I have an LD-X1 which shows strange image artifacts/jagged edges in picture areas of high chrominance. It also has trouble with the horizontal frequency response of the Snell and Wilcox test pattern. It's not the comb filter I'm using (a pioneer dvr-920h) as my other players render the diagonal section perfectly clearly.

Any ideas what is causing this? Could it be a tracking/alignment problem? I've attached a picture showing the issue, lines become distorted at the ends and the lines are very blurry.

Looks like artifacts from Y/C recombination. Also there should be a grey bit between the vertical lines and the horizontal black lines.

It's strange. The s-video out looks better but then I don't benefit from 3d comb filtering. I'd heard the X1/S2 had very clean composite output so was surprised to find this. Could alignment be an issue or is it more likely something is failing electronically?

Nope, it's just how the player's designed.

Clean is a very ambiguous term. The X1/S2 put out good composite but it's not without it's issues. For example, the TBC cuts off part of the picture information making closed captioning impossible.

Oh really? So this is just how the player performs and no tweaking will improve it in this area?

It's a shame because luma noise is basically non existent on this player.

Yeah To avoid that you either have to go low-end (a composite output-only player) or later high-end (one with a 3D comb filter) - most Pioneer players used a 2D comb filter. I forget if the typical Sony 2D one is adaptive, which would make it less obvious.

Ahh OK,

So is this effect due to the 2D comb filter in the X1 separating the Y/C and then recombining the signal poorly? So the composite output is always affected by the comb filter?

This would mean that all non-3d comb filter players exhibit similar issues. This does make sense but I always thought the recombining process created the same composite output as the source signal.

It's the same if the comb filter is accurate, which a non-adaptive 2D filter isn't. It uses information from the lines above and below from the same field, blending them even when they shouldn't be.

On balance, when watching actual movies the S2 is still better because of low noise and minimal ghosting/ringing. I can still see some, but it's extremely hard to eliminate so I can't blame it. The V8000 for instance is quite similar to the S2 and has no comb filter... but a lot of echoing, so overall it's worse.

S-video output seems better as I get horrible tearing and artifacts when feeding composite through a 3d comb filter.

Understood. Cheers happycube!

Which output composite or S-Video looks better depends on your monitor. Composite usually wins.
You should also be using the BNC output as it bypasses the on-screen character generator which improves the picture. S-Video2 also bypasses the character generator.

I'm using the BNC composite out into a Pioneer DVR (which has an excellect comb filter) then outputting component to a samsung LED which has very good deinterlacing and upscaling.

I get a "fuzzy" picture when areas of high detail are present. It's hard to explain but the image becomes almost shifted and scewed on itself, not sharp. This is shown in the s&w image above. Although the s-video is just a rainbow interference mess on this diagonal pattern, it produces a smoother image on real world material.

Anyone have experience with bypassing the comb filter on the X1/S2? Sounds like it would improve the player quite a bit!

This would be really interesting but not sure how to go about this. I imagine it's not as simple as just soldering/cabling past the comb filter chip? That's even if it is an isolated chip with no other processing built into it.

It would be incredible if it was possible to bypass it.

Looking at the service manual, I don't know if it's actually feasible at all, I don't see any "turn all Y/C processing off" lines on the custom chips. Tapping the signal before the ADC would work, but then you need to provide your own TBC.

On the later players, there is something you can ground out/pullup on the Sony chips, then you need to bypass the YNR circuit, otherwise chroma would be completely clobbered.

What you may be seeing is the lack of overly digital processing in the S2/X1. Overly digital processing adds a visual sharpness used in all the newer players. As you go back to the unprocessed picture it will appear softer and will be different. My suggestion is stop looking at these test patterns so close and see if you enjoy watching actual movies. The S2/X1 is made for low background noise and more filmlike, it was not designed for sharpest picture.

I first noticed the issue on real world footage first, then went to the test pattern to try and track down exactly what I was seeing.

You're right though Kurtis, S-Video out produces some of the least noisy images I've ever seen from Laserdisc and it's a great experience watching movies on this player. I really like how each player has its own unique look!

Happycube, How would tapping the signal before the ADC work in practice? Would you just need solder a BNC cable to the board and then input straight into a TBC? If so it wouldn't be a problem as there are lots of ex-broadcast TBC's floating around.

You would have to make sure that you're tapping a point with the right voltage level, and then output it through a buffer circuit. If I remember right, the unloaded voltage should be ~2v p-p. Then you can feed it into an external TBC. You would miss out on all the CLV special functions (no still on pause or field-by-field advance, slow motion etc).

I'd argue the S2 is doing too much digital processing, since the comb filter is digital What I'm seeing there is definitely a comb filter artifact. It was designed before LD mastering got wider color bandwidth - for pre-WEA/SuperNTSC era disks it's probably superb. Unfortunately I don't see any way to turn it all off - there are both digital and analog NR stages afterwards and none of those would work without comb filtering.

The S2 circuitry is designed to avoid the digital look by high frequency (4fsc) capturing and not having to encode the sync signal (-40IRE) as part of the capture range, effectively giving it a 9-bit digital chain. All the later players have 4fsc capture, but many do not break out the much lower sync signal.

And yup, that's how you would tap it. Not all TBC circuits can deal with it, but there's gotta be something out there that can. The results are going to have a very different feel - a lot sharper (2D combs remove that) and at least a bit noisier. If you have stuff to deal with digital data (unlikely, alas, only Glimmie over on AVS has done this), the tap point would be between IC12 and IC13 on the VSMB assembly, which is right before the Y/C sep.

OTOH, it's far easier to just use the S-Video and enjoy watching movies! (especially the pre-92ish ones the X1/S2 is born to play )

Awesome! Sounds like I need to do some reading. I'm amazed at what's possible with this format.