Super black and white pixels / lines in lower field (only) of VHS capture

[ctrl-z]

Greetings exalted ones,

I will be very grateful to anyone who can offer an explanation (and, hopefully, a fix) for the appearance of "super black" and "super white" pixels in the lower fields (field B) of my VHS and Hi8 captures. I just started using a DigiSuite ("full" model, from 2000) for these analog captures. The AG1980 (VHS/SVHS) and EV-C200 (Hi8) decks were feeding the DigiSuite's break-out box directly, with no other equipment in line.

Of course, these pixels & lines do not appear on my Sony PVM-series monitor, only on my computer screen in programs like VLC, AE, VirtualDub, etc. (see the left window in pic 1 of 2, below).

The facts behind pic 1 of 2, below:

- Captured using DigiTools "uncompressed" (huffyuv) codec. (Forgive the obvious, but DigiTools & DigiUtils are capture and configuration programs bundled with the DigiSuite, respectively.)

- All sliders in DigiUtils (including the "Levels" and "Setup" sliders) were set at default for the capture that created these unwanted lines.

- The "Broadcast" profile in DigiUtils for video levels was used, all along, which is not supposed to even allow such extreme levels of black and white (see pic 2 of 2, below).


The fix I found to avoid capturing the super black and super white pixels:


- Drastic adjustments (in my opinion) to the Levels and Setup sliders in DigiUtils did avoid the creation of these pixels & lines. (Gee-whiz, will other scenes and other tapes require *even more* boost and cut to the signal?)

Can this corruption of the lower (B) field be avoided with hardware? Or, are the "new" Levels and Setup adjustments I used the only way to prevent the capture of super black and super white?

* Lastly, could there be a filter or software fix to remove the damage, all together? *




VideoDub shows the effect of certain deinterlacing. Both bob and ELA algo's removed the unwanted pixels, when set to "Keep top field...". Can this be removed without taking such drastic steps (and losing the lower [B] field)?

For the next capture, the unwanted black & white pixels were (mostly) avoided when the Level and Setup settings were drastically adjusted (see middle panel in photo below). Is this expected when capturing old tapes?





Thanks in advance for your consideration and comments!

[ctrl-z]

(Forgot to mention that the signal from the decks to the DigiSuite was run through a high-quality Extron S-video cable, FWIW.)

[themaster1]

Have you unticked the box in the huffyuv codec for "always RGB output" ? Best stable version of huffyuy i know is 2.1.1

[NJRoadfan]

Your footage isn't ideal. Superwhites are going to be a fact of life with footage from a consumer grade 1980s tube camera. A proc-amp with some brightness/contrast adjustment would come in handy here.

Its likely the that checking the "allow super black/white" will force the card to capture 0-255 level video vs 16-235. You are likely hard clipping super blacks and whites at capture time without those boxes checked. That fancy Matrox card should have the ability to show a live waveform monitor of the input and confirm this. I'm also curious as to what the "Source has built-in TBC" option does. If the Matrox card has any sort of onboard framesync/TBC, that would eliminate the need for an external TBC.

[sanlyn]

All true. The fact that the noise is only on alternate fields makes me suspect that the 1980 needs maintenance. Do you get the same problem on all or most tapes?

[ctrl-z]

Quote:

Originally Posted by themaster1

Have you unticked the box in the huffyuv codec for "always RGB output" ? Best stable version of huffyuy i know is 2.1.1

I didn't think to look "under the hood" of the huffyuv codec that installed automatically with the DigiSuite software. I certainly will. Thanks for the suggestion about 2.1.1; I know it is available individually and is included with bundles, like the K-Lite Codec pack, ffdshow, etc. but I was afraid to install such a bundle in the DigiSuite machine, for fear of disturbing the setup. (The whole thing is running on an old NT4 machine [read: house of cards] lol )

I'll suck it up and install huffyuv 2.1.1 on the machine after checking to see whether the "always RGB output" option exists in the machine's currently-installed version.

That could sure be an easy fix -- Thanks so much!

[ctrl-z]

Quote:

Originally Posted by NJRoadfan

Your footage isn't ideal. Superwhites are going to be a fact of life with footage from a consumer grade 1980s tube camera. A proc-amp with some brightness/contrast adjustment would come in handy here.

Its likely the that checking the "allow super black/white" will force the card to capture 0-255 level video vs 16-235. You are likely hard clipping super blacks and whites at capture time without those boxes checked. That fancy Matrox card should have the ability to show a live waveform monitor of the input and confirm this. I'm also curious as to what the "Source has built-in TBC" option does. If the Matrox card has any sort of onboard framesync/TBC, that would eliminate the need for an external TBC.

You certainly know your footage, NJRoadfan : Indeed, it was shot by a high school classmate in 1989, and was a dub of the original (which was destroyed) .... Not "ideal", at all.

I guess a software-based proc amp for B/C is out of the question, since it is too late, by the time signal gets into the card? (I've tried to avoid running the signal through any external devices between the deck and the capture device, since I noticed degradation when I did so.) Turned out that the AG1980's TBC was much more effective for this tape than either the DigiSuite's built-in TBC or a Sony MPU-F100A I tried.

I just did a capture using the "Post Production" profile in DigiTools, which allows (checks both options) for super black and super white. Surprisingly, it created more of the "clipping" pixels & lines in the lower field. And, unfortunately, the Matrox utilities don't have a waveform monitor, period. I guess I'll have to compromise sharpness to prevent such clipping and add a component at the input stage. (Hey, it's gotta be better than the current results )

BTW, I also have an Elite Video BVP-4 I've never tried. (Are these any good?)

Thanks for your reply!

[ctrl-z]

Quote:

Originally Posted by sanlyn

All true. The fact that the noise is only on alternate fields makes me suspect that the 1980 needs maintenance. Do you get the same problem on all or most tapes?

Unfortunately, sanlyn, I have had the same problem with capturing Hi8 out of the EV-C200, as well. (Like the 1980, I bought the C200 new, but it has less than 20 hours of use, total. The 1980 has around 80 hrs of use and is cleaned every 10 tapes, or so.) Could it be the aged components inside these decks? They have low hours, but were manufactured in the 90's

[NJRoadfan]

Can you post a short unprocessed clip showing the artifacts? The left hand still above doesn't look like a VCR playback problem.

[msgohan]

Agreed. I think it's a problem on the capture hardware side.

[ctrl-z]

Thank you, NJRoadfan, msgohan, themaster1 and sanlyn! I am now feed-up to try your suggestions and will report back, hopefully with some encouraging news

[msgohan]

Here is another capture card with the same type of error, but creating white garbage pixels and sound distortion.

ATI TV Wonder 650 - bright screen driver bug [andy7079].jpeg