It’s so hard to make convincing static because the Archimedes can’t do a fast enough software rendering. The only convincing simulation I’ve seen is by rapidly changing the frame buffer start address, which is what Star Fighter 3000 does. Of course it makes it impossible to do in a desktop window.
I think it could be quite easy to do static even with software rendering. The original Archimedes seems very fast to do full software copy of fullscreen bitmaps as evidenced by myself and others doing in demos, so one could have scanlines of different noisy pixels, then select different ones to copy every time. I think in my tests of fast LDMIA/STMIA even in the original archimedes I could reach a good framerate a bit above 50fps. And I think if one wants to avoid, it's possible to do it easilly with palette in 4bpp mode too. Generate a screen of random pixels, your palette is greyscale gradient, but randomize every frame the gradients in the palette. At least I think this would look good and with almost no CPU wasted.
@@Optimus6128 I have tried doing it using palette swapping and my recollection is that it looked awful. The human brain is too good at pattern recognition.
I am trying to find a binary of this demo but are unable to find it anywhere. not om demozoo and not on pouet. Any chance you have it available so I can get a copy? (and maybe archive it on demozoo as well?)
This seems to be available on the icebird.org site, which is one of the few remaining archives of old Archimedes and RiscPC demos: icebird.org/demos.html It's a good point though, I'll look to grab everything from there so we can get some uploaded to scene.org etc.
The Archimedes demo-scene, where everyone goes by their real name :D
It’s so hard to make convincing static because the Archimedes can’t do a fast enough software rendering. The only convincing simulation I’ve seen is by rapidly changing the frame buffer start address, which is what Star Fighter 3000 does. Of course it makes it impossible to do in a desktop window.
I think it could be quite easy to do static even with software rendering. The original Archimedes seems very fast to do full software copy of fullscreen bitmaps as evidenced by myself and others doing in demos, so one could have scanlines of different noisy pixels, then select different ones to copy every time. I think in my tests of fast LDMIA/STMIA even in the original archimedes I could reach a good framerate a bit above 50fps. And I think if one wants to avoid, it's possible to do it easilly with palette in 4bpp mode too. Generate a screen of random pixels, your palette is greyscale gradient, but randomize every frame the gradients in the palette. At least I think this would look good and with almost no CPU wasted.
@@Optimus6128 I have tried doing it using palette swapping and my recollection is that it looked awful. The human brain is too good at pattern recognition.
@@womble1981 You might be right, but I'll try it at some point. Then second option is the fast copy of noisy blocks :)
I am trying to find a binary of this demo but are unable to find it anywhere. not om demozoo and not on pouet. Any chance you have it available so I can get a copy? (and maybe archive it on demozoo as well?)
This seems to be available on the icebird.org site, which is one of the few remaining archives of old Archimedes and RiscPC demos: icebird.org/demos.html It's a good point though, I'll look to grab everything from there so we can get some uploaded to scene.org etc.