#Color Wierdness
13 messages in this thread
I was working on my system at work and set up to try to gamma adjust the
output. I can get a good gamma adjustment (1.6 or so, I think) on my RGB
monitor, but when I ship the Targa+ composite signal either into a monitor
directly or from Targa+ to ProVAS to VO-5850 to Monitor I get a gray center
square and the outer square has a set of color rainbows moving continuously
through it.
No adjustment that I've made so far has an effect except for subcarrier phase,
and that just jumps the order of the colors.
Any ideas??
Chris
Chris – There's been quite a bit of discussion about Gamma and the Targa 32+
and I believe in the library there's a thread on it, but I haven't checked
lately. The long and the short of it is set your Gamma to 1.8 (default) to
start, don't worry about the gray boxes on the monitor for composite output, as
all that happens when you line them up is that you are effectively turning
gamma off. Take a 3ds file like ADLOGO, which is my favorite for tests as it
has a very subtle bump map on it, and the colors are excellent for testing, and
render it at various GAMMA settings till you get an NTSC output that you like.
I think the default of 1.8 is the best all around. For some projects you might
want to turn it down. 2.2 while NTSC linear loses contrast so I typically don't
go up that high.
-JE
John: Thanks. I've been following the gamma thread on and off, but the color
banding was a new one on me.
I'll try your suggestion. Thanks
Chris
Chris – maybe I misunderstood, it sounded like you were getting a
rainbow moire on your screen which can happen. But banding sounds
more like 16 bit mode which means the Targa is not set for 32 bit
mode.
John: No, the effect that I've seen is more like a rainbow moire rather than
banding – the rainbows run diagonally and shimmer with time. They repeat
several times across the outer alpha adjust square.
When I try it at home, I don't see the banding. Is there anything to adjust?
Chris
Chris – I don't recall which monitor was having the moire effect but
typically this is a result of interlaceing and if its your NTSC
monitor will be even more pronounced than an SVGA in interlace mode.
This has to do with the frequencies that each monitor employs. The
SVGA being higher than NTSC and so not as pronounced. If you are
adjusting your home monitor and it is an SVGA then it is probably not
in interlaced mode so hence you don't get the effect. On an NTSC
monitor there's really nothing you can do about it. But then there's
really no point in adjusting the NTSC monitor for Gamma correction,
as you want to adjust RGB for GAMMA so that the NTSC output is Gamma
corrected. Hope this helps.
-JE
John: Ok, thanks. In fact, I can successfully gamma correct my RGB monitors on
both systems. If I look at the gamma correction screen at work (on my expensive
Panasonic composite monitor) I see the rainbows. On my cheap Amdek monitor at
home, I don't. In either case, the outside grey levels look wierd, and don't
ever really match.
Thanks again,
Chris
In addition to what John said, it has been my experience that the VO-5850 does
some processing to the signal, if you are monitoring on the video out. For
best results look at the signal using the loop through on the input side with
terminator set to off. The signal will be much cleaner but will not change the
problems with the composite gamma adjustments. For that just follow what John
told you.
Ok, thanks. I've looked at the composite signal on either side of the VO-5850
and I get the color bands both ways. I'll have to try it using the RGB to
composite encoder on my ProVAS, and on my home system.
Chris
I just wanted to alert you about some processing that is being done
from 5850 input to output and not to trust the 5850 output as a
measurement source for the input side.
OK, thanks. I tried it on both sides of the 5850 and I still see the banding.
What does the 5850 do to the video, anyway?
Chris
First of all, what you are seeing is not called banding which
something entirely different.. What you are seeing is normal on the
gamma match boxes in 3DS. It is a moire pattern that is caused by a
fine striped pattern in the gray boxes that beats with the scan lines
of TV. It's the same thing that occurs when talent wears pin
stripped clothing on TV. A definite No-No. I wouldn't worry about
it. As Gary Yost has said on several occasions, View the gray boxes
from a distance with squinted eyes. I think what he is saying is not
to try to focus on the boxes but view them with slightly out of focus
vision. This will minimize this moire effect.
Regarding the 5850. The processing I refer to may just be poor high frequency
response in the video circuits of the 5850. I only mention it if you are doing
some critical adjustments and the 5850 output may taint the results. Best to
use the feed through on the input side for accurate measurements. If just
doing simple program monitoring it is more convenient to have the monitor on
the output side.
Thanks. I think that the only way to get far enough away is to stick my
animation rack out in the hallway! I see your point, and I've noticed the beat
effects even on the monitor that doesn't show the color bands.
I've got my system set up with a patch panel so I can look at the Targa
composite output, the signal coming out of the 5850, or the composite output of
the ProVAS. I see banding in every case. I have noticed things in the 5850
output that make me think that you are right about the high frequency
performance on that deck. Maybe I'll talk them into a better deck someday…
Chris