Thursday, October 28, 2010

How AM and FM screening equivalencies are measured


I'm often asked about what AM/XM halftone screens are equivalent to a certain FM screen - i.e. "What AM/XM screen is equivalent to a 20 micron FM screen?"

There are two ways that this halftone screening equivalency is usually measured.

One is equivalency of detail rendering - the ability of the screening to render image detail. The other is lithographic equivalency - how they perform on press lithographically. Note that in both cases, because the respective screening technology is so different, equivalency can only be an approximation.

Equivalency of detail rendering
Since halftone dots form the printed image - more dots per linear inch translates into more detail that can be rendered.

With an AM screen the detail rendering ability is specified in lpi (or lpc) - i.e. halftone dots per inch (e.g. 175 lpi or 60 lpc).

Since an FM screen has no "lines per inch" determining the equivalent detail rendering equivalency is usually done by drawing a line through the FM screen and counting how many dots are intersected (crossed) in a distance of one inch.
Measuring the relative lpi of an FM screen.

The above example shows an FM screen enlarged. The distance measured is 1/16th of an inch. In that 1/16th of an inch approximately 36 dots are intersected. So, in one inch about 576 dots would be intersected (16 x 36). Put another way, there are 576 dots per linear inch - 576 lpi - to render detail, i.e. this FM screen is equivalent to a 576 lpi AM/XM screen.

Lithographic equivalency
Lithographic equivalency is a bit more complicated to figure out. It is usually measured by counting the number of edges (transitions) in a square inch.
Measuring the number of edges of an AM/XM screen.

Measuring the number of edges of an FM screen.

Halftone screens with a similar number of edge transitions will have similar lithographic properties.

AM/XM equivalents of some popular FM screens.
Keep in mind, these are approximations only, however they do give a good indication as to screening performance.

Saturday, October 23, 2010

The Job Interview

One of the most challenging aspects of working at a hi-tech company is actually getting the job in the first place. When I was hired by the graphic arts company Creo, in 1997, I not only faced the usual human resources questions, but also a slew of engineer-developed questions intended to reveal the job applicant's thinking process and how they went about solving problems.

Here are a few of the questions that prospective Creo employees had to negotiate back in '97. Unfortunately, except for the one about drawing a bicycle, I haven't a clue what the answers to the other questions are. I wouldn't be surprised if the engineers asking the questions didn't have a clue either!

What color is crab's (or shrimp's) blood and why?
• Please draw a bicycle in as much detail as you can.
Job applicants draw a bicycle - luckily they don't have to ride it.

• How do you put a giraffe into a refrigerator? Now, how do you put an elephant into a refrigerator?
• There is a square room. In each corner is a bug. Each has a mission: to walk directly towards the bug on its right. They walk at the same speed and begin at the same time. 
a) Do they ever meet? If yes, where? 
b) What is the length of the path they travel if they do meet?
• What is the sum of all numbers between 1 and 100?
• A company grows trees. Each year the tree grows, it can be cut down and sold for $1 times the square of its age in years. At that point, the money can be put into the bank and will receive 10% interest per year. To maximize profit, should the company ever cut down and sell a tree? If so, at what age?


OK, if you’ve read this far you may be scratching your head and wondering if you should have applied somewhere else. Well, let’s look closer at one of the questions – the one about putting a giraffe in a refrigerator – to see how the engineer’s subtle mind works during the interview process.The correct answer is pretty straightforward: open the refrigerator, put in the giraffe, and close the door. Any other solution would suggest that you tend to do simple things in an overly complicated way. Now, how do you put an elephant into a refrigerator? If you said: open the refrigerator, put in the elephant, and close the door – well that would be wrong. The right answer is: open the refrigerator, take out the giraffe, put in the elephant and close the door. This tests your ability to think through the repercussions of your previous actions.

If the engineers were particularly rascally during the interview, they might follow up with another question like the one about the Lion King hosting an animal conference. Apparently all the animals attend the conference except one. So, which animal was that? Well obviously, at least for an Einstein anyway, the correct answer is the elephant since you just put him in the refrigerator!

As far as the answers to the other questions are concerned…well, if you haven’t figured them out on your own, and since this was in 1997, you'll have to wait a year for Google to be created.

Saturday, October 16, 2010

The Wayback View – 1964, The Internet imagined

1964 was the year that the Beatles first invaded America. China tested its first atomic bomb and, finally, the US Surgeon General warned against cigarette smoking.

