Jump to content
IGNORED

Hard Drivin’ on the 8 Bits?


EnderDude

Recommended Posts

I'm pretty sure the title screen can be achieved by using Rasta Converter, right?

Yes, but you run the chance of banding. I think a good pixel artist would be better off using G2F (Graph-2-Font) for this.

 

Just because, here's a WIP. Won't take long to finish this one out. I still say G2F is the way to go for this, if nothing more, than for the smaller size file we'll get.

post-650-0-65472800-1539448916_thumb.png

  • Like 3
Link to comment
Share on other sites

Yes, but you run the chance of banding. I think a good pixel artist would be better off using G2F (Graph-2-Font) for this.

 

Just because, here's a WIP. Won't take long to finish this one out. I still say G2F is the way to go for this, if nothing more, than for the smaller size file we'll get.

attachicon.gifhd_progress.png

Damn. Wouldn't you know it. Got busy on other stuff, let this run to 1.4 Billion iterations. It got down to a normal of 0.2, so this "can" be a perfect conversion. I let it run too long, so there is a single bad line in it. Also, forgot to center the image, so I will be restarting it. Restarting now. Here's screenshots of Rasta and Altirra. Notice, the bad yellow line is NOT in the Rasta middle window :(

 

post-650-0-32709200-1539468247_thumb.png

post-650-0-41997600-1539468243_thumb.png

 

  • Like 5
Link to comment
Share on other sites

Nice work! While you were working on that, I've made an optimized picture of a fan-made marquee online, slightly edited. What I've done to optimize the picture was to go into GIMP, set the color mode to the gtia palette, Ctrl+X an 8-pixel high strip, and paste it into another canvas, where I index it to a generated 8 color palette for each strip, Ctrl+X that, and paste it back into the picture, and I've repeated this process until I got to the bottom of the picture, and this was the result. I actually don't mind the stripes that much, though...

Link to comment
Share on other sites

I wonder, how long it took to create the useless C64 version of the game.I also could bet that the same "team" would have made more money, doing the game for the A8 , using the "right" graphics ;)

For those who haven't viewed the GTW page on the C64 version:

 

 

Another early version it seems existed of another Domark game, this time of the terrible Hard Drivin’ game conversion being worked on. Just how did the Spectrum receive a fantastic conversion, and we get this game? (The game was so bad that it didn’t get a full release on its own – only through compilation and budget).

 

Zzap in 1990 had the screenshot which depicted a very slightly more colourful conversion.

 

The artist behind the loading picture confirmed to GTW64 that the above screenshot from Zzap was infact an earlier version of the released game, but there was also an earlier version of Hard Drivin’ being done by a completely different programmer. The developer (Whom we are trying to get the name of) hit a lot of troubles trying to get the game finished – though apparently not that much was actually running, so they gave up and called it quits (Never to work for Domark again). This left Domark with no game, and so they drafted someone else in quickly to finish things.

 

However, advertising had already been spent we assume and there was a deadline. In typical “Total Recall” fashion, the developer had just two weeks to get it complete, and from scratch! The team was not given any code or assets to work with, the new team started from scratch.

 

The panel you see in the screenshot is an earlier staged version of the panel graphics by Andrew McCarthy, which were then changed later to a final version with hi-res elements. Apparently this version was running actually at a decent framerate (believed to be around twice the speed of the final game) with both tracks included. The problem came when the other cars on the track were introducted, which really caused issues – of which there was no time to solve as Domark was desperate to release the game.

 

The artist behind the loading screen felt that the second developer (Who although we think we know who it is, we’ll spare their blushes) got hte shitty end of the stick trying to do a complex game from scratch in two weeks. But he was young and eager and should not of taken it on. Sadly he no longer works in the industry and left in 2000 after being messed around by publishers, ridiculous requests, lack of regular funding and ever changing schedules.

 

Yes..your reading that correctly. .entire game redone from scratch in 2 weeks. ...

  • Like 3
Link to comment
Share on other sites

Ok, so a good excuse for a poor C64 version, but done to the best of their respective abilities, C64 or A8, would still only serve as an impressive technological achievement, a proof of concept for another real game that's actually fun to play. Stunt Car Racer has tech achievement and fun to play game covered in spades. So how about we take either the Spectrum or C64 versions, or both, as failed proof-of-concepts, one because of not enough custom hardware to do it justice, and one due to being a super-duper rush job. Pick up the pieces, throw them through the Atari custom graphic and faster-than-C64 CPU grinder, and spit out another 3D vector/polygon game that looks like a Hard Driv'n 3D like engine, with a speed and playability reaching for the new standard which is Stunt Car Racer.

