Colour Memory why does it move?

[freshlamb]

I looked to see if this was covered elsewhere but the search engine for the forum isn't the greatest.

So I understand why memory has to move when expansion memory (greater than 8K) is inserted (so BASIC has a contiguous space). But why does colour memory have to move?

[Mike]

Hi, freshlamb,

I answered this question in another context: viewtopic.php?t=5572&hilit=changes+base ... s&start=21.

In short: the colour RAM is actually 1K nibbles in size. It is not really moved but its use is changed as it is addressed with the lower 10 bits of the corresponding text RAM address.

That 'extra' size also is what allows for overscan displays. Though with that use case it is advisible to put the screen RAM at an address divisible by 1024. Otherwise - when the screen RAM exceeds 512 bytes - you will have to put up with a wraparound of the colour RAM ($9600..$97FF -> $9400..).

Greetings,

Michael

[freshlamb]

Thanks Mike, I knew you would have the answer!

By the way I just hooked up my VFLI vic yesterday, it turns out the 8K (16K) chips just needed to be re-seated. So it works fine. I will have to post video on YouTube when I have time!

[malcontent]

Hi, freshlamb,

I answered this question in another context: viewtopic.php?t=5572&hilit=changes+base ... s&start=21.

In short: the colour RAM is actually 1K nibbles in size. It is not really moved but its use is changed as it is addressed with the lower 10 bits of the corresponding text RAM address.

That 'extra' size also is what allows for overscan displays. Though with that use case it is advisible to put the screen RAM at an address divisible by 1024. Otherwise - when the screen RAM exceeds 512 bytes - you will have to put up with a wraparound of the colour RAM ($9600..$97FF -> $9400..).

Greetings,

Michael

This sounds like it may be a memory/code saver somehow, unless I'm reading it wrong: Could you horizontally split a screen with a repeated color memory pattern?

[Mike]

This sounds like it may be a memory/code saver somehow, [...]

Unlikely. If anything, I would categorize that behaviour as mildly annoying, but understandable due to hardware reasons and with an easy workaround - just do not put screen RAM at an odd half KB boundary with an overscan screen. After all, what is the use of an addressing irregularity mid-screen and quite likely, even mid-screenline?

Could you horizontally split a screen with a repeated color memory pattern?

I do not quite understand what you are up to, but you could always try for yourself.