[mythtv] Re: [PATCH] xvmcvld picturequality

[Terry Barnaby]

Hi Ivor,

Sound good, but are you sure ?

My, slightly further, reading into MPEG is that the RLC (run length coding)
system can be either ZIGZAG or alternate on a frame by frame basis depending
of which method produces the best compression.
The QMatrix appears to use this encoding as does the IDCT blocks.
The default encoding is ZIGZAG, but, it could use the alternate scheme setting
the alternate bit in the MPEG stream.
On none of my current MPEG source material does this happen, all frames use ZIGZAG,
but I guess some frames on other source material will with better MPEG compressors.
Did any of your source material contain frames with the alternate RLC scheme ?

In this case I would have thought we would have to apply the de-zigzag + alternate
tables as MythTv always removes the zigzag to the incoming QMatrix data ....

Terry


Ivor Hewitt wrote:
> Looks like it didn't need the other zigzag code after all. This produces a
> matrix that matches the xine one, and the picture looks fine too. Could do
> with some DVB-S people to try it out now.
> Just need to see if the chrome table makes a difference too now.
> Cheers,
> 
> 
>>-----Original Message-----
>>From: Terry Barnaby [mailto:terry1 at beam.ltd.uk] 
>>Sent: 14 September 2004 17:41
>>To: Ivor Hewitt
>>Cc: Thomas Hellström; unichrome-devel at lists.sourceforge.net; 
>>mythtv-dev at mythtv.org
>>Subject: Re: [Unichrome-devel] Re: [mythtv] [PATCH] xvmcvld 
>>picturequality
>>
>>Hi Ivor,
>>
>>Here is a patch to xvmcvldvideo based on the current Myth CVS.
>>It contains my idea to fix the QMatrix loading. This may be 
>>wrong but I thought it would be usefull for you to look at to 
>>see if it matches up with your ideas.
>>It also contains code to set the croma QMatrix.
>>
>>Terry
>>
>>Ivor Hewitt wrote:
>>
>>>Thomas Hellström wrote:
>>>
>>>
>>>>Hi!
>>>>
>>>>If you get to the bottom of this it would be interesting 
>>
>>to know. I 
>>
>>>>get similar issues with Xine with kernel 2.4. Nothing on 2.6.
>>>>
>>>> 
>>>>
>>>
>>>After I've QMatrixed tonight.... I'll add a stack load of timing to 
>>>the code and try and see what's going on. I'm displaying at 
>>
>>50Hz so it 
>>
>>>should be silky smooth.
>>>And I am on 2.6.8.
>>>
> 
> 

-- 
Dr Terry Barnaby                     BEAM Ltd
Phone: +44 1454 324512               Northavon Business Center, Dean Rd
Fax:   +44 1454 313172               Yate, Bristol, BS37 5NH, UK
Email: terry at beam.ltd.uk             Web: www.beam.ltd.uk
BEAM for: Visually Impaired X-Terminals, Parallel Processing, Software
                       "Tandems are twice the fun !"