3D in demos - F030 vs Amiga 1200

[pandy71]

Originally Posted by Mrs Beanbag

But Amiga has 4 of them, so combined you can get more than 8 bit. The easy trick is simply to play the least significant 6 bits on another channel at 1/64 volume, giving you 14 bit sound. I don't know how you would get 16 or 18 bit sound but maybe there are other ways.

No, this is not channel folding - this is noiseshaping - quantization noise s pushed to range out of human hearing threshold (or rather there when is less audible - when more than 40ksps is used energy can be shifted out of audible range - for ECS/AGA this related to current H frequency, on OCS perhaps line can be shorten by Copper/CPU or software driven - CPU/Copper mode) - higher sampling make this easier and more efficient but even with normal OCS maximum sampling rate for stereo 8 DMA sound (29073 samples per second) useable dynamics should be around 75 - 85dB which is roughly translated to 13 - 14 bits - even on PC such 8 bit audio sound reasonably nice.

[kovacm]

I do wonder, to be honest, if you only come here to say "Atari ST is better than Amiga, ner ner".

I ask simple question and get answer from few users (thanx!) but pandy71 is the one who insist that "you can not compare F030 and A1200", start talking about "tricks how to open border", "thing you need to emulate from Amiga on ST"... and post some pictures in post #14

from my point of view, he start defending (?) Amiga from every aspect, just to prove obvious thing (one that all we already know): that Amiga hardware is superrior to ST.
and thats brings us here.

my bad is that I continue to answer to him but since he obviously have knowledge, why not continue even offtopic.


anyway, I should disclosure my agenda.

(english is not my frist, not even second language so please have some apprehension )

my final goal is to made website where you could compare Amiga and Atari (and some Macs) computers with PC through achievements in time. it will include birth of today mainstream software (CuBase, Lightwave, 3D Max, Logic, Office...), and live demonstration what you could do with Amigas and Atari back in 80s... and what with PC.

to comapre them and to show how PC manage to slow down computer industries!

Douglas Adams describe this problem in one sentence:
"The idea that Bill Gates has appeared like a knight in shining armour to lead all customers out of a mire of technological chaos neatly ignores the fact that it was he, by peddling second rate technology, led them into it in the first place, and continues to do so today."



back on (off)topic:
Mrs Beanbag thanx for answering, did not know that Amiga has 4 DACs!

Plain ST support mass storage with 3.5MB/s transfer ?

yes, very interesting finding.
Atari ST have external ASCI port for almost 2MB/s. Why ASCI? In 1985. SCSI was not finished (specification was not finalize). Atari anyway built in ASCI which is very similar to final SCSI but it is not 100% compatible. Atari produced CD-ROMs, harddisks and laser printer for it (Atari SLM laser printer was the first "windows" printer - printer that has no CPU or memory).

3.5MB/s on ST?
Mr. Petari made adapter for CF Card that is connected throug ROM port (cartridge). ROM port has all MC68000 adres and data signals but it lack two signals that are important for hispeed writings. you can read more on:
http://forum.8bitchip.info/hardware-...g1064/#msg1064
http://forum.8bitchip.info/software-...ayback-on-ste/
http://forum.8bitchip.info/software-...g1007/#msg1007


btw pandy71 do you have some links regarding "noiseshaping"?

[pandy71]

Originally Posted by kovacm

I ask simple question and get answer from few users (thanx!) but pandy71 is the one who insist that "you can not compare F030 and A1200", start talking about "tricks how to open border", "thing you need to emulate from Amiga on ST"... and post some pictures in post #14

But you can't compare Falcon with factory A1200 - they can be compared when at least 68040/40MHz or perhaps even 68060 will be used on Amiga.
Do you agree for such comparison and why NOT?

Yes - most of activities on Atari ST scene is related to recreate or emulate some Amiga native capabilities - why compare something like that?
And pictures are to illustrate that some graphic capabilities are out of scope simply due of differences between hardware - different hardware and direct comparison will be unfair from Atari ST point of view - remain part is well known - 11.3% faster CPU clock for ST will make slight difference.

