Chroma delay on Laserdisc?

[metaleonid]

Quote:

Originally Posted by lordsmurf

I'm not sure what you mean, nor what I'm looking at.

If you look at the post #8, you will see 4 attachments. Open NoDelay.png and AvionicsDelay99.png. Look at the bottom center-right. You will see the square number 1.5. On NoDelay.png, there is clearly crippled chroma. No color where cyan is supposed to be. But on AvionicsDelay99.png, the chroma is in full resolution, thus the bars are cyan. In other words red bars alternate with cyan bars. But when no delay, cyan bars are shifted such that they overlap with red bars.

[metaleonid]

Hello, I'd like to resume the thread. After a thorough investigation, I noticed that if I apply chroma delay, the red colors got corrected but the rest of the colors would be shifted in wrong direction. I tried to split chroma into 2 channels (U and V) and noticed that the U channel is in place. While the V channel is the one which needs to be delayed. The question I have is how to achieve it? This needs to be done before I comb chroma and luma and feed it into motion adaptive 3D comb filter. Are there any hardware processors that can delay chroma channels rather than the entire chroma?

If I split the signal into RGB component, delay R and comb into composite, will this work? Please advise.

[themaster1]

Perhaps you need to work in RGB such as:

Quote:

avisource()
ConverttoRGB32(matrix="rec601",interlaced=true)
ShiftRedBlue(rx=0, rxd=15)##################################
### shift and/or resize Red and Blue relative to Green
##
## @ rx, ry, bx, by - "shift" red & blue position
## @ rxd, ryd, bxd, byd - "delta" width & height
##
function ShiftRedBlue(clip C,
\ float "rx", float "ry",
\ float "bx", float "by",
\ float "rxd", float "ryd",
\ float "bxd", float "byd")
{
Assert(C.IsRGB,
\ "ShiftRedBlue: source must be RGB")
rx = Float(Default(rx, 0))
ry = Float(Default(ry, 0))
bx = Float(Default(bx, 0))
by = Float(Default(by, 0))
rxd = Float(Default(rxd, 0))
ryd = Float(Default(ryd, 0))
bxd = Float(Default(bxd, 0))
byd = Float(Default(byd, 0))
C
return (C.IsRGB24)
\ ? MergeRGB(
\ ShowRed.Spline64Resize(Width, Height, -rx, -ry, Width-rxd, Height-ryd)
\ , ShowGreen
\ , ShowBlue.Spline64Resize(Width, Height, -bx, -by, Width-bxd, Height-byd)
\ )
\ : MergeARGB(
\ ShowAlpha
\ , ShowRed.Spline64Resize(Width, Height, -rx, -ry, Width-rxd, Height-ryd)
\ , ShowGreen
\ , ShowBlue.Spline64Resize(Width, Height, -bx, -by, Width-bxd, Height-byd)
\ )
}

[latreche34]

OP, Maybe you should address the root cause of the problem rather than patching the problem, First why are you capturing via S-Video, LD is composite so capture it from the composite output, If there is still a shift maybe it's a defective player or just that particular disc. If it's the player, try to get a different LD player, If it's the disc, see if you can get a copy on DVD or Blu-ray and call it a day.

[metaleonid]

Quote:

Originally Posted by themaster1

Perhaps you need to work in RGB such as:

Hi themaster1,

The thing is that I need to do this in hardware before I comb the signal into composite and feed it to the capture card. Unfortunately AVISynth TComb filter is not nearly as good as hardware motion adaptive 3D comb filter.

Ideally, I could've made a capture twice: one capture with delay and the other one without delay and then comb them with AVISynth splitting chroma into U and V. But given the number of LDs I have it is very time consuming. So the solution is either find a piece of hardware that does it or just get a different LD player (which I am considering).

Here's another question though. Does each of 3 RGB channels carry luma information? If I delay one of the channel, would it screw up the whole picture?

[metaleonid]

Quote:

Originally Posted by latreche34

OP, Maybe you should address the root cause of the problem rather than patching the problem, First why are you capturing via S-Video, LD is composite so capture it from the composite output, If there is still a shift maybe it's a defective player or just that particular disc. If it's the player, try to get a different LD player, If it's the disc, see if you can get a copy on DVD or Blu-ray and call it a day.

