Luminance Problem in VHS capture

[cwk]

Howdy,

I have a 2nd generation VHS tape that looks fairly crisp on my television--for a 2nd generation VHS tape.;) When I attempt to capture this video, it displays two problems:

1. The capture video is MUCH brighter than when viewed on television, even when I burn the resulting capture to DVD. In fact, the brighness is bleaching much of the color and detail. I can adjust the brightness in VirtualVCR, but the adjustment appears to occur after the bleaching has occurred. So although the video is noticeably dimmer, the detail is still lost.

2. In stretches, every other field is distorted and shifted 40-50 pixels to the left. Again, these 'corrupted fields' do not appear when viewing the videotape on my television played through the same VCR I am using to capture.

Here is the odd part. Problem #2 only occurs when I am using Component or S-Video output from my VCR. Using a coax cable--which appears to have poor quality otherwise compared to the other two--the fields are more stable, although the brightness problem still exists.

My setup:
VirtualVCR 2.6.9 build 6252
Leadtek WinFast TV2000XP capture card with current Leadtek drivers
WinXP SP2
p4 2.4 Ghz
512Mb Ram
2 - 120 GB drives
TMPG

I have been using this capture card for a few months and have not seen this issue before.

Has anyone seen this before? Any hints?

Regards,
cwk

[Arachnotron]

My guess is that the horizontal sync is degraded on the tape. This would affect a capture device much more than it would a TV set. Some cards have more than one way to deal with this (philips based capcards have a VCR setting you can enable), but you may need a TBC to really deal with it.

The reason for the brightness problem is that capcards in general "calibrate" the brightness by measuring the difference between the lowest part of the horizontal sync pulse and the horizontal blanking following it. (The original macrovision used to work by polluting the horizontal blanking with spikes, thus making exact brightness calibration impossible. Again, VCR's suffered by it, while TV set's were not affected)

Mind, a TBC will digitize the original video, replace the corrupted sync's with fresh ones and spit it out again as an analog signal. So it will help with #2 but I doubt if it would help you much with #1, unless the brightness shift is constant and you have a TBC model that would let you adjust brightness too.

[edit]
See the capture guide, and especially figure 3 in chapter 9.1 to see a picture of the video signal.
http://www.doom9.org/capture/analogue_video.html

The capcard uses the difference in signalstrength between "sync tip" and "blanking level" to calibrate the brightness. If the shape of Hsync is distorted, as happens when you copy tapes without using a TBC, the brightness level is distorted too. TV sets simply pass on the signal without trying to calibrate, so as long as there is some sort of Hsync in the signal they mostly do ok.

[cwk]

Thanks for your reply Arachnotron. Your explanation about the brightness makes sense. I went back to look at captures from other videos. All have been shifted brighter, just not nearly as much as this older video.

The link helps me to understand why the fields are getting distorted. As a follow-on, do you understand why this field distortion is occuring when the cable between the VCR and capture card is composite (in my original post I said component, this is incorrect) or s-video, but not when using an RF cable?

In the meantime, I will study up on TBCs.

[Arachnotron]

my guess (I am not sure ) is this:

When you use the antenna (RF), you are including two extra signal conversions. When capping RF, the tuner on the card takes RF and extracts a composite signal from it which goes into the capture chip. So you are in effect capping composite as far as the capture chip is concerned. So there are two possibilities:

- the VCR, when converting the videosignal from the tape to RF also boosts/restores the Hsync signal
or
- the tuner on the card boosts the Hsync when converting the RF to composite again.

For the field distortion, only a TBC will help (provided enough of the Hsync remains for the TBC to sync one). For the brightness problem, you could also try a simpler analog video processor that has a brightness control to adjust it before capping. Of course, once you go the TBC way, having one that can adjust that too would be the easiest.

Depending on the capture chip, playing a bit with the AGC function might also help. Normal drivers settings panels will not allow you access to that. For the BT878 you would need the BTtweaker in vdub, for philips SAA713x chips there are some options in FlyTV 2000.

I think you have a BT878 (I am not 100% sure). You may need to install the btwincap drivers instead of the leadtek ones to be able to use the tweaker. If you have a CX23881 chip I don't think there are any tweakers available. See chapter 11 (http://www.doom9.org/capture/bt_tweaker.html) of the capture guide for info on the tweaker. Options to play with are:

Full luminance range-[RANGE] (note the values are YUV values, not RGB)
This gives you more range to play with if you have to adjust the luma afterwards;

AGC crush-[CRUSH]
AGC delay-[ADELAY]
White down/up-[DNCNT and UPCNT]

All these will allow you to infulence the brightness calibration.

[cwk]

As a follow-up to this thread, I purchased a Canopus ADVC 300, which has a built-in TBC. This solved both problems I had described earlier. The TBC corrects the field distortion. The 300 also has a couple of options, called "White Step" and "Auto Gain Control", for limiting the brightness of a capture. These are helpful in keeping the luminance within range.