinstalled ffdshow and tried adding it manually to the lastbdagraph, but it reverted back to the old one.

[sub]

sub, I know ffdshow is not being chained to the graph, because I added the registry setting so that i can connect to the remote graph, and I can see that ffdshow is not being used. In ffdshow I have enabled all filters including mpeg4 and raw video filters, but gbpvr doesn't see it. I even reinstalled ffdshow and tried adding it manually to the lastbdagraph, but it reverted back to the old one.

I'm not really sure. It works ok here. Zip and attach the logs and I'll take a look.