Originally Posted by kovacm

from my point of view, he start defending (?) Amiga from every aspect, just to prove obvious thing (one that all we already know): that Amiga hardware is superrior to ST.
and thats brings us here.

my bad is that I continue to answer to him but since he obviously have knowledge, why not continue even offtopic.

Nope - i don't need to defend facts - from HW perspective Amiga is better than ST - this is fact.

Originally Posted by kovacm

anyway, I should disclosure my agenda.

(english is not my frist, not even second language so please have some apprehension )

my final goal is to made website where you could compare Amiga and Atari (and some Macs) computers with PC through achievements in time. it will include birth of today mainstream software (CuBase, Lightwave, 3D Max, Logic, Office...), and live demonstration what you could do with Amigas and Atari back in 80s... and what with PC.

to comapre them and to show how PC manage to slow down computer industries!

Then why 3D scene demos that are completely not related to applications are mentioned at the start?
You can't compare Amiga to ST as you can't compare Macintosh to ST - HW differences are to big - btw i'm not sure is there demoscene for Macintosh?
You can compare applications and "productivity" but this is completely different thing than you started this topic.

Originally Posted by kovacm

yes, very interesting finding.
Atari ST have external ASCI port for almost 2MB/s. Why ASCI? In 1985. SCSI was not finished (specification was not finalize). Atari anyway built in ASCI which is very similar to final SCSI but it is not 100% compatible. Atari produced CD-ROMs, harddisks and laser printer for it (Atari SLM laser printer was the first "windows" printer - printer that has no CPU or memory).

I know that - i like ST community and ST as a computer so this was my point - it will be unfair to ST comparing it with Amiga.

Originally Posted by kovacm

3.5MB/s on ST?
Mr. Petari made adapter for CF Card that is connected throug ROM port (cartridge). ROM port has all MC68000 adres and data signals but it lack two signals that are important for hispeed writings. you can read more on:
http://forum.8bitchip.info/hardware-16/cartridge-port-ide-if-design-finished-(project-cata)/msg1064/#msg1064
http://forum.8bitchip.info/software-...ayback-on-ste/
http://forum.8bitchip.info/software-...g1007/#msg1007

So this is additional HW - also you showing why im saying that whole topic is not fair - you can't compare peaches to pears - any HW change must have known implications to performance - only with this we can compare incomparable.
As plain Amiga have no HDD interface and A600 first with HDD interface have known limitations (PIO mode + other implications).

Originally Posted by kovacm

btw pandy71 do you have some links regarding "noiseshaping"?

There is lot links about noiseshaping
http://wiki.hydrogenaudio.org/index....=Noise_shaping
http://en.wikipedia.org/wiki/Noise_shaping

Personally i use SoX http://sox.sourceforge.net/ which provide quite nice set various flavors of noiseshaping filters.

I think some different technique can be used for PSG used in ST

"In 2006, two MSX developers created an advanced encoder that converts a wave file to optimal PSG channel transitions using a Viterbi search. They replayed a 44.1 kHz wave file on a 23 year old MSX and achieved a higher signal-to-noise ratio than an 8-bit DAC. The Viterbi search is rather CPU intensive, so even though it would have been theoretically possible to use this method already in the 80's, there were no computers powerful enough to perform the analysis required."
https://en.wikipedia.org/wiki/Genera...ced_techniques

http://www.msx.org/downloads/related...cm-encoder-001
http://map.grauw.nl/articles/psg_sample.php

Also 4 bit nonlinear DAC can be used on ST (part of PSG) but i think special tool must be created (ie noiseshaper that work with logarithmic steps not LPCM)
But with modern technology (knowledge and processing speed) should be possible.

[kovacm]

Originally Posted by pandy71

So this is additional HW - also you showing why im saying that whole topic is not fair...

enough! I asked simple thing and got answer.

You turn this topic into proving already known fact: that Amiga is better than ST.


...anyway, if somebody fish to read about "DSP on 3D" graphics, you can visit Atari-forum:
http://www.atari-forum.com/viewtopic...25174&start=25 (dml's posts)

[pandy71]

Originally Posted by kovacm

enough! I asked simple thing and got answer.

