[mythtvnz] Upgrading to 0.25 from 0.23

David Moore dmoo1790 at ihug.co.nz
Fri Apr 27 06:08:19 BST 2012


On 27/04/12 16:44, Steven Ellis wrote:
>
> On Fri, April 27, 2012 4:01 pm, David Moore wrote:
>> I was finally forced to upgrade from 0.23 due to the audio glitch on
>> Prime. Could have patched 0.23 I suppose but decided to bite the bullet
>> and go to 0.25. Works fairly well as far as I have tested so far. Thanks
>> to fellow listers for setup/bugfix clues and myth devs for rapid repairs
>> to code.
>>
>> I am seeing incorrect video lengths on some old recordings. Annoying. I
>> have noticed in the past that this could be a seek table issue. I found
>> in the past that using mythcommflag to rebuild the seek table for some
>> recordings produced incorrect time codes. Forget the details but had to
>> divide seek table entries by two to fix.
>
>
> Hi David
>
> I've been using the following to fix seektables. I've found that
> mythtranscode was a lot more reliable than mythcommflag and I now have a
> user job I can easily select when I've got a broken video file.
>
> http://www.mythtv.org/wiki/Repairing_the_Seektable#Automation_of_mythcommflag_repair_function
>
>
> --------------------------------------------
> Steven Ellis - Technical Director
> OpenMedia Limited - The Home of myPVR
> email   - steven at openmedia.co.nz
> website - http://www.openmedia.co.nz

Hi Steven, Yes, that will be my last resort method. Thing is I don't 
think it's necessary as I don't think the actual seek table values get 
updated. Makes sense because the files are seekable but just have wrong 
durations. I think the fix is in recordedmarkup but need to figure out 
how to generate correct values. Reason I think this is I noticed the 
files rebuilt had new rows in recordedmarkup. If we can work out the 
correct values then a real fast sql patchup would do the trick much 
faster than mythcommflag.



More information about the mythtvnz mailing list