In 1964, there were around 1.5 million people, in the US, using this mobile phone which ran on AT&T's network and was called their “Improved Mobile Telephone Service.”
IBM introduced the System/360 computer that was available with up to a whopping 8 MB of internal main memory.
And the Internet of the future was imagined:

Tuesday, October 12, 2010

Goodbye JPEG - hello WEBP?

Google is introducing a new open-source image format: "WebP" (pronounced ‘weppy’).

Google claims that images in the WebP image format will be close to 40 percent smaller than JPEG files while providing for images that are of higher quality by virtually eliminating the image artifacts associated with JPEG compression. At the present moment, WebP is still in a very early stage of development and hence, unlike the JPEG file format, WebP is not yet built into cameras, web browsers, image-editing programs, etc.

JPEG vs WebP compression at 100%
From left to right: Original image, JPEG compressed, WebP compressed.
JPEG vs WebP compression enlarged
From left to right: Original image, JPEG compressed, WebP compressed.

WebP uses the Y'UV color model that is used in the NTSC, PAL, and SECAM composite color video standards. It is a bit like the LAB mode color model that is used in PhotoShop, and other imaging applications, in that the Y component, like the "L" channel, determines the lightness of the color (referred to as luminance or luma), while the U and V components, like the "a" and "b" channels, determine the color itself (the chroma).
The Y'UV color model - From left to right: Original composite image, "Y" component, "U" component, "V" component.

For image compression, WebP uses the VP8 video codec - the same methodology that is used to compress keyframes in videos. This codec uses predictive coding to encode an image by using the values in neighboring blocks of pixels to predict the values in a block, and then encodes only the difference (residual) between the actual values and the prediction. The residuals typically contain many zero values, which can be compressed much more effectively.

So far, while it is still a "lossy" compression technology, WebP compressed images certainly appear to deliver a higher level of compression (and thus smaller file size), and much higher image quality than the current standard JPEG image compression method. This is a good thing, not only for images for web application, but possibly for print application as well.

Google Chrome will likely be the first consumer application to support "weppy" compression in order to provide a faster user experience on web sites while reducing bandwidth and hosting costs.

On the other hand

Since the Y'UV color model is similar in principal to the Lab color model...why not apply the same compression methodology to an image that's in Lab mode? I've done some testing with high JPEG compression of the "a" and "b" channels of a Lab image and easily achieve file sizes that are only 20% of the original file size with no apparent image degradation. In fact they look very similar to the results obtained with the WebP image format.
Left image WebP - right image Lab compressed using very high JPEG compression for the "a" and "b" channels of an Lab image.

Left image WebP - right image Lab compressed using very high JPEG compression for the "a" and "b" channels of an Lab image.

If the Y'UV compression method could be applied to Lab images then the graphic arts industry could continue to use a color model that is well understood and in use today rather than import a new color model from another industry.

Friday, October 8, 2010

CtF - Computer to Foam - Printing the cloud

"FLOGOS" - flying logos - is a biodegradable foam-helium mixture. Templates form the foam into the desired shape creating small bubble clouds which then float through the air. Output rate is approximately 100-250 FLOGOS per hour, per generator.
Yes, it's printing to the cloud.

Friday, October 1, 2010

To linearize your CtP plates or not?


A bit of background
Back in the old film to plate days the standard prepress procedure was to linearize film output. That means a specific tone request in the original file results in halftone dot in the film equal to the file tone request. So, for example, a 50% in the file became a 50% tone in the film. Linear film was the agreed standard interchange file format between prepress tradeshops, publishers and printers. At that time, the final tone on the plate was not measured. Instead, the resulting tone in the presswork was measured and deemed to be in specification, or not, relative to the supplied linear film. I.e. At 133 lpi, a 50% tone in the film resulting in a final tone of about 71% in the presswork would be considered in specification. Interestingly, although the film was linear, the resulting plates were not linear due to the dynamics of exposure in the vacuum frame.

The arrival of CtP in the late 1990s eliminated film as the intermediary. As a result, measuring tone values on the plate became a process control metric. However, CtP plates seldom have a linear response to laser exposure and if a tone reproduction curve is applied to them to make them linear - the resulting presswork is usually too "sharp" - i.e. not achieving enough dot gain.