Nope You cheated with question but this is you right

And this is proof for this:

"I would say that ST coders are far superior than Amiga 500 coders - they copy almost every amiga effect on less capable hardware!"

http://www.atari-forum.com/viewtopic...232966#p232966

Your own words... and this was real point - you not doing fair comparison, you only trying to prove some thesis at all cost even bending truth to your perspective.

Originally Posted by kovacm

You turn this topic into proving already known fact: that Amiga is better than ST.

Nope - i'm trying to be fair - compare things that can be compared and insist to be really strict to avoid any mistakes.

And seem that Atari guys are aware of the Falcon HW superiority vs plain A1200:
http://www.atari-forum.com/viewtopic...232980#p232980
http://www.atari-forum.com/viewtopic...232984#p232984

Originally Posted by kovacm

...anyway, if somebody fish to read about "DSP on 3D" graphics, you can visit Atari-forum:
http://www.atari-forum.com/viewtopic...25174&start=25 (dml's posts)

And funniest part - you judging on how smooth graphics is based on YT - which is quite funny as YT reduce (and convert) framerate to 30fps so in fact incorrectly from both machines point of view (assuming that both display 25/50fps).

But when i saw your first message i pointed you that this is simple unfair - comparing A500 to ST and A1200 to Falcon.

Be well.

[ajk]

Originally Posted by pandy71

And funniest part - you judging on how smooth graphics is based on YT - which is quite funny as YT reduce (and convert) framerate to 30fps so in fact incorrectly from both machines point of view (assuming that both display 25/50fps).

Actually YouTube won't convert 25fps material to 30fps - it will stay at 25fps. 30fps is just the maximum and anything above that will be converted. So the smoothness of those movie trailers and things like that can be displayed relatively intact. 50fps games can't be, of course.

[Steve T]

Originally Posted by ajk

Actually YouTube won't convert 25fps material to 30fps - it will stay at 25fps. 30fps is just the maximum and anything above that will be converted. So the smoothness of those movie trailers and things like that can be displayed relatively intact. 50fps games can't be, of course.

Depending on exactly how it works, there might be interpolation of the frames, so it would be more like 30fps + *proper* motion blur which would give a smoother result.

If so, a game orginally rendered and captured at 50fps would end up in motion on youtube looking smoother than a game captured at 30, perhaps not even noticably different to the original.

[ajk]

@Steve T

YouTube uses a blend conversion, ie. it will mix the various frames together at an appropriate ratio. So while you in principle get all of the images jammed in there, the result is quite blurry (has a double exposure look to it). Whether it's better than just decimating to 25fps is a matter of taste, but I'd say it isn't (30 fps is only a 20% improvement over 25, after all). Rather just have a clear 25fps YouTube video and a true 50fps version for download somewhere else.

[pandy71]

Way how YT process video depend who uploading video and how - for details you can ask Ben Waggoner who is currently Google guy (previously he work on VC-1 codec for Microsoft).

Anyway it is irrelevant as native video is not kept and final result highly depend from YT encoding, decoder and machine itself (PC have sometimes large jitter for video).


I would say that for example Vimeo is much better as it offer chance to provide original video (native resolution and framerate).

[Steve T]

Originally Posted by ajk

@Steve T

YouTube uses a blend conversion, ie. it will mix the various frames together at an appropriate ratio. So while you in principle get all of the images jammed in there, the result is quite blurry (has a double exposure look to it). Whether it's better than just decimating to 25fps is a matter of taste, but I'd say it isn't (30 fps is only a 20% improvement over 25, after all). Rather just have a clear 25fps YouTube video and a true 50fps version for download somewhere else.

that's another way of saying interpolation. You should get a much better, more accurate, result going from 60 down to 30 than 25 up to 30, either way is mixing frames.

[kovacm]

Originally Posted by pandy71

"I would say that ST coders are far superior than Amiga 500 coders - they copy almost every amiga effect on less capable hardware!"

http://www.atari-forum.com/viewtopic...232966#p232966

Your own words... and this was real point - you not doing fair comparison, you only trying to prove some thesis at all cost even bending truth to your perspective.

it is not thesis but my own opinion.

and yes, I did not post this on amiga forum, but on atari forum, because I know that here will be at least one "pandy71"

Originally Posted by pandy71

Nope - i'm trying to be fair - compare things that can be compared and insist to be really strict to avoid any mistakes.

as I said: I do not care what do you think that is "fair comparison".

