New DVB Backend wich is userfriendly and is only 2MB?!?

[cirion]

Yesterday I started playing with a backend I found in the XBMC forums.

It's only a 2MB .deb file and runs happily on my 8.10 testserver.

For me it does what VDR and MythTV never made possible... I never get them patched correctly and in the end I just run in to dependency issues. All I want is to see the channels I pay for on all my MD's.

It has a webgui for setting up, and it's just way to easy to set up... I started thinking "is this a joke?".

With only 2MB, it is able to detect my DVB card (i had installed S2API drivers), and it was identified as a DVB-C card and I could select cable providers. I selected Norway, Canal Digital. And it added a few transponders.  It did not have all (like any other software I have seen) but then things started happening in the monitoring window. It was scanning the transponders automatically, auto detecting new transponders and after 1 minute or so it had a complete list of channels and transponders?!?

On the same page there is a Probe button to add channels to the frontends, and when pressing that, it tested every channel for open or encrypted and added only those channels which where open to the frontends.

It even has a integrated Card Client, and I added my Dreambox, which is where my subscription card is located. Now I did a new probe, and it added also the channels I had on that subscription card

The EPG started filling up automatically on all channels and I could record, shows.

All this was done in the webgui. Recordings are done in MKV and play on LinuxMCE.
It even supports recording of my HD channels.

So can this be uses for LinuxMCE? The backend seems complete, the webgui is done, easy to install, no patching, works out of the box... But what about the MD's? Well, it's a backend that streams to any frontend...

It uses HTSP:
http://trac.lonelycoder.com/hts/wiki/htsp

Since there is no support for that on LinuxMCE I have tested it on XBMC and I must say I'm impressed! It works
Just added a URL and that was it, I could browse channels and start them. Pause also work.
Since XBMC has VDPAU support the streams where not using much CPU.
On SD channels there was no difference when VDPAU was off, but HD channels where unwatchable without VDPAU.

The backend is called TVHEADEND and you can find it here:
http://www.lonelycoder.com/hts/

The online manual is good and you find it here:
http://www.lonelycoder.com/hts/tvheadend/docs/2_1/index.html

Lonelycoder also makes a frontend, but I did not like that much.

Does anyone know it the HTSP streams can be added easily to LinuxMCE?

[toppot]

Yesterday I started playing with a backend I found in the XBMC forums.

It's only a 2MB .deb file and runs happily on my 8.10 testserver.

For me it does what VDR and MythTV never made possible... I never get them patched correctly and in the end I just run in to dependency issues. All I want is to see the channels I pay for on all my MD's.

It has a webgui for setting up, and it's just way to easy to set up... I started thinking "is this a joke?".

With only 2MB, it is able to detect my DVB card (i had installed S2API drivers), and it was identified as a DVB-C card and I could select cable providers. I selected Norway, Canal Digital. And it added a few transponders.  It did not have all (like any other software I have seen) but then things started happening in the monitoring window. It was scanning the transponders automatically, auto detecting new transponders and after 1 minute or so it had a complete list of channels and transponders?!?

On the same page there is a Probe button to add channels to the frontends, and when pressing that, it tested every channel for open or encrypted and added only those channels which where open to the frontends.

It even has a integrated Card Client, and I added my Dreambox, which is where my subscription card is located. Now I did a new probe, and it added also the channels I had on that subscription card

The EPG started filling up automatically on all channels and I could record, shows.

All this was done in the webgui. Recordings are done in MKV and play on LinuxMCE.
It even supports recording of my HD channels.

So can this be uses for LinuxMCE? The backend seems complete, the webgui is done, easy to install, no patching, works out of the box... But what about the MD's? Well, it's a backend that streams to any frontend...

It uses HTSP:
http://trac.lonelycoder.com/hts/wiki/htsp

Since there is no support for that on LinuxMCE I have tested it on XBMC and I must say I'm impressed! It works
Just added a URL and that was it, I could browse channels and start them. Pause also work.
Since XBMC has VDPAU support the streams where not using much CPU.
On SD channels there was no difference when VDPAU was off, but HD channels where unwatchable without VDPAU.

The backend is called TVHEADEND and you can find it here:
http://www.lonelycoder.com/hts/

The online manual is good and you find it here:
http://www.lonelycoder.com/hts/tvheadend/docs/2_1/index.html

Lonelycoder also makes a frontend, but I did not like that much.

