Hardware Scroll tutorial?

[Lonewolf10]

Hi,

Does anyone have a hardware scroll tutorial?

I am trying to scroll a 5 bitplane screen and I'm guessing that it'd take forever using the blitter, so I am attempting to do it using the hardware. The problem is I seem to only be moving bitplane 0 by 1 pixel to the left and then it stops.

I have set BPLCON1 as $0001, which I assume is why bitplane 0 has moved 1 pixel to the left. How do I move all of the bitplanes?

The Hardware Reference manual also mentions "setting the delay bits", but doesn't mention where they are! Can anyone point me in the right direction?


Regards,
Lonewolf10

[pmc]

Hey Lonewolf10

From memory: don't you just sequentially set the delay bits to scroll up to 16 pixels (which is the max you can scroll using only the delay bits IIRC) and then update the bitplane pointers to start displaying the bitplane data at the next or previous word depending on scroll direction...?

Keep doing that and you're scrolling.

Delay is set in bplcon1 IIRC.

I think you also have to do something with modulos too...?

EDIT: delay is bits 0 - 7 in bplcon1 (0 to 3 for odd bitplanes, 4 - 7 for the even ones)

[frank_b]

From memory.
You need to fetch an extra word with display dma. You need to set up the modulus registers to skip at least a word (the one which will be scrolled on screen) then set the fine scroll bits.

Something like

where x = pixels scrolled

set fine scroll (delay) inverse of x AND $f duplicated for odd, even planes.

ie $f is 0. Use a look up table to map x AND $f to set scroll for odd/even planes.

Hope this helps.

[deicidal]

Kreator / Anarchy wrote one I think many moons ago it would be in one of the early issues of Stolen Data

[frank_b]

Originally Posted by deicidal

Kreator / Anarchy wrote one I think many moons ago it would be in one of the early issues of Stolen Data

http://amigadev.elowar.com/read/ADCD.../node008B.html

You need to set the odd and even delay to the same value in bplcon1
From memory a delay of $ff means no scroll and $00 means scroll playfield left by 15 pixels. You're only setting one of the values.

[pmc]

Originally Posted by frankb

From memory a delay of $ff means no scroll and $00 means scroll playfield left by 15 pixels.

This would be the other way round wouldn't it...?

Normally the delay bits are not set so $00 == no scroll and the maximum settable value $ff == 15 bits of scroll.

You'd either decrement or increment the delay bits and bitplane pointers depending on the scroll direction required of course.

[frank_b]

Originally Posted by pmc

This would be the other way round wouldn't it...?

Normally the delay bits are not set so $00 == no scroll and the maximum settable value $ff == 15 bits of scroll.

You'd either decrement or increment the delay bits and bitplane pointers depending on the scroll direction required of course.

Possibly. It's been 15 years since I did any amiga hw scrolling code and I'm 1000 miles away from my amiga

I remember it being the opposite way around compared to the STE.
0 is no scroll on the STE and $f is view area scrolled right 15 pixels.
I have a fuzzy memory of using a look up table for the scroll values on the Amiga. Maybe I used a look up to set both values to save converting $f -> $ff when setting both playfield fine scroll settings after getting the fine scroll modulus via AND. My memory could be hazy though

http://amigadev.elowar.com/read/ADCD.../node02D0.html should clear it up

[Lonewolf10]

Originally Posted by deicidal

Kreator / Anarchy wrote one I think many moons ago it would be in one of the early issues of Stolen Data

Does anyone know if the issues of Stolen Data are stored in one place somewhere?
Googling amiga "Stolen Data" revealed multiple sites with some issues on them (e.g. most are scattered all over the internet). I'm not going to download them tonight, as I am tired.


pmc and frank_b,

I have spent the afternoon coding a simple scroll routine that moves the screen left and right by adding/subtracting 1 byte from each of the BPLxPTL pointers until the right/left edge of the screen was reached. Then it scrolls back the other way.
I should be able to add in code to use BPLCON1 now and make it a smoother scroll. Thanks for guidance. The phrase "delay bits" threw me, I was thinking in terms of time delay (e.g. set the delay bits, set BPLCON1 and the hardware would automatically scroll it every X VBL ).

