demos in preparation ??

[Galahad/FLT]

Originally Posted by Photon

To spell it out, it is because if a competing coder sees a screenshot with a new idea, he gets the new idea too - and knows the level of quality reached by the other. He therefore has an advantage - he knows 'what to beat', so to speak.

I concede that that kind of reasoning was more pertinent in the early days of A500/A1200/AGA-030/AGA060 than today.

What is relevant today is that if you do code one of the few effects that would shock and awe an audience who has seen most of the effects ever made, know the machine capabilities/limitations, and whose eyes can't be fooled by tricks, you'd be very protective of your idea until release.

Also, it's not like WIP screenshots of games, where you'd get some standard routines running and then add content. That's usually not what demos are about. A coder may spend 16 free evenings in a month working on one part of a demo, and what you would see on the screen would be

1. nothing
2. nothing, still writing perfect buffering code
3. perfectly buffered nothing (well testplots)
4. buggy lines
5. fixed lines
6. fixed small optimization that took a whole evening to implement lines
7. fillroutine, test-colors
8. presto! cube with glitches (making new ultimate polyroutine)
9. glitch free cube
10. one glitch left cube
11. light sourced cube
12. cube is now spinning around its axis. yay.
13. cube with added poly to test sorting
14. poly corner moving like a wing, small sorting glitches fixed
15. effect routines are perfect. imported dragon, all kinds of trouble, hanging and graphics trash all over the screen
16. imported flapping awesome dragon with perfectly light-sourced surfaces and glowing eyes

Just an example. The point is the other 15 WIP screenshots would be boring and not at all like game screenshots.

Another difference to WIP game screenshots is of course that you usually don't make games for prestige, and you don't compete with them. There are some demoscene gamecompos, but the prestige is not there (and even more games are made outside these compos). You make games for others to have fun playing, and you want as many as possible to play your game, it doesn't matter if it's optimally coded or if it uses new or better technology, the proof is in the fun to be had.

So exposure is good for games, bad for demos. The reason has a little bit to do with that the Amiga demoscene is (barely) alive today, and it was a long time since there was a real gamedev scene for Amiga. Back in the day, with a few exceptions, gamedevs didn't release screenshots of a game until the month before release.

Remember also that the few exceptions (gamedev diaries to drool over in Amiga magazines) were from people who coded Amiga full-time.

Today, you simply can't say that "I have 4 weeks x 5 full weekdays x 10 hours and an office to go to, to code all day until it's done".

That makes progress sporadic, so that you get in 2 weeks of work during summer and a few evenings in autumn. So even if you make a blog with the progress and screenshots, you must drool very patiently.


The complaint that AGA-060 is for the very few is very real though. A500 is still THE platform, according to me.

Not because of the number of users or because I dislike A1200/AGA-060 or because I'm stuck in the 80s (well, I may be so accused, but it's not relevant) - but because it's the "most" limited and old platform, and only the old, limited platforms can get the added appreciation that the other limited platforms get. It's simply not fantastic that an A1200 can display a converted picture made in Photoshop or Painter. Of course it can! If someone brings 32767 color modes to an A500, or even hand-paints a 64-color picture, of course it's fantastic!

Agree fully with Photon, i find productions more amazing on the A500 than an AGA Amiga, because you largely expect a great looking demo on AGA, for static pictures it can match the PC, and with a 060 and some great ability it can virtually match what the PC scene can do.

People don't expect the A500 to still shock and awe people, which is why it will always be the demo coders first port of call, because their achievements will always appear more 'wow' when its done on a lesser capable machine.

And also, the A500 is older by some years, and for someone to say "I didn't realise the Amiga could do that after nearly 25 years" makes it even better

[StingRay]

Originally Posted by Photon

The complaint that AGA-060 is for the very few is very real though. A500 is still THE platform, according to me.

Not because of the number of users or because I dislike A1200/AGA-060 or because I'm stuck in the 80s (well, I may be so accused, but it's not relevant) - but because it's the "most" limited and old platform, and only the old, limited platforms can get the added appreciation that the other limited platforms get. It's simply not fantastic that an A1200 can display a converted picture made in Photoshop or Painter. Of course it can! If someone brings 32767 color modes to an A500, or even hand-paints a 64-color picture, of course it's fantastic!