Edited by Gunstar
  • Like 2
Link to comment
Share on other sites

Yes..your reading that correctly. .entire game redone from scratch in 2 weeks. ...

A "Team" worked on it for 2 Weeks.

Sounds very coarse, doesn't it?

 

On the other hand: A "Team" managed to create the "3D" projected game in 2 Weeks. But, in another thread, people doubt about "a weekend" to have the raw engine of a "Sprite with Scrolling" game available ...

 

But that's another story.... or not?

 

As the question "why is the Spectrum Version so much better" , refers to the misunderstanding, of what the C64 is able to do.

What could it be that a 1MHz 6502 is slower than a 3.5MHz Zilog CPU?

;)

Edited by emkay
Link to comment
Share on other sites

Efficiency of instructions? The Z80 seems a lot less efficient. In fact, the A8s have more MIPS than the ZX Spectrum, multiplying 0.58 (0.58 MIPS @ 4 Mhz) by 0.875 (3.5 Mhz/4 Mhz), and I got about 0.5 MIPS. However, the Atari is actually faster at about 0.76 MIPS (0.43 MIPS x 1.79 Mhz), one and a half times the MIPS of the Spectrum, and using a 160 px mode can reduce the load on the CPU, right?

  • Like 1
Link to comment
Share on other sites

Efficiency of instructions? The Z80 seems a lot less efficient. In fact, the A8s have more MIPS than the ZX Spectrum, multiplying 0.58 (0.58 MIPS @ 4 Mhz) by 0.875 (3.5 Mhz/4 Mhz), and I got about 0.5 MIPS. However, the Atari is actually faster at about 0.76 MIPS (0.43 MIPS x 1.79 Mhz), one and a half times the MIPS of the Spectrum, and using a 160 px mode can reduce the load on the CPU, right?

It's not all about "MIPS". It's also about the power of usable commands. The ZILOG can do 16 Bit operations, which is very powerful to handle things on the screen.

Stunt Car Racer on the C64 uses the trick for faster screen updates with character mode and color ram.

But, the game of this thread uses filled polygons, which look much more polished on the C64 btw. , so the CPU has to do much more than just some calculations. The CPU could do either 3D maths OR Screen handling, but not at the same time.

The Atari has the 1.5 times faster CPU, when using Antic D , plus Antic D fills 50% of the screen and the CPU can do calculations at the same time. To do the same with just the CPU, 3MHz would be needed...

Link to comment
Share on other sites

As the question "why is the Spectrum Version so much better" , refers to the misunderstanding, of what the C64 is able to do.

What could it be that a 1MHz 6502 is slower than a 3.5MHz Zilog CPU?

;)

Besides what Enderdude said above about the Atari's 6502c 1.79MHz CPU vs. Zilog-80 3.5MHz CPU in MIPS, surround even a 1MHz 6510 with some custom chips to take the load off the CPU, and you have a more powerful system. Now take that 1.79MHz 6502C that already beats a 3.5MHz Z80 in MIPS, and surround it with custom chips to take the load off the CPU, like 15+ graphic modes to choose from, that even with low resolution, when dealing with 3D vectors/polygons, low res, but fast and smooth still looks good.

 

There are plenty demos out there that show it off, Obviously those demos are mostly showing basic shapes spin and interact, but there are some like 'Control' that are more. And such techniques could probably be further optimized. I'm not saying the way they did things for Control or other demos is the way it would be done here, they are just examples of 3D vectors/polys that programmers back in the Atari 8-bit's heyday never would have thought could be done on the Atari.

 

It's about the whole motherboard system, not just the CPU, even way back when...and Jay Miner created one hell of an 8-bit system that I have seen push the envelope far beyond what Jay ever conceived again and again, Space Harrier, AtariBlast! and now Stunt Car Racer just to name a few in the last decade, nearly 30 years after the system was designed and engineered. But again, let's not waste it on Hard Driv'n.

Edited by Gunstar
  • Like 1
Link to comment
Share on other sites

It's about the whole motherboard system, not just the CPU, even way back when...and Jay Miner created one hell of an 8-bit system that I have seen push the envelope far beyond what Jay ever conceived again and again, Space Harrier, AtariBlast! and now Stunt Car Racer just to name a few in the last decade, nearly 30 years after the system was designed and engineered. But again, let's not waste it on Hard Driv'n.

40 Years ;)

And, yes, it's about the "whole system". That's why I claim that an optimized "3D" game always could run at the triple speed, of what you see done on the C64, and the double speed of what's been done on the Spectrum.