Does anyone know it the HTSP streams can be added easily to LinuxMCE?

It is quite sad this got so old, and nobody responded... So many things are possible with LMCE, few fundamental changes (like the ones you suggest) are "easy", but to elaborate a bit further this is still worth considering - and i would like to outline 3 possible ways to move forward:

1. TvHeadend will stream via http. You can play that with anything, and have it leaving the recordings in a share that LMCE can access is something many users does already - for reasons I will elaborate a bot on below.. Further, you can easily watch liveTV through a http link. Just create a scenario that launches VLC or Xine or whatever... You would then need 1 scenario per channel.. Not optimal, but in some scenarios doable..

2. You could also use TvHeadend, and compile to HTSP addon to VLC - that would allow you to control the live streams from VLC, and having your LMCE remote mapping the proper keys.. Channel lists, channel groups are nicely integrated - EPG not so..

3. The whole project is open source - deep integration is very possible, both for retrieving the EPG, controlling the playback (if LMCE controls the EPG/Channel data, then HTTP is no setback compared to the HTSP protocol - unless the network is VERY congested), thus no work needs to be done with regards to interpreting HTSP... It is just a matter of dynamically generating an URL for Xine or VLC.. and BOOM, playback is controlled

1. and 2. can easily be done by everybody... option 3. will probably newer be done, since Thom is quite happy with MythTV... It is sad, since especially for Europeans TvH is a MUCH better alternative. I would even argue that in the long run it would be beneficial for LMCE to include support for TvH for 3 reasons:
a. It is lightweight - just like VDR
b. Is it a backend project, that for the sake of completeness also has it own frontend.. But the interest clearly lies with developing the backend. And since LMCE should always be the frontend, this is exactly what we should desire.
c. It works! Also with completely legal descrambling... (and yes I am aware of all the potential issues regarding descrambling, primarily that illegal descrambling is almost easier to get working than legal descrambling. But I don't care, I am not the police - neither is Thom, and his point of view is actually quite unamerican...

Come to think of it... Deep integration is actually just a matter of ripping the EPG info from TvHeadend - It stores it in memory, but present it both through HTSP and via its webinterface - meaning a crude solution could be to just grab it from the webif and import it in to LMCEs database.. And then on top of that, making sure that for live TV LMCE can generate an url that is properly formatted... This is actually not that hard, perhaps I should do it...

I probably won't - simply don't have the time atm., and neither the urge.. I have newer been close to using LMCE for live TV. Zapping times alone (MythTV) has always been the spoiler... And Stability in Myth (the version used in LMCE) is actually also poor. On top of that comes the poor descrambling support..). I newer got anything useful out of VDR, that is so poorly integrated that I would newer recommend it... And then I suddenly remember: Timeshift is only supported via HTSP.... And I am not going back to not having timeshift again... And I don't believe it is "just" playing a HTSP url via VLC (with the HTSP plugin compiled for it) is all that is needed, so I probably won't do it.... (just fiddled with VLC - it actually IS that easy to use HTSP via the console...)

But perhaps somebody else will fight Thom for this.. I will gladly support with DVB specific info if needed, and/or testing...

[tschak909]

Fight? Really? is that what you all think of me? I'm disgusted.

-Thom

edited:

I've been busting my ass on this project, since day one, coding, trying to make things work, making tutorials, HOSTING CONFERENCED WORKSHOPS, JUST SO MOTIVATED PEOPLE WOULD ACTUALLY TRY TO CONTRIBUTE!

If this is how I am seen, and treated, then I have nothing more for all of you users, all whiny, waaah waaah WAAAH!

This system is modular! IF YOU WANT A FEATURE, ADD IT! ASK ME HOW THE SYSTEM WORKS! PARTICIPATE! I WILL ALWAYS TEACH! ASK ANYONE WHO HAS ACTUALLY CONTRIBUTED ANYTHING TO THIS PROJECT, AND THEY WILL TELL YOU THE SAME!

toppot, it is taking every bit of my tongue right now to keep from completely unloading on you for your last statement, it is uncalled for, and I am _DEEPLY_ offended.