I have to somewhat disagree here since you make it sound like coding for/on 68060/AGA is incredibly simple which of course it isn't. It's just different to 68000 coding. You still have to optimise your code if you go for 68060 which isn't quite state of the art anymore either when it comes to speed and this can be VERY hard thanks to the cache and other stuff you don't have to deal with on 68000.

I for one like both, 68000 and 680x0 demos as they all have their very own charme. There should be more A500 demos though, that I agree with!

[britelite]

What StingRay said

[Photon]

Originally Posted by StingRay

I have to somewhat disagree here since you make it sound like coding for/on 68060/AGA is incredibly simple

I don't think that can be taken from it. Now that you mention it, a lot of problems vanish when you code on or for AGA/060, I think you'd agree.

To clarify (for others, since you know all this): I was aiming for the sort of thing when you see (for example) a simple ray-traced animation on an A500 for reasons that just don't apply to AGA/060: work was put into remapping to 31 colors to make it look not like porridge, very nice music is playing at the same time, meaning perhaps that the musician had to optimize the song size skilfully, perhaps it trackloaded quickly, and you couldn't just INCBIN a bunch of files in the binary because there isn't room for the decrunch overlap. Even though the end result is grainy graphics with a few balls reflecting each other or something, it's appreciated because more work is needed to do even simple things. Take the active C64 scene as an example: I don't like those limitations and wouldn't like to code on it, but I do understand the appreciation coming from the limitations.

Originally Posted by StingRay

I for one like both, 68000 and 680x0 demos as they all have their very own charme. There should be more A500 demos though, that I agree with!

Sounds good to my ears

[britelite]

Originally Posted by Photon

I don't think that can be taken from it. Now that you mention it, a lot of problems vanish when you code on or for AGA/060, I think you'd agree.

Some problems vanish, indeed. But you get other problems to deal with instead.

To clarify (for others, since you know all this): I was aiming for the sort of thing when you see (for example) a simple ray-traced animation on an A500 for reasons that just don't apply to AGA/060: work was put into remapping to 31 colors to make it look not like porridge, very nice music is playing at the same time, meaning perhaps that the musician had to optimize the song size skilfully, perhaps it trackloaded quickly, and you couldn't just INCBIN a bunch of files in the binary because there isn't room for the decrunch overlap. Even though the end result is grainy graphics with a few balls reflecting each other or something, it's appreciated because more work is needed to do even simple things.

But an animation is still just an animation

And "more work" doesn't equal good end result. I don't care how much (or little) time someone spent doing something, if what I'm watching looks like crap.

[StingRay]

What Britelite said.

[Amiga1992]

Originally Posted by britelite

And "more work" doesn't equal good end result. I don't care how much (or little) time someone spent doing something, if what I'm watching looks like crap.

Totally agreed

[Amiga1992]

Also: [ Show youtube player ]
Released this weekend

[Photon]

Originally Posted by britelite

Some problems vanish, indeed. But you get other problems to deal with instead.

Much, much, much fewer problems. On AGA/060, you just INCBIN whatever size files, A, WO, and done.

Maybe we're talking about different things. Can you give an example of a difficult problem you've encountered on AGA/060 (and how you solved it, if you want).

Originally Posted by britelite

But an animation is still just an animation

Indeed. My point was that a good looking long or smooth animation is harder to achieve on a limited retro platform. In the case of animation, memory is the limit. On an A500, anything above ~12 frames of fullscreen graphics takes real-time decompression or loader. The only difference between animations and effects is that it's the 7MHz and <0.5 MIPS that is the limit.

Originally Posted by britelite

And "more work" doesn't equal good end result. I don't care how much (or little) time someone spent doing something, if what I'm watching looks like crap.

Of course it doesn't, neither does "less work", as you say. I fully agree

Certainly also, it's possible to get a bright new idea that doesn't require much work and quickly code and release an intro that is appreciated. But it's pretty rare on platforms where the hardware must be used to make the effects and the hardware design therefore dictates what kinds of effects can be conceived. After a while, most of the kinds of effects have been conceived.

