More intelligent clash control for recordings. (1 Viewer)

[DragonQ]

The logic should be something like this:

- If a recording is scheduled, check if there are other recordings scheduled for the same time.
- If so, check if any of them are unable to record with default priority list.
- If so, check if a change of priority will fix the conflict.
- If not, there is a conflict so the user needs to decide what to do.

Step 3 shouldn't be terribly complicated: simply check which channels are on which cards and start with those that are mapped to the lowest number of cards. For example, if channels 1-10 are mapped to two DVB-S2 cards and two DVB-T2 cards, but channels 11-20 are only mapped to DVB-S2, and the DVB-S2 cards have priority, then the following would create a conflict:

Record channel 1 at 18:00-20:00
Record channel 4 at 18:00-20:00
Record channel 12 at 19:00-21:00

However, if we check them again but starting with the ones with the lowest number of mapped cards, then we'd find that channel 12 needs to use DVB-S2, then channel 1 can be set to use DVB-S2 and channel 4 can be set to use DVB-T2. No conflict. This does probably require each scheduled recordings to have its own card priority list though (just listing a single card won't work because it might be being used for timeshifting).

 

[kiwijunglist]

can you ignore conflict and schedule anyway?

 

[DragonQ]

can you ignore conflict and schedule anyway?

Are you responding to me? If so, ignoring scheduling conflict in the situation I gave would result in the third recording failing (with MediaPortal 1.3.0).

 

[kiwijunglist]

@elsmandino - are you able to schedule a conflicted recording or does it refuse to schedule?

 

[mm1352000]

@elsmandino
The conflict detection reflects what would happen in practise if you were to manually record the programs. In other words, the first two DVB-T tuners would be used for the first two recordings, then you'd miss the third. Creating more complex logic would involve an algorithm that runs in exponential time - scaling poorly with number of tuners and number of schedules - because each valid combination of tuner and schedule must potentially be tried.

 

[elsmandino]

@elsmandino - are you able to schedule a conflicted recording or does it refuse to schedule?

Hi,

Unforunately, I wasn't able to test it last night - there were loads of recordings last night and I couldn't find an opportunity where there was nothing happening to set up the deliberate clash. I shall be in a bit earlier tonight and shall try again.[DOUBLEPOST=1372334966][/DOUBLEPOST]

@elsmandino
The conflict detection reflects what would happen in practise if you were to manually record the programs. In other words, the first two DVB-T tuners would be used for the first two recordings, then you'd miss the third. Creating more complex logic would involve an algorithm that runs in exponential time - scaling poorly with number of tuners and number of schedules - because each valid combination of tuner and schedule must potentially be tried.

I see, so are you saying that a more complete conflict management system would indeed be feasbile but given the time it would take to go through every combination and every tuner, it is easier just to use the current method?

 

[elsmandino]

@elsmandino - are you able to schedule a conflicted recording or does it refuse to schedule?

Further to the above, I have just set a BBC1 recording, Channel 4 recording and Quest recording all for 7pm (i.e. they are all on different multiplexes).

As predicted, I was told that the Quest recording could not take place as it is only available on Freeview.

It gave me the following three options:

1. Remove the BBC1 recording to allow the Quest recording
2. Skip the Quest recording
3. Add the Quest recording and accept the conflict

I went for option three and it did accept it.

Shall be interesting to see what happens at 7pm.

 

[elsmandino]

Further to the above, I am just checking out what happened. It seems that the third recording was held off until there was an available tuner.

In this instance, the BBC One programme was only half an hour long compared to the Channel 4 programme's full hour. As soon as the BBC One programme ended, the remains of the Quest programme were recorded.

Is a shame that this issue exists - I was almost hoping that somehow TV Server was somehow going to switch the recordings around despite saying it could not.

 

[elsmandino]

@elsmandino
The conflict detection reflects what would happen in practise if you were to manually record the programs. In other words, the first two DVB-T tuners would be used for the first two recordings, then you'd miss the third. Creating more complex logic would involve an algorithm that runs in exponential time - scaling poorly with number of tuners and number of schedules - because each valid combination of tuner and schedule must potentially be tried.

Hi mm1352000 - could you clarify, exactly what you mean by the above please? Further to earlier response, though it might take a little longer for a scheduler to work out in advance exactly what recording should take place on what tuner, would there not be ample time for the Server to work this out if the recordings were made well in advance?

Or am I misunderstanding and the real problem is that the process would just be too complex to implement?

 

[mm1352000]

I see, so are you saying that a more complete conflict management system would indeed be feasbile but given the time it would take to go through every combination and every tuner, it is easier just to use the current method?

Yes, I'm saying that.
Further, little effort is being invested into improving TVE3 at present as we're focussed on developing the next generation TV Server. Anything that is more that a trivial change is unlikely to be implemented because we'd have to implement it twice (once in the current TV Server, and again in the new TV Server).