Click Here
home features news forums classifieds faqs links search
6083 members 
Amiga Q&A /  Free for All /  Emulation /  Gaming / (Latest Posts)
Login

Nickname

Password

Lost Password?

Don't have an account yet?
Register now!

Support Amigaworld.net
Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
Donate

Menu
Main sections
Home
Features
News
Forums
Classifieds
Links
Downloads
Extras
OS4 Zone
IRC Network
AmigaWorld Radio
Newsfeed
Top Members
Amiga Dealers
Information
About Us
FAQs
Advertise
Polls
Terms of Service
Search

IRC Channel
Server: irc.amigaworld.net
Ports: 1024,5555, 6665-6669
SSL port: 6697
Channel: #Amigaworld
Channel Policy and Guidelines

Who's Online
47 crawler(s) on-line.
 30 guest(s) on-line.
 0 member(s) on-line.



You are an anonymous user.
Register Now!
 bhabbott:  19 mins ago
 Rob:  44 mins ago
 DiscreetFX:  46 mins ago
 Seiya:  1 hr 10 mins ago
 Amiga4000:  1 hr 23 mins ago
 Hypex:  1 hr 28 mins ago
 agami:  1 hr 56 mins ago
 Davec555:  2 hrs 9 mins ago
 Steady:  2 hrs 47 mins ago
 Karlos:  3 hrs 41 mins ago

/  Forum Index
   /  Amiga Development
      /  Packed Versus Planar: FIGHT
Register To Post

Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 Next Page )
PosterThread
Hypex 
Re: Packed Versus Planar: FIGHT
Posted on 2-Jul-2022 16:42:34
#41 ]
Elite Member
Joined: 6-May-2007
Posts: 10832
From: Greensborough, Australia

So, an experiment I will be performing is testing how fast the Amiga could have been if had an actual packed mode. That is a packed planar mode. And how fast it would be in chip ram.

As we know the Amiga was officially rejected for Doom having no native chunky mode. But I'm going to test how fast it would be if it had had a chunky mode. By using the Doom engine to see how fast it can write into chip ram.

So with some modifcations it will skip the chunky to planar conversion and write it directly to the bitplanes. And also test rendering direct to chip ram. The result will be a confuisng mess but the point will be to test the speed! I don't recall where but someone posted that writing random data to bitplanes can damage the Amiga display controller. So it might be dangerious as well!

 Status: Offline
Profile     Report this post  
Karlos 
Re: Packed Versus Planar: FIGHT
Posted on 2-Jul-2022 16:56:51
#42 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3147
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Hypex

You can just measure the write speed of chip ram at different word sizes and work it out from that. It's been a while but if memory serves, 7MB/s for 32-bit writes from CPU to chip ram on AGA.

Back of envelope, an infinitely fast CPU doing nothing but writing 320x200 (64000) pixels using optimal, uncontended 32-bit writes at 7MB/s could manage around 109fps.

However, DOOM wrote pixel columns for walls. Without combining these into wider writes, you are looking at a quarter of that, I guess.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Karlos 
Re: Packed Versus Planar: FIGHT
Posted on 2-Jul-2022 17:14:02
#43 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3147
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Hypex

Quote:
Using vector operations should speed up any planar to framebuffer conversions.


As long as you keep the loops simple, the compiler does a reasonable job of doing that for you on x64. The problem I have is that my "display controller" has an optional beam-racer built in that works somewhat like the copper. That's not so trivially vectorisable.

Last edited by Karlos on 02-Jul-2022 at 05:18 PM.
Last edited by Karlos on 02-Jul-2022 at 05:15 PM.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Hypex 
Re: Packed Versus Planar: FIGHT
Posted on 4-Jul-2022 13:01:54
#44 ]
Elite Member
Joined: 6-May-2007
Posts: 10832
From: Greensborough, Australia

@Karlos

Quote:
You can just measure the write speed of chip ram at different word sizes and work it out from that. It's been a while but if memory serves, 7MB/s for 32-bit writes from CPU to chip ram on AGA.