The point is that if you have limitations, reaching just slightly above "crap level" in looks takes work, and this gets appreciation by people who know what's possible from the outset.

Again, you know all this, just seemed to me that you thought I meant something else.

[britelite]

Originally Posted by Photon

Much, much, much fewer problems. On AGA/060, you just INCBIN whatever size files, A, WO, and done.

You do know there's more to democoding than filesize/available memory?

Maybe we're talking about different things.

Indeed, I'm talking about coding demos, you seem to be talking about displaying animations

The point is that if you have limitations, reaching just slightly above "crap level" in looks takes work, and this gets appreciation by people who know what's possible from the outset.

If something is just slightly above "crap level" it most certainly gets zero appreciation from me, no matter how much someone spent time on it (or how hard it was to achieve)

[StingRay]

Originally Posted by Photon

Much, much, much fewer problems. On AGA/060, you just INCBIN whatever size files, A, WO, and done.

More memory is something that makes things easier, yes. But the actual code still has to be written and despite having more memory to "waste" it doesn't write itself you know.

Originally Posted by Photon

Maybe we're talking about different things. Can you give an example of a difficult problem you've encountered on AGA/060 (and how you solved it, if you want).

Try a perspective correct texture mapper without using FPU for example, I can assure you that you'll face several headaches!


Edit: Or basically, what Britelite said. Damn it.

[turrican3]

Originally Posted by Galahad/FLT

Agree fully with Photon, i find productions more amazing on the A500 than an AGA Amiga, because you largely expect a great looking demo on AGA, for static pictures it can match the PC, and with a 060 and some great ability it can virtually match what the PC scene can do.

People don't expect the A500 to still shock and awe people, which is why it will always be the demo coders first port of call, because their achievements will always appear more 'wow' when its done on a lesser capable machine.

And also, the A500 is older by some years, and for someone to say "I didn't realise the Amiga could do that after nearly 25 years" makes it even better

fairlight for which of the last demos for amiga 500 are the best ??? the ones which makes you wow ??

[Photon]

Originally Posted by britelite

You do know there's more to democoding than filesize/available memory?

You do know being able just to add stuff to your binary without even thinking about the filesize and never running out of chipmem is an incredible advantage?

Originally Posted by britelite

Indeed, I'm talking about coding demos, you seem to be talking about displaying animations

No, I told you that it was just an example. It's you who are fixating on it despite my last post.

Originally Posted by britelite

If something is just slightly above "crap level" it most certainly gets zero appreciation from me, no matter how much someone spent time on it (or how hard it was to achieve)

I stated the problem poorly. I meant to say that if you accept the limitations, you get average demo content: "1 module, 1 picture, a logo and some sprites is all that fits in chipmem after necessary buffers" - and that as soon as you want to be ambitious in some area, f.ex. high-quality samples in the module making it 300KB, you run into the limits and have to change your planning or scratch ideas.

Originally Posted by StingRay

Try a perspective correct texture mapper without using FPU for example, I can assure you that you'll face several headaches!

Um... And AGA/060 has an FPU. That makes that platform easier to code those on. That's why even simpler attempts at texture-mapping get more appreciation on limited platforms. Thanks for agreeing with me...?


This is straying off-topic, and I won't continue. I was just calling what I think is one major reason for why so A500 1MB is getting so little love from coders. You seem to want to see more demos for the A500 as well so why not suggest alternative reasons?

[StingRay]

Originally Posted by Photon

You do know being able just to add stuff to your binary without even thinking about the filesize and never running out of chipmem is an incredible advantage?

You still sound like 680x0 coding is not a challenge and incredibly easy. You are wrong!

Originally Posted by Photon

Um... And AGA/060 has an FPU. That makes that platform easier to code those on. That's why even simpler attempts at texture-mapping get more appreciation on limited platforms. Thanks for agreeing with me...?

