Removing black shadow to right of white text?

[HomeStarRunnerTron]

Here's the same VHS played in a JVC HR-S3800 (+ Sony DSR-PD170 TBC passthrough), a Funai SV2000 WV20V6 (+ Sony DSR-PD170 TBC passthrough) a Panasonic DMR-ES40V:

JVC HR-S3800, S-Video:

JVC_HR-S3800.png

Funai SV2000 WV20V6, Composite:

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

Panasonic DMR-ES40V, S-Video:

Funai_SV2000_WV20V6.png

I've also attached interlaced video copies of both sources. I really love the way JVC HR-S3800 looks, except for that terrible black shadow that forms to the right of the white text. Even when viewed without a TBC, it still looks this way... and you can see that the TBC I have isn't causing a black shadow on the Funai. The shadow also appears when using composite cables instead of S-Video. So... it's something about the JVC HR-S3800, I guess.

The Panasonic DMR-ES40V is nice except for the slightly more defined halos (and the fact that it's cutting off more of the picture, but that's not super-important)... but really, after everything I went through to get the JVC, I'd rather get an image out of that. What can be done to get rid of that smear-y shadow? Thank you so much in advance! So grateful for this incredible community.

[vwestlife]

Are you sure that's not a deliberate drop-shadow effect on the chyron? It may have been done intentionally to make the text more readable on a fuzzy CRT TV.

[un_kbron_del_664]

Your JVC HR-S3800 seems to have the correct white level (whites aren't clipped), whereas the white levels on the Funai and the Panasonic exhibit illegal values (whites are clipped). So, when you bring the white level above 100 IRE, you start loosing details (it's like burning the image), and that's why you don't see that shadow effect on the Funai and Panasonic.

Bellow, I have attached screenshots of the waveform monitor for the JVC, Funai and Panasonic.

However, given that we're dealing with VHS, this format has distortion, and said distortion can exceed 100 IRE. Therefore, I cannot guarantee which values are the correct ones. SMPTE color bars are used to calibrate video levels via a ProcAmp; that way, you can ensure that you obtain the correct levels. But I don't think your tape has color bars.

By the way, why are you capturing to DV25? DV25 uses 4:1:1 chroma subsampling (in PAL land, it uses 4:2:0). The ideal chroma subsampling is 4:2:2, as it captures color information at full resolution.

And there are many things that need to be corrected, but I'll let the community share their opinions.

[aramkolt]

I've wondered about this concept before myself. It's generally noticeable in movie credits where you'll get something like this to the right of bright white text even on a black or nearly black background. My theory is that some VCRs have a sort of "rebound" inductance in the opposite direction when the high magnetic signal of white drops off abruptly as it scans left to right. Essentially after getting such a high signal and dropping off on the sharp edge text, you get almost a "sub black" or just generally darker to the right. Could be that the other machines are just better at compensating for it in their preamps or other processing circuitry. Would be interesting to know if this happens with VHS_decode on one machine vs another where the stock circuitry causes the shadow vs a machine that does not. If VHS_decode also sees a difference, then it's down to either the characteristics of the heads (possibly wear vs differences in inductance) or the preamp circuit. If both machines look the same on decode, then it's down to how those machines process the RF differently I'd say.

All theory of course, I'm definitely not an electrical engineer.

[HomeStarRunnerTron]

Quote:

Originally Posted by vwestlife

Are you sure that's not a deliberate drop-shadow effect on the chyron? It may have been done intentionally to make the text more readable on a fuzzy CRT TV.

It has a FIRST drop-shadow, but there's that extra grayer drop-shadow ghost against the original drop-shadow-- that's what I'm trying to fix. Here's how the DVD of this same program looks, which is what I'm trying to get as close as possible to.

DVD.png

Quote:

Originally Posted by un_kbron_del_664

Your JVC HR-S3800 seems to have the correct white level (whites aren't clipped), whereas the white levels on the Funai and the Panasonic exhibit illegal values (whites are clipped). So, when you bring the white level above 100 IRE, you start loosing details (it's like burning the image), and that's why you don't see that shadow effect on the Funai and Panasonic.

That's really useful info, thank you. But man, the result of legal values = smeary shadow on all white text? That stinks.

Quote:

Originally Posted by un_kbron_del_664

By the way, why are you capturing to DV25? DV25 uses 4:1:1 chroma subsampling (in PAL land, it uses 4:2:0). The ideal chroma subsampling is 4:2:2, as it captures color information at full resolution.

If I capture my Sony camcorder through modern-day Final Cut Pro, it comes out as DV25 4:1:1 automatically. I'll see if there's another program to give 4:2:2 a try, thank you for the suggestion.

Quote:

Originally Posted by aramkolt

[...] My theory is that some VCRs have a sort of "rebound" inductance in the opposite direction when the high magnetic signal of white drops off abruptly as it scans left to right. Essentially after getting such a high signal and dropping off on the sharp edge text, you get almost a "sub black" or just generally darker to the right. Could be that the other machines are just better at compensating for it in their preamps or other processing circuitry. Would be interesting to know if this happens with VHS_decode[...]

I've felt like I was going insane looking at these streaks -- thank you for letting me know someone else has noticed this happening and has wondered how to fix it. Does that mean that the same thing happens on all of the VCRs you currently use, aramkolt? Same question to anybody else who happens upon this post. Wishing you all the best, super-appreciate your Time.

[lordsmurf]

Two concepts here:
- luma ghosting
- the effects of illegal blacks/whites

[HomeStarRunnerTron]

Quote:

Originally Posted by lordsmurf

Two concepts here:
- luma ghosting
- the effects of illegal blacks/whites

Which one is creating the double-shadow-- is it both? Is what I'm experiencing called luma ghosting, or is it something else? It sounds like most VCRs create this effect to the right of white text. For example, if I were to buy a VCR you fixed up on the marketplace, would it have this effect to the right of white text?

[lordsmurf]

Quote:

Originally Posted by HomeStarRunnerTron

Which one is creating the double-shadow-- is it both?

Both.

Quote:

Is what I'm experiencing called luma ghosting, or is it something else? It sounds like most VCRs create this effect to the right of white text. For example, if I were to buy a VCR you fixed up on the marketplace, would it have this effect to the right of white text?

This is one of the few areas where I had hoped vhs-decode would have done better. But no, it does not. As usual, that method is highly inconsistent, probably worse 50% of the time, and only slightly better the rest of the time. I think it heavily depends in the tape, and the tape signal, not entirely the fault of the VCR.

[Aya_Rei]

Guess it'd be like head switching noise where decode is dependent on the source VCR.

Anyway JVCs tend to have this smearing moreso than a Panasonic 1980 but it really depends on the tape.

You could probably reduce it by using Avisynth and edgesidebleed along with adjusting the levels so that it gets cut off.

This appears the most when you capture the blacks to be a black-ish grey in order to prevent any crushed blacks which results in loss detail.