Wilmunder's Star Raiders II released

[Heaven/TQA]

Aric... Glad to see you here... As demo coder on the 6502 I know how hard it is to pull out 3d of the 6502... Last year we recreated and reversed engineered the Fractalus line drawing routine and SR2 seem to have nice 3D routines... How many particles are you able too pull off? 128?

[Tempest]

 

Where is the "unlike" button?

 

Congrats on the Masters degree and getting married though.

 

Well in my defense at that point he hadn't found the source code yet so this was just another lost game.

[ajw]

Two years ago I spent a day with game archivists including members of this forum and brought the Corvus drive with me. It seemed to spin up just fine, but it didn't boot. We didn't know if the problem was in the controller or even the joystick ports. I did find the warranty card, but I don't think that will help much. I'll leave that project up to people in this forum and I'll work on looking for the printout of the source and floppies.

 

We had talked about releasing the game, I've always been a bit worried since this was work I did for Atari and copyrights are for 70 years. My idea was to do a presentation at the Games Developers Conference on the 3D model used in the original game, and the original game dynamics and then as a finale bring out the sequel.... Attaching the file to an email with Kevin and clicking 'Send' took a lot less effort, though I have to admit that I did hesitate a bit.

 

I have been pretty open about my work on the game and even made a reference to my working on it on my website. I am fortunate to have worked on a long list of top games and so I am usually asked about those projects and not about ones that never shipped. If you add in the release of The Last Starfighter as SRII, and unless you really know your game history, the story is a bit confusing.

[ajw]

Aric... Glad to see you here... As demo coder on the 6502 I know how hard it is to pull out 3d of the 6502... Last year we recreated and reversed engineered the Fractalus line drawing routine and SR2 seem to have nice 3D routines... How many particles are you able too pull off? 128?

 

I don't remember there being a specific count of the number of pixels since drawing the pixel is just part of the work of the CPU, the other part is determining where to draw it. The decision was based on a judgement of the framerate since this was a real-time simulation. The goal was to support up to 2 ships, the star field, an explosion, and laser fire without slowing significantly.

 

In terms of CPU usage, solid line draws are not too bad, they can be done with integer arithmetic, but the longer the line gets, the more pixels, so the longer the draw takes. I wrote the code to calculate the endpoints and tried solid lines and sure enough when the ships got closer, they slowed down which defeated the goal to have a constant and high frame rate. I needed a plan B.

 

I came up with a different approach that was both fast and easy to implement. I remembered how you could find the midpoint of a line segment by using the Midpoint Formula. This is (x1+x2)/2, (y1+y2)/2. Well the 6502 couldn't do divides, but a binary right-shift will do an integer divide by 2, so with two adds and a shift per dimension, I had the midpoint of my line. Now take that midpoint and one of the endpoints and do it again, and repeat this a few times and you get the dotted lines you see in the ships. Since the number of points don't change, the framerate stays constant.

 

I didn't work on Fractalus directly but when I was at Atari I got to know the teams working on Fractals & Ballblazer and I worked on the fractal code for Koronis Rift. There is another story behind this....

 

Loren Carpenter was doing the fractal code for the Genesis effect used in the movie Wrath of Kahn. David Fox was working on a game design and they were working in the same building. David admired Loren's work and lamented that the 800 didn't have the CPU to do fractal fly-overs. What David didn't realize was that Loren admired video game programmers because while he was using a bank of computers that might need 24 hours to generate a single video frame, we were using a 6502 and creating a new frame every 30th of a second.

 

While David thought that doing fractals would be impossible, Loren really wanted to try, and a few weeks later they had a demonstration of the code.

 

Loren and David faced a problem that the terrain rendering code worked by drawing vertical stripes so you couldn't rotate the screen. This is why you can only bank your ship a limited amount and even this was done by simply increasing or decreasing the y-position of the points based on their distance from the center of the screen. The screen is actually sheared, not rotated.

[doctorclu]

I've written to a few people that seem to have some experience with Corvus drives (fibreware, Caterpiggle, NGFrankW, joeventura, COBRAIP , Kyle22) to see if they have any advice to give about getting the Corvus drive to boot.