How many perspective mappers have you coded so far? Besides, as far as I'm concerned, there are other CPU's than 68060 that are used in AGA machines as well. And coding a fast perspective mapper even with FPU isn't done within the blink of an eye either. But that is all indeed off-topic here. Try some 680x0 coding yourself and see how "easy" everything is just because you have more memory and a faster CPU. You seem to completely forget that people expect better/more complex effects on AGA machines.

[britelite]

Originally Posted by Photon

You do know being able just to add stuff to your binary without even thinking about the filesize and never running out of chipmem is an incredible advantage?

Ahem, running out of CHIPmem is incredibly easy on AGA-machines. And just because someone has AGA and 060, you can't assume they have tons of memory either. And it's not like A500 is limited to 512k of memory either

I stated the problem poorly. I meant to say that if you accept the limitations, you get average demo content: "1 module, 1 picture, a logo and some sprites is all that fits in chipmem after necessary buffers" - and that as soon as you want to be ambitious in some area, f.ex. high-quality samples in the module making it 300KB, you run into the limits and have to change your planning or scratch ideas.

Yes, and I said coding for AGA/060 you run into DIFFERENT problems, which you seem to have a hard time accepting. I guess the problem is that you've never actually tried making any 060-demos, while both me and StingRay have done both A500 and AGA/060.

Also keep in mind that while the target platform gets more powerful, the expectations on what's getting displayed also moves way up.

Um... And AGA/060 has an FPU. That makes that platform easier to code those on. That's why even simpler attempts at texture-mapping get more appreciation on limited platforms. Thanks for agreeing with me...?

There are actually 060-accelerators without an FPU, so if you want to be sure to be really compatible, you can't use the FPU. And the FPU isn't really that fast anyway, so for speed it's better to only use integers.

I was just calling what I think is one major reason for why so A500 1MB is getting so little love from coders. You seem to want to see more demos for the A500 as well so why not suggest alternative reasons?

The Amiga in general is getting very little love, the problem being that it's such a varied platform. Some people prefer the A500, some people the 060, while most AGA-users only have an 030-board. And making A500-demos with ugly tricks aren't necessarily compatible with accelerated machines (or badly configured emulators).

[britelite]

And what StingRay said

[wXR]

What you all said. What an awesome and incredibly interesting thread this has become!

[jimbob]

Agreed wXR, except you forgot hilarious.

Originally Posted by Hewitson

I think the best way you guys can settle your minor disgreement is with a demo comp. May I suggest you call it "FAPCAM 2012 Code Off."

[UberFreak]

There is another problem with AGA/060 demos that's rarely mentioned.
As a demo fanatic, I invested in an 060 for my Amiga, so I can watch all the new demos properly (I really dislike emulators).

What I found out is, that most demos in the last 5 years or so, need more than a 50Mhz 060 to run acceptably smooth.
Think about it - how many 060 owners are out there?
How many of them have overclocked 060?
The answer is probably very little, and personally I wont risk overclocking ancient (and expensive) hardware that can die any minute.

So in the end, I still have to watch many demos (hello Elude) on an emulator, as they look more like slide-shows on a real 060 in most cases

Is there a real reason for this?
I'd really like to hear opinions of actual demo coders

[britelite]

Originally Posted by UberFreak

What I found out is, that most demos in the last 5 years or so, need more than a 50Mhz 060 to run acceptably smooth.
Think about it - how many 060 owners are out there?
How many of them have overclocked 060?
The answer is probably very little, and personally I wont risk overclocking ancient (and expensive) hardware that can die any minute.

So in the end, I still have to watch many demos (hello Elude) on an emulator, as they look more like slide-shows on a real 060 in most cases

Is there a real reason for this?
I'd really like to hear opinions of actual demo coders

Well, there are many 060-demos that are going to have bad framerate no matter what speed the cpu runs at

At least for me the problem is that I have an overclocked 060 (66MHz), so that's the only 060-board I can test my demos on. So even if my stuff runs okayish on my board, they will run worse on 50MHz. If the prices of these boards weren't so ridiculous I would buy another one and keep it at 50Mhz

Also, sometimes it feels really stupid trying to optimize your routines to the max knowing that 99% of all the viewers will watch the demo in an emulator or a video captured from an emulator.