Steven Seagal's Atari ST Web Site

Steem

"Mercenary for Emulation"

  The third screen of brags for demo fixes.

 

 
 

36.15 Gen4 by Cakeman (STF)

 
 

 
  This program failed because the VBI (VBL interrupt) and the left border trick timings are different for STF. It proves I wasn't raving about "Steem hacks". It's been fixed for v3.4, hopefully. Cool effects (especially Dino) but it's in French.

This demo is also called 3615 Gen4 like the one by ULM because it was a (French) contest. There are many more.

Note: in v3.5.1, it fails more often than works. Seems better again in v3.5.2...
Fixed v3.6.

 
 

Download 36.15 Gen4 by Cakeman

 
     
 

3615GEN4 HMD Giga (STF)

 
 

 
  This was unwittingly fixed in v3.5.1 and broken in v3.5.3, repaired more or less in v3.6.0.

v3.6.4: will work fine STF, Hacks.

v3.8.0: the hack has been removed, but now you need a patch (included) - it's a frustrating case

 
 

 http://www.atarimania.com/demo-atari-st-gen4-demo_22231.html

 
     
 

4pixel Rasters

 
 

 
  Already for Steem 3.3, CPU prefetch was fixed except for the timing placement, which I did in v3.5. I thought it was a nice effort but practically useless as Steem 3.2 already passed prefetch protections. This wasn't all for nothing though as here is the first program I know that it fixed: the rasters were 8 (?) pixels wide instead of 4 like intended in this nice technical demo!  
  Here's the same demo with wrong raster size due to lack of prefetch:  
 

 
  Gross.

Technical update: intrigued, I looked at it again. The fix is really the timing, not the value in IRC. Prefetch happens before the poke for -(An), when the program is writing on the palette registers. Generally the timing wasn't modified in v3.3, except for one or two cases in MOVE, and that's one of the cases.

 
 

 http://www.pouet.net/prod.php?which=29074

 
     
 

Beeshift

 
 

 
  This is, once again, a very technical demo by Paulo (LJBK at AF), finding ways to shift the screen at the cost of very few CPU cycles. It offers scrolling possibilities but it depends on the wake-up state.

There are several versions using different "switches", the programs recognise up to 4 wake-up (WU) states, but Steem handles only 2 WU states: 1 and 2, that fortunately are recognised.

Support in v3.5.3.
V3.5.4: Steem handles 4 states.

 
 

Download Beeshift  beesclr4

 
     
 

B.I.G. Demo

 
 

 
  "B.I.G." stands for "Biggest in Galaxy".
It is a demo by TEX, and it is of course outstanding.
In Steem SSE v3.7 the sound is closer to a real ST (more saturated).
The glitch in the 2nd screen has also been removed.

The version presented here won't crash at the 'Action-Biker #1' tune.
All those years were necessary before someone found the bug in most versions.
And they call themselves fans of the YM-2149 :(
 
 

Download BIG Demo

 
 

BIG demo STW

 
     
 

Board and Sphere (STF WU2)

 
 

 
  Part of the Ventura megademo, this is another case of destabilised Shifter. Apparently there was a bug in the code and programmers adapted the picture to the shift induced by wrong stabilisation. This wouldn't work on STE, that's why you had to press F1. Not sure it would work on all STF either. In Steem 3.6 it will display OK in WU2 mode.

Emulation is still pragmatic (hacky), but the concept used is general: when the Shifter contains X (1-3) extra words (16bit data) in its registers, there are some effects on the display, that is pixel and plane shifts. This is the same concept for cases like this one, Overdrive/Dragon, Death of the Left Border, Omega Full Overscan, and the very different "Bee" demos.

Well, there are some hacks.

 
 

 http://www.atarimania.com/demo-atari-st-ventura_22197.html

 
     
 

Cernit Trandafir (STE, 4MB)

 
 

 
  Display of one of the demo screens was wrong because CPU timings - generally accurate in Steem - were overestimated in some cases. Precisely, cycles shouldn't be rounded to 4 when source addressing mode is -(An), apparently because of the order the CPU microcodes are set up in.

