Asmusr

Awesome! I would like to see the source code.

I'm starting this thread as a one-stop resource for keeping track of ROM images for homemade cartridges. Jon's site is the most comprehensive list have found: http://www.hexbus.com/TI-99_4A_Home_Computer_Page/Downloads_and_Links.html Including: Arcturus ROM TI Never-Lander ROM RoboTron:2084 ROM TI Workshop ROM & Documentation Multicart zip file Turbo forth: http://www.turboforth.net/download.html Cortex Basic: http://www.avjd51.dsl.pipex.com/ti/ti.htm Fred Kaal's Mega Mod (128K): http://home.vodafonethuis.nl/fgkaal/Modules/modules.html Pitfall: http://atariage.com/forums/topic/158467-pitfall-binaries-and-source-code My own games: Titanium: http://atariage.com/forums/topic/215393-titanium/ TI Scramble: http://atariage.com/forums/topic/216064-ti-scramble-scramble-clone Road Hunter: http://atariage.com/forums/topic/221039-road-hunter/ Flappy Bird: flappybird.zip (also see post #2) Atarisoft muticart with Ms Pac-Man, Donkey Kong, Moon Patrol and Defender: atarisoft-multicart.zip (requires a 64K cart that starts up in the highest bank, 32K not required) Please post links to any other ROM images you are aware of.

Thanks, I wonder if anyone has made it through the first four levels.

You don't need the objects files TIC and TIO.

I'm afraid that won't work. Magellan is for the PC. The samples are graphics/map files.

Assuming you mean TI Scramble, you don't need the TISC file.

Just a little update to my demo, now with checkerboard tiles and corrected bass notes. Regarding the music, I made a small Java program that reads an ePSGMOD file and adjusts the frequency of all zero volume notes in channel 2 by 15/16 (assuming all notes used to control periodic noise have zero volume). This compensates for the difference between the SN76489 and the TMS9919 sound chips. Let me know if anyone wants it. Isoscroll.zip

Yes please.

Very impressive, and interesting finally to see the GPU pushed to its limits. With 'a better color mode' I assume you mean the bitmap layer? I was thinking about doing an image rotation demo, but I'm not sure 4 colors would be enough to create a good effect. Edit: I noticed the GPU demo doesn't work in Classic99, any particular reason?

Actually, after looking in Magellan, my map is organized like this: 00005 00045 00345 02340 12300 12000 10000 So in this case it makes sense to do what I did and store it as: 12345 12345 12345

I'm afraid I was lost already at the beginning. In which direction is the '8-pixel wide strip of the map' running, and what does the numbers mean? Sorry for being dense.

Hmm, now you mention it that does seem more logical. It would probably also be more efficient for building the name table.

Yes it would be nice with some videos for those of us who have never seen a 9938/9958 in action.

>> The bitmap layer is limited to 4 colors (AFAIK) ... As Tursi just reminded me in another thread you can get many more simultaneous colors than 4 on the F18A by changing the palette each scanline.

I don't think speed is an issue if you decided to use the GPU for the 3D/texture calculations, the problem is colors. You could either use the good old graphics 2/bitmap mode with all its limitations or the F18A bitmap layer (I don't think doing a 3D game using ECM tiles is an option). The bitmap layer is limited to 4 colors (AFAIK) but has no other limitations, so it would probably be the best choice. In order to double buffer the bitmap layer you would only use about half of the screen area. If you could accept these limitations a Wolfenstein like game is certainly possible, but only with very basic texture mapping. This is also discussed in another thread: http://atariage.com/forums/topic/211137-first-person-shooter-on-the-ti/?view=findpost&p=2736749&hl=%2B2d+%2Bshooter

In Magellan, using 24 patterns in total. The technique is basically the same as for the other scrolling routines, but in isometric scrolling you scroll 2 pixels horizontally for each 1 pixel vertically, so a transition depends on 6 tiles instead of 2 as in vertical or horizontal scrolling. E.g. if you want to move from F to T it passes through both A, B, C and D: ------------- | A | B | T | ------------- | F | C | D | ------------- I made some code in Magellan to identify and export the unique transitions. Since a transition depends on 6 tiles the 256 patterns are used very quickly, so in order to reduce transitions I copied the same blocks a few times. I haven't got a visual tool yet to assist me, and it's possible that the number of transitions could be reduced if you could see what was going on. Since I'm using almost the full character set I cannot store all pattern frames in VDP RAM, so I'm pushing them from CPU RAM for each frame. Without music it could run at 30 FPS but the music reduces this to 20 FPS. The map itself is a diagonal strip, but you can save memory by skewing it so it aligns horizontally, e.g.: 0000004444 0000333300 0022220000 1111000000 You can store as: 11223344?? ??11223344 So there's nothing actual 3D in this, it's just an illusion. But in a game you *would* have to maintain a map of the 3D locations of each block to make collision detection etc. I wonder how they did this in some of the amazing games that was made for the ZX Spectrum (Gunfright is my favourite): Any ideas how to do this with very limited memory? Edit: The bass notes in the music are made using periodic noise, but because the MOD2PSG tracker is designed for the SN76489 sound chip I had to lower the frequency by 15/16. What I did was simply to transpose each bass note down 1/2 note which explains why it's slightly off key.

I was in school

Sprites

Just let me know if you have suggestions for Magellan that would increase productivity

Impressive! And with a secret challenge...

A little demo to keep you on your toes Isoscroll.zip Thank you to Shock, whoever you are, for the music - and to Tursi for the music player. Run E/A 5 DSK1.ISO

The F18A has 12 bit colors, 4 bits for red, 4 bits for green and 4 bits for blue. 4096 colors in total.

I think it's a great idea. Let me know if there's anything I can do to help.

My guess is that you're very close to the 2 frame limit, and it just tips over on the hardware. You can measure the duration of the main loop (excluding the call to VSYNC) in Classic99 using the T(XXXX-YYYY) directive. You can also try to comment out the first line of VSYNC: MOVB @VDPSTA,R0 * Clear flag if already set Edit: My tests show that a loop takes up to 108000 clock cycles. Are you by any chance running Classic99 in 50Hz mode?