Is there anyone working on a Wolfenstein or Doom clone for the CPC?
Cheers,
Peter
Better to do a nice Xybots with the same graphics as Gryzor...
:o
I'll have that as well! :D
Quote from: ComSoft6128 on 17:47, 27 March 18
Is there anyone working on a Wolfenstein or Doom clone for the CPC?
I think, while it's interesting from technical point of view, it's not that interesting from a gameplay point of view.
There were some good tech demos from Optimus and Richard Wilson worth checking out on Youtube.
https://www.youtube.com/watch?v=COH55Uj53TY
https://www.youtube.com/watch?v=FKPiBPNG1OI
The first demo looks very slick.
Yh, the 1st one really feels like Wolfenstein
So aside from old demo's, nothing?
Cheers,
Peter
Quote from: ComSoft6128 on 06:33, 28 March 18
So aside from old demo's, nothing?
Cheers,
Peter
I think you would need a faster CPU maybe a Z80H @ 8Mhz if you could get it to run on a CPC with enough oomph to power a small windowed version of Doom.
The Amiga needed a 020 powered A1200 with RAM and Accelerator to run a game called GLOOM even then it wasn't very fast. :-[
Quote from: Vyper68 on 18:38, 28 March 18
The Amiga needed a 020 powered A1200 with RAM and Accelerator to run a game called GLOOM even then it wasn't very fast. :-[
Eh I dunno. I used to play Gloom on my non-expanded A1200 in 2 player split screen mode quite a lot so it certainly moved well enough for me to not get bored of it. And I wasn't even really into Doom/Wolfenstein clones back then. More complex ones like Alien Breed 3D though were painfully slow on stock A1200... and to be honest weren't worth the extra hardware costs anyways as the older 2D top down Alien Breeds were far better games. :D
I don't know about you guys, but both of those examples are pretty impressive, if you ask me. I am amazed that 4mhz could produce something like those. Well done. Any way of finding out "how" they were produced? (assuming it's not a trade secret and such... lol)
J
Theoretical question time:
OK,
So it looks this type of game would be difficult to program for a standard 6128 but what about the 6128 Plus?
Would the enhancements on that machine make a FPS easier to create?
Cheers,
Peter
Quote from: ComSoft6128 on 02:14, 29 March 18
Theoretical question time:
OK,
So it looks this type of game would be difficult to program for a standard 6128 but what about the 6128 Plus?
Would the enhancements on that machine make a FPS easier to create?
Cheers,
Peter
I wouldnt have thought so unless hardware sprites were used to create the effect of 3d. I think the enhanced plus features of hardware scroll, sprites and palette were more suited to the kind of games commonly played back in the day. I don't know much about 3d games but i guess it involves alot of maths and processing power.
Sent from my E5823 using Tapatalk
Quote from: Carnivius on 19:19, 28 March 18
Eh I dunno. I used to play Gloom on my non-expanded A1200 in 2 player split screen mode quite a lot so it certainly moved well enough for me to not get bored of it. And I wasn't even really into Doom/Wolfenstein clones back then. More complex ones like Alien Breed 3D though were painfully slow on stock A1200... and to be honest weren't worth the extra hardware costs anyways as the older 2D top down Alien Breeds were far better games. :D
[ot]
I might be confusing what i had to what i needed there, it was a while ago. I may have been thinking about Alien Breed 3D now you mention that game.[/ot]
Wasn't Narco Police very much indicative of a first person shooter for its time? Xybots also had potential but they went for a lazy speccy port instead.
When I bought Narco Police, my naive 15 year old self, after seeing screenshots somewhere, was expecting to see a fully 3D first person shooter like Wolfenstein on the PC. Man was I disappointed when I found it was just like Xybots, only worse lol :laugh:
At least in Xybots there was a greater range of player movement in the weird flip screen perspective view.
Back to the FPS genre... It would be nice if that full 3D effect could be used in an actual game. Perhaps an expanded CPC with extra memory could do it? Or is the processor speed itself the limiting factor, using up all available CPU power just to calculate the first person perspective in real time?
Extra memory might not be necessary if game data was drip fed by multi-load from disc.
Regarding the limitations of the CPC processor to create a viable (fast) 3D game environment, I haven't a clue. Would any programmers like to comment?
Cheers,
Peter
Quote from: blackdalek on 13:14, 30 March 18
When I bought Narco Police, my naive 15 year old self, after seeing screenshots somewhere, was expecting to see a fully 3D first person shooter like Wolfenstein on the PC. Man was I disappointed when I found it was just like Xybots, only worse lol :laugh:
At least in Xybots there was a greater range of player movement in the weird flip screen perspective view.
Oh neither were fantastic, but you gotta remember, they were also catered primarily for 64k. If the additional 64k was used, there's a ton of potential there. I'm sure something could be done in 128k.
Extra memory can help somewhat, simply because you have more opportunity to store pre-calculated results of various operations and pre-calculating is often the simplest result to speeding things up. How much more is a different question. There really is only so much you can do, pre-zooming sprites for example and building up multiplication tables to some degree. There is still a lot of work that has to be done just purely shifting data around to draw things onto the screen and ultimately that's going to be a pretty big limiting factor when it comes to a first person shooter - you aren't ever going to get high frame rates, even ignoring all the complexity of dealing with massively banked amounts of memory.
Ultimately 8-bit machines just aren't really suited to this kind of thing, really you need a minimum of fast 32-bit integer arithmetic and enough CPU grunt to push full frame redraws at an acceptable speed. The various demos and small games on the C64 and Speccy, as well as the CPC ones linked, proved it's not impossible to get something kind of workable but it's still a pretty limited experience compared to even the original PC Wolfenstein 3D.
That first video was my first preview (also used in Wolfenstrad demo (http://www.pouet.net/prod.php?which=58887) ). It's non-playable, only previews how fast the wall rendering works. It reads a sequence of precalced Z values for each wall column, in a precalced animation. It shows, if we remove the math, how fast the wall rendering wall could be (2 VBLs, vsynced at 25fps)
Few years later I went back into the engine and found some ways to make the math faster (and less buggy, not crashing under certain angle views). I have later reviews where I control the player in realtime. Maybe 10-15 fps at most. Mind you, there is a lot of work to optimize, the raycasting is Still written in C, could take the effort to rewrite it in assembly.
https://www.youtube.com/watch?v=S3hsZbS4gH4 (https://www.youtube.com/watch?v=S3hsZbS4gH4)
The next thing is that I have to write proper sprite zooming. CPC+ features wouldn't help unless you stored inbetween zoomed version of sprites but then I would be dead on memory, unless I only want to have one enemy or something. Would be more convenient to just do it with software rendering.
This is only a brute force assembly code, I have an idea for faster code and would be fine with small sprites if not zoomed too close (I could have some restrictions for the maximum). Also the math are wrong (where to place the sprite against the player camera), it was mostly a test to test sprite zooming and how slow a regular brute force routine would be. I think if I can solve the math for sprite projection against the camera (and also clipping with the wall if behind) then something is possible.
https://www.youtube.com/watch?v=V0NXGO5b25k (https://www.youtube.com/watch?v=V0NXGO5b25k)
If only I can find the motivation and time to go back to this project. I believe it's possible to have a nice game if 12-15FPS are acceptable, and all we can see is other finished attempts in other 8bits. If the C64 can do it (Mood), then I believe a Z80 computer which is slightly better with math (16bit pseudo-regs are a bliss) could do even faster, it's just that it wasn't pushed to the edge as the C64.
Thank you for the info.
Cheers,
Peter