Do you think there is anything left to "discover" about the 8bits?

[bobotech]

In regards to the stock production 8bit world of Ataris. I have wondered if there are things left to still discover. Such as a new programming technique that allows xyz in games or some new hardware trick that has never been exploited in a game.

 

I just wonder if even with the modern programming tools we have at our disposal, is there anything left to discover that makes the world of 8bit Ataris just all that much better.

[Rybags]

There's the elusive half-pixel shift in Gr. 9 / 11, although it might have been limited to certain faulty GTIAs.

 

There's "weird" behaviour by changing HSCROL register mid-line although it'll probably not be of any use.

 

There's been very little done in 480i "real interlace" vertical resolution doubling.

 

In the realm of "traditional programming" there's the brute force approach to doing softsprites and screen generation.

IIRC it's not been done much, the most notable examples being Bomb Jack and Yoomp.

With 1 Meg cartridges, it opens up the door for it to become more commonly used.

 

I've also developed a strategy for doing scrolling in bitmap modes, it might be useful for something but I've not yet tested it out.

 

There's also the trick of putting PMGs outside the normal 240 scanline area, but it's not entirely useful and wouldn't make much difference in a game.

 

One trick that's not been exploited much is dynamic generation of Display Lists to do vertical scaling effects but probably more a demo/title screen thing that something that'd be inside a game.

 

Another thing not exploited very much is changing colour mid-scanline, in theory a Gr. 9 picture could have tons of colours but there's not really tools around to make it easy.

[pavros]

There is one more thing and maybe it's good time to share it. I've discovered it in June 2009 and made several mircroprograms to check the nature of the effect. Then I've given it up because of lack of time. This year I've started again and currently working on a demo showing the effect on quite advanced picture.

The thing is related to the first entry on Rybags' list above and actually is about delaying GTIA functions of one color clock. The most important is possibility to delay CSYNC signal output. The unwanted side effect of delaying CSYNC is delaying PMG output however due to that we get possibility to check if GTIA is ready to allow control over CSYNC. We can just use collission detection agains playfield. Playfield data output is not delayed ever. The fastest method for switching GTIA to such mode takes in best cases about 40-50 minutes and it depends on temperature of the chip. By switching I mean achieving a state of stability where we have ability to delay CSYNC at any time (any scan line).

Now the most interesting. When delaying CSYNC (one color clock) every second scanline we gain average screen shift of 1/2 color clock. This is because all the video/tv devices try to smooth HSYNC out. Knowing this we can achieve interesting interlace mode by mixing two screens in 4 or 15 ANTIC mode with all PMG underlays and resulting in high resolution. I have some proposals of name of such interlace mode:

DSI - Delayed Sync Interlace

HHI - Horizontal High resolution Interlace

I will upload the demo when it's ready.

Edited November 14, 2011 by pavros

[atarixle]

There are 2^65536 things to discover :-)

[Rybags]

Delayed sync - sounds interesting.

 

Naturally, PM positions would be disturbed since Antic and GTIA would be slightly out of step.

 

The issue I foresee is that some TVs mightn't like it, but we should have plenty of volunteers to test it out.

 

Of course if we can get this kind of thing happening without needing a preheated machine, then it'd be even better.

[Tezz]

I recall hearing about this technique being discussed a while back. I'm very interested to see some demonstrations of it.

[booker]

In regards to the stock production 8bit world of Ataris. I have wondered if there are things left to still discover. Such as a new programming technique that allows xyz in games or some new hardware trick that has never been exploited in a game.

Definitely a lot room to explore still on POKEY. Raster has put very nice synth alike code in his tracker. Also very nice hacking by Analmux. You can easily tell 8 bit Atari music players are still underdeveloped, or lack certain stuff - ie. Soft Synth could have a kind of proper editor, like RMT. And so on.

 

Synth FTW!

Edited November 14, 2011 by booker

[+Stephen]

There is one more thing and maybe it's good time to share it. I've discovered it in June 2009 and made several mircroprograms to check the nature of the effect. Then I've given it up because of lack of time. This year I've started again and currently working on a demo showing the effect on quite advanced picture.

The thing is related to the first entry on Rybags' list above and actually is about delaying GTIA functions of one color clock. The most important is possibility to delay CSYNC signal output. The unwanted side effect of delaying CSYNC is delaying PMG output however due to that we get possibility to check if GTIA is ready to allow control over CSYNC. We can just use collission detection agains playfield. Playfield data output is not delayed ever. The fastest method for switching GTIA to such mode takes in best cases about 40-50 minutes and it depends on temperature of the chip. By switching I mean achieving a state of stability where we have ability to delay CSYNC at any time (any scan line).

Now the most interesting. When delaying CSYNC (one color clock) every second scanline we gain average screen shift of 1/2 color clock. This is because all the video/tv devices try to smooth HSYNC out. Knowing this we can achieve interesting interlace mode by mixing two screens in 4 or 15 ANTIC mode with all PMG underlays and resulting in high resolution. I have some proposals of name of such interlace mode:

DSI - Delayed Sync Interlace

HHI - Horizontal High resolution Interlace

I will upload the demo when it's ready.

Sounds interesting.

[emkay]

In regards to the stock production 8bit world of Ataris. I have wondered if there are things left to still discover. Such as a new programming technique that allows xyz in games or some new hardware trick that has never been exploited in a game.