That can be done yes. But I'm interested in a real-time scenario, with a live 3d renderer, as well as a copy of screen map to bitmap data. And also speed of rendering 3d images direct to bitmap.

Quote:
Back of envelope, an infinitely fast CPU doing nothing but writing 320x200 (64000) pixels using optimal, uncontended 32-bit writes at 7MB/s could manage around 109fps.


Sounds great!

Quote:
However, DOOM wrote pixel columns for walls. Without combining these into wider writes, you are looking at a quarter of that, I guess.


Yes I've read that. VGA would have helped as the underlying planar organisation would have allowed multiple columns to be written. Though it could have unchained the planes but I can't find any clean info on what mode Doom used.

In any case, I've taken a look at the ADoom source, where a month was initially spent on porting it to the Amiga. And it looks like the wall routines were replaced by optimised 68K ASM. Aside from that multiple pixel column writes can be moved on the 68K but of course alignment needs to be taken into account.

Quote:
As long as you keep the loops simple, the compiler does a reasonable job of doing that for you on x64. The problem I have is that my "display controller" has an optional beam-racer built in that works somewhat like the copper. That's not so trivially vectorisable.


I've read discussions on being able to simulate that. Popular ideas are using 3d shaders. I suppose rendering a whole llne would be acceptable. But rendering a whole frame would be desirable.

 Status: Offline
Profile     Report this post  
Karlos 
Re: Packed Versus Planar: FIGHT
Posted on 5-Jul-2022 0:19:22
#45 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3147
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Hypex

Shaders are the answer for everything is a common starting gambit... Except when you come to actually use them for something like this. Generally, conditional branching code is bad for shaders. My beam racer allows any number of modifications to the controller (register and palette manipulations) at arbitrary beam positions. I called the thing FILTH (Finite Interpreted Line Transfer Handler) as it's dirty and even allows self-modification.

Reimplementing what it does in 3D, using shaders is tricky, even some of the simplest aspects. For example, if you were dealing with, for example an 8 bit display mode with a 1D texture sampler for the palette, then you get extra complications when trying to do stuff like update the palette when a particular coordinate is hit. There are workarounds but for the sort of thing I'm doing right now, a total software implementation (with GL just for scaling of low resolution modes) is ideal.

When the time comes that I want to do any sort of advanced hardware rendering, I'll just provide a completely new ABI for it to call from the VM side.

Last edited by Karlos on 05-Jul-2022 at 12:20 AM.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Hypex 
Re: Packed Versus Planar: FIGHT
Posted on 6-Jul-2022 14:24:08
#46 ]
Elite Member
Joined: 6-May-2007
Posts: 10832
From: Greensborough, Australia

@Karlos

FILTH. That's funny. But self modification sounds useful.

Hardware is good. Like functions shaders can provide. But if it makes it too hard to manage then software wise would be a better solution.

As an aside, I was looking up VGA modes and found this old post from John Carmack. He doesn't mention the Amiga specifically but he thinks bitplanes stink.

https://groups.google.com/g/alt.games.doom/c/3tMB2UmEBK0/m/m1VR6LiJRQMJ

 Status: Offline
Profile     Report this post  
bhabbott 
Re: Packed Versus Planar: FIGHT
Posted on 8-Jul-2022 3:02:22
#47 ]
Regular Member
Joined: 6-Jun-2018
Posts: 230
From: Aotearoa

@Hypex

Quote:

Hypex wrote:
So, an experiment I will be performing is testing how fast the Amiga could have been if had an actual packed mode. That is a packed planar mode. And how fast it would be in chip ram.

I look forward to seeing that.

Quote:
As we know the Amiga was officially rejected for Doom having no native chunky mode.

Yes, but even if the Amiga did have chunky mode it wouldn't have helped. The only stock machine with the required specs (256 colors, 8 MB RAM, hard drive, fast CPU) was the A4000, which sold in such small numbers that an Amiga version of Doom wasn't worth developing. There were also a number of graphics cards sold for the A2000 etc. which had chunky mode, but again the potential market was too small.

