Moderator: Forum Moderators
Unfortunately I had not much luck. After having found out the playlist-format (by reading the source of HLSReader and M3U) i threw together a small php-script that generates a compatible playlist on the fly. With this I got really close, but could not get a fronted to "LiveTV" a stream from the box. It looks like there is no data arriving in the reader. (It always says "Reading 0 of 0 bytes").
Here is the used meta-playlist: https://gist.github.com/Shyru/f5573dcd9 ... aylist-m3u
And here the used m3u8-playlist for the channel 101: https://gist.github.com/Shyru/f5573dcd9 ... ylist-m3u8
And this is the log-output from mythbackend: https://gist.github.com/Shyru/f5573dcd9d48555a143e
I'm using mythubuntu with mythtv v0.27-193-g8ee257c.
Has anyone got a clue what could be wrong? I tried to load the stream url from the log with wget and it worked (endless download) and vlc played it happily. So the Digibit R1 seems to work correctly. But what goes wrong in the HLS code that recording does not work? Is there anything that I missed?
Any help would be greatly appreciated!
Then adjust your main playlist to look like this, without fake m3u8 playlists.
Code: Select all
#EXTM3U #EXTINF:0,101 - Das Erste http://192.168.100.82/?src=1&freq=11837&sr=27500&pol=h&msys=dvbs&pids=0,100,101,102,103,106
Code: Select all
#EXTM3U #EXTINF:0,101 - Das Erste #EXTVLCOPT:program=28106 http://192.168.100.82/?src=1&freq=11837&sr=27500&pol=h&msys=dvbs&pids=0,100,101,102,103,106
Code: Select all
2015-01-18 13:34:24.552554 E DTVSigMon(): Program #10301 not found in PAT! Program Association Section PSIP tableID(0x0) length(37) extension(0x44d) version(3) current(1) section(0) last_section(0) tsid(1101) programCount(7) program number 0 has PID 0x0010 program number 28106 has PID 0x0064 program number 28107 has PID 0x00c8 program number 28108 has PID 0x012c program number 28110 has PID 0x01f4 program number 28111 has PID 0x0258 program number 28113 has PID 0x0320 2015-01-18 13:34:25.030160 E MPEGStream(0x7f64f401a340): ProcessPAT: Program not found in PAT. Rescan your transports. 2015-01-18 13:34:25.030175 E MPEGStream(0x7f64f401a340): Desired program #10301 not found in PAT. Cannot create single program PAT.
It appears as if you edited the URL by hand, setting service_id and PIDs, but didn't change the frequency, symbol rate and modulation system to the other transponder. So you ended up looking for a service from transponder B on transponder A.Shyru wrote:However trying out Das Erste HD (which has Service-ID 10301) did not work. I got the following log output:
http://192.168.100.82/?src=1&freq=[b]11 ... ,5101,5106
(that example is an untested educated guess only, i don't know how to signal DVB-S2 etc)
Code: Select all
2015-01-18 13:34:24.313399 I Loaded [data]http://192.168.100.82/?src=1&freq=11494&sr=22000&pol=h&msys=dvbs2&pids=0,5100,5101,5102,5103,5106[fectype][fec0][fec1] for 1001
I came across https://github.com/catalinii/minisatip today, which claims to be a SatIP server that exposes a Linux DVB device - it might be useful as a dev reference. An alternative approach seems to be trying to connect the SatIP device using vtuner - https://code.google.com/p/satip/ claims to do that, although I've not got it compiling cleanly yet.
I did take a look at the code with a view to developing a SatIP recorder - it looked like there'd need to be a reasonable amount of refactoring of the current DVB driver, to allow elements of it to be common regardless of whether the underlying device is network based or local DVB - I suspect it would be too ambitious a project for a beginner on the codebase (especially one like me without a background in C++). Implementing MPEG2TS over HTTP would probably be a lot easier (but would still need care around the HLS code). I think a dedicated recorder type would be needed to take advantage of Multirec, and definitely would for channel scanning (which you need if the channel list on the device is incorrect).
As a gotcha for anyone who does attempt it - the "pids=all" element of the spec does not appear to work for my device - it passes no data unless you specify a definitive list of PIDs.
In the short term, a work around may be set up VLC to relay the stream from HTTP to RTP as an on-demand feed.
So far I've only tested using the scan command, so I can't confirm it works properly within Myth. The dvb-utils scan command does seem to work properly though, so I'm hopeful that it has a good chance of working with Myth.
Update: Having now tried setting up a slave backend for this, you can scan for channels successfully, but the vtunerc driver always reports a zero signal strength, and so Myth never detects a tuning lock.
It depends on what you want to do.wurststulle wrote:is there any progress on this?
Recording via the M3U Playlist and HTTP works on master. (see #12168) The HTTP recorder part has not yet been ported to fixes/0.27 but is available in devel/027candidates for testing. Just needs someone to find the time and test it (with channel changing in LiveTV) so it can be merged.
Also recording via the M3U playlist and RTSP/RTP should work on master and fixes/0.27.
As far as I know no work has been done towards a proper SAT>IP recorder with SSDP autodiscovery, DiSEqC configuration, channel scan, etc.
If you want to play with this stuff, other example applications including a SAT>IP server are available, too.
It looks like quite a bit of work has been put into this but I've had a hard time getting anything much to work. Too little information is output if something is wrong. For example, satipscan can take up to 31 parameters but any mistake merely prints the usage info. There's no clue at all about which parameter, or parameter combination is incorrect or missing.anstein wrote:To be able to do so I had to write proper 'middleware' in form of a CUSE based virtual DVB device daemon. You can find the code at https://github.com/not1337/libsatip
It's also really hard to understand the code because there are no comments and too many single letter variables.
I've no doubt made a mistake somewhere with the configuration file(s) but even spending some time adding debug output and single-stepping hasn't let me figure out exactly what the problem is, and that's already way more effort than I'm guessing most people are going to invest.
Like I said, you've obviously put a significant amount of work into the functionality, but the error reporting needs improving because it just seems to be too hard to set up at the moment without understanding the code in detail.
Now I want to add my SAT>IP (OctopusNET http://www.digital-devices.eu/shop/de/n ... zwerktuner) to mythTV. How all write here in the post, I also try to generate a m3u like this
Code: Select all
#EXTM3U #EXTINF:0,103 - Das Erste HD #EXTMYTHTV:xmltvid=DasErsteHD.de rtsp://10.10.9.111/?src=1&freq=11494&pol=h&msys=dvbs2&sr=22000&pids=0,5100,5101,5104,5102,5103,5106,5101
I've been a MythTV user for 9 years and recently got a SAT>IP box in aid of simplifying my setup, removing the need of internal cards. I've tried moving over to Tvheadend but it is a bit too limiting due to the lack of recording groups (which I use to separate ours and the kids TV shows). The Kodi support for MythTV is a lot better than Tvheadend for my needs, top level directories (recording groups) and artwork, not to mention the recording rules... You don't know what you need until it's gone
From my point of view integrating most of anstein's library code into a proper form in MythTV itself (because a library or daemon that is not supported by most distros is practically useless) might be the best way to go. Also having a look at the VDR Sat>IP Plugin might help if something is unclear. Unfortunately I don't have a Sat>IP Server as of yet and since this is not a pressing matter for me yet, it'll most likely take a while before I can get myself to dig through all the code. Maybe in the meantime someone else has done it.. there has already been a lot of stuff done, but it seems clear that most of it is incomplete and just puzzle pieces.. would be great if people pooled their resources to the benefit of all (you know OSS n' stuff) .