Can anyone try on a TRUE msx1 on a TV this simple program and tell me if he can see something strange?

10 screen 2
20 for c1=0 to 15
30 for c2=0 to 15
40 x = c1*8: y=c2*12
50 for yy=y to y+11 step 2
60 line (x,yy)-(x+7,yy),c1
70 line (x,yy+1)-(x+7,yy+1),c2
80 next 
90  next 
100 next
110 goto 110

the trick should be that there are some square off the diagonal that look like they have a uniform color
(a NEW color) different from the standard VDP colors.

Let me know if it works and with which colors the trick works

I remember that mixing some green and magenta gives the best results

Hm, dithering may give you a better result that straight alternating lines:

01010101            00000000
01010101            11111111
01010101            00000000
01010101            11111111
10101010 instead of 00000000 
01010101            11111111
01010101            00000000
10101010            11111111

Just 2 cts

But you loose the foreground color to play with...

I do not mean dithering !
I mean a special effect due to the analog circutry of the MSX1
The resulting colors should be uniform except for the borders

A possible explanation of the effect could be that the color signal is composed of 3 components (Y,U,V)

The effect could be due to the facte that the PAL systems do the averaging of V color component of two successive lines, so, when you change the color in two successive scan lines the HW system does an average.

In any case, before any explanation, I would like to know if someone has seen such an effect.
I have seen this on my HB75P in 1985 and it is 20 years since I do not turn on an MSX1 so maybe I could be wrong.

I start remembering that at that time I was using not only a PAL TV but also an Amiga monitor...
In this case my explanation of the color effect sounds a bit strange as probably I was using an RGB connector (?!?!?!?!?)

BTW I am almost sure this effect exists but I cannot say really how to reproduce it nor I can do real tests due to the lack of HW.
Thanks to anyone will do the test on its MSX1 and a PAL TV (or a RGB analog monitir (?))

Let me know if my monitor had an hw problem and i was only thinking to have found a trick

Hard to say and very subjetive; but here's some info:

Taking this coordinate system...

A B C ... O P
01
02
03
.
.
15
16

Only perfect/near perfect shades (at least to my eyes):

1E : False color! Due to 0+4 looks like 4
3D : Light green
3H : Light blue
4C: Light green
4H: Light blue (near perfect shade)
4O: Light gray (near perfect shade)
5A : Same as 1E
7I : Dark pale red
7N : Pale purple
8C : Light cyan
8D : Lighter cyan
11D : Yellowish light green
11L : Light yellow
11O : Light gray
12D : Light green (near perfect shade)
12K : Yellowish green
12O : Light gray
13C : Medium green
14G : Pale medium purple
15D : Pale turquoise(?)
15H : Lighter cyan
15K : Yellowish gray
15L : Light yellowish gray

The TV used is a Grundig 21" PAL unit.

Hope this helps!

Cool!, the MSX1 tested is a Sony HB75P too...

Oops, and I forgot to say that I'm using a euroconnector (RGB).

check The Cure... it also has 'new colors' by exchanging 2 colors.

@jltursan
thanks a lot for testing this!
My amiga monitor had an RGB out, so, any explanation?

I want to try with a TFT unit. Maybe the effect could be still visible...
So, it's an optical effect due to the mixing of the luminosity of the two colors? (I've not checked every pair of colors; but seems that both always are bright ones, or at least one of them), or is really a tube artifact?.

Anyone knows about the curious graphic modes obtained with ANTIC, the graphic chip of ATARI series?, there's an incredible amount of color and resolution artifacts there

I'll give a look at the cure; but where exactly are they exchanging 2 colors?. I can't remember that...

So, have anyone else done the test ?

Does it happen only on HB75P and RGB connectors?

Does any emulator reflect this behaviour ?

And more important, does anyone have an explanation?

Maybe this demo has larger areas and the effect is more visible

10 screen 2
20 for c1=0 to 15
30 for c2=0 to 15
40 x = c1*16: y=c2*12
50 for yy=y to y+11 step 2
60 line (x,yy)-(x+15,yy),c1
70 line (x,yy+1)-(x+15,yy+1),c2
80 next 
90  next 
100 next
110 goto 110

Blur

On my vg8020 its possible to see the lines on all the different color combinations. Some combinations seems to blur it a bit more and some less but in all cases its possible to see the lines of the two colors.