 

What AJW said previously on the Corvus drive:

"It seemed to spin up just fine, but it didn't boot. We didn't know if the problem was in the controller or even the joystick ports. "

 

I know Fibreware was developing Corvus drivers and interfaces a few years back. Maybe between trying different interfaces and maybe even other Atari 8 bit computers (for the joystick question) we can narrow down what the problem is.

 

AJW said previously (while I was writing the above bit)

"Loren Carpenter was doing the fractal code for the Genesis effect used in the movie Wrath of Kahn. David Fox was working on a game design and they were working in the same building. David admired Loren's work and lamented that the 800 didn't have the CPU to do fractal fly-overs. What David didn't realize was that Loren admired video game programmers because while he was using a bank of computers that might need 24 hours to generate a single video frame, we were using a 6502 and creating a new frame every 30th of a second. "

 

Love that sequence in Star Trek 2!

 

https://www.youtube.com/watch?v=52XlyMbxxh8

 

24 hours for a single video frame? Wow! Yeh I guess getting what was accomplished in Rescue on Fractilus and here in Star Raiders 2 would be extremely impressive. I guess as a kid I would see scenes like those in Wrath of Khan and the more simple versions on the Atari 8 bit computers and never realized how amazing that was at a simplier scale. Now with what you said I am even more amazed.

Edited December 13, 2015 by doctorclu

[Heaven/TQA]

 

I don't remember there being a specific count of the number of pixels since drawing the pixel is just part of the work of the CPU, the other part is determining where to draw it. The decision was based on a judgement of the framerate since this was a real-time simulation. The goal was to support up to 2 ships, the star field, an explosion, and laser fire without slowing significantly.

 

In terms of CPU usage, solid line draws are not too bad, they can be done with integer arithmetic, but the longer the line gets, the more pixels, so the longer the draw takes. I wrote the code to calculate the endpoints and tried solid lines and sure enough when the ships got closer, they slowed down which defeated the goal to have a constant and high frame rate. I needed a plan B.

 

I came up with a different approach that was both fast and easy to implement. I remembered how you could find the midpoint of a line segment by using the Midpoint Formula. This is (x1+x2)/2, (y1+y2)/2. Well the 6502 couldn't do divides, but a binary right-shift will do an integer divide by 2, so with two adds and a shift per dimension, I had the midpoint of my line. Now take that midpoint and one of the endpoints and do it again, and repeat this a few times and you get the dotted lines you see in the ships. Since the number of points don't change, the framerate stays constant.

 

I didn't work on Fractalus directly but when I was at Atari I got to know the teams working on Fractals & Ballblazer and I worked on the fractal code for Koronis Rift. There is another story behind this....

 

Loren Carpenter was doing the fractal code for the Genesis effect used in the movie Wrath of Kahn. David Fox was working on a game design and they were working in the same building. David admired Loren's work and lamented that the 800 didn't have the CPU to do fractal fly-overs. What David didn't realize was that Loren admired video game programmers because while he was using a bank of computers that might need 24 hours to generate a single video frame, we were using a 6502 and creating a new frame every 30th of a second.

 

While David thought that doing fractals would be impossible, Loren really wanted to try, and a few weeks later they had a demonstration of the code.

 

Loren and David faced a problem that the terrain rendering code worked by drawing vertical stripes so you couldn't rotate the screen. This is why you can only bank your ship a limited amount and even this was done by simply increasing or decreasing the y-position of the points based on their distance from the center of the screen. The screen is actually sheared, not rotated.

Exactly what we found when reverse engineering and understanding RoF. As it only has 48 lines of view... People hardly realize that moving up/down only "scroll" and left/right do tilting by adding offsets across the line... So no rotation but as you are in a 160x48 viewport you don't realize that's a fake. Add the thrill action and sound and voila... People are distracted that's not "real"... But RoF and Koronis rift (and SR...) give you a "realistic world feeling" that your mind makes you feel immersive....

