PDA

View Full Version : [ET9x00] Problem with playback when still recording same channel using OSCam



purenirvana
15-02-13, 15:47
Hi,

I think this is my first post here, please don’t take my first posting as a rant or complaint. I think the work you guys do is great, it’s just that I’ve never had the need to post before!

Box type:

Clarke Tech/Xtrend ET9x00

Image:

Vix 3.0.507 and Vix 3.0.600 (maybe other versions too)

Incompatibility issue:

Playback of a file from an encrypted channel on 28.2E whilst it’s still being recording using OSCam (all versions). Problem also exists in OpenAAF 2.0 but not in OpenPLi 3.0. If using CCCam 2.3.0 or earlier the problem doesn’t exist regardless of the box image. Using an official local card or proxy makes no difference, the problem still exists. I have ruled out it being OSCam at fault as it works perfectly with OpenPLi.

Description:

As said, if I’m recording a program off an encrypted channel on 28.2E (not tested other sats yet) and I begin to play it back whilst it is still recording I get a significant problem. It can be seen initially if viewing the channel and I press record, the picture freezes briefly then continues as normal. If I view OSCam’s log file it shows that at that moment of pressing record it requests the ECM from the cache rather than the card. Not a big issue in itself as it is a fairly brief interruption to the picture.

The main problem is when you fast forward or rewind (with the number buttons or arrow buttons), what happens here is that the playback moves as requested, but whatever part of the programme is being recorded at the time you press the keys asks for the ECMs from the cache rather than the card. This causes the ECM requests to get out of sync with the recording. It’s almost like the playback takes over the timing of the ECM requests rather than the recording as should be the case. This means that the recording is forced to wait a second or two to get the ECM from the cache as the playback ECM timing is taking priority from the card, the result of course is a frozen picture for a few seconds when playback catches up to the affected part of the recording.

I would think that playback doesn’t need to request an ECM? The recording aspect should request the ECM, maintain the timing of the requests, decrypt the picture and move on until the next ECM. Playback should just play the file as recorded. If I use CCCam this problem does not exist, nor does it exist in OpenPLi using OSCam. The problem also exists with OpenAAF. Something isn’t right with the playback whilst still recording the same channel aspect of the media player with these images. OpenPLi’s version doesn’t suffer this problem.

If I use the timeshift pause, and skip around back and forth with the rewind, the ECMs continue to be requested by the live channel rather than the timeshift. It’s just playback whilst still recording when using OSCAM for the ECM requests that is affected.

Sorry to ramble on, it’s just something that I feel needs looking into. I love the OpenViX images and would hate to be forced to use CCCam or OpenPLi to get around the problem, as in my opinion both of these are inferior to OSCam and OpenViX. Hopefully my explanation is clear enough to understand so the issue can be fixed. I will keep experimenting and see if anything else of interest crops up that might help. I’m going to go back through the OpenViX images to see in what build the problem first arises.

On that final point, how far back through the images can I go without ‘bricking’ my ET9x00? I read somewhere you shouldn’t go too far back as it can ‘brick’ the box? Can I go back to the 2.x versions?

Many thanks in advance to anyone that can help.

pn

Oh yes, the problem seems to have been partially discussed in the thread linked below. Hopefully the little extra detail provided above will help.

http://www.world-of-satellite.com/showthread.php?25083-Issue-with-recordings

judge
15-02-13, 15:59
when you set a timer, you have the option of changing the recording type.
Have you tried any of the other options there to see if it makes a difference?

purenirvana
15-02-13, 16:06
Thanks for the reply judge,

Not tried it from a recording started with a timer. I will try that and see if the problem arises. The issue I'm having is when hitting record mid-programme and then say watching the recorded file from the beginning again and then rewinding or fast-forwarding whilst said recording is still ongoing.

What options did you have in mind to try changing?

judge
15-02-13, 17:09
What options did you have in mind to try changing?

Try all 3 recording types & see if any work as you expect.

purenirvana
16-02-13, 02:10
Right, thanks again judge. I think you've helped me to begin to crack this! :thumbsup:

If I set a recording with the timer and select the record type to 'descramble and record ecm' or 'don't descramble, record ecm' the problem occurs as described in my first post. However, if I select the record type as 'normal' the problem goes away and everything works as I would expect i.e. no ECM request from the cache is made by OSCam in sync with playback (rather than the recording) when replaying a recording that is still in progress. No picture break up, no freezing, everything is spot on.

I'd therefore deduce that we need an option to set the default 'record type' when hitting the record button mid-programme and for any timer or autotimers that are set up. Much like hollinshead is saying in this thread:

http://www.world-of-satellite.com/showthread.php?23557-Recording-type-descramble-and-record-ecm

Despite what is said in hollinshead's thread, it is a requirement to select the default record type as two of the three record options are causing, and will continue to cause an issue until the 'record ecm' and OSCam conflict is resolved. Maybe it should still be an option to select the default record type regardless of whether the conflict is resolved? :)

I assume OpenPLi doesn't have the 'record the ecm' options, hence why I couldn't replicate the issue with their image? Also CCCam seems to work with or ignore the 'record ecm' options, again, this is presumably why I could not replicate the issue with this cam either?

