| |
ws
Registered: Apr 2012 Posts: 229 |
4x4 FLI questions..
So, Onslaught Design uses a gfx mode which I have never consciously observed before, and i was amazed that it actually seems to be all characters, no bitmap, no sprites.
So, do i get this right: like this you can get any of 16 colours next to eachother in a 4x4 pix matrix with only 2x 1k (minus the fli bug gap) ? Or are there neighbouring restrictions, too?
Question2: is there a crossplatform converter for this already, like PNG or BMP to 4x4 Charmode FLI?
Question3: can you point me at a plasma that uses this mode?
Thanks! |
|
... 13 posts hidden. Click here to view all posts.... |
| |
chatGPZ
Registered: Dec 2001 Posts: 11147 |
Stupid terms in the scene? No way :) |
| |
DKT
Registered: Aug 2004 Posts: 96 |
zoom the pixels in any gfx editor 4 times and you will figure out why it's called zoom4 ;) |
| |
Krill
Registered: Apr 2002 Posts: 2854 |
Quoting wsSo, So, do i get this right: like this you can get any of 16 colours next to eachother in a 4x4 pix matrix with only 2x 1k (minus the fli bug gap) ? Or are there neighbouring restrictions, too? Quoting wsOh no. I just understood it actually DOES use bitmap. My bad. Nevermind! Yes, bitmap mode is required to set two independent colours per 8 pixels wide tile.
To minimise memory usage, you could use the repeated-badline technique rather than extra DMA by FLI.
This would limit the bitmap portion to the first 320 bitmap bytes, but spread out the colour map all over RAM in chunks of 40 bytes.
Interrupt overhead is higher (need an interrupt every 4 rather than 8 rasterlines), but there is no FLI bug and thus full width. |
| |
Oswald
Registered: Apr 2002 Posts: 5028 |
Recently I was wondering if repeating first line of bitmap 4x4 is possible, thanks for the answer :) hadnt had a use case yet tho. |
| |
chatGPZ
Registered: Dec 2001 Posts: 11147 |
The way it scatters the VRAM over memory is a bit... annoying :) |
| |
HCL
Registered: Feb 2003 Posts: 717 |
Yes, the memory usage is very fragmented.. but the interrupt-overhead is probably *not* higher, like Krill suggests, perhaps even lower.. Since there is no FLI-bug involved, the interrupt is not timing critical, and you can also trigger two badlines within the same interrupt.. |
| |
Copyfault
Registered: Dec 2001 Posts: 467 |
Well, initially I wanted to post something very close to what Krill wrote, but then I decided to refrain from it, mainly due to that awful memory scattering already mentioned.
And cheering in to HCL's post: the approach even allows for getting along without any IRQ at all - code interleaving is your friend ;) |
| |
chatGPZ
Registered: Dec 2001 Posts: 11147 |
and that's where "annoying" turns into "plain ugly" :) |
| |
ws
Registered: Apr 2012 Posts: 229 |
that reminds me of a question in another thread, if there would be any desirable additional feature for the VICII, and yes, that would def be a switchable/banked color ram. but since it is hard-wired, that will remain an eternal desire. |
| |
ChristopherJam
Registered: Aug 2004 Posts: 1381 |
<c128> Heyyyy, I've got two color RAMs! |
Previous - 1 | 2 | 3 - Next |