 

Add on top the clever Eor filler (which was discovered later mainly with release of the Amiga Chipset mainly Blitter) and the one used in Koronis Rift (using 80x64 res) and works well...

 

I showed Loren Carpenter via Facebook the results but obviously he did not responded but our fractal line code does have the same "jaggies" and "visible errors" like the ones in RoF and Koronis Rift...

 

But back to SR2.... The midline division approach sounds very interesting... saved into my notebook . Never thought that this would be faster than actually drawing the lines... (Using eg Bresenham et al).

[Heaven/TQA]

don't want to sidetrack... but Ajw... this should remind you Koronis Rift somehow... (just start that in Altirra Atari Emulator PAL preferred and press START... (you can not do anything... just proto for the line algo).

 

 

fractaltest.zip

[spookt]

Just got around to playing on my 800. Awesome! I did find the ground attack section harder than the space combat but a really great game. Such a huge pity this never saw the light of day at the time. I for one would have snapped it up. Thanks for sharing it with the community.

[Kyle22]

Hi Aric. Great game, and welcome to AtariAge!

 

Now, let's get your Corvus working. It would be helpful if you could provide as much detail as possible about the drive, interface, and any software you may have for it (diagnostics, utilities, etc.).

 

Any and all info (model #s, serial #s) is helpful.

 

-K

 

Edit: Also, please describe the action of the ready light.

Edited December 13, 2015 by Kyle22

[Lynxpro]

Great interview! Finally finished it.

 

I find it amusing there were secret teams at Atari Coin that were doing home computer games. If I recall, Ed Logg said in another interview that they were all supposedly banned from doing such work - especially after the royalties for home conversions were finally hammered out - but he had done so on the 2600 and later the Famicom with some semi-official managerial support.

 

Mr. Wilmunder, is it possible you might've worked on finishing up the Atari 7800 version of Ballblazer while at Lucasfilm Games? Although I guess with that date, it might've been making sure the existing Atari 8-bit version worked in cartridge form for the XE Game System since that was released in circa 1987. Actually, I think both the 7800 and XEGS versions were released in that year.

 

As for polishing up old games, there's a lot of that going on these days in these parts. Adding Trak-Ball support to a lot of 2600 games that should've had it from day one has been happening feverishly here in the past couple of months. And then there's Keithen Hayenga who finished up 5200 Tempest decades later after the original source code was discovered. Speaking of, just as you worked on the Danger/T-Mobile Sidekick, he worked at Palm/PalmSource/HP on PalmOS and later webOS. Throw in Joe Decuir and you have 3 ex-Atari people who have definitely contributed greatly to our modern smart phone world!

 

It was interesting to hear about your interactions with the Tramiels. So, post-July 1984, did you move over to Tramiel's Atari Corp from Atari Coin/Games when you were mentioning your work on Star Raiders 2 or was this during the time period where Corp and Games were both sharing a building and you were thus an Atari Games - aka the real Atari Inc still - employee asking if they'd still be interested in releasing the game? If you were still at Games then I'd imagine Games would've had the rights to the code and game although the Star Raiders brand obviously was retained with Corp. If you were a Corp employee at that point, then the game rights would probably still be with the current "Atari" *cough* Infogrames *cough* but if you were with Games, then the rights would probably be with WB Games today. The plot thickens!

[Level42]

Simply awesome !!!!

 

My son loves SR, he's 13

[joeventura]

I sold all my Corvus stuff to Fiberwire, who sold it to someone else a year or so ago it had the hardware and all the software and was 100% functional

[Kyle22]

Aric, I forgot to mention earlier that for normal operation, the switches on the front of the drive must all be in the left position.

[ajw]

Hi Aric. Great game, and welcome to AtariAge!

 

Now, let's get your Corvus working. It would be helpful if you could provide as much detail as possible about the drive, interface, and any software you may have for it (diagnostics, utilities, etc.).

 

Any and all info (model #s, serial #s) is helpful.

 

-K

 

Edit: Also, please describe the action of the ready light.

 

Please contact TheRedEye.... He currently has the drive.

