So...
After some real life entanglements and some scary issues with nasty 'chocolate pin' DIP risers (!) we finally continue the journey.
As a recap, the RGB line resistors have been bypassed to boost colour output.
Going back to the test program I made which ONLY displays the (unfortunately LSB so darkest colour bits) top 4 of the 16 bits of the palette colour data, which neither the standard ST, or the Ste can use, to show some of the new colours, we are going to look at colour differences before/after.
So what appeared previously as:
- HICONST.JPG (105.76 KiB) Viewed 1899 times
Now shows a better colour representation overall:
- P1040263.JPG (72.74 KiB) Viewed 1899 times
As can be seen, colours have a better definition overall, with less vague murkiness but the 'chroma/luma' vertical lines are still there. Note that contrast/brightness have been turned up (as can be seen from the background brightness) but remember the same was done (above) originally too.
In a normally lit room, we can now see the colours better too:
- P1040264.JPG (90.7 KiB) Viewed 1899 times
Now onto the greyscale test images supplied by
@sporniket designed to show the STe palette shades and any differences with regards to the regular ST palette shades.
The first one, I will show another before/after set.
This was before:
- P1040237.JPG (131.2 KiB) Viewed 1899 times
And now:
- P1040272.JPG (100.33 KiB) Viewed 1899 times
I guess the levels are harder to differentiate when all 3 RGB channels are involved and at higher values, but I believe there is slightly more of a clearer channel value definition on the later image.
Now, all seems very good so far. It's just this last image that puzzles me. Apparently it has the STe colours in columns below the white dots:
- P1040270.JPG (190.51 KiB) Viewed 1899 times
But for the life of me, I cannot see anything below the dots. Here's an even closer image that even shows the dots of the electron guns:
- P1040269.JPG (247.34 KiB) Viewed 1899 times
...but still no visible greys.
You may be seeing a slight 'something' there, but unfortunately I BELIEVE this is a screen display error due to the 'chroma/luma' bars. Perhaps changing to a 15Khz capable TFT screen may eliminate some of the more subtle image problems and provide further enlightenment?
While I'm at it, I must apologise for the sometimes ropey image quality. Sometimes the best colour separation images were taken with the screen scan halfway down the screen, sometimes I got a bad image with a perfectly off screen scanline. I presented here the best of a non perfect bunch.
This is what you get when crossing an iffy camera with a 30+ year old CRT!
Anyhow, it looks to be a generally positive result so far, apart from the puzzling last test image, which (it looked better with my eyes than it did in the 2 images above, but STILL I couldn't see anything) I'm not sure what to make of yet.
The other issue seems to be the C64c/128 style screen bars (as
@stephen_usher seems to be suffering from). As far as I'm aware, on the C= machines, these are caused at least partly by memory accesses, so it will be interesting to see how to tackle these at a later point.
But for now, I think generally a thumbs up for progress made.