Update v3.6: because the general fix would break other programs (Ishar3 cracktro, Fullparts), we fall back on the specific fixes on ADDL.L -(An),D and ADDA.L -(An),A.
v3.6.4, another little change.
v3.7, many CPU timing were reviewed at the light of the Yacht document, this is now handled, like so many other cases, in a general and systematic way, it's just correct.

 
 

http://www.atarimania.com/demo-atari-st-cernit-trandafir_23297.html (HDD)

 
     
 

Circus back≤STage (STE)

 
 

 
  [If the pic doesn't show up, it may be because it's confused with some NSFW material by your filters - there's nothing like that in the demo]

The second screen of this demo was probably fixed at the same time as Riverside. Anyway now it works.

Some glitches still appear but they do on my STE too.

 
 

 http://www.pouet.net/prod.php?which=61178

 
     
 

Dragon by Ghost (STF WU2)

 
 

 
  This is the very last fix for Steem 3.4.0, in fact an ugly hack as I think emulating this kind of shifter quirk isn't trivial and I wasn't aware of the issue until very recently. At least you get a good rendering of the cool demo. It seems the coder couldn't see all that happens in the borders. Even in Normal display size mode there is border trash (note this comment isn't political). This will work only in STF2 mode. It is part of Phalanx/Overdrive.

For v3.5.2, I changed the way, it's less a hack. Rendering is worse though.

In v3.5.3, because of LJBK 's Beeshift demo, we now can try to explain it:

At the end of the frame there's a savage and useless removal of the left border, not followed with a "stabiliser" switch. The effect is dramatic! When the left border is removed, the MMU fetches 26 more bytes than in a regular scanline. As the 4 shifter IR input registers are 16bit, they hold 8 bytes max. A scanline divisible by 8 will empty the shifter registers.The scanline without the left border has 160+26 bytes, or 160+24+2. Two bytes remain in the shifter by the end of the scanline, and the frame. As the shifter gets multiples of 8 bytes, there will remain 2 bytes at the end of each scanline. They're still there at the start of next frame, and as long as there's no border effect, in the top part.

The effect that must be emulated is that each of those scanlines is shifted, both the planes and the pixels. The shift in pixels is exactly 4, just like in the bottom lines, that are shifted because of the left border removal. Those lines are stabilised.

The nice thing: the shifts are compatible with those we use for the bee (see above), so it all holds up together. And it is a nice demo, properly rendered again.

Option wake-up state. If set to "ignore", no shifter "preload" (still hesitant how we must name this) is emulated.

v3.6.0: WU2 for this one.

 
 

 http://www.atarimania.com/demo-atari-st-overdrive-demo-_22265.html

 
     
 

Hier sind noch zwei Plštze frei

 
 

 
  Since v3.4.1, the plasma works with 6301 option checked STF or STE.
Without it, too short ACIA timings literally prevent the screen from being cleaned.
 
 

Download Hier sind noch zwei Plštze frei

 
     
 

HighResMode (STE)

 
 

!

 
  Argh! I wasn't aware that this 29.791 colours demo (their brag; such pics look better live due to the trick) used to perfectly run in Steem 3.2. I wasn't aware that it was broken in that buggy SSE 3.3. In fact I thought it was a monochrome demo and left it aside until now! Fixed again in v3.4. There was also an issue with the disk image. Because of this I make the demo available here.

Note that the demo runs at 60hz to reduce flicker, so it may not look smooth on a 50hz screen. Like others it's perfectly emulated with VSync. In fact, you may verify this easier with this 60hz demo than with a 50hz one if your display is 60hz.