OTOH Doom was ported to the SNES, even though it didn't have a suitable graphics mode (most modes use bitplane graphics, and while mode 7 does have 256 color 'chunky' pixels there are only enough tiles for a 128x128 pixel bitmap). The solution on the SNES was a special cartridge with it's own graphics processor and RAM, and custom rendering code.

So the Amiga 500 could have had Doom if someone was willing to make the effort, and anyone was willing to pay for it. But of course they weren't. If a stock A500 or A1200 couldn't do it (which would be the case even if they did have chunky mode), we weren't going to get it.

Quote:
But I'm going to test how fast it would be if it had had a chunky mode. By using the Doom engine to see how fast it can write into chip ram.

So with some modifcations it will skip the chunky to planar conversion and write it directly to the bitplanes. And also test rendering direct to chip ram. The result will be a confuisng mess but the point will be to test the speed! I don't recall where but someone posted that writing random data to bitplanes can damage the Amiga display controller. So it might be dangerious as well!

This is exactly what I was thinking of doing. It will be interesting to compare rendering chunky directly to chipram vs rendering to fastram and then doing C2P to chipram.

For comparison, here are some results of running Quake on a PC with Athlon XP 2200+ CPU and various graphics cards:-

Quake 1 software rendering 320x240

Trident TVGA8800C (8 bit ISA bus) 8.7 fps
Trident TVGA9000i (16 bit ISA bus) 13.6 fps
ATI Rage 128 Pro (AGP4x) 205.6 fps

The frame rate is obviously severely constrained by graphics card bus bandwidth.

 Status: Offline
Profile     Report this post  
Hypex 
Re: Packed Versus Planar: FIGHT
Posted on 8-Jul-2022 12:13:34
#48 ]
Elite Member
Joined: 6-May-2007
Posts: 10832
From: Greensborough, Australia

@bhabbott

Quote:
I look forward to seeing that.


Thanks. Just something for a little fun. Hopefully post some test results this weekend.

Quote:
Yes, but even if the Amiga did have chunky mode it wouldn't have helped. The only stock machine with the required specs (256 colors, 8 MB RAM, hard drive, fast CPU) was the A4000, which sold in such small numbers that an Amiga version of Doom wasn't worth developing. There were also a number of graphics cards sold for the A2000 etc. which had chunky mode, but again the potential market was too small.


I think at this point it would have needed to target the A1200 even if the A500 had a chunky mode. The new A1200 set a new baseline. When you look at future releases like AB3D and slogans about blood flying faster in a blizzard it reveals a new gaming era that needed more power. And the way things were moving demanded CPU power.

Quote:
OTOH Doom was ported to the SNES, even though it didn't have a suitable graphics mode (most modes use bitplane graphics, and while mode 7 does have 256 color 'chunky' pixels there are only enough tiles for a 128x128 pixel bitmap). The solution on the SNES was a special cartridge with it's own graphics processor and RAM, and custom rendering code.


That showed how unsuitable it was for the hardware but at least it could be embedded in a cartridge.

Quote:
So the Amiga 500 could have had Doom if someone was willing to make the effort, and anyone was willing to pay for it. But of course they weren't. If a stock A500 or A1200 couldn't do it (which would be the case even if they did have chunky mode), we weren't going to get it.


I don't think an A500 would have been suitable. It may have came out the same time as VGA but VGA was superior in comparable aspects. Also, the average PC in 1993 was more powerful than an A500. More expensive as well. So the most affordable option for all practical purposes would have been an A1200 with 030 and fast ram at least I think.

Quote:
This is exactly what I was thinking of doing. It will be interesting to compare rendering chunky directly to chipram vs rendering to fastram and then doing C2P to chipram.


I plan to do all that as well as compare to RTG alike with both direct rendering and writing a chunky buffer to RTG.

Quote:
For comparison, here are some results of running Quake on a PC with Athlon XP 2200+ CPU and various graphics cards:-


Ouch.

