add some noise. Filter->Noide->Add Noise. Use non-gaussian, about 1, then again (Control+F).
A slight variation: Duplicate the layer that has the banding, view the individual channels, usually it’s one or two channels doing the banding.
Apply the noise to those channels. Then apply a layer mask and mask out the areas of white.
Applying noise will result in a scum dot in the whites, that’s why I’m recommending the layer mask that keeps those areas white..
What is the CMYK build that is going to white, and what is the distance over which the change takes place?
It sounds like you have previously used vector-based blends (ID or QXP) and doing so to a ps level2 rip means bands. Period. If the bands are thin enough, there is no problem. If they are much wider than say 1/16 inch they could cause, a problem.
My guess is if you use the Photoshop gradient tool, and make sure that "dither" is checked off in the tool’s options, you can simply create a gradation that will not be banded (by virtue of dithering). You should not have to add noise, the dithering will breakup the bands.
I can guarantee the result will be an improvement over the vector-based solutions.
Also the PS gradation will transition to white much more smoothly than the vector gradation which is purely linear and ends rather abruptly. To make the transition even smoother, you can extend the canvas in PS and add a layer that will be for a white-to-transparency gradation right on the very end, where the color is going to white. You can transform this layer and play with the position of it to assure the smoothest transition to white, though granted it may be difficult to see. You might add a temporary curves adjustment layer that imparts a very severe darkening effect on the color gradation to highten the visibility of what happens in the transition-to-white zone.
or work with the lights off, which really helps…adding a mask with noise on it where it hits white helps as well…basically adding noise helps gradients in CYMK and its not even noticable below 10% even on very good print…
John,
I’m confused-I always thought that "dither" reduced banding and, in fact, hovering over the dither option states "Dither to reduce banding"
Please elaborate.
Barry:
I’m not sure where the misunderstanding lies, either in my writing or your reading, but I never intended to imply that dithering induces banding… quite the contrary.
My writing style tends to confuse… perhaps that is the problem.
basically adding noise helps gradients in CYMK
But as alluded to earlier, when performed on a CMYK file will add a uniform, albeit slightly noisey, 1% dot in all 4 channels.
Beside "dirtying" up those areas which should be pure white, this also adds a uniform highlight dot to the black channel, even in a gradation that involved no black to begin with.
In the printing game, black in the highlights has never been, and never will be, a good thing.
It’s getting a little clearer now, I think.
Thanks.
I thought John was quite clear, you shouldn’t need noise if you are using dithering.
My poor addled little brain is having difficulty wrapping around this (from post #3): "My guess is if you use the Photoshop gradient tool, and make sure that "dither" is checked OFF in the tool’s options,
you can simply create a gradation that will not be banded (by virtue of dithering). You should not have to add noise,
the dithering will breakup the bands."
….now I see the misunderstanding.
When I wrote that, I meant "checked off" as in "selected", like checking off a selection on a ballot.
I never meant that the option should be turned off.
One could easily interpret that to mean the opposite of what I intended.
Barry, you and I have always been on the same page, I think. Let us remain so.
You only have 256 levels of gray to work with in each channel. So if the gradient is long enough, banding will result (the same is true in masks). Leave dithering checked, but you will probably have to go into the channels and apply noise. As little as possible. Usually 1-2% gets rid of it. If the gradient is really large, more might be needed. Where it fades to white, I would try the mask suggestion. Don’t use vector gradients. GOOD LUCK!
peace
