[Openchrome-users] SP13000 + fuzzy vga output at high resolutions

[Luc Verhaegen]

On Thu, Jan 18, 2007 at 10:02:14AM +0100, Xavier Bachelot wrote:
> Luc Verhaegen wrote:
> >On Wed, Jan 17, 2007 at 03:13:38PM +0100, Xavier Bachelot wrote:
> >
> >>brightman wrote:
> >>
> >>>SP13000 + fuzzy vga output at high resolutions
> >>>I was thinking about buying an SP13000 board and driving the VGA output 
> >>>at the same resolution as my LCD which would mean a resolution of 1680 x 
> >>>1050. 
> >>>Firstly is this reasonably straighforward using the openchrome drivers 
> >>>(is there an easy way to fathom out a suitable modeline, I've used 
> >>>powerstrip in the past?)
> >>>
> >>
> >>The modeline won't be hard to get. Just run 'gtf 1680 1050 60 -x' and 
> >>stick the output into your xorg conf.
> >>
> >>
> >>>Secondly I've been reading on a couple of VIA related forums about 
> >>>problems with the VGA output on SP13000 boards at resolutions of 1280 x 
> >>>1024 (and higher). 
> >>
> >>However, and I don't know if it's related to the above mentioned 
> >>fuzziness pb, high resolutions will require a lot of bandwidth, so the 
> >>ram type you are using might come into the game.
> >>
> >>Regards,
> >>Xavier
> >>
> >
> >In this case, using cvt -r might relieve the problem. 1680x1050 is a 
> >resolution for a TFT based monitor, so reduced blanking should be 
> >preferred if you're providing modelines anyway.
> >
> >Luc Verhaegen.
> >http://unichrome.sf.net/
> 
> Thx for the info, Luc. What are the differences between gtf and cvt 
> apart from the reduced blanking switch ?
> 
> Is there a good explanation of reduced blanking somewhere ? I think I 
> got it mostly, but would be unable to properly explain it to someone 
> else. And just to be sure, can reduced blanking be used on any TFT w/o 
> pb ? Using reduced blanking on a CRT simply won't work or will it also 
> damage the screen ?
> 
> Regards,
> Xavier
> 
Normal timing is just a simplified calculation.

I'm not sure, it's pretty straightforward. No need to shift an electron 
beam, no need to spend as much time on sync, and, you also don't have 
big edges hidden behind the monitors frame.

You can try reduced blanking on a CRT reasonably safely. You just have 
the edges wrapped a bit. Nasty, but not that unsafe. You just don't want 
to use it all the time, which is logical for a sane person, as it really 
is a buggered mode. I think that fear of monitor damage is quite 
overrated. And i _have_ killed at least one CRT by testing; it's about 
extensive repitition, not about accidentally doing the wrong thing a few 
times.

It should work quite happily on a definitive number of TFT based 
screens (there's no 100% certainty of course), whether they use a 
digital link or not. The big problem is that VESA EDID does tell you 
that you have an analog or digital link, but it doesn't bother with 
telling you whether it allows reduced blanking or not.

Most people treat the digital bit as definitive, but it isn't. CRTs do 
exist with DVI-D.

Luc Verhaegen.