-Thom

[golgoj4]

Some of us just stopped being snarky and started coding. So choose which kinda person you are.


Good Day sir.

[toppot]

Fight? Really? is that what you all think of me? I'm disgusted.

-Thom

Glad I was able to provoke a bit….
This is most likely NOT how “all of us” think of you… Actually, what you got from my post is not even how I think of you. The provocation was intended, but believe me when I openly state that I am a very big fan of you, and deeply impressed with all the hard work you (and others, none mentioned, none forgot – but this whole project lives and dies with you 5-7 die hards that have been putting an incredible effort in to the system) have lend to this project over the years..

edited:

I've been busting my ass on this project, since day one, coding, trying to make things work, making tutorials, HOSTING CONFERENCED WORKSHOPS, JUST SO MOTIVATED PEOPLE WOULD ACTUALLY TRY TO CONTRIBUTE!

If this is how I am seen, and treated, then I have nothing more for all of you users, all whiny, waaah waaah WAAAH!

Would be a loss for all..

This system is modular! IF YOU WANT A FEATURE, ADD IT! ASK ME HOW THE SYSTEM WORKS! PARTICIPATE! I WILL ALWAYS TEACH! ASK ANYONE WHO HAS ACTUALLY CONTRIBUTED ANYTHING TO THIS PROJECT, AND THEY WILL TELL YOU THE SAME!

hmm, over the years I could probably find a soul or 2 being put off by harsh comments (because they see the world slightly different than you) - but you are obviously right that if you only look a those that managed to get code folded in to the project, they will all agree that you are most helpful.

toppot, it is taking every bit of my tongue right now to keep from completely unloading on you for your last statement, it is uncalled for, and I am _DEEPLY_ offended.

-Thom

You shouldn't be offended: Try reading again... When I state "option 3. will probably newer be done, since Thom is quite happy with MythTV..." all I am really saying is, that I recognize the fact that you carry a tremendous load in this project, and if you don't do it - it will most likely never be done, simple as that!

That said #####################################################################################################
###########################################################################################################
##################################### CENSORED ##############################################################
###########################################################################################################
Long piece elaborating on the way this project work with references to historic threads ending in utter flame wars - I censored this, since this is the last I desire to accomplish ############################################################################################################
############################################################################################################
############################################################################################################
#########################

To get back on track!
Even with a modular system like LMCE, sometimes choices becomes exclusive - either for technical reasons or due to the desire we all should have NOT to fork or diverge the effort towards a stable and usable system. In some of these debates you have been ... Lets just say firm  in some of these situations. On many occasions with the sound reasoning behind that you are one of the few pillars that will continue working on the system, meaning a choice that is favoring a developer that is "in and out" in 6 months, or half a release cycle, could potentially be harmful in the long run... But sometimes good ideas have been shot down, even if they actually ARE good...

So to be very concrete:

1. Apologies if I offended you (or any other of you guys doing so much for the community). That was unintentional!!
2. The provocation was intentional. Now that I have your ears:
2.a.: Could you give some advice to get descrambling (legally - obviously) working under Myth? We are not all living in US, so some of us rely on subscriptions to scrambled DVB signals for TV... (OK, ignore this - it was just another tease...)
2.b.: You actually managed to get me provoked. I would like to integrate the TvHeadend backend server in to LMCE. I am willing to put the programming effort in to this. It is, conceptually, dead easy. The package must be compiled for 10.04. The VLC HTSP plugin compiled for 10.04. The controls should be controlled via the VLC console - all should be small changes/updates to the controls for VDR, right? (this actually support absolute positioning in playback of live TV very nicely - pulling out "get_length" and using "seek X" to position...
All I would need to do further is to enable hooks to grab the EPG data from TvH in to LMCE - and then code a bit to generate proper commands to schedule recordings and control VLC...

Can you point me a bit in the right direction?? And how do you feel about adding this "fork"/module?? Are you all for catering for us Europeans, or would you prefer sticking to Myth (and the somewhat stale VDR integration)? EDIT: Just read i bit in old threads.. I realized that you Thom actually do not favor Myth.. Just that you would start working on an alternative route. Has some of this work started?? If not, then I volunteer to this. My suggestion is not to take an integrated system like Myth and bolt that on to LMCE (heavy and slow), nor taking a single-system like VDR and try to force it to be server-client like... Rather: a backend that wants to be a backend (like TvH) and a frontend that aims to be exactly that... could be VLC (du to the HTSP integration) or Xine, Mplayer - whatever...

