| |
Paulko64
Registered: Jul 2010
Posts: 24 |
asymmetric 38 column border?
Is there a particular reason why, if I switch to 38 column mode, the side border is not centered anymore? I.e. the left border comes in 7 pixels, and the right border 9 pixels! Why not 8 pixels on both sides?
If I want to center the screen in 38 column mode, I first have to copy everything one character to the left, and then shift the X fine scroll register to it's max value... |
|
| |
Oswald
Registered: Apr 2002
Posts: 5094 |
also why its not 39 column ? |
| |
JackAsser
Registered: Jun 2002
Posts: 2014 |
Quote: Is there a particular reason why, if I switch to 38 column mode, the side border is not centered anymore? I.e. the left border comes in 7 pixels, and the right border 9 pixels! Why not 8 pixels on both sides?
If I want to center the screen in 38 column mode, I first have to copy everything one character to the left, and then shift the X fine scroll register to it's max value...
Never noticed, but you're right. Poor you. :D |
| |
Smasher
Registered: Feb 2003
Posts: 520 |
lft knows. the rest just trolls (like me). |
| |
ChristopherJam
Registered: Aug 2004
Posts: 1409 |
Quoting Paulko64Is there a particular reason why, if I switch to 38 column mode, the side border is not centered anymore?
My best guess is the engineers weren't thinking straight when were under the hammer to get stuff working in time for CES, then had SFA time to revise the design before they went into mass production. TBH it's amazing anything worked; check out any of the histories of the company. |
| |
Oswald
Registered: Apr 2002
Posts: 5094 |
Quote: Quoting Paulko64Is there a particular reason why, if I switch to 38 column mode, the side border is not centered anymore?
My best guess is the engineers weren't thinking straight when were under the hammer to get stuff working in time for CES, then had SFA time to revise the design before they went into mass production. TBH it's amazing anything worked; check out any of the histories of the company.
according to "on the edge" Al Carpentier's VICII worked flawlessly for the first time trying, except 1-2 buggy sprites. My bet would be on transistor sparing, less gates, more simple design, etc. |
| |
oziphantom
Registered: Oct 2014
Posts: 490 |
And getting the wrong number of clocks per line... |
| |
ChristopherJam
Registered: Aug 2004
Posts: 1409 |
Quoting Oswaldaccording to "on the edge" Al Carpentier's VICII worked flawlessly for the first time trying, except 1-2 buggy sprites
Damned impressive work. |
| |
Jammer
Registered: Nov 2002
Posts: 1335 |
Hah, I came across that coding my thing ;) |
| |
lft
Registered: Jul 2007
Posts: 369 |
I'm puzzled by this too. If they hade used 8 pixels on both sides, they could have used 1 MHz logic for the entire border unit, improving yield.
I've got a theory anyway, although it's pure speculation. The current approach creates a symmetry between yscroll and xscroll: The number written into the register is one less than the number of revealed pixels in the top row or left column respectively.
Still, if they wanted symmetry, they should have used 39 columns, and moved both borders three pixels to the right in the wide configuration.
But we also have to be careful about what we wish for: The way it's done in the VIC chip, with a 9-pixel difference in the right border position, virtually guarantees that there is a cycle in which we can open the sideborder. A 40/39-column system would only have a 4-pixel difference, so d016 might well get sampled at two instances within the same clock cycle. Then the sideborder would remain closed forever. |
| |
Compyx
Registered: Jan 2005
Posts: 631 |
Aargh, scary. I need my open borders. Please, somebody close this thread ;) |
... 7 posts hidden. Click here to view all posts....
|
| Previous - 1 | 2 - Next |
