Physics Talk Physics Talk    FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in

Education Forums

Trapping, Neutral Density Goto page 1, 2  Next

Physics Talk Forum Index -> Color and Light

View previous topic :: View next topic
Author	Message
Gernot Hoffmann Guest	Posted: Wed Mar 26, 2008 3:17 pm    Post subject: Trapping, Neutral Density I've some questions in the context of trapping, as described here: http://www.fho-emden.de/~hoffmann/overprint.pdf Is Neutral Density the same as ND = -log(Reflectance) ? If so, how could this happen for two samples 1,2 ? L1*>L2* and ND1*>ND2* Best regards --Gernot Hoffmann
Back to top
	Ads
Advertising Sponsor
Gernot Hoffmann Guest	Posted: Fri Mar 28, 2008 1:41 pm    Post subject: Re: Trapping, Neutral Density Roger Breton schrieb: Quote: Density is following some Status response function. Density is the reciprocal of opacity, I believe. Neutral Density? What could be "neutral" in Density? Is it that the same as ISO Visual Density? Roger Roger, you're right: density is normally based on measuring through a status filter. Neutral Density (ND) is calculated without such a filter, for reflective prints the 'reciprocal' of reflectance. The word 'neutral' is used as well for optical filters: flat transmission spectrum. I'm wondering why I get by DTP22 and ColorShop Lab luminances and NDs which seem to be in contra- diction to each other. Do you know the ISO definition of Visual Desnsity ? This might help. Best regards --Gernot Hoffmann
Back to top
	Ads
Advertising Sponsor
Roger Breton Guest	Posted: Fri Mar 28, 2008 4:35 pm    Post subject: Re: Trapping, Neutral Density Density is following some Status response function. Density is the reciprocal of opacity, I believe. Neutral Density? What could be "neutral" in Density? Is it that the same as ISO Visual Density? Roger Quote: I've some questions in the context of trapping, as described here: http://www.fho-emden.de/~hoffmann/overprint.pdf Is Neutral Density the same as ND = -log(Reflectance) ? If so, how could this happen for two samples 1,2 ? L1*>L2* and ND1*>ND2* Best regards --Gernot Hoffmann
Back to top
	Ads
Advertising Sponsor
Mike Russell Guest	Posted: Sat Mar 29, 2008 11:24 am    Post subject: Re: Trapping, Neutral Density Gernot said: Quote: I've some questions in the context of trapping, as described here: http://www.fho-emden.de/~hoffmann/overprint.pdf Is Neutral Density the same as ND = -log(Reflectance) ? If so, how could this happen for two samples 1,2 ? L1*>L2* and ND1*>ND2* Another deceptively simple question, Gernot :-) As you know, the Adobe docs define the term "ND", or "neutral density" as equivalent to XRite's "Visual Density" or "V". I found this statment by Brian Dieckman in colorforums.com: " With a standard reflectance densitometer, you can get density readings through a Red, Green or Blue filter. (For Cyan, Magenta and Yellow readings, respectively) Also, a third reading is derived from the other three, used to measure black and spot colors. That density is referred to as "visual" by the X-Rite instruments I use." So neutral density is a nominal density value, derived from the CMY density values. The motivation for this calculation is almost certainly to simulate a neutral densitometer, based on the CMY values. (The C, M, Y, and V density values, incidentally, are displayed in ColorShop X's density scratchpad. Adding to the confusion, the poor V value is labeled K.) Neutral density may or may not be a simple dot product of the CMY with a constant vector, similar to the RGB luminance calculation, and this may explain the discrepancy you are seeing. On to your example image. It's a devilishly worst-case image, in that trapping and not trapping produce equally ugly results. From a pragmatic standpoint, a designer might choose from several alternatives to improve things. 1) Change the design. Stroke the interior of each letter with a black or white border, and use no trap at all. 2) Separate the gray background as a mixture of CMY, so that misaligned areas are less apparent 3) In the pdf you ask which stroke color would should be chosen. You may see an improvement using a blend, instead of a solid color, for the stroke. In effect this shares the ugliness between the aligned and non-aligned versions of the image. -- Mike Russell - www.curvemeister.com