 
 

Download HighResMode

 
     
 

LoSTE Screens

 
 

 
  This new demo by Sync is at the basis of a few shifter trick fixes for: 0 byte lines, lines +2, lines -106! V3.5.3.  
 

 http://demozoo.org/productions/88297/

 
     
 

The Musical Wonder 1990

 
 

 
  This demo makes your screen borders go to hell. In 416 x 286 display size mode, Steem 3.5.3 is the biggest of all Atari ST emulators. It will display 36 "top border" lines instead of 30. This is because some palette effects target those lines, like obviously here. In any smaller resolution, the 'Offbeat' is cut, and we're really off the beat.  
 

 http://demozoo.org/productions/73274/

 
     
 

Overdrive Demo

 
 

 
  Integrating IPF support in Steem has had some positive consequences: insights gained through reading the WD1772 disk drive controller doc and the CAPS source - not to mention Hatari - may be useful in Steem's native drive emulation as well. Steem 3.4 already loaded Overdrive by Phalanx (here the infamous boot screen on which Steem 3.2 was stuck) only thanks to sordid hacks.

In v3.5 we go a step ahead. Now the Disk manager option Accurate Disk Access Times (ADAT for those in the known) is even more... accurate.

In true HD6301 emu mode, the spaceship didn't show on the menu (but you could make it appear). Fixed in v3.5.1

V3.5.2: fixed some ACIA trouble (keyboard stuck). I hope this time it's final.

 
 

 http://www.atarimania.com/demo-atari-st-overdrive-demo-_22265.html

 
 

Overdrive STW

 
     
 

Riverside (STE)

 
 

 
  Riverside, a recent 64KB demo by DHS, didn't display correctly in Steem.

It's because the STE-only technique used to remove the left border, line+20 bytes, causes an 8-pixels shift of the scanline which wasn't emulated yet.

Don't cry rivers, this is fixed for v3.5.2, but because the core video rendering system of Steem is in assembly and modifying it is involved, I used a hack instead, which makes the display of some other demos using the same trick (eg More or Less Zero) less elegant in the borders, only in very large display size mode. The importance of this is more or less zero, just mentioning it.

There was another problem in the lower border due to some tests not being done if the scanline wasn't supposed to be done yet even though cycles were ahead (eg. we're 516 cycles into scanline). Pure internal issue. This latter fix could very well pollute other programs, so some testing is needed.

 
 

 http://www.pouet.net/prod.php?which=61514

 
     
 

SSE3.5 Demo

 
 

 
  This uses the Super Neo Demo Show to display some cool NEO pics and a scroller. The demo was included in the download package of Steem SSE 3.5.0.  
 

Download SSE3.5 Demo

 
     
 

ST NICCC v2

 
 

 
  In STF mode, this demo had two problems: drive speed was wrong, and sampled
music wasn't well rendered.
 
 

 http://leonard.oxg.free.fr/demos/atari/ST-NICC2.zip

 
 

ST NICCC STW

 
     
 

Summer Delights (STE, 4MB)

 
 

 
  Emulation of this demo in Steem progressed with the versions, with some screens fixed in v3.3 and others in v3.5.  
 

http://demozoo.org/productions/61454/ (HDD)

 
     
 

Sunny demo by Sphere (STE)

 
 

 
  I don't know why exactly but it's a fact that the scroller of this demo isn't jerky anymore in v.3.4, so why not a screenshot on this sunny day for emulation?  
 

 http://www.atarimania.com/demo-atari-st-sunny-demo_26169.html

 
     
 

XMAS-2004 (STE)

 
 

 
  This demo would suddenly stop in Steem because of an unexpected bug in... Microwire (volume & bass control) emulation.

When this was fixed, there remained a confusing bad screen caused by conditions we're still examining: STE HSCROLL, writing the video counter, etc.

This is "fixed" in v3.6 thanks to hacks.

 
 

 http://www.atarimania.com/demo-atari-st-xmas-2004-demo-_23688.html

 
     

 

Previous screen

Next screen

Games (1) (2)

Demos (1) (2) (3) (4)

Apps

Steem

Other downloads