Jump to content


malakudi

Member Since 23 Feb 2010
Offline Last Active 04 Oct 2020 06:45
-----

#1259125 SAT>IP support improvements

Posted by malakudi on 30 September 2020 - 16:22

Hello everyone, I've been away from this forum for quite a while but I am happy to be back.

 

The last weeks I have been using sat>ip support in OpenPLI to receive signals from around the world (WAN access). Server used is the minisatip open source.

There are various issues trying to use such setup. Some of them can be fixed and I expect to send required patches in the next days or weeks. But some issues needs improvements in the enigma2 core.

 

- As it is, you can only "discover" LAN broadcasted SAT>IP servers from the enigma2 menu system. You are not allowed to enter the connection details yourself (unless you manually edit vtuner.conf). I am planning to fix this.

- While satipclient has been updated to understand the "tcpdata:1" option to enable TCP RTSP access (which really helps WAN setups), the enigma2 plugin does not. I am planning to fix this as well.

- satipclient has a fixed port for RTSP (554) and it is not configurable. I am planning to add a "port:XXX" option to also support different RTSP port for SAT>IP server

- DVB-T2 is not supported from satipclient, because it does not export "FE_CAN_2G_MODULATION" capability for the vtuner. This can be easily fixed.

 

Those are kind of easy fixes and I will send patches soon. More difficult issues include the following:

- You cannot have two (or more) DVB-T SAT>IP tuners from different areas (probably same applies to DVB-C but haven't tested it). For example I have one local SAT>IP DVB-T server for my local DVB-T channels and I also have another DVB-T SAT>IP server for the DVB-T channels in UK. As soon as I have both enabled, whenever I tune a DVB-T local channel, it tries to tune the UK server and fails. It never tries the second tuner, unless I setup a PiP or recording etc. I guess this needs a support inside enigma2, maybe to give IDs to DVB-T tuners and this tuner ID should follow the service scanned, in order to only tune to this ID. Something like "satellite position" for DVB-T.

- You can have a SAT>IP server that can have both DVB-T and DVB-S. But satipclient does not allow to have a "hybrid" tuner. Don't know if this is possible with vtuner, but since hybrid tuners are supported in enigma2 (C/T or even S/C/T) I think we could configure this.

- You can have SAT>IP server with 2 or 4 or 8 tuners (or more) but you can only use one in enigma2 (unless you setup a second connection to same server for a 2nd vtuner). I don't know if this can be fixed/improved somehow since the concept of one tuner = one frontend in enigma2 is very low level.

- Since most STBs have a limited amount of vtuners and satipclient does the assignment statically outside of enigma2, in order to add support for more SAT>IP server sessions maybe we need "on-demand" assignment of servers to vtuners. But this to work has to be managed inside enigma2 code.

 

In the era of fast internet, sharing DVB-T/DVB-C or DVB-S tuners over WAN with a friend is a great feature.

 

Waiting for your thoughts on this.

 




#540311 VPN throughput not sufficient for HD streaming

Posted by malakudi on 1 April 2016 - 19:21

Hardware	type		  16 bytes     64 bytes    256 bytes   1024 bytes   8192 bytes
vusolo		aes-128-cbc       3588.79k     4173.91k     4331.52k     4364.21k     4398.43k
et5x00		aes-128-cbc       3987.17k     4735.01k     4948.96k     5011.07k     5011.09k
vuultimo	aes-128-cbc       4023.21k     4752.94k     5005.40k     5082.67k     5101.51k
et6x00		aes-128-cbc       4062.41k     4810.82k     5032.19k     5098.50k     5120.00k
vuduo		aes-128-cbc       4098.91k     4844.15k     5015.83k     5117.15k     5131.38k
vuzero		aes-128-cbc       6970.90k     8203.70k     8433.32k     8610.53k     8607.10k
xpeedc		aes-128-cbc       7221.40k     8643.96k     9039.02k     9184.49k     9169.58k
et8x00		aes-128-cbc       7521.94k     8905.47k     9316.78k     9439.91k     9472.68k
osmini		aes-128-cbc       7610.81k     8956.64k     9312.99k     9532.51k     9513.20k
gigabluequad	aes-128-cbc      16434.86k    19588.33k    20610.75k    20902.66k    20944.58k
et10000		aes-128-cbc      18283.89k    20361.19k    20967.85k    21166.42k    21195.43k
Atom D525	aes-128-cbc      24423.47k    27090.75k    27799.89k    28091.85k    28090.37k
vusolo4k	aes-128-cbc      59945.33k    68150.09k    71901.95k    72968.19k    73277.44k
Xeon E3-1275 v5	aes-128-cbc    1284881.69k  1410752.96k  1469046.15k  1480765.10k  1485348.86k

Atom D525 (my router) and Xeon E3-1275 v5 are for reference :)




#540197 Openpli-5 (still next master)

Posted by malakudi on 1 April 2016 - 11:55

The libtalloc issue is waiting to be fixed upstream, till then a quick fix is:

cd meta-openembedded/meta-oe/files/waf-cross-answers

cp cross-answers-mips.txt cross-answers-mipsel.txt

gst 1.6.3 of master-next compiles fine here.