  • Like 1
Link to comment
Share on other sites

A "Team" worked on it for 2 Weeks.Sounds very coarse, doesn't it?On the other hand: A "Team" managed to create the "3D" projected game in 2 Weeks. But, in another thread, people doubt about "a weekend" to have the raw engine of a "Sprite with Scrolling" game available ...But that's another story.... or not?As the question "why is the Spectrum Version so much better" , refers to the misunderstanding, of what the C64 is able to do.What could it be that a 1MHz 6502 is slower than a 3.5MHz Zilog CPU? ;)

You'd have to take the points about Frank Gasking's (GTW site owner and researcher) throw away comments,up with him on the comments section of C64 Hard Drivin on the site.

 

I suspect, knowing Frank, they were very tongue in cheek and just a gentle ribbing of the Spectrum and a nod to the classic UK playground battles between ZX Spectrum and C64 owners.

 

They aren't meant to be taken seriously.

 

The information GTW uncovered is the serious content.

 

Oh and 2 weeks is a luxury for UK Coders in the industry back then. ..

 

People had been told if they couldn't port an Amiga game to the Acorn A3000 in under a week, not to bother.

 

They did it

Edited by Lost Dragon
  • Like 1
Link to comment
Share on other sites

40 Years ;)

And, yes, it's about the "whole system". That's why I claim that an optimized "3D" game always could run at the triple speed, of what you see done on the C64, and the double speed of what's been done on the Spectrum.

uh...of course, 40 years, and that's what it was suppose to say, a typo from typing in the dark and bad eyesight...time to get a PC keyboard that's back-lit... ;)

  • Like 1
Link to comment
Share on other sites

the z80 beats the 6502 imho in terms for 3d calcs due to 16bit reg pairs etc...

I can't disagree with that, but my point is custom support chips can help out, even if it's using 2D tricks that look 3D in some areas or objects, mixed with 3D with a better end result, even if it's not all true 3D. Also, as I mentioned, being able to use certain Atari graphic modes that can multiply the speed of the 3D engine over the Spectrum's Z80 speed in a higher resolution requiring more pixel pushing than say one of the Atari's GTIA modes, at like half or quarter resolution, but with more colors too.

  • Like 1
Link to comment
Share on other sites

I really can only encourage to have a look again on the table I've posted.

 

Now take that 1.79MHz 6502C that already beats a 3.5MHz Z80 in MIPS, and surround it with custom chips to take the load off the CPU, like 15+ graphic modes to choose from, that even with low resolution, when dealing with 3D vectors/polygons, low res, but fast and smooth still looks good.

 

Actually the other way around is true: the custom chip (ANTIC) generates load on the bus and that slows the CPU down. You can try (f.e. by using a more load friendly graphics mode (the often mentioned "D" here)) to reduce the load, but it's not a question to "speed-up the CPU" but to minimize the slow down.

 

...

The Atari has almost the same "CPU power" as the NES. Using Antic Mode D leaves approximately 1.5MHz for calculations.
...
Polygons have been created in the half time , necessary, additional objects, like bridges, could be done with PMg.
...

 

Examine the bold part: With PMG-DMA you're actually putting more load on the bus which will reduce your (quite optimistic) 1.5 MHz equivalent by about 0.07 points... (=> 1.4 MHz) ...and you need at least a bit of "CPU" to "service" the additional PMG data.

 

It is what it is: Without tricks (like the color ram background in SCR or my proposal for color span selection/movement via display list and HSCROL) you'll likely end up in creating an interactive slide show...

 

  • Like 3
Link to comment
Share on other sites

In my experience, based on profiling my software in Altirra, DMA takes about 40% of the time in 200 line modes, about 20% in 100 lines modes. Of course, it all depends. If you only do code in VBI, it will take none. If you only do code when the image is being generated, it will be more.

Going for 100 lines is good idea IMHP, it does something C64 or Spectrum can't do. The CPU time itself wont improve much .. you will go from 60% to 80%. The main speedup comes from the fact you have to process half amount of data. That's the cool stuff.

Also that might or might not be the reason my Rescue demo uses 100 lines mode ;-)

 

But .. it won't help improving Hard drivin. Both C64 and Spectrum versions are in hires, 1bpp.

You would have to convert the code to 2bpp, half res. That would require deep insight into somebody else's 3D engine. Just typing that makes mi chuckle. Good luck with that.

Than you would have to make sure, the physics engine would handle it correctly. And after all this, you might just end up with shitty game. Faster than C64 version, but still shitty game.

  • Like 1
Link to comment
Share on other sites

The star wars games like that play way better on the real thing. centering for the trench is even better. Emu's have difficulty with these titles. like how the death star died when it clealy missed in the middle of that video?.... terrible. on real Atari 8 bit you have to get it for real and it lines up for real...

the game could be improved with todays tools and programmers btw.. faster math pack wouldn't hurt either.

Edited by _The Doctor__
  • Like 3
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...