[opendtv] Re: New Chips Improve Color TV Dramatically

  • From: "John Golitsis" <john@xxxxxxxxxxxxxxxxxx>
  • To: <opendtv@xxxxxxxxxxxxx>
  • Date: Thu, 1 Jul 2004 10:53:22 -0400

First of all, doing a DVE calibration without having the grey scale
calibrated first is a recipe for 'disaster'.  I've actually seen it make the
picture look much WORSE than it was with the factory settings.  All those
poor suckers out there that think their TV looks better because they've used
DVE, etc!  Without knowing more about what you're doing, It's really only
good for setting the black level and maybe the hue.

The green and red filters are used to check the colour decoder's accuracy.
Ideally, the set will have service-level controls for Gain and Axis of the
red and green channels.  Gain is pretty self-explanatory, and Axis is
basically like a hue control for that colour channel.  So, in Doug's case,
the green filter should show that the pure green bar is way out of whack in
gain as compared to white, so he'd lower the gain until the green bar
matches the intensity of white.  Then he'd compare the intensity of green
and white against cyan and yellow, and adjust the axis to that they all
match.  The green filter isn't terribly 'accurate' though, so yellow and
cyan actually end up looking more subdued that white and green, but as long
as they match each other, you're okay.  I should mention that different TVs
have different colour decoder controls, or none at all.  Hitachi and I
believe Sony have green&red gain/axis, Panasonic's CRT RP sets have Y-b/Y-r
controls and Toshiba's have no controls at all (but accurate decoders,

I'm certainly no video engineer, but I can't think of any reason why this
wouldn't work on a set with wider colour gamut, unless the colour filters
themselves become inaccurate as a result.

----- Original Message ----- 
From: <jeroen.stessen@xxxxxxxxxxx>
To: <opendtv@xxxxxxxxxxxxx>
Cc: "Joe Kane" <joekane@xxxxxxx>
Sent: Thursday, July 01, 2004 4:58 AM
Subject: [opendtv] Re: New Chips Improve Color TV Dramatically

> Hello,
> Doug McDonald has said some very interesting things:
> > All I know is that when one views the SMPTE bars
> > from Digital Video Essentials, with the color matrix
> > set the way I have it set, one gets pure Red, Green,
> > and Blue ... no mixtures. There is also no mixture
> > with a computer input to the RBG input.
> and
> > AH! but, as Jeroen so ellipically put, with primaries
> > that are "purer" than the official HDTV standard,
> > they (the bars) perhaps actually should. Since
> > the green on my set is greener than the standard, and
> > more saturated, a substantial amount of red and
> > a bit of blue should be added to make it the
> > official color. This is not done. (My set is not CRT,
> > but the same logic applies.)
> Whoa ! Do you know what this means ? It means that
> the "Joe Kane method" for calibrating the saturation
> can not be applied to wide-gamut displays like Doug's !
> Let me explain.
> With "Digital Video Essentials" you get 3 pieces of
> colour filter: red, green and blue. If you follow the
> calibration instructions, then when you are done you
> should see from a colour bar picture: only the red,
> magenta, yellow and white bars through the red filter,
> the green, cyan, yellow and white bars through the
> green filter, and the blue, cyan, magenta and white
> bars through the blue filter. This means, I believe,
> that you have adjusted your display for _cutoff_ of
> each channel for the 3 colours that should not be
> seen. There will be no red in the black, blue, green
> and cyan bars, no green in the black, blue, red and
> magenta bars, and no blue in the black, red, green
> and yellow bars. And for a traditional display that
> has native standard (EBU, RP.145, Rec.709) primaries,
> this will be correct. So far so good.
> But what about a wide-gamut display ?!
> As Doug has correctly worded, if some of the colours
> are better than standard then in order to display the
> correct standard colours it is necessary to mix in a
> fraction of the other colours. (An over-saturated
> display is calibrated by a reduction of saturation.)
> So, in his case there should be a bit of red and/or
> blue mixed into his super green to give standard
> green. The red and/or blue should not be cut off.
> You can never do a correct colour calibration for a
> wide-gamut display based on the cutoff method. You
> can only do it with a professional colour analyzer !
> If we may assume that Doug has calibrated his display
> with the (D)VE disc and the filter(s), then he has
> actually calibrated it wrong. He will be rewarded
> with a slightly over-saturated picture. This may be
> pleasing for him, but it is not "correct" !
> And what about multi-primary displays ? How are you
> going to "calibrate" a 5-primary display (e.g. red,
> yellow, green, cyan, blue), or a 4-primary display
> (e.g. red, green, blue, white) with a colour bar with
> only 6 colours, and a maximum of 3 (RGB) filters ?
> Answer: you can't. Not without a real colour analyzer.
> Joe, be warned: you're going to mess up the high-end
> display market if you advise your customers to do
> their own calibration of wide-gamut displays !
> Disclaimer: I may have a wrong interpretation of the
> method of calibrating the saturation with the SMPTE
> colour bar picture. This is where you view a blue
> bar next to a white bar, through a blue filter.
> You should adjust the 2 bars for equal brightness.
> At over-saturation, the red and green will clip to
> zero and the blue output will be too bright.
> At under-saturation, there will be red and green
> mixed in and the blue output will be too dark.
> But you can not see the contribution from red and
> green at all, only the brightness of the blue.
> I think that it will go wrong then. Am I right ?
> Greetings,
> -- Jeroen.
> |-----------------------------+---------------------------------------|
> | From:     Jeroen H. Stessen | E-mail:   Jeroen.Stessen@xxxxxxxxxxx  |
> |-----------------------------+---------------------------------------|
> | Building: SFJ-5.22 Eindhoven| Philips Digital Systems Laboratories  |
> |-----------------------------+---------------------------------------|
> | Phone:    ++  | Visiting & mail address: Glaslaan 2   |
> |-----------------------------+---------------------------------------|
> | Fax:      ++  | NL 5616 LW Eindhoven, the Netherlands |
> |-----------------------------+---------------------------------------|
> | Pager:    ++31.6.6513.3818  | Visit us: http://www.pdsl.philips.com/|
> |-----------------------------+---------------------------------------|
> ----------------------------------------------------------------------
> You can UNSUBSCRIBE from the OpenDTV list in two ways:
> - Using the UNSUBSCRIBE command in your user configuration settings at
> - By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.

You can UNSUBSCRIBE from the OpenDTV list in two ways:

- Using the UNSUBSCRIBE command in your user configuration settings at 

- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts: