MagicEngine

[Tomaitheous]

Chris, I checked out your B&W demo with an oscilloscope today. Bit 7 of $400 removes the Color Burst signal on the front porch of all scanlines, but the subcarrier is still intact. As far as I can tell it's still a perfectly legal B&W signal that all standard NTSC US sets must still adhere by.

I wrote a demo that allow you to change the RCR on which the VCE would swtich from B&W to color mode, by using the D-pad. On a standard set, it made a pretty neat transparent vertical block of fading hue stretching down the screen - overlaying the onscreen graphics. Both the capture card and the HDTV had different effects though

I also found the VCE responds to writing bits 3-6 (not sure exactly which ones) of $400 during active display - in some unusual way. Some sort of shifting patterns( 8 of them I think) for the line the RCR routine is called on and the following line. The patterns appear to repeat within a $00-$FF indicating not all the bits effect the display. Also - from one line to the next, the pattern block is in a different bit order than the previous value. I don't fully understand it yet, but I do know it only effects the current line and/or the next line depending on the pattern - all other proceeding lines go back to normal.

And the last test. Setting the display terminal and external sync bits do not appear to effect the VCE output signal. I check this on the o-scope and the scanlines remained intact. The screen will wave and/or roll horizontally, but only because the data being sent to the VCE by the VDC is not insync. It rolls and waves with-in the active display of the VCE, and all this with a static boarder on the top and bottom of the screen

More tests in the works.


-Rich
_________________
www.pcedev.net

[ccovell]

Cool. Pics and videos, please!

Yes, the B&W mode of my demo works fine on my TV & capture card... but the pausing in Shapeshifter (goes B&W) messes up on my TV (rainbowy hues, regular brightness). A different set of graphics, I guess.

Interesting about bits 3-6 (that would give 16 steps, unless one of the bits actually does nothing.) Hmm... maybe you've found the tile-flipping mode that we've been looking for over the years!
_________________
http://www.chrismcovell.com

[Tomaitheous]

I'll see if I can find a way to post some pics and video. I'm working on a few SGX demos that show some cool tricks with the VPC - so it's currently taking priority.

***Run the demos below at your own risk. I take no responsibility***

Here's an ISO/CUE of the VCE test code - here. Up/Down sets the RCR line and left/right set the value (and #$78 to keep the res ,etc) to the VCE. Try line $42 and cycle through the values to see the different shift patterns. If you can see into the overscan area, you can see even the boarder is shifted and not clipped.

Here's the HUE test as well - here.

NOTE: All above demo's were run on an SGX+SCD.

-Rich


Edit: Here's a pic of your B&W demo Vs the SCD boot screen.

(missing the color burst signal)


(and color burst signal)

_________________
www.pcedev.net