#537377 VU Uno: not able to record on Eurosport 2 HD | 13E | 11258H

Posted by malakudi on 17 March 2016 - 16:34

I think it is quite obvious: Eurosport has many ECM pids (for many different packages due to simulcrypt). Adding ECMs to the recording hits a limit in the drivers of max streams per demuxer. Multiple ECM pids + multiple audio pids triggers the issue.




#525950 e2 eDebugNoLock function.

Posted by malakudi on 19 January 2016 - 08:36

I think we should not go for commented eDebug lines and quickly apply a quick fix for something that is in there for years.... I think it is better to invest our energy to really solve this by adding debug levels.... So you can enable/dispable it e.g. with a parameter when enigma2 is executed... And put e.g. servicemp3.cpp eDebugs to a level that is off by default...
I thinkk a quick apply makes not many sense since we already commented out the most importee showstopper.... Or do we really have a big argument with the other ones?

 

After trying hard to follow all chrisstophe's posts, it seems the other two are also important (one for subtitles and the other I don't remember).

Of course the correct way of fixing this is to implement a whole new debug/trace system with levels etc. But applying a two line comment patch will not affect the maturity level of the code. Anyone with commit access, just apply it, is not a big deal.




#525878 e2 eDebugNoLock function.

Posted by malakudi on 18 January 2016 - 22:26

I have to say in favor of christophecvr that we are making a way big deal for a simple two line comment patch. We should skip this "bureaucracy" for such a simple patch. 




#516931 VU Solo 4k support

Posted by malakudi on 1 December 2015 - 18:27

What is wrong with our epg language selection???

 

Ours works fine. However, OpenATV picked the commits and imported them into their enigma2 codebase, and it doesn't work. It doesn't keep the setting after enigma2 restart.

 

edit: this was my report on the issue on OpenATV forum. http://www.opena.tv/...ng-restart.html

I also reported them another issue about DM800SE images, here: http://www.opena.tv/...l-flashing.html

No developer replied. I guess they either do not have developers or they don't watch the english section (openatv is mostly german speaking). After those two reports going completely unnoticed, I decided that I will not report anything else back.




#512836 New VU+ Solo 4K

Posted by malakudi on 8 November 2015 - 20:29

Try outputting to mpeg-ts file. It might be playable realtime. mp4 files can't be played back while they are recorded because stream metadata in mp4 files is in the start or end of the file, and in order to play it back full metadata is needed. mpeg-ts can be played from any position.

Of course, in order to work, ffmpeg should not read lock the written file. This remains to be seen, I haven't tried it.




#511074 VU Solo 4k support

Posted by malakudi on 28 October 2015 - 12:23

 

... If you replace your LNB's with especially suited ones, you even get four individual virtual tuners per physical tuner.

 

I've seen this comment before (I think from betacentauri). What is this "especially suited one" LNB? Full band LNB on coax? (so far I have only seen optical full band LNB). Something else? Is there a specification/product link somewhere?

 

 

To answer my own question, after some searching it seems that the twin FBC tuners can work as eight independent tuners when used with Unicable or JESS (aka Unicable 2) LNBs and/or multiswitches. The biggest limitation of Unicable is that it only support two satellite positions and although JESS (Unicable 2) have increased this limit to 64 satellites, there is no multiswitch in the market to support more than four (4) satellites.




#505298 Vu+ Zero & Rec ?

Posted by malakudi on 16 September 2015 - 15:29

The "democratic" way is the lobbists from MAFIAA paying money to politicians in order to promote their interests? That's no democracy my friend. You live in an illusion of democracy. Banks and big companies rule the world. There is no democracy, either in Europe or in USA. You fail to pay a tax, you might go to prison. Bankers destroy global economy? They get funding. Sure, this is democracy.

 

Anyway, we are way off topic and the forum doesn't like political discussion. So I end it here.




#505142 Vu+ Zero & Rec ?

Posted by malakudi on 15 September 2015 - 20:14

So, you suggest people in China to not try avoiding the internet filters their goverment enforces. Or you suggest that what Juliane Assange or Edward Snowden did was wrong. They broke laws in order to present to the public the bigger picture.