Those lower cards are real slow. It must have taken ages testing all those cards. Given less powerful PC hardware had fast 2d action games would suggest the CPU is being taxed as well.

 Status: Offline
Profile     Report this post  
Karlos 
Re: Packed Versus Planar: FIGHT
Posted on 8-Jul-2022 16:05:22
#49 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3147
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Hypex

Quote:
I don't recall where but someone posted that writing random data to bitplanes can damage the Amiga display controller.


That sounds like nonsense. Writing random data to the control registers could have undesirable wide effects but writing random stuff to o bitplanes should just result in random visible output.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
bhabbott 
Re: Packed Versus Planar: FIGHT
Posted on 9-Jul-2022 3:11:34
#50 ]
Regular Member
Joined: 6-Jun-2018
Posts: 230
From: Aotearoa

@Hypex

Quote:

Hypex wrote:

I don't think an A500 would have been suitable. It may have came out the same time as VGA but VGA was superior in comparable aspects. Also, the average PC in 1993 was more powerful than an A500. More expensive as well. So the most affordable option for all practical purposes would have been an A1200 with 030 and fast ram at least I think.

A stock A500 would not be suitable, but an expanded one might. GVP introduced the A530 in 1992, which had a 40MHz 030, hard drive and up to 8MB RAM. Graphics would have to be cut down to 64 or 32 colors so it wouldn't look so good, but c2p could be faster. But the vast majority of A500 owners weren't interested in spending any money on upgrading their machines just to play games.

An A1200 with 8MB FastRAM and a hard drive can run Doom. Not very well, but at least playable in a smaller windows size - similar to a 386SX-16. My A1200 with 50MHz 030 runs it quite well. An A4000-40 is more than capable. Even though the recommended setup for Doom was a 486, many 386SX owners also tried it out on their machines - because they could.

Carmack's decision to not support the Amiga meant that even those of us who did have powerful enough Amigas couldn't run Doom on our machines. He dismissed the Amiga as 'stink' because its bitplane graphics weren't optimal for his game. But being 'sub-optimal' hadn't stopped games being ported to different platforms in the past. The real issue was a changing attitude towards ports. Instead of thinking they had to reach as wide a market as possible, developers like Carmack figured - correctly - that they could just concentrate on high-end PCs.

That attitude of only supporting the best platform continues today except in one field - retro computing. Rather than forgetting about creating 3D games on the Amiga because it doesn't have chunky pixels, we enjoy the challenge of doing it even when the results are 'sub-optimal'. We are like artists working with a restrictive medium - it makes us more creative, not less.



 Status: Offline
Profile     Report this post  
Karlos 
Re: Packed Versus Planar: FIGHT
Posted on 9-Jul-2022 6:50:39
#51 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3147
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@bhabbott

Did anyone try to run the Apple version of Doom under emulation back in the day?

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Hypex 
Re: Packed Versus Planar: FIGHT
Posted on 26-Jul-2022 8:47:14
#52 ]
Elite Member
Joined: 6-May-2007
Posts: 10832
From: Greensborough, Australia

@Karlos

Quote:
That sounds like nonsense. Writing random data to the control registers could have undesirable wide effects but writing random stuff to o bitplanes should just result in random visible output.


That's what I thought. I don't know if it was understood what I was saying. But yes, I was just intending on writing random data or in reality confused data by disorder into the bitplanes. I'm not aware of the display controller actually caring that it makes sense. It just deals with bits, data and logic and converts it into electrical impulses representing colour.

And, it's happened before. OCS games on ECS could scramble the screen. Just as it happened for ECS on AGA. And today I still see it, when there is some mismatch in a driver, that doesn't render an image correctly the lines are all skewed.

