Why is Digital8 camcorder passthru output different than its digital tape output?

I've been using a Digital8 camcorder to do some video capture using DV
format, this case capturing 8mm film, going in to the computer via Firewire.
I'm running into a problem.

I find that when I output through the camcorder directly to the capture
program without recording to tape first, I get an odd vertical banding
artifact when the resulting avi file is rendered for tape output or DVD. The
banding isn't extremely pronounced but it's there and enough to be
distracting when viewing. I'm basing this on what I see on a tv monitor,
i.e. outputting to the tv instead of to a VHS tape, or playing a test DVD
made from the file on a tv. However, this doesn't occur when recording to
tape first in the Digital8 format and then outputting that to the computer
and then rendering..

This banding also occurs when playing analog tapes through the Dig8
camcorder (this camcorder is backwards compatible) to the computer.

It doesn't occur when going through composite cables and capturing as an
mjpeg avi file instead of the DV format.

I thought it would be the same, but apparently there's some difference in
the DV output from the camcorder when doing it on the fly as opposed to
going from DV tape output.

Any ideas why this is happening, why it would be different and how it can
fixed? Capturing on the fly in the DV format is preferable for various
reasons. Time savings and the size of the resulting files among them.

Using WinXP Home, Sony TRV-240 Dig8 cam, Pinnacle Studio 9 as my capture
program. (Also tried using AMCap, with no difference in the results)

Thanks for all input.

 

You probably have web space somewhere where you can copy files for
others to download? A short downloadable test clip (say, a second's
worth of original, unprocessed DV AVI material that displays the
problem) would probably help in investigating the phenomenon.

 

A second isn't enough, you really need to see a fair amount of the video and
see it on a tv to see it and the webspace I have available via my internet
provider isn't large enough hold several seconds of raw video. The lesser
resolution of even higher quality Windows media or Realmedia files obscures
it.

I did put together a site where you can see it a bit through stills if
you've got a decent monitor. Don't be misled by the presence of black
borders on some of them. Some of them had had a border applied and the ones
that are mjpeg hadn't. Doesn't have any effect on the banding artifact.


http://home.mpinet.net/~docsavage20/examples_of_video_banding/index.html


Since I've tracked down the fact that it's there when doing it "on the fly",
but not if doing playback from a Dig8 tape, finding out what's different
about the output would probably yield some answers.

 

It seems to me you have a 60Hz hum when connecting to the computer. Try to
run the camcorder only on batteries and use the same power strip for all
connected gear, especially the computer and VCR.

 

"Ban" wrote ...

Powerline mains hum would appear as horizontal "banding".
Furthermore, here in NTSC-land, the hum bars move slowly
up the frame, taking ~10 seconds to make it through the whole
frame. (This is because of the 59.94 field rate vs. our 60 Hz
power freq.)

"Doc" said he was seeing "vertical" bands. Or perhaps he
should clarify so we are all on the same page. Vertical bands
is not a very common video artifact. I've never seen such a
thing. Of course, a picture is worth a thousand Usenet
postings and followups.

 

I read in sci.electronics.design that Richard Crowley
about 'Why is Digital8 camcorder passthru output different than its
digital tape output?', on Sun, 11 Sep 2005:

Thin vertical lines used to be a sign of parasitic oscillation in the
equipment, not necessarily in the video chain. Horizontal/line output
tubes/valves used to suffer from Barkhausen oscillation, which produced
quite a crop of vertical lines.

Broad brightens variations, more 'stripes' than 'lines' can be due to
velocity modulation of the horizontal scan.

 

"John Woodgate" wrote ...

Right. I remember both from back in the earliest video
equipment I worked with (the last days of firebottles).
But I've never seen anything like either of those in any
solid state equipment (and certainly not in digital).

I'm guessing he means horizontal bands moving vertically.

 

http://home.mpinet.net/~docsavage20/examples_of_video_banding/index.html

Personally, I'd be much more worried about the drastic color difference
and not so concerned about these edges!

Guess: The passthru is going through an analog stage, probably to show
OSD information. and you're seeing ringing on high-frequency components
(vertical edges). The tape output does not go through the OSD stage.

 

http://home.mpinet.net/~docsavage20/examples_of_video_banding/index.html

You have been through here before with this same "problem"
and these same pictures. I'd guess that you may not get any
better answer in this round.

When I crank my imagination up to 11, I think I can see your
"vertical lines". At best they are what I would call a "3rd
order effect". There are certainly much more deserving
parts of the video that could use your attention.

 

Not really a big point of concern. Different lighting situations, different
projector and the mjpeg capture allows you to adjust the color balance as
you capture. I just made the mjpeg captures quickly to see if the banding
artifact was there.

Hmm. Why would it not go through this stage with the tape?

 

Since I have new information - i.e. noticing the difference between
passthough and tape output, and posting to additional forums, I figured it
would be worth putting it out there again. Plus, there's always a chance of
getting it in front of different eyes.

They're there. No imagination needed. Those vertical lines that I described
on the webpage are them, though they're more obvious when you actually see
the video on a tv. This is the best I can do with limited webspace
bandwidth. I'm more concerned with getting rid of this artifact than focus
or whatever issues at the moment.

 

If by "he" you mean me, I mean vertical bands. It's as if the image were
being viewed through many, many thin vertical strips of glass whose edges
are *almost* invisible but not quite.

 

Guess: The passthru is going through an analog stage, probably to show

On Screen Display.

Perhaps because the passthrough is assumed to be a quick preview thing
and the tape output is for production. Note, I did say it was a guess.
My camcorders both show different information on the LCD vs. the
composite output, and different information again on the DV stream. The
LCD shows battery information and a bunch of other stuff; time,
position on tape, etc. The composite output only shows an icon
indicating the current function (PLAY, REC, etc). The DV output doesn't
have any superimposed info.

 

Essentially the same on mine. You can include the info on the LCD display on
the composite output if you want (not sure why anyone would) but not on the
DV output.

However, not sure about the "passthrough simply a preview tool" thing. Going
by the manual, it's supposed to work, but clearly there's *something*
different about the 2 different types of DV output.