Back to top
	Ads
Advertising Sponsor
Gernot Hoffmann Guest	Posted: Sat Mar 29, 2008 3:30 pm    Post subject: Re: Trapping, Neutral Density Mike Russell schrieb: Quote: Gernot said: I've some questions in the context of trapping, as described here: http://www.fho-emden.de/~hoffmann/overprint.pdf Is Neutral Density the same as ND = -log(Reflectance) ? If so, how could this happen for two samples 1,2 ? L1*>L2* and ND1*>ND2* Another deceptively simple question, Gernot :-) As you know, the Adobe docs define the term "ND", or "neutral density" as equivalent to XRite's "Visual Density" or "V". I found this statment by Brian Dieckman in colorforums.com: " With a standard reflectance densitometer, you can get density readings through a Red, Green or Blue filter. (For Cyan, Magenta and Yellow readings, respectively) Also, a third reading is derived from the other three, used to measure black and spot colors. That density is referred to as "visual" by the X-Rite instruments I use." So neutral density is a nominal density value, derived from the CMY density values. The motivation for this calculation is almost certainly to simulate a neutral densitometer, based on the CMY values. (The C, M, Y, and V density values, incidentally, are displayed in ColorShop X's density scratchpad. Adding to the confusion, the poor V value is labeled K.) Neutral density may or may not be a simple dot product of the CMY with a constant vector, similar to the RGB luminance calculation, and this may explain the discrepancy you are seeing. On to your example image. It's a devilishly worst-case image, in that trapping and not trapping produce equally ugly results. From a pragmatic standpoint, a designer might choose from several alternatives to improve things. 1) Change the design. Stroke the interior of each letter with a black or white border, and use no trap at all. 2) Separate the gray background as a mixture of CMY, so that misaligned areas are less apparent 3) In the pdf you ask which stroke color would should be chosen. You may see an improvement using a blend, instead of a solid color, for the stroke. In effect this shares the ugliness between the aligned and non-aligned versions of the image. -- Mike Russell - www.curvemeister.com Mike, thanks for your analysis. I think it's still not clear how ND is defined. 1) I have to respect a design, which works well for large format inkjet posters (but not for the toner printer): http://www.fho-emden.de/~hoffmann/hagiabook.pdf Please have a look at the big numbers top right. 2) The background is K-only gray. The reason for this decision: If the printers are gradually drfting, then the neutral areas would be affect visually very early. 3) By 'blend', do you mean a visual mixture of background and character color ? This wouldn't work (in my understanding) because this stroke pseudo-trapping are has to be overprinting. For some combination it looks better, if I use only half the charakter color values for the stroke, but I can't see a general law. Best regards --Gernot Hoffmann
Back to top
	Ads
Advertising Sponsor
Mike Russell Guest	Posted: Sun Mar 30, 2008 1:49 am    Post subject: Re: Trapping, Neutral Density "Gernot Hoffmann" <hoffmann@fho-emden.de> wrote: Quote: thanks for your analysis. I think it's still not clear how ND is defined. I have not been able to find where XRite has not published this. It would be nice to know for sure. The goal is to mathmatically simulate older densitometers that used red, green, and blue filters, and calculated ND from that. If they are using the full spectro data, and similating the filter and detector response of legacy equipment, the calculation could be fairly complex, even involving nonlinear terms. This is really just a curiosity item, though, isn't it? This is why I think the numbers are nominal. Adobe provides the ability to redefine ND for each ink, for purposes of controlling trap. In some cases, such as for metallics, they