[mythtv-users] Overlapping recordings on SAME channel w/ one tuner

[Michael T. Dean]

On 03/27/2010 01:30 PM, Bill Bogstad wrote:
> On Thu, Mar 25, 2010 at 7:48 PM, Michael T. Dean wrote:
>    
>> On 03/25/2010 07:41 PM, Jeff Evans wrote:
>>      
>>>> Given that virtual tuners were already built to solve this
>>>> problem, the myth devs probably consider a solution to your
>>>> question as a corner case.
>>>>          
>>> Ouch Chris! Surely there are many out there still using analogue and
>>> other non-DVB capture devices?
>>>        
>> IMHO, your buying additional capture cards is /immensely/ cheaper than the
>> amount of code required to do what you're asking.
>>
>> If you don't have enough capture cards to record what you want, you need
>> more capture cards.
>>      
> And why wasn't this argument made when multirec/virtual tuners was
> proposed/implemented?
>    

It was--for years.  Then some developers finally decided to implement 
it--and got paid to do so.  (Granted, what they got paid was 
significantly less than the time it took was worth, but I'm sure it 
played a part in their decision about whether implementing the 
functionality was worth their time.  And, no, I am /not/ trying to 
encourage some bounty approach for trying to convince a dev to write it 
for you.)

> I don't see how anything that multirec does couldn't have been done
> with more physical tuners.
>    

Right, everything multirec does could have been accomplished with more 
physical tuners.

The difference between multirec and analog recording, however, is that 
it seems no one actually cares enough to implement the functionality for 
analog recording.

> I don't currently have a multirec installation so I have a technical
> question.  The wiki page
>
> http://www.mythtv.org/wiki/record_multiple_channels_from_one_multiplex
>
> states:
>
> Multiplex recording also allows overlapping of sequential recordings
> of the same program. That avoids the annoying problem with the end of
> a broadcast being in the next recording or the beginning of a
> broadcast being at the end of a previous recording, as is lately
> popular with syndicated reruns of programs like Scrubs and ER.
>
> It sounds like this means that tuners that allow virtual tuners can
> copy a single "program" stream
> into multiple files.  (i.e. Chop a single stream into potentially
> overlapping in time recording files of the same single continuous
> video stream.)
>    

No, there are multiple "tuners" with multirec--they're just virtual.  
Each "tuner" gives output that goes to only one file.  Basically, AIUI, 
we're not reading once and writing many, we're reading the "same" data 
many times and writing it many times (assuming that multiple virtual 
tuners are tuned to the same logical channel at the time).

> Is there any reason that single-stream only tuners couldn't be handled
> as if they just happened to be a multi-stream capable tuner where
> there was never more then one actual program stream available per
> frequency that was tuned?   Could this be implemented as a truly
> virtual tuner which simply acted as a front end to the underlying real
> tuner and presented a multirec capable view to the rest of the system
> (i.e. scheduler).   The underlying real tuners would be configured
> into the system, but would never have an input connection made.   Each
> truly virtual tuner would have a dedicated physical tuner, be
> configured to have two (probably all that would be required) virtual
> tuners (in the current sense) tuners and would have the appropriate
> input connection.  Doing it this way might add some minor overhead to
> recordings but would allow any single video stream physical tuner to
> handle overlapping in time recordings.
>
> Or is my total lack of knowledge of  the implementation of multirec
> mean that I'm missing some major problem with the above suggestion?
>    

Having only glanced at the workarounds suggested, they all suffer from 
the fact that they either wrongly tell MythTV it can record multiple 
channels from the same physical analog capture device or--if you use 
input groups--only one of the "virtual" analog capture devices would be 
usable at a time (meaning you have the exact situation you have with a 
single analog capture device with a "proper"/supported configuration).

The multirec implementation laid the ground work for a real ability to 
do what you all seem to want.  We just don't seem to have anyone who's 
willing to do the work of implementing it properly (just a bunch of 
people who seem to think that work should be done /for/ them :).  If 
someone implemented the solution and did it properly, the code would 
likely be added to MythTV.  If the patch isn't done right, however, I 
can pretty much guarantee that it won't be added.

Mike