 Status: Offline
Profile     Report this post  
Hypex 
Re: Packed Versus Planar: FIGHT
Posted on 26-Jul-2022 9:45:33
#53 ]
Elite Member
Joined: 6-May-2007
Posts: 10832
From: Greensborough, Australia

@bhabbott

Quote:
A stock A500 would not be suitable, but an expanded one might. GVP introduced the A530 in 1992, which had a 40MHz 030, hard drive and up to 8MB RAM. Graphics would have to be cut down to 64 or 32 colors so it wouldn't look so good, but c2p could be faster. But the vast majority of A500 owners weren't interested in spending any money on upgrading their machines just to play games.


That true. Indeed, because I was an "advanced user" I had an A590 with 20MB HDD. But they cost money and so out of reach of most people unless they really used their A500 for practical purposes as well. I survived on floppy for ages before the A590 and most productivity was not written for it or assumed the user had an external floppy which was more common. Swapping between 3 disks 20 times just to load KindWords was not a user friendly experience!

In any case it certainly would have needed a HDD on an A500. The original relied on it so no way could it easily work without it. ADoom has ECS modes using EHB for 64 colours. On lower CPUs it also uses assisted blitter for c2p conversion. A lot of work was put into that. But, I think if work was put into using the Amiga features where possible, it could have worked better. For example, surfaces like ceiling and floors have gradients. So, they would have a section in the palette for this. I haven't looked into it but it's not unreasonable I think to use a base palette of 16 colours and then spread the palette into 16 shades of light to dark or even 8 shades of 32 colours. Upper nibble of pixel could have gradient and lower colour or vice versa.

My point here is that the Amiga was suited to gradients without even thinking. Floor and ceilings would have reserved colours in palette. A copper list then changes palette every 8 lines or so there is a change in gradient. This reduces the amount of data needed to be written into biplanes just like traditional games while keeping higher colour content. As much as it could be reduced while still needing other objects in foreground. The panels need not be chunky and could be a direct bitmap to be optimal. As well as using sprites for the big gun. Just the main view would need to be rendered as chunky. Possibly dual play could could be used so background and foreground can be rendered independently, which would be more suited to a 2d background scrolling sky, but it would reduce colours and need more management.

Quote:
An A1200 with 8MB FastRAM and a hard drive can run Doom. Not very well, but at least playable in a smaller windows size - similar to a 386SX-16. My A1200 with 50MHz 030 runs it quite well. An A4000-40 is more than capable. Even though the recommended setup for Doom was a 486, many 386SX owners also tried it out on their machines - because they could.


And I note here the lowest is 386. 286 didn't cut it. 286 is like A500. In fact 286 could be better at 14Mhz. IIRC even a 286 couldn't handle Wolfenstein or they at least decided it needed VGA hardware. What's interesting is I found Doom has an actual low res mode of 160 pixels in 320 width screen. I didn't know this and thought only the Amiga port has this lower res speed up trick. I'm sure DoomAttack had this trick and did it to the whole screen. That was my preferred port for my A1200/030. But in Doom itself, the lower res mode only renders the game view at half resolution, while any borders and bottom panel are still at full resolution.

Quote:
Carmack's decision to not support the Amiga meant that even those of us who did have powerful enough Amigas couldn't run Doom on our machines. He dismissed the Amiga as 'stink' because its bitplane graphics weren't optimal for his game. But being 'sub-optimal' hadn't stopped games being ported to different platforms in the past. The real issue was a changing attitude towards ports. Instead of thinking they had to reach as wide a market as possible, developers like Carmack figured - correctly - that they could just concentrate on high-end PCs.


Even though it was common earlier for a host of platforms to have their own conversion I can understand him not wanting it. For one thing the Amiga was out of favour and was last decades pin up girl, so to speak, lol. Also, Doom was not only written for, but designed to use VGA hardware features. It's probably worse than taking a C16 game, with poor sound and no sprites, and porting it to a C64 with better sound, actual sprites but less colours. In a manner of speaking, it would have been like taking a high speed AGA multi parallax scrolling game with copper overload, and trying to port that to VGA. It can be done, but the result suffers, since VGA can't compete with a live 24 bit copper list. Being VGA was 18 bit colour with 256 colours max. But, the PC did have other games the Amiga had, like Oscar.

