No subject

Fri Aug 3 13:33:49 UTC 2007

perspective) fine tuning is simply requesting that the hardware tuner
module tune a slightly different frequency.  Remember, though that
kernel tuner module units are 62.5kHz, so most fine-tuning winds up
being more of a placebo effect than anything.

It used to be that the US cable/OTA frequency tables used frequencies
that were exactly those kernel unit boundaries.  This has since been
changed such that values are now "exact" (as per the specifications) in
trunk, and--while fixing issues with the channel scanner--Jarod
backported these changes to 0.20-fixes before the 0.20.2 release.

However, when the tuner module is asked to tune a channel, all of this
"exactness" gets converted to kernel units by the kernel tuner module. 
These kernel units have much lower resolution, so often the fine tune is
rounded or truncated off.

>>> I tried using the Channel editor page on MythWeb, and that knocked the
>>> backend over; I gather that's a known problem in 20.1.
>>> I succeeded in using PHPMyAdmin to do it, but it's a much taller pain
>>> than just using '<' and '>' on the keyboard...
>> The mythfrontend channel editor also allows you to specify a finetune
>> value (which is, at least, safer than direct DB edits).
> Sure, but again; not interactive.


>>> I'd be happy to write a patch myself, but I'll probably need some
>>> pointers...
>> There's nothing in there now.  As a matter of fact, due to the
>> implementation of the tuner hardware, most likely--to truly be able to
>> test the new finetune value--you'd have to switch to a channel that's
>> far-enough separated from the current channel (frequency-wise--which
>> doesn't necessarily correspond to channel numbers) and possibly even to
>> a channel in another band (depending on tuner hardware--and bands,
>> themselves, are tuner specific) and then change back to the channel
>> whose finetune you're testing.  So, > would result in two channel
>> changes, then the next > results in two, ...  The hardest part would be
>> figuring out what "other" channel to use when testing the current
>> channel's finetune.
> Hmmm.  Sounds like the tuner phase-locks to the carrier it finds,

Right.  At least for all the tuner modules I've seen.

>  and
> finetuning doesn't do much of anything.

Again, finetuning--from the application/kernel level--just specifies a
different initial frequency and the hardware's fine tuning mechanism
takes it from there.

>   Although that doesn't match
> what I actually observed, so I'm probably wrong.
> Can you spare me two more sentences on *why* that procedure would be
> necessary,

The re-tuning would be necessary because that's the only mechanism by
which application-level finetune is applied (by specifying a different
initial frequency).  Changing to another channel may be necessary
depending on the hardware implementation, but I have to admit, I don't
know the hardware implementations, so I'm not the right guy to go into
details on this one.

>  and whether it's true of all tuner modules?

I definitely can't say that it's true of all hardware tuner modules.

I hope this information is useful.  Feel free to follow up with someone
more "in the know" than I.  Those on #v4l would probably be far more
helpful (and definitely far more certain of the details).


More information about the mythtv-users mailing list