[mythtv-users] Has anyone on the list used the TBS 5220 USB DVB-T2 tuner?

[Stephen Worthington]

On Mon, 11 May 2015 16:43:18 +0100, you wrote:

>On 11 May 2015 at 11:49, Stephen Worthington <stephen_agent at jsw.gen.nz> wrote:
>> On Mon, 11 May 2015 10:35:01 +0100, you wrote:
>>
>>>Any feedback?
>>>
>>>Is it stable......are the drivers up to date (their website states
>>>kernel 4 is supported).
>>>
>>>Does it support multi recordings off one mux?
>>
>> All DVB-S/S2 and DVB-T/T2 support multirec.  The support is more in
>> MythTV than the drivers - if the drivers work with MythTV, then it can
>> do multirec.  The drivers can have an extra option that allows them or
>> the device itself to do the filtering of what data streams are sent to
>> the CPU, reducing the load on the CPU a bit, but it is really not
>> necessary to have that option as the CPU does not have to do much to
>> record data from a DVB tuner.
>>
>>>If not any other suggestions for the 'best' (I realise it's
>>>subjective) USB DVB-T2 (and any USB DVB-S2) units out
>>>there......Kernel drivers preferred.
>>>
>>>Thanks
>>>
>>>Tony
>Sorry Stephen I should have been clearer......
>
>I meant will the hardware cope with it, I realise the software
>actually does the demuxing of the complete signal stream.....but
>sometimes if the hardware is close to tolerance...or gets heated it
>won't transmit the whole stream cleanly and you sometimes only see it
>if trying to record more then one stream.
>
>Have you played with that unit though?
>
>Tony

No, I have never heard of that happening.  Disruptions in digital TV
signals are normally because of reception problems (signal level or
quality), not the tuner overheating.  Considering how they work, it is
not at all likely that running multirec on a tuner will cause it any
problems at all.  If it can receive and process any channels at all, a
tuner has to be able to receive and process all the digital data on
the multiplex.  Just then selecting out some of the streams to pass on
to the CPU is actually likely to reduce its workload, not increase it.
There are quite a few tuners that do not do multirec processing
themselves and just send all the data for the entire multiplex to the
CPU all the time.

The situation where multirec recordings can cause problems has nothing
to do with the tuners.  That is where you do not have enough hard
drives to support all of the recordings that are being done at
simultaneously.  If you have too many recordings being written to one
hard drive, the heads on the drive will be spending most of their time
moving back and forth between the positions each of the recording
files are being written to, not leaving enough time for all the data
to be written to the disk.  So the RAM buffers build up and eventually
overflow and some of the recording is lost.  It is most likely to
happen when the post-roll of one set of programs overlaps the pre-roll
of the next set of programs, so if you are recording say four programs
at once on two hard disks (two to each) and another four programs
start recording for an overlap period of say four minutes, for that
overlap time you will have eight programs recording, and that is
probably too many for just two hard drives.  Remember that the head
movement speed of hard drives has not increased nearly as much as the
drive capacity has increased.  The speed of writing once on track has
increased dramatically with each new technological refinement to
increase the drive sizes, hence the sequential write speeds seen now
are pretty impressive numbers, but that only works when the heads only
have to move just one step to get to the next track.  When they have
to move 160 tracks to the next file, another 100 to the third, and
then back 260 to the first file again, the write speeds achieved drop
dramatically.  So the solution to this problem is simply to add
another hard disk.  I currently have seven recording drives, and do
not have to worry about this problem.  Back when I only had two drives
(one with the system on it also), it happened quite frequently until I
worked out what was happening and made sure to avoid scheduling too
many recordings at once.

And no, I do not have that model of TBS tuner, but I do have a TBS
5922 DVB-S2 USB tuner, and it works well.  It is run with TBS supplied
drivers, although I believe that there is also now an in-kernel driver
available which should be in new kernels at some point soon.  It has
been available by compiling the latest V4L source for a while.