Ray Tracing

Rick has been busy writing about “Ray Tracing” in BASIC on his site.

The code with small mods has been run on a Beeb!

Like this? 1600×900 png

A few lines of Basic and 4 whole colours.

Rick has been busy writing about “Ray Tracing” in BASIC on his site.

It’s actually ray casting. Slightly different beast, I did explain the difference.

Like this?

Pretty nifty. Thank you for demonstrating my point about people who create ray tracers liking reflective spheres. ;)

Like this? 1600×900 png

Is it my eyes, the screen I’m viewing on, the browser, or is the chequerboard representation of a planar surface slightly distorted?
This is the bit where someone says “most of the above” :(

Is it my eyes?

Please feel free to improve it:

Yes, it is an adapted BBC Micro program

[You won’t be able to change mode so easily unless you have my VDUMode module though.]

The actual f……..

I know you said

A few lines of Basic

but that’s taking the…….

that’s taking the…

I once won an Acorn User (I think – might have been a different mag) competition with a 3 line (iirc – could have been 4 or 5, I don’t have a copy to check, memory might be playing tricks but I don’t think so) BASIC program that converted a string representing an integer in any base*, to another string representing the same number in any other base*. 3 (or however many it was) LINES, but a LOT of instructions. You can even cheat on the 255 character limit on the length of a line, by putting the tokens in instead of the human-readable versions.

* Up to some limit, possibly 256, I forget.

Up to some limit

Length byte maximum is 255 of course but it, the CR and the line number take up 4 bytes, so 251 is the maximum number of actual Basic chrs/tokens.

Please feel free to improve it:

Shouldn’t that be: Please feel free to get it run?
I was hoping that selecting text and dragging it into a basic editor would suffice, but it turns out to be a PNG, so I started copying character by character. Then I spent an hour or so to discover all of my typos. When it finally ran in a TaskWindow, it didn’t do anything. Admittedly, I’m one of those persons with overwhelming daft, so I really need some additional instructions.

some additional instructions

Not in a TaskWindow will help, for sure. Save it out and run it.

but it turns out to be a PNG

Of course. Just providing the code makes it too easy. ;)

Save it out and run it.

Note also nemo’s reference to a custom mode. You’ll want to tweak that to be something else.

custom

It’s not a custom mode but a custom API.

It has irked me that the introduction of deep modes, mode specifiers and WimpMode strings in RO3.5 was not extended to the actual VDU, so whereas VDU22,<num> continued to select numbered modes, there was no way to select anything else.

VDU23;22;<width|0>;<height|0>;<log2bpp|9>; is implemented by the VduMode module and I’d forgotten that isn’t standard, sorry. The silly numbers select the existing values, so it’s possible to change BPP without caring what your screen size is, for example.

And then because this is an emulator, there’s another module called AnyMode that enables any screen geometry to be selected (subject to RO granularity restrictions).

Fun fact: This monitor is 900 lines high and RISC OS 4.39 doesn’t totally get scrolling right when it isn’t a multiple of eight, so leaves some lines of junk at the bottom of the screen, uncleared, when hardware scrolling. Anyway, enough of my agonies.

Paul typed in a listing and then tried to debug it like it’s 1982 muttering

I spent an hour

I am very sorry. I didn’t think anyone actually would. Here’s the code as it is:

10  REM€˜†ž$Œ*’‰!ƒ›-•'
20  VDU23;22;1600;900;1;:OFF:S=800:w=1600:h=900:x=799.5:y=449.5:A=2
30  FORN=0TOh-1:FORM=0TOw-1
40  X=0:Y=-.1:Z=3:U=(M-x)/S:V=(N-y)/S:W=1/SQR(U*U+V*V+1):U=U*W:V=V*W:I=SGNU:G=1
50  E=X-I:F=Y-I:P=U*E+V*F-W*Z:D=P*P-E*E-F*F-Z*Z+1:IFD>0T=-P-SQRD:IFT>0X=X+T*U:Y=Y+T*V:Z=Z-T*W:E=X-I:F=Y-I:G=Z:P=2*(U*E+V*F-W*G):U=U-P*E:V=V-P*F:W=W+P*G:I=-I:GOTO50
60  IFV<0P=(Y+2)/V:V=-V*((INT(X-U*P)+INT(Z-W*P)AND1)/2+.3)+.2
70  GCOL0,3-(48*SQRV+((?(PAGE+5+M MOD4+N MOD4*4))AND127)/3)DIV16
80  PLOT69,A*M,A*N:NEXT,

However, the bytes in the REM statement at the start (I’m sorry, this was a BBC Micro program I adapted slightly) will probably get mangled by the internet pixies. In hex they are: 80 98 86 9E 24 8C 2A 92 89 21 83 9B 2D 95 27 8F which you could stick in a DATA statement, read into a DIM’d block and then replace the PAGE+5 by that dim.

