Changing of RGB

With the change of how RGB is defined with the Ti – is there anyway of testing software – to see if they display any problems – using pre Ti machines?

Give the software a RISC OS 5 format sprite in (a) BGR and (b) RGB colour order, and see what comes out. There are not many colour processing apps that are fully compatible yet. Remember – if the app is simply displaying a sprite using official OS calls, then it is down to the version of the OS as to what happens.

The other thing to note is that the IGEPv5 board is the same as the Ti – both have R/B swapped.

64K colour modes on a Pi also uses RGB modes!

Which version of RO5 should I be using?
With the 64K colour mode on the Pi – are the Desktop colours shown correctly- and if the program uses the OS SWIs?

Could the emulators be made to reflect this change – so software could be tested?

Which version of RO5 should I be using?

Whichever one has the 64K option available :-)

I’m not sure offhand whether it’s available in RC14.

With the 64K colour mode on the Pi – are the Desktop colours shown correctly

Yes

and if the program uses the OS SWIs?

Yes, if it uses the correct OS SWIs.

Part of the problem is that the original OS_SpriteOp calls (e.g. 28) don’t perform any colour translation – they just copy bytes directly to the screen, without producing any error if the formats don’t match. It’s only the calls offered by SpriteExtend (e.g. 52) which will correctly deal with paletted sprites, different colour depths, RGB orders, etc.

OMAP3 and OMAP4 machines are also capable of using the new screen modes if you’re using recent-ish development ROMs (support went in about a year ago). If you go into the display manager and add “LTRGB” to the mode string then that will switch to a red-blue swapped mode.

If you go into the display manager and add “LTRGB” to the mode string then that will switch to a red-blue swapped mode.

Well I never. Knowing that would have been useful in the past when testing. Does the inverse work on a IGEPv5 or Ti?

Does the inverse work on a IGEPv5 or Ti?

Nope. If they supported the old RGB order then they’d be using that by default and this would effectively be a non-issue.

For completeness, if you have 5.22 on an Iyonix then you might find that some modes are red-blue swapped there (depending on graphics card and *Configure NVidia setting). But it’s probably a poor choice of machine to use for testing because the hardware generally only supports one red/blue order for each mode – so if you configure it for the wrong order you’ll end up with the wrong colours on screen and will have to look for things that are “right” to detect buggy software instead of looking for things that are “wrong”! (*Configure NVidia exists because RISC OS can’t always tell what order the card supports)

I was looking to see if programs like !TextEase and DAPicture would work on the new machines?