I care to know which AGA demo works on stock A1200 PLUS FastRAM.

---
btw
Atari demos could be watch at http://dhs.nu/video.php
ST PhotoChrome video could be found at http://8bitchip.info/forum/

[pandy71]

Originally Posted by kovacm

it is not thesis but my own opinion.

and yes, I did not post this on amiga forum, but on atari forum, because I know that here will be at least one "pandy71"

haha, good joke - blame me for everything if this make you more happy, i don't care.

Originally Posted by kovacm

as I said: I do not care what do you think that is "fair comparison".

I care to know which AGA demo works on stock A1200 PLUS FastRAM.

Yes, i know, you don't care about comparison.
Demo requirements are usually provided with demo itself - check http://pouet.net/ before starting "comparison".
Be well.

[ajk]

Originally Posted by Steve T

that's another way of saying interpolation. You should get a much better, more accurate, result going from 60 down to 30 than 25 up to 30, either way is mixing frames.

Interpolation can mean many things. Modern standards conversions usually do actual motion interpolation, where entirely new frames are created based on motion vectors, not by blending together two existing ones. Motion interpolation can also be used to simulate motion blur. But YouTube does not do that, and it's anyway usually not that great for "unnatural" material like our games or demos.

Going from 60fps to 30fps (or 50 to 25) with frame blending just means that you have a 50/50 mix of two frames each time. I usually just drop 50fps to 25fps by simply leaving out the other frames - motion is not as smooth but frames are left perfectly clear.

Some games present a problem when doing this, however, since they use a 50Hz flicker to simulate transparency or other effects (eg. Turrican, Superfog and many others). If you just drop frames here, the effect is lost (either you get a missing player character or no transparency/flicker at all). When uploading such material, I have used a custom blending filter, which attempts to retain the effect where it exists without blending entire frames. [ Show youtube player ], Turrican 2 and Lionheart are processed like that, others are just decimated to 25fps.

@pandy71

Unfortunately Vimeo doesn't keep anything above 30fps either, it will get converted.

[pandy71]

Originally Posted by ajk

@pandy71

Unfortunately Vimeo doesn't keep anything above 30fps either, it will get converted.

yep, seems i was wrong - i've check all vimeo files i downloaded and all of them are bellow 30 fps (with sometimes strange framerates but anyway 30 or less). I need to check vimeo by some synthetic video - some up-loaders allow to download original file (mostly quicktime so it looks like untouched by vimeo).

[Photon]

Originally Posted by NovaCoder

I think almost all demos are pre-calculated.

Originally Posted by StingRay

That is, quite frankly, far away from truth, most demos are 100% realtime.

When you watch a demo and see an effect that looks much more impressive than the others, it has very likely been precalced/prerendered. Often, demos also have animations of course, but for the most part it's clear to the viewer that it's an animation.

If you download Vision, you can see the 400KB animation file, but of course you don't see the files or even the filesize at compo time (or on Youtube). So in this case it's a bit devious since the framerate is so slow. And so you think they could maybe have done the impossible.

Of course, it all becomes clear if you run it on a real Amiga or emu matching compomachine specs.

There are various ways to make impossible things possible, f.ex. animation, precalc, prerender etc. There are some famous ones, I don't know if Alpha and Omega was first. It's obvious to a good coder that it's precalced, but that's usually a tiny part of the audience.

You can test it out in WinUAE by doubling or halving the CPU speed and see if the framerate changes.

There are two schools of thoughts on the realtime thingy. One is that realtime should be realtime and the other is "who cares?". Mostly, the realtime phalanx accepts general-purpose tables (ie. not made for the effect) on lower-spec machines.

Also check out Dual Crew/Shining (DCS) and Skarla.

From what I gather A1200+Fastmem was not the stock platform like OCS/1MB is, 030 cards arrived pretty quickly and was accepted as demoscene platform.

