The Amiga 1000 could of done a game like Wolfenstein in 1985 - shock!

[AndNN]

Originally Posted by Samurai_Crow

The Copper can stretch pixels vertically by alternating modulo register values so the chunky to planar conversion doesn't have to stretch in that direction.

I'm curious. I use the Copper to alter the bit-plane pointers so I repeat the line and double up the pixel that way.

Forgive me if I should know this technique but I have not programmed an Amiga since the mid-ninties (just got back into it the last couple of months using emulators) so I probably don't know all the tricks for this. What's the technique alternating the modulo register?

[Samurai_Crow]

The modulo register is added to each plane pointer at the end of each row of pixels so alternating between negative one times the plane width and zero every row of pixels will cause the display DMA to reread every other row. It's similar to yours but only requires one Copper write per row and is not timing sensitive since the modulo always triggers at the end of the row regardless of when it's written.

[AndNN]

Originally Posted by Samurai_Crow

The modulo register is added to each plane pointer at the end of each row of pixels so alternating between negative one times the plane width and zero every row of pixels will cause the display DMA to reread every other row. It's similar to yours but only requires one Copper write per row and is not timing sensitive since the modulo always triggers at the end of the row regardless of when it's written.

That's sounds like a better technique! I'll give that a go. Cheers

[sandruzzo]

Why don't put the effort to support as my pixel mix as possible in order to do some good test? 1x1, 2x1, 2x2 1x2, would be great.

As I said 16 colors, will left us with so many dma cycles

[sandruzzo]

Is it possible to do ray cating by row instead by colum?

[desiv]

Originally Posted by chb

But of course later on, most machines, including most A1000s, the kickstart was put into ROM,

Not sure about that...
I know you could get a kickstart in ROM with a 3rd party board )that I believe required some soldering???), but I don't believe any A1000's came with a kickstart ROM..

That being said, I know Dragon's Lair used the kickstart memory to be able to allow the game to run on an A1000 with only 512k (otherwise, it required 1M on any other Amiga).

desiv

[robinsonb5]

Originally Posted by sandruzzo

Is it possible to do ray cating by row instead by colum?

Short answer: no.

Long answer: yes, as long as your game world consists of horizontal surfaces rather than vertical walls, and doesn't allow you to steer left and right, only up and down. Perhaps a spaceship game that would be to Descent what Wolfenstein 3D is to Quake?

[AndNN]

Yep, totally agree. For walls constant z is on the vertical. I do remember a technique that was used on the Game Boy Advance - you rendered horizontal but rotated the screen 90 degrees. Unfortunately the Amiga does not do screen rotation.

[Amigajay]

Originally Posted by AndNN

The Atari ST version of Wolfenstein uses 16 colours with 2x2 pixels and the framerate is not that good (don't get me wrong the guy that did that went through hoops to get that to work) but the Amiga can do so much better.

Was on the 1989 STE hardware, would be even more impressive on an 1985 ST!

[hooverphonique]

Originally Posted by AndNN

Yep, totally agree. For walls constant z is on the vertical. I do remember a technique that was used on the Game Boy Advance - you rendered horizontal but rotated the screen 90 degrees. Unfortunately the Amiga does not do screen rotation.

we could instruct people to physically rotate their monitors

[frank_b]

Originally Posted by Amigajay

Was on the 1989 STE hardware, would be even more impressive on an 1985 ST!

It runs on a standard ST too. The ST's screen layout is good for C2p.

[ Show youtube player ]

[Amigajay]

Originally Posted by frank_b

It runs on a standard ST too. The ST's screen layout is good for C2p.

[ Show youtube player ]

That's pretty decent for a regular ST then.

[sandruzzo]

Can someone provide me a good tutorial about Ray casting, I'll try to write down some test code?