Definitely a lot room to explore still on POKEY. Raster has put very nice synth alike code in his tracker. Also very nice hacking by Analmux. You can easily tell 8 bit Atari music players are still underdeveloped, or lack certain stuff - ie. Soft Synth could have a kind of proper editor, like RMT. And so on.

 

Synth FTW!

 

Wow. After all that time, someone else read my words?

[emkay]

What's even not fully discovered is the balanced usage of all chips in the A8.

You know, you could do a game like Rescue on Fractalus and play a digi track from disk at the same time.

[emkay]

 

Synth FTW!

 

Btw: Now that some real hardsynth stuff exists, optimized for the real thing, people now can load the XEX file and compare how different the emulations sound.

Possibly someone is able now to adjust the timing of the emulations.

[+Stephen]

In regards to the stock production 8bit world of Ataris. I have wondered if there are things left to still discover. Such as a new programming technique that allows xyz in games or some new hardware trick that has never been exploited in a game.

Definitely a lot room to explore still on POKEY. Raster has put very nice synth alike code in his tracker. Also very nice hacking by Analmux. You can easily tell 8 bit Atari music players are still underdeveloped, or lack certain stuff - ie. Soft Synth could have a kind of proper editor, like RMT. And so on.

 

Synth FTW!

That sounds great! Hard to believe that no samples are being played back!

[emkay]

Is there an MP3 download of all the tunes available?

[booker]

That sounds great! Hard to believe that no samples are being played back!

Thx. And no samples.

 

The main idea there was to write POKEY STEREO tune different. Instead of trying to simulate echoing on both chips and go for a well known sounds I focused on more synthish sounds, and this sometimes mean single stuff played on left POKEY. So, the second half of the effect here is when we apply some low stereo separation (30% in this case).

 

However under real POKEY there are slight differencies. Also many times tune done under RMT detunes much on real machine.

 

My tune is just a simple example. IMHO much more can be done with better player code - in this case I mean mostly synth-alike code. Look on it this way - there's already SID emulation functional for small Atari. Probably not many people found it revolutional, but it is. Using such techinques but for native POKEY music player will cost less rastertime, and with two POKEYs and a single frame player execution you might get very good results and even use it in a demo.

Edited November 15, 2011 by booker

[booker]

Is there an MP3 download of all the tunes available?

So far only

#1 place Edited November 15, 2011 by booker

[Tezz]

Look on it this way - there's already SID emulation functional for small Atari. Probably not many people found it revolutional, but it is. Using such techinques but for native POKEY music player will cost less rastertime, and with two POKEYs and a single frame player execution you might get very good results and even use it in a demo.

I agree, I have some wip done earlier in the year with the technique and intend to go into it fuly and develop further.

[Bryan]

I think there's all kinds of stuff Pokey can do, but since it's based on finding useful patterns in complex behaviors it takes brute force work to discover it all. The best tool would be a gate-level simulator on the PC that would allow for rapid experimentation.

 

Bill Williams was a master at coaxing new sounds out of Pokey. I still want to know how he did some of those effects.

[pavros]

Here it is. The demo of effect based on shifting CSYNC. There are several graphic modes we can gain out of it. The whole family I called DGF - Delayed Gtia Function. The hi-res one is DGI - Delayed Gtia Interlace. I'm planning some simple demo for the others too. In the Knight demo press 1 or 2 for subpictures, 3 for both interlaced - DGI, 4 or 5 for DGX (DGI + real tv interlace).

CRT tv and s-video connection is strongly recommended.

First you need to run DGF INIT and warm up the chips. Don't hesitate to use hair dryer for fastest effect (3 minutes).

I'm attaching source codes for mads.

DGF_demo.zip

DGF_Init_src.zip

Knight_src.zip

[Rybags]

Cool, I've been waiting for this. Running Init now on my 130XE - S-video to 1084s monitor.

 

AC turned off (ambient ~ 24 Deg C), put a tablet cover over the vent to speed up heating.

Edited December 20, 2012 by Rybags

[+Stephen]

I wonder at what temp this kicks in - and what signal / timing exactly changes?

[Rybags]

45 minutes - nothing's changed.

 

What it affects - I've not looked at the demo code yet so can't put in a guess.

 

One thing I did notice - GTIA runs funny if the CPU isn't there. So it could even be something external like relative timings of the system clocks.

[Rybags]

I'm wondering - maybe there's some easy way to insulate GTIA to hold the heat in to speed this process up. 50 minutes, still nothing.

 

ed now at about 90 minutes, still nothing.

 

Looks like I might have to look for a hairdryer when I go out to buy pizza later on. I've got a heat gun but that's massive overkill and I'd probably end up melting something.

Edited December 20, 2012 by Rybags

[Rybags]

Almost 6 hours and not a thing. Although the computer is probably enjoying 18 Deg or so ambient as it's not far from the AC.

 

Got stuff to do, but might drag out the XEGS later and put a heater near it or something.

[+MrFish]

Propane torch?

[pavros]

Surrounding temperature must be at least 25 C deg. In such temparature it takes about 2-4 hours to warm up. In 29 C deg only 30 minutes. BTW It's actually surrounding air you need to warm up. The chips warm up themselves.