recommended setting an arbitrarily high value, to prevent any trapping from occuring. Quote: 1) I have to respect a design, which works well for large format inkjet posters (but not for the toner printer): http://www.fho-emden.de/~hoffmann/hagiabook.pdf Please have a look at the big numbers top right. Hmmm - yes. I see what you mean. Can you talk the designer into a thin black stroke around each letter? (looks like a very interesting book, BTW). Quote: 2) The background is K-only gray. The reason for this decision: If the printers are gradually drfting, then the neutral areas would be affect visually very early. You can go a middle route, using enough K to stabilize the neutrals, and get some relief from the trap situation by using a mixed K value for your background gray - something like CMYK(25,25,25,25). I notice the other gray areas have dark borders, so fringing will not be an issue. Quote: 3) By 'blend', do you mean a visual mixture of background and character color ? Yes, for example a 25% stroke of the color. Quote: This wouldn't work (in my understanding) because this stroke pseudo-trapping are has to be overprinting. For some combination it looks better, if I use only half the charakter color values for the stroke, but I can't see a general law. It's subjective, balancing the dark fringe that you will always get against the lighter one that you sometimes get. You may also see an improvement by using a smaller trap value. If this is a finite run on a single unit, it may be only one specific plate that is out of alignment - mechanical adjustment or shifting the plate in Photoshop might simply eliminate the problem. I suppose you could even use a blurred stroke, to get a progressively lighter trap - this would have the advantage of eliminating the sudden white border, and at the same time soften the edges of the dark border. I notice that you had ND values for the mixtures. The ND values are defined for each plate, not for each mixture of inks. The magenta and yellow plates will trap independently for CMYK(0,100,100,0), with the yellow flowing in some situations where the magenta would not. More curiosity items, as I've been looking into this. From experimenting in Photoshop, it appears that the lighter ink will always be the one to flow, even if it is printed darker than the ink it is flowing under. This seems wrong. For example, a CMYK(0,0,100,0) will flow under CMYK(0,0,0,10), using a 10 percent blend of CMYK(0,0,10,0). The exact value of the blend is dependent on the ink - I'm guessing this is tied to the ND value, so ND affects not only which colors trap, but how much of that color is used when flowing under a darker ink that is being printed lighter. The trap generation in Photoshop seems extremely crude. I hope Illustrator, Indesign, and other programs do a better job. -- Mike Russell - www.curvemeister.com
Back to top
	Ads
Advertising Sponsor
Graeme Gill Guest	Posted: Sun Mar 30, 2008 6:53 am    Post subject: Re: Trapping, Neutral Density Mike Russell wrote: Quote: " With a standard reflectance densitometer, you can get density readings through a Red, Green or Blue filter. (For Cyan, Magenta and Yellow readings, respectively) Also, a third reading is derived from the other three, used to measure black and spot colors. That density is referred to as "visual" by the X-Rite instruments I use." Visual density basically uses the visual response curve (i.e. the CIE Y matching curve, or the V-Lambda function). See Table H-1 of the ANSI CGATS.5-1993 (Or equivalent later document) for the exact definition of spectral responses for Visual and other status measurements. Graeme Gill.
Back to top
	Ads
Advertising Sponsor