Rob van der Does
16-02-13, 09:41
I assume OpenPLi doesn't have the 'record the ecm' options,
OpenPLi has exactly the same options, but the default is 'normal' (i.e. no ECM data is being recorded).

purenirvana
16-02-13, 10:53
OpenPLi has exactly the same options, but the default is 'normal' (i.e. no ECM data is being recorded).

Thanks for clarifying that Rob. Do you think one of the programmers will be able to comment on whether OpenViX can be made to have either 'normal' as the default too or an option for the user to define what the default record type is?

Either that or programming the record ECM options to sync with the recording rather than the playback as is what happens at present?

purenirvana
16-02-13, 22:25
Does anybody know in what OpenViX version or build the two 'record ecm' record options came in? I'm just thinking that in lieu of a fix at the moment, I could roll back to an older image version.

judge
16-02-13, 22:44
Some time before 399, not sure exact build. Sorry.

purenirvana
16-02-13, 23:06
Some time before 399, not sure exact build. Sorry.

Thanks judge, that narrows things down quite a bit! I've had a look here:

http://www.world-of-satellite.com/archive/index.php/t-21343.html

It would seem that the first mention of recording ECMs is on build 395 under the 'OpenViX 3.0-399 AVAILABLE NOW' post (post #19). So, to me, build 394 or earlier should be used as a roll back in lieu of a fix for now. Would you agree?

judge
16-02-13, 23:27
Thanks judge, that narrows things down quite a bit! I've had a look here:

http://www.world-of-satellite.com/archive/index.php/t-21343.html

It would seem that the first mention of recording ECMs is on build 395 under the 'OpenViX 3.0-399 AVAILABLE NOW' post (post #19). So, to me, build 394 or earlier should be used as a roll back in lieu of a fix for now. Would you agree?

Yes, that looks about right, not sure if it will fix your current issues, but no harm in testing a build under 394 & letting us know if it fixes them.

Larry-G
17-02-13, 00:32
it was build 395


openvix: build 395
[RecordTimer] record ECM data in recordings as default, this should allow decoding of file if softcam fails at time of recording, and cleanup.
[RecordTimer] fix possible BSOD
[VideoMode] fix for 1080p
epgcache: Save EPG in flash when no harddisk is available
[Translations] ET update, thanks Rimas.
Fix a bunch of compiler warnings
Add sLiveStreamDemuxId getter (for VU+ HBBTV plugin)
Add doxygen documentation system
InfoBarGenerics: Make blue button in Numberzap toggable
[About] remove then need for WirelessLan plugin even if not use WLAN usb stick.

purenirvana
17-02-13, 13:15
not sure if it will fix your current issues, but no harm in testing a build under 394 & letting us know if it fixes them.

Ok, I've tested a build under 394, I've tried OpenViX 3.0-393 as I already had it archived. I can't test 394 as I couldn't find it anywhere that didn't require a registration to download it. If somebody could post it here I'll happily experiment with it.

Anyway, the upshot of it all is that build 393 fixes 99% of the issue raised. I can playback a file still being recorded, rewind and fast forward back and forth, no problems, no freezing or picture blocking at all whilst using OSCam as previously. OSCam doesn't go to the cache like it does in later OpenViX builds, hence the picture being recorded 'live' stays stable.

The only slight 1% niggle is that when you first press record mid-programme, the picture freezes and stutters at the moment you press record. If I look at OSCam's log, for some reason the ECM is requested from the cache at this point rather than the card, hence the picture freezes briefly. I can more than live with this minor thing, I just thought I'd mention it as it may help to get to the root of the problem. :)

Anyway, an option to select the default 'recorded type' to normal in later builds seems necessary to me (at least for now) if the user wishes to use OSCam (as I do). The record ECM options are a great feature, have great value and are certainly worth pursuing IMHO, there is just an issue with OSCam and calling the ECM from the cache that needs ironing out.

I'm happy to help out further where I can with any experimentation, unfortunately I'm no programmer though, so can't help with that aspect! :)

Like I said, if anyone has build 394 for me to try I'll give that a go too.

purenirvana
18-02-13, 16:14
Ok, just to update this further, I've upgraded back to build 600 and have used the very useful information posted in hollinshead's post here:

http://www.world-of-satellite.com/showthread.php?23557-Recording-type-descramble-and-record-ecm&p=178389&viewfull=1#post178389

Telnet into your box and type the following:

opkg install enigma2-src

<then press enter>

opkg install enigma2-plugin-vix-core-src

<then press enter>

sed -i 's/descramble = True, record_ecm = True/descramble = True, record_ecm = False/g' /usr/lib/enigma2/python/RecordTimer.py

<then press enter>

reboot

<then press enter>

Your box will now reboot and the default 'record type' will revert to 'normal' as in builds 394 and prior. I thought this might be useful to others that would like to upgrade to build 600, but have the problem I have experienced.

From my experiments the telnet commands above are still safe to use with build 600. Whenever I upgrade my box I will try the commands so that it can be recorded at what point the commands become ineffective or unsafe to use.

Very many thanks to hollinshead for the telnet commands :)