QTR-Quadtone RIP

> 
> The idea came from the way I decided to go about creating my basic
> curves. Firstly, the Carbon curve at 1 extreme and the Cool curve at
> the other extreme, so in my mind by using the blending option in QTR
> I should theoretically move up and down the Lab B axis (give or take)
> and then if I spend time producing the best standalone neutral curve
> that I can for my setup, I should in theory, be able to move along 
> the Lab a axis from a roughly neutral position to where ever I 
> wanted to go (obviously image dependant) using the Lab a axis toner 
> curve. Maybe my thinking is to simplistic and someone could jump in 
> with a few ideas,
> 
> Jamie.
>

Jamie and Joost,

I've followed this correpondence and thought now would be the time to 
jump in. Like Jamie, I'm playing with UT-3D inks in a 1290 and have 
recently started using QTR. Like both of you, I've felt that 
mixing 'standard' curves was the way to go for producing a full range 
of tones. I've implemented this approach - kind of - in photoshop, and 
am working on it for QTR also.

I've spent quite a bit of time in the last few weeks developing 
Photoshop curves for this inkset, and have generally been pleased with 
the results. Although it's quite fiddly and time-consuming (especially 
as I'm using a flatbed scanner to read the test strips) I've been able 
to produce pretty much exactly what I've wanted, at least within the 
accuracy of the scanner and the gamut of the inks. The approach I've 
taken has been a little different to what I understand the two of you 
have done, in that I've aimed for starting curves that sit on the a- 
and b- axes as far as possible. I found that printing 'fully cold' 
produced a noticable greenish cast in some tones, so my 'cold' curve 
has some a-axis toner in it. Similarly, my neutral curve is tweaked to 
come as close as I can get it to pure grey. 

It seems that the gamut of the inkset makes an on-axis curve harder to 
achieve in the warm tones, although I did manage to produce a very 
slightly warm (lab b around +2)curve with lab a around 0 in the 
midtones and highlights. (I was really more interested in getting the 
green out of the cool tones, as I quite like the natural carbon tone). 

I've had some success 'interpolating' between my neutral, cool, warm 
and carbon curves by blending in Photoshop. To do this I applied the 
curves to separate copies of the image, flattened each and overlaid 
them with opacity between 0 and 100%. This seems to produce broadly 
what I was after (I linearised my original curves 'the hard way') but 
I'm sure creating a new ICC for the blend would do a better job. 

I didn't follow this any further because the limitations of the Epson 
driver were becoming apparent, with horizontal banding evident in 
spite of my best alignment efforts and no way to print at 2880 on 
heavyweight matte paper. So I dived into QTR.

At this point, I'm just getting started, but I've had a look at 
Joost's curves and re-linearised the carbon one for my printer, just 
for fun. I then made a pure carbon curve from scratch, mostly as a 
learning exercise but also to form a view on the relative densities 
and best ink limits for the carbon inks for this printer. So far so 
good.

My intention is to follow the same approach as I used with Photoshop - 
I'm working on a neutral curve at present. I'm keen to get into the +a 
tones also. A thought I had looking at Joost's 'selenium' curve: the 
Eboni ink seems to have a natural +a / -b tone, with suggests to me 
that an approach with just Eboni, a-axis toner and one or both Carbons 
to bridge the density difference, could be worth trying. I'll let you 
know how I get on with it - unless someone else has been there first!

I hope at least some of that is interesting and/or useful. My PS 
curves are available to anyone who wants them, and I look forward to 
further developments.


All the best,
Ben Connor