In a previous thread, I got some good advice about problems I was
having capturing S-VHS into a PC...

http://groups.google.co.uk/group/rec...5808cb12bc77d8

Now I have the ADVC110, it seems that the levels really are a little
too high on the tape, and are clipped at the input to the ADVC110.

Decoding the DV AVIs to the range 16-235 helps, because there is
information above 235 (which gets clipped during a typical YV>RGB
conversion which maps 235 to 255).

However, there's still some video information being lost at the very
top of the range.

Can anyone suggest a suitable Proc Amp or other method of bringing the
levels down a little? I'm tempted to try a parallel resistor before
the input to the ADVC110, though I can think of at least two reasons
why that shouldn't work properly!

Cheers,
David.

"2Bdecided" wrote ...
Can we assume that when you view this video on a *good
picture monitor* (not just any old TV screen), the video is
NOT clipped?

Did you set the switch (on the bottom of the ADVC) for
black level properly? Did you try it in both positions?

It certainly wouldn't hurt to try simply attenuating the video
signal. The worst that could happen is that it wouldn't work.

In article <61tr92F20rahmU1@mid.individual.net>, "Richard Crowley" <rcrowley@xp7rt.net> wrote:
levels.

On 18 Feb, 16:49, "Richard Crowley" <rcrow...@xp7rt.net> wrote:
I'm not sure I have a good picture monitor, but there is more range on
the tape than the ADVC110 can cope with. (The HV20 and various DVD
recorders faithfully preserve more).

Irrelevant for PAL (that switch selects PAL/SECAM for 625-line
signals)

I expect the wrong termination (which is what I'd be introducing) can
set up standing waves on the cable, but maybe they won't be severe
enough to be visible - I'll try.

Cheers,
David.

"2Bdecided" wrote ...
That seems highly unlikely unless you ADVC110 is broken,
or you have some non-standard source. Video is pretty well
defined for voltage levels. I've used Canopus ADVC boxes
with lots of different equipment and never encountered such
an issue.

Theoretically. So you are saying that you did not try it?

You would be introducing *additional* termination
(resistance lower than 75 ohms). Probably not the best
experiment you could run. If I wanted to attenuate the
signal, I would rather use a *series* resistance which
would form an "L-pad" with the ADVC termination.
I'd try something like a 500-ohm pot wired as a
series rheostat.

No. Not at video frequencies, and for cable lengths of a few
meters or less. Your presenting symptom seems unrelated
to "standing waves".

Visible where? You said that you don't have a proper
picture monitor, so you don't really know what your
video really looks like. Your average consumer TV
screen is uncalibrated, and you know (don't you?)
that computer screens have a very different gamma
and do not display television images properly.

On 18 Feb, 18:44, "Richard Crowley" <rcrow...@xp7rt.net> wrote:
I've looked on a proper scope now. The _video_ levels are fine (702mV
max blank > peak white), but the sync voltage is a little too low
(258mV blank > sync tip). This is PAL, so it should be 700/300.

That could cause the problem if the box amplifies the whole signal to
get the sync back to 300mV. It would make it 16% too big, which is
about right.

HOWEVER, it turns out that the ADVC110 _does_ capture the YUV data
without sever clipping. The YUV data itself is nearly fine (very
slight clipping at peak white, but almost right), it's the YUV>RGB
conversion to display it on my PC which clips it, by expanding 16-235
If I keep the full range (0-255 > 0-255) then things are much better,
and I can bring the whites down to 235 if I want to.

THERE'S STILL A MAJOR PROBLEM: there are bright areas with strong
colours which cannot be represented in RGB colour space. For example,
there are near-peak whites, which have a strong blue component. This
causes severe clipping of the blue channel in RGB colour space.


My understanding is that the issue of legal vs valid colours is well
known, and that these saturated colours are invalid. The problem is
that I don't want to reduce the saturation, because these videos are
already slightly desaturated (I've always had to increase the colour
control on the TV to watch these tapes).

The only alternative appears to be to reduce the luma (Y) before
converting to RGB - that gives an unclipped, but comparatively dark
image. After this, the saturation is about where it should be - less
luma plus same chroma = greater saturation, so there's no need to
tweak it further.


I don't _have_ to convert it into RGB at all, but it seems a bad idea
to put "invalid" colours on a DVD - especially as most European DVD
players have RGB outputs, which are going to clip these colours at
that point anyway; and many people now watch on PCs, which will clip
these colours for display.

Does anyone have any experience with this issue? What do people
normally do in this situation? Or is it a strange situation?


I think you misunderstood me. In NTSC mode, that switch changes the
set up level. In PAL mode, it switches to SECAM. You cannot capture
PAL tapes in SECAM mode - the result is black and white.

I'm going to try something like that to remove the slight luma
clipping I'm seeing, though as the voltage of the picture information
is already OK but the sync is not, I'm not hopeful! It's nice cheap
and easy to try though.

Cheers,
David.

"2Bdecided" wrote ...
Are you saying that it looks "comparatively dark" when
viewed on a computer screen, or on a real television screen?
Remember that ALL video looks "dark" on a computer
screen. That is normal and expected. And that is why we
use a real television screen to judge any subjective picture
adjustments, NOT a computer screen.

If you have source video that is THAT much out-of-spec,
and you want good quality out of it, may be time to take it
to a professional with the video processing and monitoring
equipment to do a proper job of it.

On 26 Feb, 16:44, "Richard Crowley" <rcrow...@xp7rt.net> wrote:
Peak white would have to be below 195 on a scale of 0-255. That's
"comparatively dark" whatever you view it on!

I don't think being a professional will make the impossible, possible
- it seems the combination of luma and chroma on the tape cannot exist
in RGB colour space. I'm just looking for suggestions (or reports from
experience) on how to square the circle!

I'm not looking to pay someone else to do it. I want to learn through
doing it myself. It is the challenge, as well as actually preserving
the footage in the digital domain, that motivates me. If I can't do it
myself, it won't get done.

Cheers,
David.