Problems with YUV video from VG-8020

Page 2/2
1 |

By hit9918

Prophet (2868)

hit9918's picture

11-06-2013, 10:01

@ace,
I heard of the issue from emulator software side, getting palette problems with the voltage values from TI manual,
and it is very interesting that a hardware maker too gets palette problems,
so the odd values seem to really come out of the chip.

Look at the voltage values for black listed in the TI manual:

Color   Y    R-Y  B-Y
1 Black 0.00 0.47 0.47

Looks like the chip does 0.47 volts too much on the U and V pins.
Comment in emulator code is: "I assumed the "zero" for R-Y and B-Y was 0.47V."
A guess: maybe connect the 0-volts-wires of U and V to 0.47 volts.

By HansO

Paladin (672)

HansO's picture

11-06-2013, 15:26

The Coleco Vision adapter designer states: and here

[i]The ColecoVision outputs Y, B-Y, and R-Y natively, but at different levels than standard YPbPr/component. When you bring the "B" and "R" levels down to where they need to be, the associated "-Y" levels also change. This needs to be compensated for, otherwise the colours, especially the yellows, will be muted. The "Y" compensation on the "B-Y" and "R-Y" lines is the essence of the design, which makes it different than the ones I've seen pictures of.

Anyway, the 3 required signals are Y (green cable) which is the luminance signal (and sync), Pb which is blue minus Y, and Pr which is red minus Y. With the blue too bright, I turned down the Pb (B-Y) signal to compensate. Unfortunately, this not only turned down the B of B-Y, but it also turned down the "-Y". By subtracting some Y from the B-Y signal, I was able to get a signal that worked much better. I did the same procedure to the R-Y signal, but the change required for that signal is minimal.
This explains that he had the same problem and solved it with the aforementioned 0.47V shifts. This means voltage shift, not amplification. When I look at the pcb, it looks like he used a LMH6644 to do the buffering and level shifting.

Also have a look at this http://spatula-city.org/~im14u2c/vdp-99xx/e3/SPPA004A_9928-2... whre DC restoration is discussed.

Hans

By mohai

Paladin (839)

mohai's picture

18-03-2018, 17:00

Hello,
Did anybody get success in getting correct component output?

I have an 8020 and would like to get component (YUV) output from it.

By mohai

Paladin (839)

mohai's picture

25-05-2018, 18:44

Hello,

I am investigating on how to get real component signals from an VDP chip (TMS9928a, TMS9929a).
I read several other forums and it seems somebody found the solution.
It is based on TI documents here: http://spatula-city.org/~im14u2c/vdp-99xx/e3/SPPA004A_9928-29_9118-28_Interface_to_Monitors.pdf
And this project here: https://hackaday.io/project/13056-tms9929a-rgb-and-component-adapter

The main problem is that VDP delivers component output but slightly modified in order to make to composite encoding or RGB easier .

In this forum (read from page 6 is the interesting part) there is a solution: https://www.classic-computers.org.nz/forums/viewtopic.php?f=11&t=960&sid=d54e66f22b0b769dc6ed50d60b0f30d6&start=75&mobile=on

I want to do something more simple than those projects.
I already chose the components and designed a circuit I will build.
I will let you know when it is ready and how it works.

Page 2/2
1 |