I am a strong believer in adhering to the design concept of the framework - meaning it should be possible to use this independent of how a implementation is made, or how (orbiters etc.) it is controlled..

With sincere hopes that I didn't ruin your day - but also with hopes for an open debate on this specific topic. Luckily I CAN code - I realize that otherwise you are not very welcome to make suggestions in this project. Contribution is in my book not just getting lines of code in the repo - but there we just differ...

BR

-Toppot

[Marie.O]

Things that need to be done for another media provider:

* Create a plugin for it, which provides data to LinuxMCEs data grid providers
** Create the datagrid provider for EPG
** Create the datagrid provider for recordings (probably not needed when you directly store in the video subdirs)
** Create the datagrid provider for timer
* Create a player which talks to the plugin, and controls the frontend for the media provider.
* Create frontend screens for media playback control, including pages to add/modify/delete timers and view EPG data.
* Create setup pages for the 'Sarah Wizard' to setup devices and channel configuration

You can probably use the VDR and VDRPlugin codes quite verbatim as a good starting point.

I actually use VDR within LinuxMCE, and agree that the MythTV integration is much tighter.

Our current development is happening for 1204 in svn/trunk.

[toppot]

Much appreciated feedback!

Things that need to be done for another media provider:

* Create a plugin for it, which provides data to LinuxMCEs data grid providers

Check - part of 1. priority (needed to get anything working)

** Create the datagrid provider for EPG

Same as above..

** Create the datagrid provider for recordings (probably not needed when you directly store in the video subdirs)

Right - but would you recommend to just dump the recordings and let LMCE discover them, or would you (from design policies) recommend to add them directly in the DB??

** Create the datagrid provider for timer

Should be pure copy paste - and part of 1. priority

* Create a player which talks to the plugin, and controls the frontend for the media provider.

Some work needs to be done here - as far as I remember it is already possible to use VLC for media playback? Would you recommend sticking to a generic implementation allowing for selection between Xine and VLC (or mplayer), thus limiting functionality (timeshift)? Or would it be an OK approach to require VLC as the frontend player?
Actually not sure what you mean by "create a player"... I should just ensure that the controls from the plugin can work with something lightweight like VLC or another off the shelve player, right?

* Create frontend screens for media playback control, including pages to add/modify/delete timers and view EPG data.

Copy paste from existing frontend screens, unless I  have completely misses the point of the modular design??? It is via the plugin that pressing the button ensures that the proper thing will happen in the router??? Or??

* Create setup pages for the 'Sarah Wizard' to setup devices and channel configuration

Really??? The current implementation of this is almost never plug&play, unless you live in US, and subscribe to one of the supported XMLTV providers... Raise of hands: I there a single user of TV in LMCE that hasn't been using either the backend setup of Myth or VDR??? I doubt so - this is clearly prio 5... But more important - it should be possible to choose TvH/VLC as the TV provider both during setup and in the webinterface..

You can probably use the VDR and VDRPlugin codes quite verbatim as a good starting point.

I also plan to do so... Only caveat here is, that I actually never got a really working setup out of neither VDR nor Myth... But I have seen moving images (brief, oftens pixelated - and crashing after a short while, usually before the lag from the slow Myth channel change was over 

I actually use VDR within LinuxMCE, and agree that the MythTV integration is much tighter.

Are you satisfied with the current state of the VDR implementation? Or put in other words, in no way trying to be "smart": Does it work well, or is it (more or less) a daily source of frustration? Feel free to elaborate, I am not venturing in to this, unless I have a strong feeling that I can create an improvement...

Our current development is happening for 1204 in svn/trunk.

Then I will work from there...
One question remains: Are anybody else working on TV atm. ?? And should there be more than one backend available (remembering how Thom aired frustration over the Myth team, and how some of their design decisions makes our lives difficult - he used the words: create a new solution for TV altogether)??

In my book, the virtue of LMCE is NOT developing things from scratch - it is integrating existing solutions i an extremely clever way... Unless there is nothing that can solve a problem for us, then we can be forced to develop..