 

http://atariage.com/forums/topic/246591-wilmunders-star-raiders-ii-released/?p=3389084

[ajw]

Great interview! Finally finished it.

 

I find it amusing there were secret teams at Atari Coin that were doing home computer games. If I recall, Ed Logg said in another interview that they were all supposedly banned from doing such work - especially after the royalties for home conversions were finally hammered out - but he had done so on the 2600 and later the Famicom with some semi-official managerial support.

 

Mr. Wilmunder, is it possible you might've worked on finishing up the Atari 7800 version of Ballblazer while at Lucasfilm Games? Although I guess with that date, it might've been making sure the existing Atari 8-bit version worked in cartridge form for the XE Game System since that was released in circa 1987. Actually, I think both the 7800 and XEGS versions were released in that year.

 

As for polishing up old games, there's a lot of that going on these days in these parts. Adding Trak-Ball support to a lot of 2600 games that should've had it from day one has been happening feverishly here in the past couple of months. And then there's Keithen Hayenga who finished up 5200 Tempest decades later after the original source code was discovered. Speaking of, just as you worked on the Danger/T-Mobile Sidekick, he worked at Palm/PalmSource/HP on PalmOS and later webOS. Throw in Joe Decuir and you have 3 ex-Atari people who have definitely contributed greatly to our modern smart phone world!

 

It was interesting to hear about your interactions with the Tramiels. So, post-July 1984, did you move over to Tramiel's Atari Corp from Atari Coin/Games when you were mentioning your work on Star Raiders 2 or was this during the time period where Corp and Games were both sharing a building and you were thus an Atari Games - aka the real Atari Inc still - employee asking if they'd still be interested in releasing the game? If you were still at Games then I'd imagine Games would've had the rights to the code and game although the Star Raiders brand obviously was retained with Corp. If you were a Corp employee at that point, then the game rights would probably still be with the current "Atari" *cough* Infogrames *cough* but if you were with Games, then the rights would probably be with WB Games today. The plot thickens!

I was at Corp Research and frustrated with the project, Gossip, I was asked to implement so when I finished the work I resigned. Before I left, Chris Horseman, who worked for Lyle Rains in Coin-Op, hired me to implement SRII. Our team started in an office in a warehouse where the games were stored before trucks came in to pick them up. I wish I had a photo since it was not unlike the warehouse at the end of Raiders of the Lost Ark, but with rows of coin-op machines for over a hundred yards. Later we moved into the main Coin-Op building in what I think was Milpitas.

 

While I don't remember Chris Horseman providing significant input on the design of SRII, I remember putting his name before mine on the design document since he was my boss. I believe that Chris' was married to Kim Whitmore. Her background was in optics, and my boss at Corp Research didn't think my implementation of Gossip was correct, so he hired Kim to go through the code and check my math. I heard later that Kim couldn't find any problems with my implementation, just that the design was flawed from the beginning.

 

Re: secret projects, someone described working for a company where the right hand didn't know what the left hand was doing. At Atari, I think it was more the case that the right hand didn't even know that there was a left hand.

 

I didn't work on the 7800, but after the original Ballblazer programmer left Lucasfilm Games, we had a request to build cartridges for what I think was called the 65XE. I was asked to put this together and to the original coder's credit, I don't remember the changes being particularly difficult.

 

Ballblazer was a case of the engineer creating the special effect, and then developing a game to match. The engineer came up with the grid using a specific coding trick on the graphics card. Then he needed a game to put on top, and I heard that other designers were brought in to help. Since the game was so dependent on this graphic trick, implementing Ballblazer on other platforms was really difficult.

 

Re: Tramiels, I was in Coin-Op when I was Riffed. There had been 8,000 employees when I started, and fewer than 500 when I left, so I didn't take it too personally. I only remember meeting with the Tramiel's twice. Once shortly after leaving Coin-Op to get permission to continue working on the game. About a month later I met with them again to show them the progress and to sort out the business side, and it became clear that they weren't going to agree to anything. Shortly after this I found myself invited to meet with the team at Lucasfilm Games. I do believe that at the second meeting they showed me a prototype of the Atari ST.