At the same time that CtP was rapidly being adopted, printers also began to use finer halftone screens, including FM screens, which had very different dot gain characteristics compared to the old published standards. Printers began to leverage the flexibility that CtP provided in being able to apply different tone reproduction curves to their CtP plates to achieve the tone reproduction on press that they required.

So the question for the printer becomes: should prepress first apply a curve to linearize the plate and then, if needed, apply another curve on top of the first to achieve the desired final press tone response?

I was shocked
So, just to confirm that the method that I have been using for the past 13 years was indeed the standard method used in the industry, I posed the question to an internet printer's forum: "Do you linearize your plates before applying a press curve (a two curve workflow - e.g. one to linearize the plate followed by another one to compensate for dot gain) or do you only apply a press curve to the uncalibrated plate (a one curve workflow - e.g. one to compensate for dot gain)?"

The response shocked me - a whopping 70% said they first linearized the plate with a curve and then applied a press curve while only 30% responded that they simply applied a press curve to the uncalibrated (natural state) plate.

70% using one curve on top of another? That makes no sense to me at all.


In a film to plate workflow, linear film is exposed to the plate in a vacuum frame. The function of the plate exposure is to reproduce the halftone dots in the film as consistently as possible across the surface of the plate, and perhaps more importantly, to create a robust halftone dot on the plate that will maintain its integrity on press. However, although the film may be linear, the resulting plates are not linear due to the dynamics of exposure in the vacuum frame. In North America using negative film there is typically a 2%-5% dot gain on plate at 50% (i.e. 50% in the film creates about a 54% on the plate) while in Europe and Asia where positive film was used there is typically be a 2%-5% tone loss at 50%.

In a CtP workflow, as with a film to plate workflow, the important thing is to set laser exposure and processing (or lack thereof) to the manufacturer's specifications so that the result is a robust halftone dot on the plate that maintains its integrity on press. However, as with a film workflow, the resulting plates are typically not linear due to the dynamics of laser exposure, individual plate characteristics, and processing.

In this example, the thick line that dips below the 0 line is the natural uncalibrated plate curve after the engineer has done their work setting up exposure and processing for the most robust dot possible.With this particular positive thermal plate the uncalibrated plate curve results in a negative value through the tones. The bottom numbers in the graphic are the requested tone values in the file - 5%, 10%, 20%.... 90%, 100%. The "0" line represents linearity. I.e. if the plate was linear then that 0 line would be straight and be the "plate curve". But, in this case, a 50% request has resulted in about a 47% on plate. This is fairly typical - a well and properly exposed CtP plate does not have a linear response (i.e. a straight line). Also note that it is typically not a classic Bell curve - there is no symmetry. Different CtP/plate combinations will each have their own characteristic natural curves.

So, from a CtP vendor engineer's perspective, it does not matter whether the result of their setup is a linear plate or not since a tone reproduction curve can always be applied to achieve whatever tones are required on plate - including linearizing the plate. What's important is that the exposed dot is robust and that the plate imaging is consistent across the plate and repeatable from plate to plate.Put another way - the key criteria is that when properly set up the plate will have a characteristic non-linear tone response. And that's fine - as long as the plate responds the same - i.e. delivers the same non-linear tone response – every time because without that consistency it is not possible to build any tone reproduction curves at all.

Some definitions

These definitions are not "official" however they are useful to keeping the issues and discussions clear.

A "plate curve" is a tone reproduction curve that is applied in the workflow to a plate in order to have it render tone values that are different from those it delivers when the laser exposure and processing (or lack thereof) have been set to the manufacturer's specifications. So, applying a linearizing curve that makes an inherently non-linear plate linear is an example of the use of a plate curve.

A "press curve" is a tone reproduction curve that is applied in the workflow to a plate in order to have it render tone values that are required to deliver a specific tone response on press. The assumption is that the laser exposure and processing (or lack thereof) have been set to the manufacturer's specifications.

By this definition, if only a linearizing curve is applied because a linear plate is needed to deliver the correct tone response on press then that linearizing curve is a press curve.

A plate curve in this sense is not related to tone reproduction on press. It is effectively a calibration curve. It brings the plate to a known condition. However, in a CtP environment, the manufacturer's setup of laser exposure intensity, processing chemistry, and processing time effectively calibrates the plate plate to a known condition. It might not be linear but it is known. There is no need to recalibrate by applying a plate curve to what is already calibrated.

