[mythtvnz] 0.25 seektables recording lengths..

[dmoo1790 at ihug.co.nz]

On Sat, 19 May 2012 22:26:44 +1200, Brett <mythicalbeast at slingshot.co.nz>
wrote:
> On Sat, 2012-05-19 at 19:41 +1200, David Moore wrote:
>> On 19/05/2012, at 11:38 AM, Brett <mythicalbeast at slingshot.co.nz> wrote:
>> 
>> > Anyone else notice that with 0.25+fixes all the old TVNZ recordings
>> > needed their seek-tables rebuilt ?
>> > 
>> > Thanks,
>> > Brett
>> > 
>> > 
>> No. In fact I have checked several old TV1 recordings and they are good
>> except that the duration is wrong. The seek tables are fine because
>> seeking works ok. The incorrect duration is not due to seek table
errors.
>> _______________________________________________
> I never said the seek tables were bad...just needed to be rebuilt (to
> fix recording lengths implied by subject)
> And I don't think that just because you can seek in the recording means
> they are okay/correct.
> 
> IMO It is safer to use mythcommflag --rebuild --file & easy for a novice
> to make a script from ls -m1 etc.
> 
> Most of my TVNZ recordings are off tvnz7 so they process at 6000fps.
> 
> So why does this only affect TVNZ multiplex ?
> 
> Brett
> 
> 
My bad. I'll pay more attention to the subject next time.

I do believe the seek tables are ok or at least as ok as they were before
upgrading. Why would the upgrade trash the tables? They're just an index of
key frames vs file offsets which would only change if the files changed.

I suspect TVNZ is/was doing something not quite standard with timing which
may upset myth. As an example I found that mythcommflag did not correctly
rebuild seek tables in the past. I had to fix some rebuilds by dividing, if
memory serves me correctly, the frame number by two. Something like that
anyway. It was a while ago. :)