So, at this stage of the game, porting Doom to the Amiga wasn't as simple as any other port. Earlier ports would have used a baseline. When the PC was still EGA porting was easy to Amiga. Convert graphics and go. In fact, EGA had bitplanes as well, so that would make it trivial if the format was compatible. For a proper port to the Amiga platform, they would have needed to code it to make use of the platform strengths. Making use of copper, sprites and other tricks. Carmack would have needed to off load it to an Amiga coder who would put in all the work to do that. And by then there was no money in the Amiga game scene. Plus using Doom Carmack killed it.

Quote:
That attitude of only supporting the best platform continues today except in one field - retro computing. Rather than forgetting about creating 3D games on the Amiga because it doesn't have chunky pixels, we enjoy the challenge of doing it even when the results are 'sub-optimal'. We are like artists working with a restrictive medium - it makes us more creative, not less.


I agree. I don't quite recall what I was thinking years ago, but I made up this saying, which reflects that: "Through limitations comes innovations."

Perhaps it''s the challenge of it all. Human nature. When told it can't be done we want to set out to prove otherwise. Especially with an Amiga Doom. Don't quite know why it was open sourced so soon if at all, but they indirectly gave the Amiga a hallpass in doing so.

My dream is, or was in my 20's, to create an Amiga 3d game without the chunky pixels. That is native 3d using hardware features as much as possible. And only resort to chunky if needed. But not the traditional polygon games. Something way better.

I have a friend who doesn't like the modern demos. Whenever he sees rotating 3d, especially looking chunky, he thinks it's just trying to be a PC. Pity that, when the Amiga does some 3d, it can look like it's just copying a PC. And not be an Amiga effect.

Last edited by Hypex on 26-Jul-2022 at 09:53 AM.

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: Packed Versus Planar: FIGHT
Posted on 26-Jul-2022 11:20:28
#54 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12344
From: Norway

@Hypex

Doom was actually not a commercial game, it was a promotion game, to promote id software and give the company recognition.

The source code to the game was released as late as 1997. Of Couse we already played game to boredom on 486 by that time, Quake was already out.

Last edited by NutsAboutAmiga on 26-Jul-2022 at 11:22 AM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
Karlos 
Re: Packed Versus Planar: FIGHT
Posted on 27-Jul-2022 14:06:14
#55 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3147
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Hypex

Quote:
My dream is, or was in my 20's, to create an Amiga 3d game without the chunky pixels. That is native 3d using hardware features as much as possible. And only resort to chunky if needed.


What do/did you have in mind?

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Hypex 
Re: Packed Versus Planar: FIGHT
Posted on 30-Jul-2022 15:28:06
#56 ]
Elite Member
Joined: 6-May-2007
Posts: 10832
From: Greensborough, Australia

@NutsAboutAmiga

Yes it had some kind of shareware model. The most popular variant must be the free level one I imagine. I've still not even completed that.

I tested it on some PC's just to see what the fuss was about. But only had Amiga at home which didn't have it. I wasn't impressed by the blockiness when going up against the wall, didn't know what all the crap was about when it looked crap to me, then went back to my Amiga!

 Status: Offline
Profile     Report this post  
Hypex 
Re: Packed Versus Planar: FIGHT
Posted on 30-Jul-2022 16:09:49
#57 ]
Elite Member
Joined: 6-May-2007
Posts: 10832
From: Greensborough, Australia

@Karlos

Quote:
What do/did you have in mind?


What I wanted to do was target AGA with a car game and also had ideas for fighting game using same engine to display 3d effects.

So what I was thinking was using a dual playfield for having split screen parallax scrolling. Though it does reduce colours. In the top background would be sky and a blue gradient copper. The clouds would be 3d rendered and as they scroll along a different frame would be displayed for a scrolling 3d effect. Needs more memory but would add depth with a 3d look.

