How to produce an effective/accurate dot gain curve
2012-11-05 by lougray1234
Yahoo Groups archive
Index last updated: 2026-04-28 23:12 UTC
Thread
2012-11-05 by lougray1234
Hi Guys, I'm trying to work out the procedure for producing an effective / accurate dot gain curve. I have access to an i1 spectro. I have read the User Guide from the QTR website, which suggests starting with values 6 for highlight and shadow, and Gamma=1. However, there does not appear to be any information on how to go about refining the curve empirically (e.g. patch target to use + calculations etc...). Any information at all would be greatly appreciated :-) Lou
2012-11-05 by Paul Roark
One write-up of how I use or recommend people start with QTR starts at page 11 of http://www.paulroark.com/BW-Info/Eboni-4-Plus.pdf . I think the approach would apply to most inksets. I rarely use the tabs to control the inks aside from the linearization tab. Usually, the default QTR partitioning curve plus linearization is all that is needed. Determining the individual ink limits and densities are the key inputs that you'll need to determine. Even there, QTR seems relatively tolerant of variation. I do tend to use a black ink limit that is slightly less than where the dmax was -- using the Black Boost for the end of the black curve. But this is not necessary either. I'd recommend you start simple. Paul www.PaulRoark.com [Non-text portions of this message have been removed]
2012-11-07 by lougray1234
Thanks for your response, Paul. That info is very handy. I managed to also find a previous thread that you wrote in, describing how to determine max black and "Black Boost" in a more empirical way. However, I'm still a little confused. The QTR manual offers a way to compensate for dot gain, yet below, you seem to be suggesting that the dot gain curve in QTR is not a very significant quality factor. I am trying to understand why this is the case? Is it because the black / gray curve linearization process is a method in itself for compensating for dot gain? Or, is it more a case that dot gain on an inkjet (using inkjet paper) is so minimal that it is an insignificant issue? Many thanks again, and apologies for my annoying questions. --- In QuadtoneRIP@yahoogroups.com, Paul Roark <roark.paul@...> wrote:
> > One write-up of how I use or recommend people start with QTR starts at page > 11 of http://www.paulroark.com/BW-Info/Eboni-4-Plus.pdf . I think the > approach would apply to most inksets. > > I rarely use the tabs to control the inks aside from the linearization tab. > Usually, the default QTR partitioning curve plus linearization is all that > is needed. > > Determining the individual ink limits and densities are the key inputs that > you'll need to determine. Even there, QTR seems relatively tolerant of > variation. > > I do tend to use a black ink limit that is slightly less than where the > dmax was -- using the Black Boost for the end of the black curve. But this > is not necessary either. > > I'd recommend you start simple. > > Paul > www.PaulRoark.com > > > [Non-text portions of this message have been removed] >
2012-11-07 by Paul Roark
lougray1234 <lougray1234@...> wrote: > ** > > > ... The QTR manual offers a way to compensate for dot gain, yet below, you > seem to be suggesting that the dot gain curve in QTR is not a very > significant quality factor. > ... > I ignore the concept of dot gain. Since I work with actual densities, I don't care how much of that actual density comes from the spread of the ink as opposed to some theoretical density calculated by the size of the dot with no spread. I think you'll find that if you use the simple approach I use, the pre-linearization curves will be well within what QTR can successfully linearize. I do recommend that you graph the pre-linearization Lab L values. (I open the spectro results in Excel to do this.) Don't expect the curve to be perfectly linear. Smooth curves linearize the best. A sharp turn at a point is the most likely to cause trouble. These usually do not occur. Good luck -- just dive in and give it a try. Paul www.PaulRoark.com [Non-text portions of this message have been removed]