[TI ASM] Splash Screen In Greyscale

kalan_vod · Post by **kalan_vod** » Sat 22 Apr, 2006 1:50 am

Have you looked at the code of Jims package or Ducks?

Post by **L4E_WakaMol-King** » Sat 22 Apr, 2006 1:52 am

No. I'm not sure where to get them... but even if I did, I have a feeling it would be a bit over my head. I'm *hoping* there is some kind of simple solution to remove the flicker... though I could be in for a nasty suprize.

kalan_vod · Post by **kalan_vod** » Sat 22 Apr, 2006 1:57 am

Are you using interrupts?

Post by **L4E_WakaMol-King** » Sat 22 Apr, 2006 2:36 am

No, but if I can avoid using them, I would like to. I plan to add music behind this, and I think I need the interrupts for that. I don't know if I can use interrupts for both music and greyscale

.

kalan_vod · Post by **kalan_vod** » Sat 22 Apr, 2006 2:40 am

I think it is possible, but maybe not done yet?

Liazon · Post by **Liazon** » Sat 22 Apr, 2006 4:21 am

L4E_WakaMol-King wrote:No, but if I can avoid using them, I would like to. I plan to add music behind this, and I think I need the interrupts for that. I don't know if I can use interrupts for both music and greyscale .

how many interrupts are on the 8x+ series calcs?

Post by **L4E_WakaMol-King** » Sat 22 Apr, 2006 5:29 am

Good question. There's only one useful one that I know of. But I'm not a very experienced ASM programmer, so I am not the one to ask.

threefingeredguy · Post by **threefingeredguy** » Sat 22 Apr, 2006 5:41 am

On the bare minimum 83+, you can have 3. With all the spiffy timers on the SE and the 84+s you can have lots.

Spencer · Post by **Spencer** » Sat 22 Apr, 2006 8:30 am

There's only one useful maskable interrupt on the hardware/software setup we're using. Normally with Z80 the interrupting device puts a byte that refers to its service routine on the data bus, but for us, that byte is random (that is, there's no distinction between what generated the interrupt). So we fill all slots for service routines with references to our interrupt code.

You will not have enough time for sound and grayscale.

_grbufcpy is TI-OS's safe routine to copy plots to the LCD. The minimum delay and the nature of the LCD driver prevents simple page flipping from looking good on the 83 series. The most commonly used method of grayscale is merging the two video buffers in the correct proportion and sending them to the LCD together every display frame, rather than the 2-1-2-1 with page flipping.

Post by **L4E_WakaMol-King** » Sat 22 Apr, 2006 9:12 am

Ehh, that sounds a bit complicated. Could you show me what you mean in code? For what it's worth, I'm using ION, and can use IonFastCopy if that will help at all.

Here's what I've got so far:

Code: Select all

lp_greyscale_loop:

   ld hl,Splash1
   call disp_pic
   ld hl,Splash2
   call disp_pic
   ld hl,Splash2
   call disp_pic

   bcall(_getcsc)
   cp $38
   jr nz, lp_greyscale_loop

   bcall(_getkey)
   ret

disp_pic:
   ld de,PLOTSSCREEN
   ld bc,768
   ldir
   bcall(_GrBufCpy)
   ret

Timendus · Post by **Timendus** » Sat 22 Apr, 2006 3:39 pm

Yes, ionFastCopy will probably already improve the quality. What Spencer meant is that you can mix the layers on the fly in an intelligent way to reduce the flicker and create some noise in the image instead. It usually looks better, but it's quite a bit more complicated to write. But if I were you I'd do it the other way around. Just use Jim or Duck's greyscale package, use that to show the splash screen with the interrupt, and have the main program produce some music.

CoBB · Post by **CoBB** » Sat 22 Apr, 2006 3:57 pm

How would that work? Timing is critical to produce the proper sounds, and a greyscale interrupt ruins it completely, in a nondeterministic way.

Jim e · Post by **Jim e** » Sat 22 Apr, 2006 5:05 pm

Yeah you cant do one with out the other suffering.

But for gray really timing is key, not speed. what you want is to use a delay after each lcd update. If you are using ionfastcopy you probably want a delaying routine in the nieborhood of 50,000 tstates. You'd have to play with that number but its possible to get it looking good without interlacing.

Timendus · Post by **Timendus** » Sat 22 Apr, 2006 5:44 pm

CoBB wrote:a greyscale interrupt ruins it completely, in a nondeterministic way.

Ah, yes, I see your point. I figured a little slowdown wouldn't matter, you'd just increase the frequency a bit to counter that, but that's really stupid reasoning... I'm curious what sound greyscale would make though

Perhaps you don't even notice it. But it'll probably just sound like noise...

CoBB · Post by **CoBB** » Sat 22 Apr, 2006 5:59 pm

Jim e wrote:But for gray really timing is key, not speed. what you want is to use a delay after each lcd update. If you are using ionfastcopy you probably want a delaying routine in the nieborhood of 50,000 tstates. You'd have to play with that number but its possible to get it looking good without interlacing.

The problem is that you inevitably have to stop all sound processing while the screen is refreshed, which is quite a nasty square wave modulation... Okay, you can theoretically use the wasted cycles as well, but are they still only accessible at a frequency that doesn't play nice with various tones.