litho negs

Hi,

I had a 31 step tablet made on litho film and I printed it using VDB and I
notice something that may (or not) help others that intend to have some
litho negs done. Practically half the steps printed to the darkest value
possible and then slowly lighten to paper white without very good separation
while a standard Stouffer step tablet printed on the same print showed a
more even (linear) decreasing in tonality starting from step 1 up to step
11. I mention these observations because it became clear to me that the
machine used, (I'm not sure it's an image setter) produce a linear opacity
increase while a Stouffer asn't a linear increase in opacity, the difference
between each step is always a factor of 2 and it makes a world of
difference. To have a litho produce a similar response to what a Stouffer
does here is what you could do. Thinking in terms of opacity, you can check
my thinking, say a square inch of the darkest value on the litho film is
made up of 2400 by 2400 dots, the number of DPI doesn't matter in the end,
now you want the next lighter step to be 5% less opaque, it would need to
have 95% of this 2400*2400 number of dots and we would need to no how many
dots that make on each side of the square thus sqrt(0.95 * 2400 * 2400)(1)
and this is equal to sqrt(0.95) * sqrt(2400*2400)(2). We could then find out
how to translate our 0-255 pixel values by scaling equation (1) or (2) which
I'll rewrite below.

(3) sqrt(opacity%) * sqrt(DPI*DPI) / DPI * 255

The part sqrt(DPI*DPI) / DPI simplifies to 1 because sqrt(DPI*DPI) = DPI
then DPI/DPI = 1 so we are left with equation 4.

(4) sqrt(opacity%) * 255

To convert this to negative value which we need, we substract the result of
equ (4) from 255 and round it up to have pixel value that relate to opacity.

(5) round( 255 - sqrt(opacity%) * 255)

 Here is a table you could use to create the curve to use in photoshop or
whatever you're using. It was made using equation 5 for opacity varying from
0% to 100% in 5% increments. Actually the step vary from 0 to 1 in step of
0.05.

step input output
1 255 255
2 242 198
3 230 174
4 217 156
5 204 141
6 191 128
7 178 115
8 166 104
9 153 94
10 140 84
11 128 75
12 115 66
13 102 57
14 89 49
15 76 42
16 64 34
17 51 27
18 38 20
19 26 13
20 13 6
21 0 0

If you plot these values you would notice that from step 5 and up the curve
is almost a strait line meaning you probably don't need to use all 21 values
to create your curve but it's possible you may need to compute a couple more
points the first few step depending on how your software smooth the curve in
this area.

I assume something similar was done in the print engine of our computer
otherwise the prints would be awful.

Regards
Yves
Received on Fri Apr 7 07:09:34 2006