Guidelines for Creating Good Halftone Images?

[Criminy]

Are there any short and sweet guidelines for me to share with my clients regarding the creation or manipulation of halftone images? What will render well and what won't?

For instance, I've always heard it's good to keep your highlights no lower than 3-4% and your shadows no darker than 94%. Of course, compensating for paper stock is a given...

Is there a standard I could adopt as good common practice? MUCH THANKS!

 

[Lukas Engqvist]

Well that is true for film and if you have your photo's in a frame. MAny of todays design need to burn out to paper colour, and a CTP can usually create a point of 1%. I would say those values are old school. I would be careful with the shadows though because dotgain still means that the shadows risk loosing detail.
On the other hand I'd normally work in RGB and let ICC profiles take care of the translation to %. For me then what needs to be white (or transparent, I'd be sure it has not as much as a decimal % coverage) and that dark tones have full range but no flat areas in the shadows. Is that too phragmatic?

 

[Criminy]

Thanks for your input! Yeah, the values I stated were probably vestigial prepress teachings. Just thought I'd throw them out there to see if they still held weight. I've been asked to get a consensus of opinions. I've had to lighten and sharpen images for a client, whose work prints on newsprint. The stock is so gray that their images appear muddy or downright dark. Compensating for the stock often means pumping light into their images, and adding detail by using the unsharp mask feature in Photoshop.

 

[gordo]

[SNIP]I've had to lighten and sharpen images for a client, whose work prints on newsprint. The stock is so gray that their images appear muddy or downright dark. Compensating for the stock often means pumping light into their images, and adding detail by using the unsharp mask feature in Photoshop.

Using unsharp masking does not add detail. It actually destroys detail. What it does do is the "secret" to getting good B&W presswork on newsprint.
What the USM filter does is increase contrast where there is a tone change. The key is contrast. Whenever possible you should try and increase the contrast between tone areas. For example, in this image:

the original at top will not reproduce as well as the edited image below it where tone transitions have been exagerrated.

Rent the video "I, Robot" (starring Will Smith) and study the lighting that was used. You will see this technique applied in many scenes done with careful lighting and shadow placement - always dark against light rather than low contrast tones that will blend together. This gives some scenes an almost 3D appearance.

Also, tone transitions within images should be short and not subtle. The best place to see examples of that technique is to look at car and movie advertising in your local newspaper. You'll find that the ones that look best are the ones that have very few greys and show these short, abrupt, tone transitions.

So, it's not so much about making images lighter - but of increasing local contrast.

You might also try going to a coarser halftone screen - e.g. 85 lpi will look contrastier/snappier than 100 lpi.

best, gordon p

 

[leetaylor]

Using unsharp masking does not add detail. It actually destroys detail. What it does do is the "secret" to getting good B&W presswork on newsprint.
What the USM filter does is increase contrast where there is a tone change. The key is contrast. Whenever possible you should try and increase the contrast between tone areas. For example, in this image:

the original at top will not reproduce as well as the edited image below it where tone transitions have been exagerrated.

Rent the video "I, Robot" (starring Will Smith) and study the lighting that was used. You will see this technique applied in many scenes done with careful lighting and shadow placement - always dark against light rather than low contrast tones that will blend together. This gives some scenes an almost 3D appearance.

Also, tone transitions within images should be short and not subtle. The best place to see examples of that technique is to look at car and movie advertising in your local newspaper. You'll find that the ones that look best are the ones that have very few greys and show these short, abrupt, tone transitions.

So, it's not so much about making images lighter - but of increasing local contrast.

You might also try going to a coarser halftone screen - e.g. 85 lpi will look contrastier/snappier than 100 lpi.

best, gordon p

Excellent, accurate advice.

 

[Criminy]

Sweet! Thanks, Gordo for the helpful tips. This is exactly what I was hoping for. I am by no means a Photoshop guru. I can make minor color corrections using levels and brightness/contrast adjustments. To date, I've been able to make changes to specific trouble spots to the client's satisfaction. However, it would be nice to be able to tell THEM how to make these adjustments to their own work prior to giving us files for output. I'm going to start experimenting a bit more, and dig out a few more reference guides to gain a better understanding of the behavior of shadowing.

I've got a copy of "I Robot," and will be reviewing it for the scenes you described!

 

[gordo]

I've got a copy of "I Robot," and will be reviewing it for the scenes you described!

Great, here are a few scenes from I, Robot, to show you what I'm talking about. Although it's a color movie, it's colors have been reduced to near neutrals i.e. B&W, so the the cinematographer used dark against light to create contrast and separate the figure from the background and add dimension and clarity to the scenes.

The Japanese have a word for it: "Notan" (which you can look up on the interweb). It's a technique that has been used in western art since the popularization of Japanese art in the 19th century.
If you ever watch old episodes of the original Star Trek TV show you'll see a similar use of lighting for the same purpose. Because they couldn't guarantee that the show would be seen on a color TV, rather than black and white, they used lighting to create this kind of contrast to clarify the image rather than relying on color.

Same thing for newspaper work - you've realistically only got about 6-10 levels of grey to work with to define the image so these "tricks" help to deliver an image that "pops" off the page. If you can't get it in the original shot, then a bit of post-processing in PShop can help.

There, I bet you didn't think you could learn about newspaper reproduction by watching vintage TV shows! LOL!

best, gordon p

 

[otherthoughts]

Firstly, Great posts as always Gordo!

Just to amplify somewhat and add my own perspectives to the thread from a scanning point of view.

I always liked to use the analogy that "Unsharp masking is like a Magician's illusion".

The Magician's secret must not be revealed to the audience, or the illusion will be spoiled. In other words, if the unsharp masking levels your thinking of using become too obvious to your audience, then your illusion is revealed. So don't get too carried away with it.

For a bit of background.
The reason unsharp masking was initially introduced, back in the long forgotten days of camera separations way before computers, was to help compensate for the vast gulf in dynamic ranges between film positive originals(3.5 DMaxx) eg, Kodachromes, Ektachromes, etc. and what the printing process was capable of(2.10 DMaxx, on a good sheet-fed #1 coated sheet day). With Newsprint printing being the most challenged of all.

Fast forward about 3 decades and not all that much has improved regarding offset printing's dynamic ranges.

Like Gordo said "Contrast is the key". Unsharp masking is one way to provide the illusion of more contrast than is actually present.

Perhaps a more significant area of contrast to pay attention to is the gradation/curves/F-stops/aperture/shuttter-speed used to acquire/scan/shoot your images.

For example, you wouldn't use the same curves to scan a product shot of(highlight to 1/4 tone) white/cream/eggshell colored fabrics as you would for an image of(3/4 tone to shadows) with a black-cat, dark-suits and black shoes in a dark closet would you? Most basic scanner curves devote the bulk of their tonal range/contrast to the mid-tones. Use the type of curve that will serve each image best.

If you scan the image with the wrong curve or shoot the subject with the wrong exposure. You will never be able to color correct/retouch the images as well as if you had done so.

As an extreme example, one company that I had the privilege to visit, often developed their Kodoachrome film positives in such a way that their DMaxx reached a 4.0 level. The scanner operators that I trained there, often scanned the same image three times using three different curves in order to capture every level of tonality within the image. Then the prepress operators cut and spliced portions of the three scans back together using the best segments of the three scans in order to exaggerate the contrast in each of the three tonal ranges and enhance the most captivating portions of the image.

Best Regards
OT