[Marco!]

WOW! I'd never see the day that there'd be a sequel to this game, SR was one of the four cartridges that I bought together with my Atari 800XL in 1984 (the others were Ms. Pac Man, Donkey Kong, and Galaxian). Who cared if it said ©1979, it was the most awesome game ever as I used to play it on the 2600.

 

As for the botched/strange introduction, perhaps at one point this can be turned in to an official disk-version which would allow for a proper intro which can be (part of) the loader. This may -appropriately inspired by the reimagined Battlestar Galactica- involve the "old graphics" to tell the story of the first human-Zylon war, then set the story for the second Zylon invasion using the "new graphics". Maybe we can also find out what happened to Star Cruisers 1 through 6, or perhaps that best left in reserve in case a retro programmer decides to create a Star Raiders Zero (prequel) game, but I can also see a Star Raiders III (not necessarily for that Atari 8-bit platform) featuring multiple galactic charts (one per sector) where you use Hyperwarp to travel within the sector and Transwarp to travel between sectors (there's some strategy stuff that can be built there, e.g. limiting the number of Transwarp jumps or special requirements to perform one, like having 9500+ energy, then end up with so low energy you need to dock at a Star-Base ASAP), and new formidable enemies, the Grob (Borg spelled backwards), the Gon-Klin, and Morulan Star Empire. I like the idea of having both lasers and photon torpedoes. Perhaps the number of torpedoes can be limited to a certain number, after which you can resupply at a Star-Base (kinda like Blue Max works, first energy, then torpedoes, then damage, and you can leave early --with the shuttle still en route, perhaps use different colors for the different functions-- to gain a strategic time advantage ).

[doctorclu]

I'm just wondering what is the story with the Tie Fighter and the sleek ship. Are they designed to take multiple hits, or are you trying to hit them in a direct spot? Can the Tie fighter only be hit in the mid section? How about the Sleek ship?

 

The bombing run seems the only effective way to take out a base is when they are far off in the distance.

 

And how many ships around a starbase is still a surround threat? Is it the number of squadrons or number of total ships?

[ajw]

I'm just wondering what is the story with the Tie Fighter and the sleek ship. Are they designed to take multiple hits, or are you trying to hit them in a direct spot? Can the Tie fighter only be hit in the mid section? How about the Sleek ship?

 

The bombing run seems the only effective way to take out a base is when they are far off in the distance.

 

And how many ships around a starbase is still a surround threat? Is it the number of squadrons or number of total ships?

 

I believe that the original Star Raiders collision detection was done in the core 3D spacial coordinates. The SRII Lasers are done in screen coordinates and I'd have to look at the code to see if I used a simple scale for the target size. I would also want to check the code to see the # of required hits.

 

It has been many years, but I think that the ships were done in a cylindrical space using sines and cosines but using 0-255 degrees. The 3D star field is using a different coordinate system that allows you to rotate a point around the center with a simple addition and then a divide by 256. Using 16 bit values, you get the divide for free by just taking the high byte. This method does deteriorate over time, but I had thought about doing the ships in this format which would have allowed rotations in any dimension. That might have allowed for more realistic movement, but I would have had to find a way to reset the ships on occasion since the rotate is destructive over time. The best way to view this rotate is to go into the long range scan, set your engine speed to 0, and just rotate. If instead of stars, these were the endpoints of the ships, you could do rotations in all 3 dimensions.

 

The file upload system doesn't like spreadsheets soI have attached an image of one I created that shows the rotation. Note the loss of accuracy after the full rotation.

 

[billkendrick]

Loren Carpenter was doing the fractal code for the Genesis effect used in the movie Wrath of Kahn. David Fox was working on a game design and they were working in the same building. David admired Loren's work and lamented that the 800 didn't have the CPU to do fractal fly-overs. What David didn't realize was that Loren admired video game programmers because while he was using a bank of computers that might need 24 hours to generate a single video frame, we were using a 6502 and creating a new frame every 30th of a second.