[alpine9000]

Originally Posted by sandruzzo

Can someone provide me a good tutorial about Ray casting, I'll try to write down some test code?

http://lodev.org/cgtutor/raycasting.html

[DanScott]

Yes, the amiga 1000 could probably do the equivalent of that ST wolfenstein, but probably not much more

Of course, there are a few things that you "could" do, to make it run just as fast in a 1x1 pixel screen.. but I will leave it for others to discover (hint, may require a couple of megs of fast ram)

[DanScott]

second hint.. you don't need C2P

[AndNN]

Originally Posted by DanScott

second hint.. you don't need C2P

I'm intrigued, spill the beans

Barring having all the textures defined at all the possible sizes then using the blitter to blit one vertical line from the correct texture size, then I cannot think of a good way to do this. Would the blitter have enough time to blit 320 lines at full screen over 3 frames? It probably does not but may be worth a check. And would all the textures at predefined sizes fit into memory?

With the tests that I've done on bandwidth it's possible to do Wolfenstein at 2x2 pixels with 64 colours running at 17fps which will make the Amiga version better than the ST version for sure.

The bandwidth tests also prove what the cpu is capable of doing to write 1 byte(pixel) into a chunky buffer. And the 68000 running from fast, reading and writing does not have enough time for 1x1 pixels.

DMA contention is not really an issue for EHB mode when the cpu is allowed to run from fast memory. Also the blitter has enough time to do the c2p conversion spread over 3 frames.

So in theory we have one method that can work for definite! It will be interesting to see if it's now possible to get a higher resolution pixel than 2x2.

Edit

When I originally set out to do this I was hoping for 2x1 pixel resolution but the bandwidth forced me to degrade to 2x2 pixels. We also have to remember that it's not about drawing the world but sprites need to be considered into the equation - there's no point in doing a tech demo that cannot handle gameplay. The sprites can be done more efficiently than the world because constant z can be on the horizontal (basically sprite scaling) and the predefined instruction sequence can be geared to writing out DWORDS. Anyway, sprites are another side of the testing

[AndNN]

Originally Posted by hooverphonique

we could instruct people to physically rotate their monitors

But that's cheating

It does make me think about the old arcade machines where portrait mode was basically the engineers mounting the monitors at 90 degrees in the cabinets.

Here's some pure fantasy thought. Someone in 1985 got a game like Wolfenstein running on Amiga hardware. They upgraded to a faster processor and licensed out the chipset from Commodore and produced an arcade version. All in the realms of possibility but of course pure fantasy.

Anyway, less fantasy... can 2x2 pixels be improved upon - that is the real question now.

[sandruzzo]

Originally Posted by DanScott

second hint.. you don't need C2P

Can you helps us here, so we can do a great demo?

[AndNN]

Originally Posted by AndNN

I'm intrigued, spill the beans

Barring having all the textures defined at all the possible sizes then using the blitter to blit one vertical line from the correct texture size, then I cannot think of a good way to do this. Would the blitter have enough time to blit 320 lines at full screen over 3 frames? It probably does not but may be worth a check. And would all the textures at predefined sizes fit into memory?

With the tests that I've done on bandwidth it's possible to do Wolfenstein at 2x2 pixels with 64 colours running at 17fps which will make the Amiga version better than the ST version for sure.

The bandwidth tests also prove what the cpu is capable of doing to write 1 byte(pixel) into a chunky buffer. And the 68000 running from fast, reading and writing does not have enough time for 1x1 pixels.

DMA contention is not really an issue for EHB mode when the cpu is allowed to run from fast memory. Also the blitter has enough time to do the c2p conversion spread over 3 frames.

So in theory we have one method that can work for definite! It will be interesting to see if it's now possible to get a higher resolution pixel than 2x2.

Edit

When I originally set out to do this I was hoping for 2x1 pixel resolution but the bandwidth forced me to degrade to 2x2 pixels. We also have to remember that it's not about drawing the world but sprites need to be considered into the equation - there's no point in doing a tech demo that cannot handle gameplay. The sprites can be done more efficiently than the world because constant z can be on the horizontal (basically sprite scaling) and the predefined instruction sequence can be geared to writing out DWORDS. Anyway, sprites are another side of the testing

Actually there's not enough chip ram to do that by using the blitter to blit predefined graphic slices. I'm pretty intrigued by this claim.