 

I will say it once more. People should (and have to) disobey laws that go against the "common sense of justice". The "common sense of justice" is not about my interests. It's what people believe is right and wrong. It is about ethics. And laws cannot change ethics.




#496533 Decided to solve the Greek Debt Crisis via crowdfund...

Posted by malakudi on 2 July 2015 - 15:41

This fund raising is stupid and insulting. First of all, Greek citizens don't owe anything to the IMF. The IMF program (and the EFSF program also) didn't give any money to Greek citizens. Those money were given to save the German and French banks in 2010-2011. Greece bankrupt in 2010, not now. Those programs only helped German and French banks to hide their exposure to Greek bonds. Those programs only helped those bankers to not register losses and keep getting their bonuses. The debt that bankers created, since they were the ones that kept buying bonds of a state that everyone knew it could not pay off those bonds. So, those programs converted banker losses to debt of Greece to the IMF, EFSF and through it, to several European countries.

 

Instead of cutting the value of those bonds to 50 or 60% (after all this was their value in exchange markets), and then maybe help those banks by giving them money, they chose to lend Greece money to pay those bonds, and in order to bring down the percentage of national debt vs GDP, they chose to cut up to 90% the bonds that the greek banks and greek insurance institutions were holding. And of course this led to even more problems - now IMF says you can't pay pentions because your insurance institutions don't have money! But it was their decision to cut the bonds to 90% - while paying the bonds of foreign banks.

 

I doubt these are known to our european friends. I guess what they know from their media is what Mr. Dijsselbloem or Mr. Schäuble say. That Greeks are lazy and corrupted and they want money for nothing. Money for nothing was given to bankers. There are people here that search the garbage to get something to eat. There is 30% unemployment, more than 50% in young people. There are people who commit suicide in their desperation.

 

Europe should either become the USE (United States of Europe) or drop the euro currency completely. In USA, if California bankrupts. the central goverment does not close the banks of California. This is what happens now in Greece. The ECB is denying liquidity in Greek banks, because we are no longer on a "program" and greek bonds don't have any value in exchange market.

 

We will vote "NO" to the referendum this Sunday. NO to the eternal austerity for our children and grandchildren and many more generations to come. NO to the concept of saving banks and bankers, and letting people die or live in poverty. We say NO to the ultimatum of  Mr. Dijsselbloem or Mr. Schäuble. We don't say NO to Europe, but we can't live like this any more. European leadership should decide if they want to help Greek people find their way inside the eurozone or if they want to humiliate us.




#492825 New OpenPli based image

Posted by malakudi on 27 May 2015 - 18:28

Around September 2010, DM has changed license for Enigma2 project making it CLOSED SOURCE and cutting ut GPL part.

This isn't a violation of GPL License? It's quite strange that a GPL Project can be converted as CLOSED but probably I'm not the right person and a lot of lawyers can detail this strange license change.

 

 

No it is not. The license allows them to continue closed source. The last GPL version is still GPL and cannot be closed. People that contributed on this double license scheme knew this could happen from the beginning.

Code that VU+ used on their "dvbapp" is derived from this last GPL version. Improvements/additions that were made by VTI and/or Blackhole on this codebase is still GPL. Since you distribute BINARIES of this GPL code, you need to have a way to deliver the sources to anyone might ask them. You are not required to make them available online, you are not required to give them for free. You can charge media and shipment costs. You could burn them on a CD or print them in a book. But you have to be able to deliver the changes to anyone that is asking. That is how GPL works.

 

I want to ask though people complaining, what are those changes/features seen in original Blackhole that you would like to become public? I remember (haven't used Blackhole for quite some time) that Angelofsky has made some work to bring additional DVB-T sticks supported. I also remember having this multiboot-multiple worlds kind of image - don't know if it is still supported. What else (that is not already present on the new OpenBlackhole) you would like to be released?

 

edit: @bacicciosat: Moderators might haven't seen this topic yet. You can report an offending post, it is in the bottom right corner. Vandyke is not OpenPLi.




#489047 VU+ Zero

Posted by malakudi on 27 April 2015 - 18:16

Although I don't want to defend VU+ for what they did on SoloSE and Zero, the "PVR tax" is something any EU citizen should shout against. Especially since most of these money go to MAFIAA representatives in european countries. We pay our subscriptions to our content providers and we have the right to record - applying a tax on a basic usage of a service is ridiculous.




#488626 CA Module CI+ Vu+ ?

Posted by malakudi on 23 April 2015 - 17:01

My Max TV CI+ module could decode 3 different channels in CI mode. With new OpenATV it can only decode 1 - same as VTi. Also enigma2 from OpenATV 5.0 does create /etc/auth.dat - same as VTi. However, I cannot find any such code in OpenATV's git.

 

@shilev: I will check later tonight on Solo2.