In front of clouds would be mountains using same technique to produce a moving 3d mountain range. With minimum frames used to generate a realistic effect. They would be produced from a pan of left to right.

On the back bottom would be the road. Which would use more complicated scrolling. This would be an anim of a road moving forward from a 3d render. The bends would use the copper to shift the lines across. But bumps in the road would use copper to skip lines so it waves up and down. I actually experimented with this and used DPaint to produce a 3d road render. Just check pattern IIRC. I then wrote my own RLE codec to unpack the frames on the fly and two commands to pack and unpack frames. FlattenFile and FatttenFile. It was optimised for 010/020 loop mode. I embedded it all into a demo than ran the anim while using copper to bend the road. I called it RoadSlides which I thought would also make a cool name for a car game.

On front bottom would be trees on the side and cars in main view. The trees I thought I would scale in realtime if possible but also using the pre panned effect. The fake flat 3d look of later games annoyed me. I wanted to do better. Vertical scaling is easy but horizontally is more work. Still trees like pines with triangular shapes can be optimised with shorter repeating patterns.

On top bottom. The cars I wasn't sure. They needed spinning tyres. But also needed rotation as well as panning for effect. I considered how to do this using less time as possible. Since cars were the most complex to deal with. Possibly might have needed chunky rendering for ease then convert to planar in bitmap.

This was all on a bog standard A1200 with no fast but I did have a Seagate 40MB IDE. Actually I might have had my Quantum 500MB IDE by then. I later upgraded to a DKB1240 with a 40Mhz 030. Yes DKB models were confusing. Unfortunately the "better" CPU messed up my 020 optimised code. I may have messed up vertical blanks or cache. But it just stalled the anim routine which went choppy. I never recovered from that and the project stalled.

 Status: Offline
Profile     Report this post  
Karlos 
Re: Packed Versus Planar: FIGHT
Posted on 30-Jul-2022 19:37:59
#58 ]
Elite Member
Joined: 24-Aug-2003
Posts: 3147
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Hypex

Hmm. I'm not sure about "3D" mountain ranges per se, I mean if you are racing you are usually facing forwards and move sideways a bit, the distance means there's little parallax. It's in part due to this why a lot of games still use 2D billboarding for distant models. Maybe something like a skybox could work well with the basic sky colour being a copper gradient and the distant horizon being the textured part.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Hammer 
Re: Packed Versus Planar: FIGHT
Posted on 31-Jul-2022 0:28:06
#59 ]
Elite Member
Joined: 9-Mar-2003
Posts: 4600
From: Australia

@matthey

Chip's transistor budget was also a major factor, hence CSG's chip fabrication capabilities have to be factored in.

16 million color palette is not free and it consumes a higher transistor count.

CBM created two 256-color (8-bit) capable chipsets i.e. C65 (with 4096 color palette) and AGA (with 16 million color palette). AGA needs 3rd party chip fabrication support.

Ranger, AAA, and Hombre were uncompleted.

_________________
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 3080 Ti
Ryzen 9 7900X, DDR5-5600 32 GB RAM, GeForce RTX 3080 Ti
Amiga 1200 (rev 1D1, KS 3.2, TF1260, 68060 @ 63 Mhz, 128 MB)
Amiga 500 (rev 6A, KS 3.2, PiStorm/RPi3a/Emu68)

 Status: Offline
Profile     Report this post  
MEGA_RJ_MICAL 
Re: Packed Versus Planar: FIGHT
Posted on 31-Jul-2022 1:28:14
#60 ]
Super Member
Joined: 13-Dec-2019
Posts: 1200
From: AMIGAWORLD.NET WAS ORIGINALLY FOUNDED BY DAVID DOYLE

PADDING, OF COURSE

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK

 Status: Offline
Profile     Report this post  
Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 Next Page )

[ home ][ about us ][ privacy ] [ forums ][ classifieds ] [ links ][ news archive ] [ link to us ][ user account ]
Copyright (C) 2000 - 2019 Amigaworld.net.
Amigaworld.net was originally founded by David Doyle