 

I went to Sonoma State University, and Tom Duff (and someone else, maybe it was Loren?) spoke at a computer science symposium one day. They talked about the struggles to get the Genesis effect going, including having to (really slowly) re-render a segment of the scene, because (due to the randomness(?) of the terrain) at one point the camera slammed through a mountain. (They solved it by 'forcing the mountain to be flat' or something )

 

Anyway, the best memory I have of this talk was the complaint that, after all the time they spent perfecting this amazing 100% computer graphics scene, the editors or director or someone decided to cut away to show Kirk & Spock "looking thoughtful" for a few seconds, basically ruining the "from dead planet to full-of-life" effect. D'oh!

[SoundGammon]

For those who don't have an Atari 8-bit computer, there's WARPSPEED by Accolade for the Sega Genesis and Super Nintendo.

 

Played it a little a few years ago on an emulator. A lot like Star Raiders.

[Tickled_Pink]

 

The original Star Raiders came out in 1979 for the Atari 400/800. It was groundbreaking. I first saw it at an electronic store, that was using it to sell their big screen TVs. I was guffawed, slack-jawed, jizzing myself, all at the same time.

 

I recall Star Raiders was a release title for the 5200. The knock-offs (Starmaster, Star Voyageur, Phaser Patrol, etc.) didn't start coming out until 2 or 3 years later.

 

The primary inspiration for Star Raiders was the text-base game, Star Trek, first coded in 1971. You can play Stellar Track, for the 2600, to get a feel for that game.

 

I think that's the issue I have. I didn't pick up an XL until 1985, by which time other systems had games like Elite in their library.

 

The first time I saw Star Raiders was on a cousin's 2600, probably around '81, and I'm pretty sure I was impressed by it back then. I have a copy that I bought NIB a few years ago. I've broken open the shrinkwrap but have yet to plug it in or even open the box containing the keypad.

[Mclaneinc]

AJW, thank you for releasing the game even though you had some worries over doing it, its pure Star Raiders and then some. But I also want to thank you for the stories re the game and Atari in general, I'm very in to film SFX (my collection of Cinefex proves that) and the talk of the genesis effect was very interesting considering it also linked in the Atari, I love all these titbits, what for you are memories, are grand stories for us lot.

 

Thank you...

[almightytodd]

The video looks awesome but I'm having trouble figuring out the right settings to run it in Altirra... ...maybe it's time to RTFM...

[doctorclu]

Screen Shot 2015-12-13 at 10.45.30 PM.jpg

 

I believe that the original Star Raiders collision detection was done in the core 3D spacial coordinates. The SRII Lasers are done in screen coordinates and I'd have to look at the code to see if I used a simple scale for the target size. I would also want to check the code to see the # of required hits.

 

It has been many years, but I think that the ships were done in a cylindrical space using sines and cosines but using 0-255 degrees. The 3D star field is using a different coordinate system that allows you to rotate a point around the center with a simple addition and then a divide by 256. Using 16 bit values, you get the divide for free by just taking the high byte. This method does deteriorate over time, but I had thought about doing the ships in this format which would have allowed rotations in any dimension. That might have allowed for more realistic movement, but I would have had to find a way to reset the ships on occasion since the rotate is destructive over time. The best way to view this rotate is to go into the long range scan, set your engine speed to 0, and just rotate. If instead of stars, these were the endpoints of the ships, you could do rotations in all 3 dimensions.

 

That modification sounds great. If we ever get the source that would be something worth looking at. Thanks for explaining how this Star Raiders and the original Star Raiders work for hit detection.

[Irgendwer]

In the meantime, can anyone figure out the Altirra settings for the touch tablet? There aren't any presets, and I can't seem to make anything work with the touchpad.

 

http://atariage.com/forums/topic/218202-altirra-240-final-out/page-12?do=findComment&comment=2968761

 

I really wonder why phaeron doesn't supply this configuration with his emulator. The CX77 was an Atari device and Atari800 has a simple to apply "Touch Tablet"-setting too.

Maybe he needs more feedback to include such configuration...

 

I now have to test SR2 with the CMI08...

Edited December 14, 2015 by Irgendwer