Gernot Hoffmann Guest	Posted: Sun Mar 30, 2008 7:32 am    Post subject: Re: Trapping, Neutral Density Graeme Gill schrieb: Quote: Mike Russell wrote: " With a standard reflectance densitometer, you can get density readings through a Red, Green or Blue filter. (For Cyan, Magenta and Yellow readings, respectively) Also, a third reading is derived from the other three, used to measure black and spot colors. That density is referred to as "visual" by the X-Rite instruments I use." Visual density basically uses the visual response curve (i.e. the CIE Y matching curve, or the V-Lambda function). See Table H-1 of the ANSI CGATS.5-1993 (Or equivalent later document) for the exact definition of spectral responses for Visual and other status measurements. Graeme Gill. Then it should not happen that I get by DTP-22 and ColorShop for two samples 1,2: L1>L2, V1>V2, where V means Colorshop's Visual Density. I have Y1>Y2, L1>L2, and expect V1<V2. Best regards --Gernot Hoffmann
Back to top
	Ads
Advertising Sponsor
Roger Breton Guest	Posted: Thu Apr 03, 2008 4:38 am    Post subject: Re: Trapping, Neutral Density Gernot, Quote: Then it should not happen that I get by DTP-22 and ColorShop for two samples 1,2: L1>L2, V1>V2, where V means Colorshop's Visual Density. So you have two samples. Sample A has a higher Lightness than Sample B, while at the same time Sample V has a higher Status T (or E) Visual density than Sample B. This makes sense. Quote: I have Y1>Y2, L1>L2, and expect V1<V2. Isnt' Visual density the contrary of L* and Y? As L* increases from 0 to 100 so does Y (assuming normalized Y, ie 100). But Visual density will decrease as the sample gets lighter. Quote: Best regards --Gernot Hoffmann Am I confusing the issue more? MfG, Roger
Back to top
	Ads
Advertising Sponsor
Gernot Hoffmann Guest	Posted: Thu Apr 03, 2008 8:09 am    Post subject: Re: Trapping, Neutral Density Roger Breton schrieb: Quote: Gernot, Then it should not happen that I get by DTP-22 and ColorShop for two samples 1,2: L1>L2, V1>V2, where V means Colorshop's Visual Density. So you have two samples. Sample A has a higher Lightness than Sample B, while at the same time Sample V has a higher Status T (or E) Visual density than Sample B. This makes sense. I have Y1>Y2, L1>L2, and expect V1<V2. Isnt' Visual density the contrary of L* and Y? As L* increases from 0 to 100 so does Y (assuming normalized Y, ie 100). But Visual density will decrease as the sample gets lighter. Best regards --Gernot Hoffmann Am I confusing the issue more? MfG, Roger Roger, my samples are not A and B but 1 and 2. Sample 1 has Y1, L1 and V1. Sample 2 has Y2, L2 and V2. Y is luminance, L lightness and V visual density. Best regards --Gernot Hoffmann
Back to top
	Ads
Advertising Sponsor
Gernot Hoffmann Guest	Posted: Thu Apr 03, 2008 1:27 pm    Post subject: Re: Trapping, Neutral Density Now I can add a little more confusion, concerning Neutral Density: Edmund Optics about Kodak Wratten Neutral Density Filters: 'These filters are used to reduce light intensity across the visible spectrum without altering the spectral profile.' The transmittance(lambda) function is almost flat. Density Transmittance/% 1.0 10 2.0 1.0 4.0 0.01 A status filter may have (as Graeme says) a spectral transmittance like the V(lambda) curve. That seems to be so in Henry R.Kangs book Computational Color Technology, fig. 15.1, Status T Neutral filter, but the curve is somewhat different. The maximum is shifted to larger wavelengths. Perhaps only a wrong handdrawing ? So far we have two VERY DIFFERENT interpretations for ND. My instrumend X-Rite DTP-22 is basically a spectroradiometer. Everything is calculated based on spectral data. If the visual density was derived by applying the y-bar CMF, then I would get monotonously decreasing density values for monotonously increasing Y and L* values. But that's not the case. G.H.
Back to top
	Ads
Advertising Sponsor
Gerhard Fuernkranz Guest	Posted: Fri Apr 04, 2008 1:30 am    Post subject: Re: Trapping, Neutral Density Gernot Hoffmann wrote: Quote: My instrumend X-Rite DTP-22 is basically a spectroradiometer. Everything is calculated based on spectral data. If the visual density was derived by applying the y-bar CMF, then I would get monotonously decreasing density values for monotonously increasing Y and L* values. But that's not the case. But I guess that the Y and L* numbers are likely computed for a D50 illuminant, while the density is computed from the reflectance only, w/o assuming any particular illuminant (or equivalently, an equal-energy light source, with intensity 1.0 at each wavelength). Regards, Gerhard
