Re: How to translate log density readings to percent?

From: Dave Soemarko ^lt;[email protected]>
Date: 09/08/04-07:04:12 AM Z
Message-id: <001b01c495a4$568a5ce0$0a808080@wds>

Hi Loris,

I understand what you said. When I said the imagesetter negative can be
considered "ideal," I didn't mean that the tonality of the print would be
right, and that's why we need calibration of the print. What I meant was the
negative itself would be right (that is, the supposably 50% area would
indeed be 50% in the negative).

The problem that we see with rough paper is understandable. Because of light
scattering, the digital becomes analog.

Undercutting is when even the negative and emulsion are in close contact,
you still might not get perfectly binary effect because the light is
diffused (not collimated), so the light doesn't go through the dots in
straight-line fashion but diffuse in every direction. I am not good at
giving formal definition, but hopefully it makes some sense.

Dave S

----- Original Message -----
From: "Loris Medici" <>
To: <>
Sent: Wednesday, September 08, 2004 7:00 AM
Subject: RE: How to translate log density readings to percent?

> Hi Dave,
> I use a 225lpi negative on a 3600dpi imagesetter (225 x 16 = 3600). The
> source file resolution is 450dpi (2x factor). I agree that "logically",
> an imagesetter negative should do fine for my purposes "without" severe
> curve adjustments... BUT "pratically" this is not the case (at least for
> me).
> Let me describe my conditions:
> I use textured paper (cold press) and when I contact print an
> imagesetter negative emulsion to emulsion I get strange patterns on my
> print. The patterns are caused by the fact that the portions of the
> paper that are in close contact with the negative (lets name them "the
> hills") clearly show halftone dots (with a loupe of course) but the
> portions that are not in close contact (lets name them "the valleys")
> don't show halftone dots - because of light
> scattering/halation/diffusion??? whatever. Therefore, the tone of the
> "hills" and "valleys" differ considerably given the same negative dot
> percentage.
> I overcame this problem by printing emulsion up (yes, I loose sharpness
> but not much - and I can compensate for this by oversharpening the file
> in Photoshop). That way, I use the thickness of the negative to turn
> hills to valleys, and valleys to deeper valleys so there are no strange
> patterns and tone difference (because there's no "true" contact)
> anymore. When I print emulsion up, the density change - somehow - is not
> "linear". That's why I opt to design a curve by careful measuring.
> Will keep all what you said in my mind when the densitometer arrives.
> Thanks again.
> Regards,
> Loris.
> BTW, What is "undercutting"?
> ----- Original Message -----
> From: Dave Soemarko
> To:
> Sent: Wednesday, September 08, 2004 12:21 AM
> Subject: Re: How to translate log density readings to percent?
> Loris,
> Your method is not wrong but might be unnecessarily complicated. If you
> are using an imagesetter negative, you almost don't need a densitometer.
> Assuming your service bureau is doing a reasonable job of maintaining
> the chemical and calibrating the system regularly, the negative itself
> should be fine. When you print, you want to be able to have the 5% tone
> and 95% tone printed as suggested before. If that is not possible, it
> means that for your printing process, you need to use a negative of
> lower resolution (or alternatively, you could print from 15% to 85%, for
> example, and adjust the curve accordingly; but what that means is that
> you lose some tonal values because you are not using the whole available
> range).
> What you are doing is that you are trying to match percent dots to
> densities and compared that with the theoretical values and try to
> adjust the curve to get to this theoretical value, but this is almost
> like going through a detour. Your imagesetter negative can be considered
> "ideal" negative (though not perfect), so if you are able to achieve the
> 5% and 10% tone, everything else should fall correctly theoretically,
> but in reality there are some othe factors involved. Years ago I
> mentioned about dot gain. Some people think there is no dot gain because
> there is no bleeding as in mechanical printing. Actually while our
> mechanism is different from traditional printing, the effect of dot gain
> does occur because of undercutting, for example.
> But what applies to us is that you could look at the 50% dot and see if
> it printed too light or too dark, and when you need to adjust it, you
> can look at the printed patches and determine the adjustment. You could
> then adjust the curve.
> I don't know if I can make myself clear in a short email, but you should
> probably try to adjust the 5%, 50% and 95% tones to get a reasonable
> print. Once you get the feel of it (and understand this calibration
> process completely), you can fine tune and do the 25% and 75% tone.
> Dave S
Received on Wed Sep 8 07:04:31 2004

This archive was generated by hypermail 2.1.8 : 10/01/04-09:17:55 AM Z CST