Yes, I'm aware that the signal is composite. However, in order to delay chroma channel, I split the signal into luma and chroma right out of S-Video out and delay chroma.

It turns out that all line of LD-S2 players are like that. I was told about the issue by Anthony Cuozzo who has the same player. But I investigated further and found out that only V channel of chroma is off.

The discs are not defective. My other player Pioneer CLD-D703 doesn't exhibit the delay. However, the other player is noisy (grainy) compared to LD-S2 (which is considered 2nd best player). I am capturing the discs that are not available on DVDs/BDs. If they were, I probably wouldn't even bother.

As far as getting a different player - it maybe an option. But the one I'm seeing on eBay HLD-X0 goes for $3000. Not sure it is worth it. Also not sure if this player exhibits similar problems.

PS. In fact a lot of people in the past have been pointing out that the picture on LD-S2 is soft while less noisy. When I delayed chroma, I get a lot of sharper image.

[latreche34]

Grainy doesn't always mean a bad thing, Most of the time it means that you are getting a sharper not washed out picture, If I was in your shoes I would take the grainy source and apply a light DNR filter in post digitally than fiddling with chroma in the analog domain, Big difference.

[metaleonid]

Quote:

Originally Posted by latreche34

Grainy doesn't always mean a bad thing, Most of the time it means that you are getting a sharper not washed out picture, If I was in your shoes I would take the grainy source and apply a light DNR filter in post digitally than fiddling with chroma in the analog domain, Big difference.

This is possible too. I'm also trying to investigate if the less noisy but softer picture from LD-S2 is simply the result of chroma channel being not in sync.

BTW, care for screenshots later in the day?

[latreche34]

Sure, But a simple video from both players would be preferable.

[metaleonid]

Hello, here're the test captures from Video Essentials LaserDisc. It has bars and Snell&Wilcox. I used 2 players:
1. LD-S2 with no chroma delay and with chroma delay 99 nanoseconds
2. CLD-D703

Later on I intend to post 2 more samples of the real capture (not the test LaserDisc).

[metaleonid]

Quote:

Originally Posted by latreche34

Sure, But a simple video from both players would be preferable.

What's your opinion on the sample video clips that I have posted so far?

I have exercised many options up to purchasing S-Video to Component converter and Component to Composite converter. Unfortunately putting them in the workflow significantly degrades quality.

I'm left with 3 options
1. Use just S-Video.
2. Not using LD-S2 at all and stick with CLD-D703
3. Make 2 captures: with delay and without delay and then combine U and V channels in AVISynth.

[latreche34]

I didn't see anything wrong with Pioneer CLD-D703 samples, I would use it for capturing using composite out, In my opinion you are looking for a solution to a problem that doesn't exist. You should focus your effort in getting a capture device with the best comb filter, that can make a lot of difference.

[metaleonid]

Quote:

Originally Posted by latreche34

I didn't see anything wrong with Pioneer CLD-D703 samples, I would use it for capturing using composite out, In my opinion you are looking for a solution to a problem that doesn't exist. You should focus your effort in getting a capture device with the best comb filter, that can make a lot of difference.

The only problem with CLD-D703 is that the footage is noisier (has more grain) than that of LD-S2. But yes, it is a lot sharper and no colors are lost.

I have 2 capture devices with best motion adaptive 3D comb filters - so this part is not an issue.

[latreche34]

From the samples I don't see a problem with CLD-D703, The problem you are trying to solve is created by the LD-S2, But this is my personal opinion.

[metaleonid]

Quote:

Originally Posted by latreche34

From the samples I don't see a problem with CLD-D703, The problem you are trying to solve is created by the LD-S2, But this is my personal opinion.

Thanks for the feedback. As you see I seriously do need a second opinion to decide which equipment to use.

PS. By the way as far as motion adaptive comb filter goes, I did previously capture some of the LDs using CLD-D703. I was doing 2 captures simultaneously: one via S-Video (no 3D filter) using https://www.diamondmm.com/product/diamond-ati-theater-hd-750-usb-tv-tuner/ and the other via composite using https://www.diamondmm.com/product/diamond-ati-theater-750-pcie-hd-tv-tuner-card/ (with 3d filter on). Honestly, I didn't notice any difference. The only difference there was during rare moment of static displays. That's it. During motions 3D filter acts like 2D one. That PCIe card has one of the best 3D comb filters BTW.