s-video mod, slight shadows, is there a fix?

[dblondin]

Hi, first post! Ok I recently completed a 2600 s-video mod using electronic the electronic sentimentalities mod kit. All went smoothly. However, there one thing that seems slightly off. The bottom image shows the problem. There is a slight shadow on everything. This is on a Commodore 1702 with chroma and luma from the modded 2600. Composite does the same thing, although it is worse. I used shielded RCA cable for the 3 video signals. There are 75 ohm resistors between ground and composite and luma out and ground. Is there a way to get rid of this shadow effect? It's not a huge deal but I was expecting a sharper image.

 

Edit: Actually, I'm seeing this with other devices plugged into the 1702 as well. So it's the monitor! I'm still clueless as to how I might fix it but looking into it. If you have any hints... It appears this guy had the exact same problem but no solution was posted.

 

http://atariage.com/forums/topic/170884-commodore-1702-ghosting/

 

 

Edited January 18, 2015 by dblondin

[dblondin]

I tried a couple of other things:

 

different cable: no difference

different gaming console: no difference

different monitor (Sharp LC20B6U-S LCD display): no difference. wait, what?

 

I'm seeing this shadow with entirely different devices, even an LCD monitor. Is this just a normal artifact of analog video signals? Do I need some special type of cable? It's true I can get it to go away with some games, certain color combinations and shapes. But generally, mostly, the ghosting effect is always there.

 

I have no idea what is going on at this point!

Edited January 18, 2015 by dblondin

[Mef]

Hi There,

 

Not to worry, there will always be some slight shadowing with S-Video. It's all down to combination of colors. Some transitions will give you almost RGB-like quality look, like those details here.:

 

 

 

Some will produce a strong shadow/light banding:

 

 

 

Note how white and pink doesn't give eny banding here, but the brown-on-blue does:

 

 

 

Some programmers use certain color combinations to purposedly trigger this artifact, to cheaply simulate "3d shadow" of maze walls and the like.

 

 

As for this particular game, here's how it looks on my setup:

 

 

 

I dunno how much are your picture's colors true to what is showing on the monitor, but it looks as if you've got much more saturation. First off, I'd remove the 75 ohm resistors from the cables, these aren't needed at all for Commodore monitors, as the signals on Commodore and Atari 8-bit stuff weren't up to spec with 75 ohm impendance anyway. Then I'd use the trimpots on the mod board to bring the color levels down a bit, it's the swings in Chroma that produce this shadowing*, so with weaker signal I guess that should be less noticable. Can't really say much more without looking at the schematics of this particular mod you're using and comparing it to mine, but it looks like you've got an awfull lot of capacitors in there, which migh introduce needless delays in the signal, making the artifacting more harsh. Might just as well be that PAL is less susceptible to this behaviour (because of its nature), so I wouldn't worry about comparing your picture to mine.

 

*Try running some games with Color/Black&White switch in B&W position, or simply detach the Chroma cable - you'll see no banding/shadows. It's also a good way to check the monitor's convergence - there shouldn't be a tiniest hint of color showing in B&W picture.

Edited January 18, 2015 by Mef

[Mef]

P.S. Couldn't edit the previous post anymore...

I'm using a later 1084S model (and darkened the pics a bit as they were taken with a phone camera and horrible light, making them grey and flat), so that might add to the differences as well. The main idea was to prove that it's just a matter of the colors' combinations and S-video will always give those artifacts, to some extent. It's just the way it is, nothing really wrong with your setup. I've also noticed that these effects are much less noticable on my C64, propably thanks to much higher horizontal resolution (faster color clock?) and limited palette.

Edited January 18, 2015 by Mef

[+SpiceWare]

Yep, the 3rd color between transitions is normal. A while back I posted some images about it in the discussion where Omegamatrix came up with the flicker-free14 digit score/lives display that I used in Space Rocks.

[dblondin]

Thanks guys. I had no idea this was normal. Well, maybe my problem is *somewhat* normal. I removed the 75 ohm resistors, and toned down the color saturation (making this adjustment on the 1702 produces better results than with the mod trimpots). It's not nearly as good as the photos Star Raider has shown, but it is improved. Yes, it goes away entirely when unplugging the chroma cable.

 

But. There is something endearing about this artifact and now that I have more information I guess I'm ok with it. One reason I'm still playing with these old devices is because I appreciate the crude imperfections.

[Mef]

Good to hear that. Btw, Star Raider is just a default forum status, not a nickname.

 

That'd be Mef.

[dblondin]

Ah, sorry Mef, hehe.

[Bryan]

You have to remember that color comes from a PLL circuit which determines the phase difference between two waves. Since the 2600 (and computers in general) generate very abrupt color changes, it takes some time for the color decoder to re-detect the phase and swing to the new position. If the phases are close, there will be less of an error band. I suppose a fully digital approach could extend the eventually decoded color back to the edge of the luminance transition.