32768 color dual Shifter ST mod - schematics lost?

[mrbombermillzy]

Yes, that's the spirit, with legacy STe bits being the Most Significant Bits though, meaning the final mix that I devised would be, with your notations :

Code: Select all

R_analog = 321043
G_analog = 321045
B_analog = 321043

Sorry, I don't quite follow.

I guess you have left in the 2 obsolete extra (18-bit) bits in the mix..(?)

So, do you mean (positionally and bit significantly; 5/4=MSB, 0=LSB):

Code: Select all

R_analog = -3--2104
G_analog = 3-2-----1045
B_analog = ---3--------2104

[sporniket]

Code: Select all

           GRGBRRRRGGGGBBBB
R_analog = -1--2543
G_analog = 0-1-----2543
B_analog = ---1--------2543

Now as you noticed, I did a trick to infer a 6th bit (LSB) for R and B that are coded as 5-bits values. In other words I have to scale R and B to get the same range of values than G.

The rules to follow for this 6th bit are :

• The 5bits value '0' must give a 6bits value '0', so that black is coded with RVB565 (0,0,0)
• The 5bits value '31' must give a 6bits value of '63', so that white is coded with RVB565 (31,63,31)
• The 6bits values inferred from the 5bits values are evenly spaced most of the time. In other words, most of the times the difference between 2 successives 6bits values is 2

That's how I ended up re-using the MSB of the 5bits values as the LSB of the 6bits values

Code: Select all

y = (x < 16) ? 2 * x : 2 * x + 1 ;

[mrbombermillzy]

So, in essence, if I'm seeing this right, you are using the 5 MSB of the R and B colour channels to match the intensity with G (sorry V in France! :lol: ).

Then instead of the 64 intensity levels per channel, you have 32 which match every 2nd intensity value step, EXCEPT for the highest and lowest value (0 and 31) which must be 'hardwired' to match the Green intensity values (0 and 63).

Is this correct?

[sporniket]

So, in essence, if I'm seeing this right, you are using the 5 MSB of the R and B colour channels to match the intensity with G (sorry V in France! :lol: ).

Correct

Then instead of the 64 intensity levels per channel, you have 32 which match every 2nd intensity value step, EXCEPT for the highest and lowest value (0 and 31) which must be 'hardwired' to match the Green intensity values (0 and 63).

Is this correct?

Almost. R/B values of 0..15 are matched with even values of green (0,2,4,..,30) ; R/B values of 16..31 are matched with odd values of green (33,35,37,...,63)

[mrbombermillzy]

Almost. R/B values of 0..15 are matched with even values of green (0,2,4,..,30) ; R/B values of 16..31 ar matched with odd values of green (33,35,37,...,63)

Got it! :thumbup:

Thanks for your patience.

[mrbombermillzy]

So looking at it...

the colour channels cannot be pulled and converted to planar directly from a (e.g.) 16bpp 565 bitmap image (except green) and must have nearest neighbour interpolation conversion done in high or low half (the step is more evident in the middle odd/even crossover which is 3 steps wide; 30<>33).

Not criticising, just working out how to reassemble the image when I code an image viewer! :D

Again, you've done great work here. Thank you.

[sporniket]

A reference BOM as a mouser project link : https://www.mouser.com/ProjectManager/P ... 4f4fb391a8

EDIT : the gerber

rgb565-experiment-v1.0.0-xprc2.zip

EDIT 2 : the github release with the schematics and the gerber

[sporniket]

Test pcb in progress...

2022-01-12--pcb-in-progress.png

[mrbombermillzy]

Looking good. Not long now. :)

Unless you have done so already, I'm going to try to make some time to write an image viewer to go with this for you.

As time is of the essence ATM, I can base it off of my TT 16 colour viewer, but will have to remove all the 030 bitfield code for the STFM. (I don't suppose ticking 'compile for 68000' in Devpac will quite do it! lol)

[sporniket]

For the record, I will only be able to check that the pcb fit inside the can. The STF I have is the donor machine, and I already have removed some key parts that make it unable to run. On the other end, my H5 is not ready to run either. Are you at ease with soldering ? I can send you one PCB when I receive them.