FYI, the screen contains the logo for my demo group (is it a group if there is only one member?!), which was kindly created by Greyfox a few months back.


Regards,
Lonewolf10

[Codetapper]

Try here: http://bitworld.bitfellas.org/demo.php?id=14886

[pmc]

Originally Posted by Lonewolf10

is it a group if there is only one member?!

Well... it works for me!

My "group" is me, myself and I. It's always nice to work with other people though.

Regarding the scrolling, it sounds like you're nearly there to be honest. Once you're using the delay bits to fine scroll as well it'll look fine.

[TheDarkCoder]

Originally Posted by Lonewolf10

pmc and frank_b,

I have spent the afternoon coding a simple scroll routine that moves the screen left and right by adding/subtracting 1 byte from each of the BPLxPTL pointers until the right/left edge of the screen was reached. Then it scrolls back the other way.
I should be able to add in code to use BPLCON1 now and make it a smoother scroll. Thanks for guidance. The phrase "delay bits" threw me, I was thinking in terms of time delay (e.g. set the delay bits, set BPLCON1 and the hardware would automatically scroll it every X VBL ).

FYI, the screen contains the logo for my demo group (is it a group if there is only one member?!), which was kindly created by Greyfox a few months back.


Regards,
Lonewolf10

As far as I remember, the values stored in the bitplane pointers, should be even addresses. Usually ones add/subb 2 everytime the BPLMOD switches from $f to 0 (or the converse)

Shoudn't this thread go into the Coding Tutorial sub-forum ?

[TCD]

Originally Posted by TheDarkCoder

Shoudn't this thread go into the Coding Tutorial sub-forum ?

Your wish is my command

[pmc]

Originally Posted by TheDarkCoder

As far as I remember, the values stored in the bitplane pointers, should be even addresses.

You're right, the pointers should point at even addresses (word aligned).

[deicidal]

Originally Posted by Lonewolf10

Does anyone know if the issues of Stolen Data are stored in one place somewhere?
Googling amiga "Stolen Data" revealed multiple sites with some issues on them (e.g. most are scattered all over the internet). I'm not going to download them tonight, as I am tired.

Just to add I am going to make an educated guess that it will be in one of the first 5 issues

[Lonewolf10]

Originally Posted by TheDarkCoder

As far as I remember, the values stored in the bitplane pointers, should be even addresses. Usually ones add/subb 2 everytime the BPLMOD switches from $f to 0 (or the converse)

That depends whether you fully use BPLCON1, hehe. Still not had a chance to add in the BPLCON1 code today, will hopefully do that tomorrow - unless I get distracted by Dune II (EAB/Lemon Amiga competition)

Originally Posted by TheDarkCoder

Shoudn't this thread go into the Coding Tutorial sub-forum ?

Yes, I was a douche and started it in the wrong forum

Originally Posted by pmc

You're right, the pointers should point at even addresses (word aligned).

Maybe I got it wrong and I did it every 2 bytes. I'll check tomorrow.

Originally Posted by deicidal

Just to add I am going to make an educated guess that it will be in one of the first 5 issues

I will check them first, thanks.


Added, 24th Jan 2011:

TheDarkCoder and pmc,

I just checked my code, and you guys were right. I was sticking to even addresses - I got it wrong, despite it being my own code!!
I'm still struggling with using BPLCON1 though - I can still only shift bitplane 0, instead of planes 0-4. I will read up on the guides mention in previous replies.


Added later on 24th Jan:

Originally Posted by frank_b

You need to set the odd and even delay to the same value in bplcon1
From memory a delay of $ff means no scroll and $00 means scroll playfield left by 15 pixels. You're only setting one of the values.

Urgh! How did I miss this the first time? Nevermind. I have got the scrolling working properly now


Added on 22nd March:

Originally Posted by Codetapper

Try here: http://bitworld.bitfellas.org/demo.php?id=14886

Thanks for the link. I have just started reading through the issues
The examples in issue 2 don't seem to work. I have managed to get the examples from issues 1 and 3 to work though.
Already learnt quite a bit, now to read the other issues


Regards,
Lonewolf10