Back to top
	Ads
Advertising Sponsor
Roger Breton Guest	Posted: Fri Apr 04, 2008 4:17 am    Post subject: Re: Trapping, Neutral Density Yetz, I understand what you mean by Neutral Density filter. Of course, it has nothing to do with ISO Visual filter which uses V-lambda. Quote: ... If the visual density was derived by applying the y-bar CMF, then I would get monotonously decreasing density values for monotonously increasing Y and L* values. But that's not the case. Interesting. Quote: Then it should not happen that I get by DTP-22 and ColorShop for two samples 1,2: L1>L2, V1>V2, where V means Colorshop's Visual Density. I have Y1>Y2, L1>L2, and expect V1<V2. Assume the following: Sample 1 (L1 =90) (Y1=90) (V1=0.2) Sample 2 (L2=85) (Y2=85) (V2=0.3) This is how I expect the relationship to go, Gernot. L and Y and directly proportional. V is inversely proportional. What sort of numbers are you getting? MfG / Roger
Back to top
	Ads
Advertising Sponsor
Gernot Hoffmann Guest	Posted: Fri Apr 04, 2008 7:54 am    Post subject: Re: Trapping, Neutral Density Roger Breton schrieb: Quote: Yetz, I understand what you mean by Neutral Density filter. Of course, it has nothing to do with ISO Visual filter which uses V-lambda. ... If the visual density was derived by applying the y-bar CMF, then I would get monotonously decreasing density values for monotonously increasing Y and L* values. But that's not the case. Interesting. Then it should not happen that I get by DTP-22 and ColorShop for two samples 1,2: L1>L2, V1>V2, where V means Colorshop's Visual Density. I have Y1>Y2, L1>L2, and expect V1<V2. Assume the following: Sample 1 (L1 =90) (Y1=90) (V1=0.2) Sample 2 (L2=85) (Y2=85) (V2=0.3) This is how I expect the relationship to go, Gernot. L and Y and directly proportional. V is inversely proportional. What sort of numbers are you getting? MfG / Roger Roger, the numbers are here, page 1, bottom: http://www.fho-emden.de/~hoffmann/overprint.pdf V is probably related to L, Y or to the reflectance factor (as Gerhard says) by a negative logarithm. Best regards --Gernot Hoffmann
Back to top
	Ads
Advertising Sponsor
Gernot Hoffmann Guest	Posted: Sat Apr 05, 2008 11:48 am    Post subject: Re: Trapping, Neutral Density Gerhard Fuernkranz schrieb: Quote: Gernot Hoffmann wrote: My instrumend X-Rite DTP-22 is basically a spectroradiometer. Everything is calculated based on spectral data. If the visual density was derived by applying the y-bar CMF, then I would get monotonously decreasing density values for monotonously increasing Y and L* values. But that's not the case. But I guess that the Y and L* numbers are likely computed for a D50 illuminant, while the density is computed from the reflectance only, w/o assuming any particular illuminant (or equivalently, an equal-energy light source, with intensity 1.0 at each wavelength). Regards, Gerhard Gerhard, I think you're right: The indicated Visual Density doesn't depend on the illuminant and it doesn't depend on the status filter (ColorShop). A model for the calculation (correct ?): 1. Measure spectrum 2. Divide by instrument's light spectrum 3. Multiply by y-bar and integrate for Y 4. In mode Absolute use Y=1 for the ideal reflector (calibration by white reference) 5. Calculate V=-log(Y) 6. Rename V --> ND Remark: this is obviously not a VISUAL density. Best regards --Gernot Hoffmann
Back to top
	Ads
Advertising Sponsor
Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First

	Physics Talk Forum Index -> Color and Light	All times are GMT Goto page 1, 2  Next
Page 1 of 2

Jump to: Select a forum Physics Talk----------------Physics Talk-- BiophysicsMechanicsColor and LightRelativity  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum

Australian Debt Consolidation Experts medical insurance Wedding Ring Annunci di escort, accompagnatrici e massaggiatrici Naughty British Wives Sex Contacts Ruby Make Your Own Website Free calls to Poland Cleaning Service mold killer UK Swingers Genuine Contacts Site Janitorial Supplies Hoover Vacuum Parts

93 Attacks blocked

Powered by phpBB © 2001, 2005 phpBB Group