Colored Vertical Line When Ouputting 8 mm Camcorder Footage

[echo1434]

I picked up a Sony DCR-TRV480E Digital 8 camcorder for the purpose of transferring some old 8 mm tapes to digital format. For the most part I'm very impressed with the quality. I suppose it's largely a result of the footage being "captured" within the camera before it's output to DV.

Everything seems pretty good, except one strange problem. A colored line will appear at the right of the screen when given certain lighting conditions. Reddish colors bring it out the most (producing a green line), and pink colors produce a strong blue, etc. At first I thought this was just part of the tape, but then I captured the video with another device and there were no lines (but the image quality was poorer).

Since then I've noticed other capture devices producing small "problem areas" on the left or right of the screen, while others don't. So what I'm wondering here is if its an inherent design flaw in the camcorder's output process. Like I said, capturing with other devices doesn't cause this colored line to appear, but I still prefer the camcorder method for quality.

The availability of Video 8-to-DV devices these days is pretty low, especially when you're working with the PAL format, so I can't just try out a bunch of different units. I'm trying to live with this problem now, but I was just wondering if anyone has some insight. I'm curious as to what might cause this, etc.

Below is a very clear example from the last video I captured. In the end, I'm just very confused about all this. Any thoughts or ideas will be greatly appreciated. Thanks in advance.

[edDV]

This picture is shifted left. The green was put there by the equipment to fill the space. There is no luma in the gap, just green chroma.



The picture is also grossly oversaturated with chroma as shown here. This will not record properly to tape.



PS: the arrow at the top should be pointed at 120 IRE not 100 IRE.

[echo1434]

Interesting. May I ask what program you're using to get those neato graphs?

Fault of the playback device in shifting the image, or... ?

[edDV]

Sony Vegas.

Is this just VCR into the DCR-TRV480E D8? Normally the chroma would be clipping at those extremes. The problem is either the chroma saturation is high on VCR playback or it was recorded that way.

[edDV]

I switched Vegas to PAL DV project settings and added the vectorscop display. The chroma is clipping in the yellow as seen. This means the VCR is oututting PAL chroma saturation too high or the tape was recorded that way from the input.

H shift is still unexplained but most likely that was due to upstream equipment to the recording VCR. Luminance levels appear to be set well but the spotlight is in saturation at 108 IRE. This is OK for 8 bits.

The blanking appearing below 0 % indicates this may have been NTSC source recorded to PAL. That may have somethig to do with it. Or it is just 0-255 scaling for the frame capture.

[echo1434]

Thanks for all the info.

However, it seems I'm more confused than ever. The below pic seems to counter the idea that the image is shifted to the left. In this case, you can clearly see something behind the pinkish-purple line on the right.




To clarify what's involved here: This is an 8 mm PAL master tape being played back on my TRV480E camcorder and output to PC via FireWire. There's no analog generation, and it's 100% PAL.

Surely the colors themselves are a little screwy, but it's the line itself that's really driving me crazy. I suppose I could fix the colors through editing, but the right-side line will always be there unless I add a huge black border to mask it. Since you can "see behind" the line in the above instance, I'm thinking now it maybe be a "colored shadow", for lack of a better term. I really have no clue...

[JohnnyMalaria]

My old Sony Video8 PAL camcorder does the same thing - and that was many moons ago using a FAST AV Master for analog capture.

It's the nature of beast - a limitation of the consumer technology.

The troublesome area is not visible on an ordinary display since it falls in the overscan area. On the computer, you are seeing the whole frame. If you look at the bottom of the frame, you'll also see some other undesirable material - the head-switching noise.

If you intend to view your edited footage on a standard TV (i.e., with overscanning), don't worry about these - you won't see them.

[edDV]

Were these as shot from the camcorder CCD or is the camcorder recording an external input? This looks over processed.

Whichever source, the color spike is in the digital video itself. The clipped color channels are causing the strange composite encoding problems. This is the video equivalent of driving audio levels into distortion.

The first shot shows the green spike is in the video


The second shot shows the red spike is in the video


Whatever caused it, it is recorded into the video.

[echo1434]

Curious, does this look any better to you?

[edDV]

Yes, less clipping. What are you adjusting? What does the captured DV-AVI look like before processing?

[echo1434]

That's from the original DV AVI (output by the TRV480E).

The full size screenshots I showed before are from the DVD I made, which I did some "adjustments" on prior to the MPEG encoding. Well, now I know that I didn't do a very good job there!

I know there's still some clipping, but I think this is a source problem, like you said. Surely better now, though.

Thanks a lot, really. I'm more of an audio guy who eventually made his way into video. So I'm still learning...

[edDV]

For DV and DVD normal black is level 16 and normal white is level 235. The levels above (i.e. 236-255) are supposed to be for "headroom" or overshoots when following broadcast practice. Consumer camcorders often use levels up to 255 to improve signal to noise but this often results in clipping at 255. Pro cameras have a "knee" that controls whites to around level 235 only letting true highlights like that spot light hit the 255 ceiling.

[JohnnyMalaria]

Here's an image from an analog PAL Video8 tape, captured via DV using a Sony DSR-11.

You can see on the left that the chroma is weak (i.e., the left part of the video frame looks black-and-white). You can also see from the vectorscope that, generally, the saturation is very high - quite typical of the consumer analog PAL equipment I've used.