Another way to look at the question

Let's suppose that a linear plate provided the tone response on press that we need. Would it make sense to then use two curves - one to linearize the plate (a plate curve) and a second curve (a press curve) to linearize the linearized plate? I doubt it. Makes more sense to just apply the one linearizing curve - based on the uncalibrated natural condition of the plate.

So, if that logic makes sense, why wouldn't it make equal sense if we needed a non-linear press curve? Just apply the one non-linear press curve based on the uncalibrated natural non-linear condition of the plate.

As long as the plate's tone response is consistent then it can be the basis on which to build press curves. However, if the plate is inconsistent in its tone response then the use of linearizing plate curves as well as the use of press curves will fail. You cannot use curves, plate or press, on a device that is inconsistent.

What the "authorities" have said*Some quotes on this topic from the Idealliance G7 guides:

6.2 Origin of NPDC curves
To determine the 'natural' NPDC curves of commercial CtP-based printing, G7 research analyzed numerous press runs made with ISO-standard ink and paper, and a variety of plate types imaged on “un-calibrated” CtP systems (no RIP curves applied, not even to “linearize” the plate).

5.4 Set up the RIP
Set up the plate making RIP exactly as you would for a normal job, but clear out any values in the current calibration table, or begin with a new, empty table. The first press run is best made with ‘un-calibrated’ plates – i.e. no calibration values in the RIP.
IMPORTANT: Do NOT linearize the plate-setter so that measured dot values on plate exactly match original file percentages. Contrary to common belief, this may reduce accuracy of subsequent steps.

a. PRINTING IDEALIZED TARGETS VALUES - Achieving calibration condition with raw or linear plates, not requiring a curve, is an ideal situation.

*A note about authorities. I had trepidations about including these points from G7 because I do not believe that people should blindly do what some authority says they should do. It is not enough to say "Do it this way because I say it should be done this way." If the authority cannot explain exactly why one way is wrong and another right then it is just an opinion and without evidence to back it up it is not a credible opinion. I included these quotes only because they may carry credibility for some readers of this post.

Scenarios“We’ve always done it this way!” or “This way works just fine!” Even when we have the time to think about how or why we do things a certain way, our thoughts are often clouded by that kind of thinking. However, it can make it easier to understand the merits of a one curve workflow compared with a two curve workflow if one breaks down the sequence of steps required to get a plate into the press room. Given the same final result, the fewer the steps - the better the workflow since it provides fewer opportunities for error.

Here are some examples of workflow scenarios to see what happens with a one curve workflow vs a two curve workflow:

One CtP & one plate shop - to achieve the same final result on press:
One curve workflow: one press curve = one curve total.
Two curve workflow: one linearization plate curve plus one press curve = two curves total.

One CtP & one plate shop using three different curves to optimize for three different papers. To achieve the same final result on press:
One curve workflow: one press curve per paper type = three curves total.
Two curve workflow: one linearization plate curve plus one press curve per paper type = four curves total.

One CtP & two plate shop - to achieve the same final result on press:
One curve workflow: one press curve per plate type = two curves total.
Two curve workflow: two linearization plate curves plus one press curve = three curves total.

One CtP & two press shop - to achieve the same final result on two presses:
One curve workflow: one press curve per press = two curves total.
Two curve workflow: one linearization plate curve plus two press curves = three curves total.

One CtP & one plate shop - what happens if a new batch of plates do not perform as the previous batch did:
One curve workflow: modify one press curve so that the plate tones are the same as the previous plate batch = one modified curve total.
Two curve workflow: modify one linearization plate curve plus apply the standard press curve so that the final plate tones are the same as the previous plate batch = two curves total.

One CtP & one plate shop - what happens if the press curve needs to be tweaked/adjusted:
One curve workflow: modify one press curve to achieve the required tone reproduction on press = one modified curve total.
Two curve workflow: one linearization plate curve plus modify one press curve to achieve the required tone reproduction on press = two curves total.

One CtP & one plate shop - what happens if the CtP device is replaced:
One curve workflow: measure the new plate output and modify one press curve to achieve the same tone reproduction/dots on plate as with previous CtP = one modified curve total.
Two curve workflow: measure the new plate output and modify the linearization plate curve to linearize the plate then apply the existing press curve = one modified curve for two curves total.

Looked at this way, the linearization plate curve, in the vast majority of cases, is redundant. It serves no useful purpose except to add complexity and another point of failure.