How fanciful is this kind of mockup?

[Rob Beschizza]

Amstrad games being childhood faves (and unrequited dreams of game development being an adult preoccupation) I find myself constantly doodling mockups of remakes. How credible might this mockup of IK+ be? It struck me that you could use up to 16 colors, and an interrupt for more legible text in the empty middle of the screen, and maybe make the backdrop wider too. The original sprites appear to be only 3 colors, which I suppose lets them be defined with 2 bits per pixel. So would it be a brick wall trying to hack in more colorful sprites? Might I pay someone to attempt it?

You cannot view this attachment.

The OG masterpiece, for reference...

You cannot view this attachment.

[andycadley]

I would imagine that all that dithering behind the characters is likely to be a problem, but it rather depends on how IK+ does its sprite masking. The fact the background is normally a solid colour may not be an accident, there are various tricks to accelerate masking in that case (although players overlap so who knows)

The mid screen text thing is probably a no go without a major overhaul of the game. CPC interrupts occur at fixed positions on the screen so where you can easily do a Mode change is (effectively) fixed.

I'd also guess the sprites come from the C64 and are therefore probably just stored as three colour images that get palette swapped as they render, so that might also limit things without major rework. But you never know with these things until you prod around in the code.

I've always been curious what a Plus version could look like. The sprites are probably too big to be entirely down with HW sprites, but you could maybe do something like the C64 where one fighter is drawn to the screen and the others are sprites. That would give you a lot more flexibility in terms of use of colours etc (and would make the mid screen Mode split fairly simple too). Would be a major undertaking though. 

[Anthony Flack]

I doubt very much you could achieve this by hacking the original game unless you completely pulled it apart, but if you rewrote it from scratch it should be perfectly possible.

[arnoldemu]

@Rob Beschizza:

Take a look at this post and feel free to run the dsks in an emulator:

https://www.cpcwiki.eu/forum/programming/interrupt-positions-with-various-sized-screens/

This will help show where the interrupt positions are for an Amstrad CPC.

If targetting an Amstrad Plus machine then the interrupt positions can be programmed to any line.

If you can position mode changes on or close to the lines where the colours change in the border then it wastes less CPU time.

EDIT: It's also worth, if not already done, checking the following:
- your mockup uses only CPC palette colours
- colour changes happen at interrupt points (see linked dsk)
- in any area which is mode 0 there is only 16 colours used (because changing colours multiple times takes CPU time and the game would have to be completely re-written to account for any of that and it may not be possible),
- that any area which is mode 1 is 4 colours,
- screen is a multiple of 4 pixels in mode 0 and 8 pixels in mode 1 width (as this corresponds to sizes that are possible by programming the CRTC.
- height of screen is a multiple of 8 scanlines.
- screen is no wider than 192 in mode 0, 384 in mode 1
- screen is no taller than 272 lines
- any animation or movement mocked up happens on 'byte boundaries' so horizontally every 2 pixels in mode 0, every 1 pixel in mode 1.

All of these make it more like how a real CPC game would be done and fit within proper CPC limits)

[arnoldemu]

This version from the 90s show part of what you suggest is possible:
https://www.cpc-power.com/index.php?page=detail&num=3625

In the above graphics ripped from IK+ and Barbarian II together but sprites are over a background.

If new ram expansions were used to give more breathing space for new data then a version of your mockup is possible.

I say a version because when it comes to implementing need to be mindful of cpu time so this means some might need to be adapted and you then get a version of what you designed.

For example if the background animated a lot and the only way to do that was to spend cpu time drawing the changes, then that would impact frame time and if it's too much then other methods would need be found or some animation dropped.

EDIT: But don't let any of this put you off mocking up graphics because it is always nice to see what could have been and there are times when the mockups do become a reality.

[Rob Beschizza]

I doubt very much you could achieve this by hacking the original game unless you completely pulled it apart, but if you rewrote it from scratch it should be perfectly possible.

I would imagine that all that dithering behind the characters is likely to be a problem, but it rather depends on how IK+ does its sprite masking. The fact the background is normally a solid colour may not be an accident, there are various tricks to accelerate masking in that case (although players overlap so who knows)

Someone  suggested I use a modern engine or library (Telera is one that comes up and looks promising) rather than go mad trying to learn and hack z80 and that looks like the consensus! I guess there's some nostalgic bug imagining what might have been done had a team had years to work on something instead of months.

*tricks and tips*

That was fascinating, because I knew the basics (specifics of the pallette restrictions) and other practical limits (don't scroll the background or animate it much), but not the detail-devils like those interrupt timing hoops. I guess the sun rippling effect perhaps involves that. A much more desirable effect than switching to mode 1 for some text.

EDIT: But don't let any of this put you off mocking up graphics because it is always nice to see what could have been and there are times when the mockups do become a reality.

That Barbarian/IK+ mashup is outstanding! And funnily enough another one I mocked up. All have the same inspo