QTR density inconsistencies in blocks of tone

[davido]

hello

I've been working away learning QTR. After printing many step tablets, I've realized that I'm getting density inconsistencies within blocks of similar tone.
I'm using a 3800 printer and printing on OHP.
The step tablet has a 90% dark border (K in photoshop). When measured with my x-rite 361-T with the UV setting, I'm getting a range of measurements sometimes .10 log apart.
The denser part of the border is that which is parallel with the print head and, therefore, the longest length of printing. The border which is perpendicular to the print head has .05 to .10 less density. I'm not sure if there is connection?

I'm not sure what is causing this? I have changed from the Uni-directional to the bi-directional, which doesn't seem to make a difference.
I have the dither set at "ordered".

Could this perhaps have something to do with the OHP having inconsistencies in the ceramic coating?

thanks
david

 

[gmikol]

David--

First thought, could this be the result of banding from a clogged nozzle? In the long dimension (direction of print head travel) you might be measuring in an area where the clogged nozzle doesn't cover, but where you're measuring in the short dimension (direction of paper feed), you might be measuring in an area where there is banding.

Second, wow wide is the border around the step tablet? The 361-T typically uses a 3mm measurement aperture, so you need an absolute minimum of 3mm (1/8") of area to measure, and ideally, 1.5x to 2x that to take into account positioning inaccuracy, flare, and light piping through the substrate. Is the rubber on the bottom of the measurement head (that makes contact with the OHP), in good condition, or is it hard or cracked?

Also, have you taken repeated readings from the exact same place? How much variation there? Is it possible you've got some electrical noise problem that's causing the light source to vary or the noise is coupling into the electronics?

Just some more things to look into...

Good luck.

--Greg

 

[davido]

Thanks for your reply, Greg.

I have been taking three readings of each spot, they are consistent.
The border is 1/2 inch thick.
The border is wide enough that, I would think, banding would be fairly evident.
I think it might have something to do with the dither: I do recall reading something that large areas of the same tone can come out mottled ie. skies.

david

 

[sanking]

David,

I don't think this is all that unusual. When printing with either the Epson driver or QTR I have also observed a density difference of up to log 0.10 when measuring different areas of the same tonal block. And I see this with two different densitometers, a Gretag D200 and an X-Rite 361T. I don't really do anything to counter it, just measure in the center and record the reading as is. If one of the blocks is out of line with its neighbors I will measure it several times and take an average of all of the readings.

Sandy

 

[Ben Altman]

I've had my suspicions about this for some time and have been meaning to do some tests. In particular I'm wondering if the variations happen more near the edges, where the dither math gets tricky. In other words, due to the width of the head not all nozzles may be available at the end of the carriage. That would be consistent with your observation that the parts printed at the ends of the head movement are different, David.

Interesting, Sandy, that you see this with the Epson driver too - is that anywhere on the sheet?

If it is an edge effect, or any predictably recurrent effect, the consequences for profiling are not altogether trivial. It may be worth printing a wider border around test wedges, or at least positioning the edge of the print away from the limits of the head travel.

On a related topic, when exposing step wedges - which are typically small - uneven edges of the coating and differential humidity changes during exposure near the edges are possible problems.

I'm moving towards more randomized type step wedges with multiple scattered patches of each step to try to reduce some of these errors.

 

[PVia]

The question should be "How does the print look?" in the particular process you're using. Can you see the difference in a print?

Also, this is the difference between an inkjet neg and the beautiful continuous tone of film. IMO, digital inkjet negs are amazing and convenient, but film is amazing-er. ;-)

 

[R Shaffer]

The question should be "How does the print look?" in the particular process you're using. Can you see the difference in a print?

Fur sure... the proof is in the pudding.

I would have to assume, based upon everyone's input, that my digi-negs are equally inconsistent, although I have no way of measuring it. But my prints and test strips appear to be quite uniform in tone.

I think that the heavy texture of Rives BFK contributes a lot more to non-uniform appearance than the negative. But it is the texture that I love so much.

 

[sanking]

Fur sure... the proof is in the pudding.

I would have to assume, based upon everyone's input, that my digi-negs are equally inconsistent, although I have no way of measuring it. But my prints and test strips appear to be quite uniform in tone.

I think the slight difference in density that we are observing is kind of a bogus issue that results from the fact that the UV densitometers use a blocking filter that measures light in a fairly small bandwidth, about 20-30 nanometers on each side of the central bandwidth of about 370 nanometers. That excludes a lot of radiation to which the alternative processes are sensitive, especially radiation in the far visible range of 400-450 nanometer range.

A UV densitometer is a useful tool but in my experience the density range that it suggests with our negatives is not always identical to the actual printing exposure scale, and sometimes it is off by quite a bit. I recall one case with either the Epson 2200 or HP 9180 where the UV density jumped from log 2.5 to log 5.0 from on block to the next, but the two blocks looked and printed almost identically.

Sandy

 

[Ben Altman]

Agree completely - the densities of the print are what matter and don't by any means always track with the measured UV densities of the neg.

OTOH when using a step tablet that is broken into three rows I sometimes see suspicious flat spots and jumps in the readings (from the print) at the ends of the rows. May be just my faulty printing technique and nothing to do with the neg...