POINT will be minutely faster than PLOT69, too.

w,h are the size of your screen in pixels, S is effectively the focal scaling parameter, x,y are the pixel coordinates of the centre of a pixel in the middle of the screen (so (w OR1)/2 in effect) [the ½ is to avoid unpleasant zeros], and A is the Eigenvalue to convert from pixels to OS coords (and you’d need two for a non-square mode), and the rest of it is obvious [this is a joke]. HTH.

And here’s a version that uses 16 colours, a nice palette, a standard MODE, and no BBC Micro tricks like REM-peeking or GOTO:

DIME%15:FORZ%=0TO15:READE%?Z%:NEXT
MODE27:OFF:w=640:h=512:A=2
S=(w+h)/4:x=(1ORw)/2:y=(1ORh)/2
FORZ%=0TO7:COLOURZ%,0,Z%*48,Z%*8:COLOUR8+Z%,Z%*16,Z%*32,240:NEXT
FORN=0TOh-1:FORM=0TOw-1
X=0:Y=-.1:Z=3:U=(M-x)/S:V=(N-y)/S:W=1/SQR(U*U+V*V+1):U=U*W:V=V*W:I=SGNU:G=1
REPEATE=X-I:F=Y-I:P=U*E+V*F-W*Z:D=P*P-E*E-F*F-Z*Z+1
IFD>0T=-P-SQRD:IFT>0X=X+T*U:Y=Y+T*V:Z=Z-T*W:E=X-I:F=Y-I:G=Z:P=2*(U*E+V*F-W*G):U=U-P*E:V=V-P*F:W=W+P*G:I=-I:UNTILFALSE
UNTILTRUE:B%=0:IFV<0P=(Y+2)/V:V=-V*((INT(X-U*P)+INT(Z-W*P)AND1)/2+.3)+.2:ELSEB%=8
GCOL0,B%+7-(48*SQRV+((?(E%+M MOD4+N MOD4*4))AND127)/3)DIV8
POINTA*M,A*N:NEXT,
DATA&80,&98,&86,&9E,&24,&8C,&2A,&92,&89,&21,&83,&9B,&2D,&95,&27,&8F

so whereas VDU22, continued to select numbered modes, there was no way to select anything else.

Acorn probably figured we should be using the shiny new SWI and not spitting bytes at the VDU driver.
Unfortunately they forgot to implement SWIs for all the other VDU things like the graphics cursor, text sizes, etc etc.

This monitor is 900 lines high

Whoops! I have one that’s… 1440×900? Something like that. I use it on the PC.

tried to debug it like it’s 1982

Oi! That’s bleeding edge around these parts. Lots of SWI 298, and you’ll know how far it got by counting the asterisks on the screen.

The advantage of the VDU interface is that it can be captured by *Spool and indirection, can be spat out by a SWI OS_WriteN, and is automatically directed to all Printer Drivers (which won’t magically gain support for new VDU sequences, but where’s the Printer Driver intercept for OS_ScreenMode etc?).

all the other

VDU is a bit olde worlde because of its one-byte-at-a-time API, though UKVDU23V filters that for implementors somewhat. But the Printer Driver issue is a real one. Especially in an OS that traditionally has been phobicly resistant to coining new vectors.

And here’s a version that uses 16 colours, a nice palette, a standard MODE, and no BBC Micro tricks like REM-peeking or GOTO:

Also works a treat in a Graphic TaskWindow

Years ago the Arc plus Colour Card gave a mode of “MODE 120” 1024 × 768 16col.

Long time ago I tried to make a module to recreate MODE 120 on newer machines without success.

Has anyone got some ideas – Tks

The answer is simple. The 1024×768 is not relevant. You just want a 16 colour MODE number in place of 120 in the sprite.

1024 × 768

Do you mean “How can I get 1024 × 768” or “how can I make ‘MODE120’ select that”?

To select Mode 120 in a graphics taskwindow !Smalltsk – which has number modes as a input.

Try this and see if it works.

If not I’ll have to put more numbers in it.

Just found where I downloaded Nemo’s Mode 120 module 1024 × 768 16col – many thanks Nemo – seems to work ok with !SmallTsk on VRPc DL RO 5.x the Iyonix is playing up at the mo – USB being temperamental :-(

Any explanation of how the 120 Module works :-)

FYI my Iyonix USB (front panel) went high-resistance due to frequent keyboard swap-out. A quick resolder on the earth lug cleared the problem :)

My Iyonix double boots and then sits there looking at me no keyboard no mouse – auto booting no keyboard :-(

Is there a prog that gives the USB a kick if it is not working?