The substantial advantage of Falcon to A1200+Fast only is that chunky mode and playing wav comes for free, so the CPU can just render. A 68030 vs 68EC020 at the same bus width and frequency is roughly 2.15x faster, with the faster clock that would be 2.46x faster.

[carls]

Originally Posted by kovacm

Hi,
I am long time Atari user (since 1986.) and I was wonder what are best Amiga 1200 demos with most complex 3D scenes BUT for STOCK A1200 (14MHz - with FastRAM).

Thanks in advance!

At the moment, I'm very into A1200+fastmem demos.

Below are a few with good 3D or other complex effects. Most of the videos were probably recorded on something faster, but I've watched all of them on real hardware (I've got both Hawk and Microbotics cards, pure memory expansions without faster CPU) recently and I've been impressed by their speed on A1200+fastmem.

Wit Premium by Freezers is crazy fast:
[ Show youtube player ]

Mindflow by Stellar was designed for A1200 with FastMem
[ Show youtube player ]

The 4K intro Dawn by Artwork contains a really fast 3D engine which is impressively fast:
[ Show youtube player ]

G-Force by Pygmy Projects features shaded 3D without texturing and it's quite snappy:
[ Show youtube player ]

Motion by Bomb also contains some fairly complex 3D and runs smoothly - EXCEPT the final Doom world, which is quite jerky. I think the video is recorded using an 030 turbo.
[ Show youtube player ]

Nexus 7 by Andromeda runs fine on even an A1200 without fastmem and is still very impressive:
[ Show youtube player ]

Syndrome by Balance is another demo that is nice and fast with some 3D:
[ Show youtube player ]

Dove by Abyss is another nice one with a super funky soundtrack to boot:
[ Show youtube player ]

[kovacm]

thank you!!

yes, main problem is that many videos of demos are recorded on faster hardware or on WinUAE with "unlimited" speed

I also have A1200 (need to but FastRAM card) so it is good to know which one works on this hardware!

[carls]

Originally Posted by Photon

From what I gather A1200+Fastmem was not the stock platform like OCS/1MB is, 030 cards arrived pretty quickly and was accepted as demoscene platform.

That's true, the "vanilla" A1200, even when accepting fastmem as "vanilla", had a very short scene life, basically most of the good releases for this platform were made in 1994. In 1995 you really wanted a 68030 running at 40 MHz or faster to be able to enjoy the winning demos from the big parties.

Originally Posted by kovacm

thank you!!

yes, main problem is that many videos of demos are recorded on faster hardware or on WinUAE with "unlimited" speed

I also have A1200 (need to but FastRAM card) so it is good to know which one works on this hardware!

I'm a firm believer in that demos should be watched and recorded on their target machine. There's nothing impressive in a demo being fast on a 50 MHz 030 if it was intended for a 14 MHz 020.

Make sure you get a good RAM board - the Hawk (and clones) is a real slouch. Microbotics MBX 1200z, Power Computing PC1204/PC1208 and the Blizzard 1200/4 all have better memory architecture.

[kovacm]

^
I got A1200 with Blizzard PPC, low end model.
Can you explain me some more things:
Is it possible to turn off 040 but use fast memory (and would speed be comparable to plain fast ram upgrade)?

[carls]

Originally Posted by kovacm

^
I got A1200 with Blizzard PPC, low end model.
Can you explain me some more things:
Is it possible to turn off 040 but use fast memory (and would speed be comparable to plain fast ram upgrade)?

Nope, I don't know of any CPU card that can do this. I guess that on a processor card, the RAM is directly wired to the faster CPU and thus inaccessible to the 68020 on the MoBo.

Originally Posted by Frog

A1200 with FastRAM

Crazy Sexy Cool / ESSENCE
Makaveli / ESSENCE

Glow / TBL
Tint / TBL
Goa / TBL

Closer / CNCD

Phi / ARTWORK
The Gate / ARTWORK
Megademo IV / ARTWORK

Grid 2 / NATURE
Zeon / NATURE

My Kingdom / SCX & HJB

Fear Factory / ARSENIC

While some of these demos can be run on A1200+fastmem, they'll look like slideshows. The target platform for most of them is 030/25 MHz or faster.