-Toppot

[Marie.O]

I am okay with VDR. SWMBO uses it, and only complains when I stupid enough to update stuff. The VDR implementation itself rarely has to do with the frustration, but that doesn't matter.

The goal for a new TV media provider must be to be better than the existing ones. This means, setup must be possible from Sarah, timer screens must exist, and EPG screens must exist.

[toppot]

IThe goal for a new TV media provider must be to be better than the existing ones. This means, setup must be possible from Sarah, timer screens must exist, and EPG screens must exist.

Actually, to venture in to this, my personal goals must be a tad more ambitious. For me the experience from daily usage is by factor 2000 more important than ease of setup etc... That is exactly my biggest problem with how TV is working currently. Easy to setup (not even that for 95% of users), distress to use (probably true for 80%+ of users)... My order of priority:

1. Flawless LiveTV viewing. (I really mean flawless!!!! Like you would have on a highend STB). Subtitles, audio language and audio track selection should "just work".
2. Intergration of EPG - and timer management (including repeated and series recording).
3. Playback of recording (this should work OOB, but please see #6). Also the pre-selection of Subs and audio should persist.
---------- Cut line, until we are here it should never be released!!! --------------
4. Flawless Timeshift (probably meaning that LiveTV will be through VLC).
5. Integration of TvH management in the LMCE web-admin
6. Commercial detection (using Comskip for Linux, and ensuring player will use cutlist - I know Mplayer support this, but not if it is enabled - not sure about VLC). The design question this opens is, should we just dump the recordings for LMCE to figure out, or use the HTSP transport from TvH... I still await some feedback to earlier questions before I can form my opinion here.. But I lean towards letting LMCE do the work, granted cutlist will be considered either through setting or as standard..
7. Ease of setup through the Sarah-Wizard
8. Integration of Comskip management in the LMCE web-admin
..
..
10012. Transcoding of LiveTV for playback on Orbiters (personally I believe playback on an Orbiter is misunderstood... I prefer splitting between Orbiters, and Tablet/Phone MDs with on-screen orbiter, although they should be a special case of the MD concept, since they will be a frontend application rather than a unit controlled by the Core)
..
..
15003. Automatic installation of new TV adapters, including transponder/channel search - EPG mapping etc. etc... I don't see the need to enter a separate web-admin and controlling TvH that way as a big issue, but for the ultimate Plug&Play experience I will keep it on the list
..

But please - I could have some priorities skewed here - provide your input to how priorities should be... The more opinions I get here, the better..

-Toppot

[toppot]

Our current development is happening for 1204 in svn/trunk.

Never got any input in to if anybody is working on TV in 1204... But looking around in svn revealed a bit for me...

There's no trace of Myth (neither in branches 1004 or trunk) - it does exist in 810.. odd..
In /external I can find "VDR.Oct-10-2005/"... Will consider that not to be under heavy development??? (The current version of VDR is 2.0.2, but it is still from the design concept meant to be a stand alone application, and only support the server-client needs we have via plugins - to the best of my knowledge).

So input on what's going on, if anything, would be great...

-Toppot

[Marie.O]

We take the verbatim VDR packages. No need to modify them.

[golgoj4]

I am an F-B-I Agent....


oh wait wrong thread, proceed.

[hari]

@toppot: just hot air from you, as I did expect..

[toppot]

@toppot: just hot air from you, as I did expect..

Good relevant input. Really pointing me in the proper direction with regards to which design choices to make in this early stages... And obviously, something that has been "broken"/stale since 2004 can't take more than 10 days to solve...

Nice also to see the detailed roadmap and milestones plan in the same link... Meaning everybody now knows what we are working towards, and have a proper sense on the urgency towards relevant dependencies..

Super :-)

[jamo]

@toppot: just hot air from you, as I did expect..

Even if it turns out to just be that.... I for one appreciate the effort.
Discussions like this teach me a huge amount about the system (not to mention the personalities involved  ) and that is really valuable. It gives me keywords to google, search for, helps me to dig in the code and to understand the architecture. If someone is looking at doing something similar then searching the forum is always a good place to start and if some of the groundwork has been done that can only help.

@toppot - keep at it and I hope you are able to get something going here. I'm sure it will be of great value if you are and I hope to be able to help in some way.