New Demo release - MemoPad 480i

[Rybags]

You mean narrow DMA mode?

The required timings are the same no matter what you're doing.  The exploit does it's thing on certain scanlines such that fake sync pulses are generated at the right time halfway through the scanline.

It's in the VBlank/VSync portion of the screen so no display activity is going on and you have plenty of time to set whatever graphics environment is needed in the normal display area.

From memory I'm fairly sure we need to be in wide mode at the point of scanline 240 otherwise we get unwanted sync pulses in the leadup to VBlank (which is what causes screen warping)

Edited February 10, 2023 by Rybags

[woj]

Yes, I meant the narrow DMA mode, that's good to hear, I asked to make sure that if anything goes wrong with my planned experiment then it is not due to VBLANK/VSYNC code timing. The 240-th scanline is something I wondered about too, I do not need all 240 scanlines, by 232 I'd be done with what I need to have on the screen, but here I can figure out things by experimentation, no need to ask pointless questions just yet

[Rybags]

It's been a while since I've played with this but from memory...

I think there's a way to activate the bug without actually displaying something on scanline 240.  Possibly I did it by having an Antic 2 line in wide mode then set DMACTL=00 at the end of it then set wide DMA the next scanline.

 

Regardless - you can equally use a Mode 3 or F line instead.  If you don't want stuff displayed just have a DLI beforehand that sets PF1 and PF2 to black (you generally need to do that anyway)

[woj]

I will just leave it here for now, this is what I have in mind 😀:

 

Preliminary research tells me that nothing prevents me from doing 1:1 clone of the original arcade graphics. Now I have a long way ahead of me doing a proper engine for this, there is hardly anything that scares me programming wise, but I am not a game developer 😞

 

What is really annoying is that two out of three times it starts with fields reversed, I was hoping that the vblank code is constructed to have the sequencing foolproof... Will need a key to flip this it seems.

 

TBC (elsewhere)

[Rybags]

I put field reverse key into most of the stuff I did.  Though I don't think it was needed at least for the TVs I used.

Also - some modern LCDs seem to just assume you're doing interlace even if it's not a valid signal and will interlace the fields whether you want it or not.

[woj]

On 2/14/2023 at 12:59 AM, Rybags said:

I put field reverse key into most of the stuff I did.  Though I don't think it was needed at least for the TVs I used.

Also - some modern LCDs seem to just assume you're doing interlace even if it's not a valid signal and will interlace the fields whether you want it or not.

 

I am slowly starting to get the hang of it, the amount of things I learn on the way is just immense.

 

As you noted, indeed by total accident (left a bug in the code for several iterations that was effectively disabling the interlace vblank code) I discovered that my LCD is of this kind, it just interlaces the picture without me asking for it. This is actually really bad, because I will have zero means to test whether I start on the right foot with the odd/even field, and otherwise if I do not loose the effect in a proper CRT somewhere in the process. Picture (Altirra frameshot, it does not give away the beauty of a fully hires PopEye) where I am at the moment  This will get proper colors, and all of them, once I sort out the DLIs, but since I also need a GTIA PM overlay and reposition things there in the DLIs too, it will be a while before I get the list of DLIs I need.

 

[Rybags]

You'd probably need a CRT device.

There's also capture cards, USB devices, DVD/HDD recorders that have A/V input but they'll often just assume it's an interlaced signal as well.

But generally I found that although I provided function to reverse the fields it wasn't needed.