IPTV/DVB-IP in MP: Support for http, rtp and udp (6 Viewers)

[velis]

Rtv slo 2
This is hard to duplicate - Kids have been watching TV the whole morning and it locked up only that once

Can you post the latest sources? I'll take a look and see if there's something semi-obvious

Edit: Got it again. This time on HD channel. Hope these logs help more.

 

[georgius]

Hi, can you post also logs from OldLogs directory? Logs are again overwritten.

Rtv slo 2
This is hard to duplicate - Kids have been watching TV the whole morning and it locked up only that once

Can you post the latest sources? I'll take a look and see if there's something semi-obvious

Edit: Got it again. This time on HD channel. Hope these logs help more.

Also attached new version of filter. Version, which you're using, has a bug (which occurs when buffer is full) which cause not releasing critical section and buffer cannot get empty => video stops and TvService goes to 100% CPU utilization.

I didn't test your version ... I hope that I find some time to test ASAP.

 

[georgius]

Hi,

I have tested your version of filter and here are results:

• SD channel
  • TvService 50%
  • MediaPortal <10%
• HD channel
  • TvService 52 - 55%
  • MediaPortal 30 - 40%

Logs are attached.

OK, I did a bit of comparison testing.
On my HTPC both filters do their job well, without any visible stuttering.

The only difference is that TV Service takes about 1% CPU more with my filter. Seems like all that synchronization takes its toll

Logs + latest version of my filter included (adjusted priorities for mm thread as well - seems it helps with stutter).

I'm currently time testing Georgius's filter since it seems the better solution. Will see in a day or so how it fares long-term.

 

[georgius]

Hi,

here is newer version of filter with working UDP and HTTP streams (I did my best) The biggest change from previous post is removed cyclic buffer and replaced with linear buffer.

Please, test and let me know.

 

[micmanm]

Hi georgius. Question for http part of filter.
Work in VLC and MP:
http://tv.x-lan.net.ua:8013/1pus1
http://tv.x-lan.net.ua:8013/2plus2

100% work in VLC, but not work in MP:
http://tv.x-lan.net.ua:8015/ictv
http://tv.x-lan.net.ua:8015/1nacionalniy
http://tv.x-lan.net.ua:8015/tet
http://tv.x-lan.net.ua:8015/trk-ukraina

What wrong?

 

[georgius]

Hi,

the main problem with not working HTTP streams is that server on other side don't accept standard HTTP/1.1 requests - it returns nothing. If you click on not working links, you get empty results. If you click to working links than your browser should show you some response.

I have to study VLC source code to determine how it works. When I simulate communication with sockets I was unsuccessful - I don't know what to send to server. If I send standard HTTP/1.1 request I get no response.

By studying VLC source code I found way how to get these streams working. I hope that I will have new release tomorrow.

 

[velis]

Georgius:
Yesterday we enjoyed the whole evening stutter free
Using my version of filter - I kinda gave up on yours after that last lock-up.
Seems that you and I are converging towards the same algorithm, both employing a linear buffer now
Today I won't have that much time to take another look at your latest iteration

Attached is the newest iteration of my modifications that uses bigger grab buffers == lower CPU usage. This means my filter now takes as little CPU as your version.

I added logging to determine why you get 100% CPU with my filter. This is rather CPU intensive (adds ~10% CPU), but it should at least give me an idea where to look...

Can you please take a look?

 

[georgius]

Hi,

I'll test your version today evening. Don't matter about MP CPU utilization, it's not caused by filter - I have similar results on my version.

When I test my filter I realize that cyclic functionality never been used, so I decided to replace it with much simplier linear buffer - as you wrote it seems to be working.

Georgius:
Yesterday we enjoyed the whole evening stutter free
Using my version of filter - I kinda gave up on yours after that last lock-up.
Seems that you and I are converging towards the same algorithm, both employing a linear buffer now
Today I won't have that much time to take another look at your latest iteration

Attached is the newest iteration of my modifications that uses bigger grab buffers == lower CPU usage. This means my filter now takes as little CPU as your version.

I added logging to determine why you get 100% CPU with my filter. This is rather CPU intensive (adds ~10% CPU), but it should at least give me an idea where to look...

Can you please take a look?

Tested your version, logs attached.

 

[georgius]

Hi,

new version of filter with more working HTTP streams. Please, test and let me know how it's working.

 

[SMak044]

I have a link to stream, it is normally played in VLC but do not want in MP
http://89.232.125.188:3526/stream.flv
What's the problem?