Pages: [1] 2 3 ... 5   Go Down

Author Topic: The terms "linearization" vs "calibration"  (Read 28951 times)

Stephen Ray

  • Full Member
  • ***
  • Offline Offline
  • Posts: 217
The terms "linearization" vs "calibration"
« on: May 02, 2016, 01:17:08 am »

I’m researching printer calibration. As I pore over books, manuals, and guidelines from various manufacturers of RIP software and printing machines, I’m finding many seem to use the terms “calibration” and “linearization” interchangeably.

Do the experts use these terms as entirely interchangeable nowadays or are they differentiated somehow?

Comments, clarifications anyone?
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: The terms "linearization" vs "calibration"
« Reply #1 on: May 02, 2016, 04:12:00 am »

I’m researching printer calibration. As I pore over books, manuals, and guidelines from various manufacturers of RIP software and printing machines, I’m finding many seem to use the terms “calibration” and “linearization” interchangeably.

Do the experts use these terms as entirely interchangeable nowadays or are they differentiated somehow?

Comments, clarifications anyone?

Hi Stephen,

Calibration establishes the relationship between input and output. It then allows to convert from one (the reference or standard) to the other (the unit under test), or vice versa.

The meaning of Linearization on the other hand depends on context, IMHO (I'm not an expert though). It modifies the relationship that was found by calibration into uniform intervals (in the context of numerical, or perceptual, or ...). So that would require a modification in the weighting parameters, especially when interpolating between known reference values.

Cheers,
Bart

P.S. Although analogies are often shaky, maybe something like this will clarify.
Input  Output
0 <--> 0
1 <--> 1
2 <--> 4
3 <--> 9
That reveals a power of 2 calibration. To linearize the output, one takes the square root of the output.
« Last Edit: May 02, 2016, 04:28:08 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 608
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: The terms "linearization" vs "calibration"
« Reply #2 on: May 02, 2016, 09:15:14 am »

I’m finding many seem to use the terms “calibration” and “linearization” interchangeably.

"Calibration" has some agreement as to what it means - adjusting the behavior of the device. "Linearisation" is typically more of a hand-waving term for "making the device be better behaved". Most print processes use screening to convert continuous tone images into discrete dots of ink, and screening introduces dot gain. Some dot gain can be beneficial in giving the printing process a more perceptual response (that's where the Mac's 1.8 display gamma came from), but too much (because of a high DPI and/or lots of dot overlap) can make it very perceptually non-linear, and hard to control. So applying suitable per-channel calibration curves "to linearize the response" can tame this.
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The terms "linearization" vs "calibration"
« Reply #3 on: May 02, 2016, 09:58:42 am »

Calibration might contain a step that one would call linearization and it might not. Some devices can be linearized, some can't (due to the driver sending data to the printer, an Epson with the Epson driver comes to mind and yes, it could use better linearization). Calibration is an attempt to put the device in a desirable condition so it's a broad term where linearization is far more specific.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 18090
  • When everyone thinks the same, nobody thinks
    • My website
Re: The terms "linearization" vs "calibration"
« Reply #4 on: May 02, 2016, 10:08:50 am »

Does anyone speak English around here?

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: The terms "linearization" vs "calibration"
« Reply #5 on: May 02, 2016, 10:14:25 am »

I’m researching printer calibration. As I pore over books, manuals, and guidelines from various manufacturers of RIP software and printing machines, I’m finding many seem to use the terms “calibration” and “linearization” interchangeably.

Do the experts use these terms as entirely interchangeable nowadays or are they differentiated somehow?

Comments, clarifications anyone?

Based on a very succinct explanation of linearization given by Harald Johnson on page 134 of his "Mastering Digital Printing", linearization is a particular type of printer calibration designed to align input values and output values to be the same by controlling the density of ink the printer lays on the paper. A perfect alignment would exactly print the same ink density as that required. For example, it would output 25%, 50%, 75% Black if the Input (the reference values) were each 25%, 50%, 75% Black (i.e. a straight line relationship between the value required and the value printed). Dot gain and printer performance drifting over time can cause these output values to diverge from their input or required values, in which case the printer performance is non-linear. Linearization is the process that brings them back to convergence. As explained by Amadou Diallo on page 93 of his "Mastering Digital Black and White", linearization aims to eliminate abrupt shifts in density and produce a smooth gradation of tones, by adjusting the printer's ink output to compensate for non-linear printer behaviour. We linearize the printer to assure its basic linear behaviour before creating specific paper/printer profiles, as the profiling of papers should start on the basis of the printer being able to lay down the right amount of ink for aligning input with output values.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The terms "linearization" vs "calibration"
« Reply #6 on: May 02, 2016, 10:41:27 am »

Does anyone speak English around here?
That question begs another; do you understand English? IF one of the answer's isn't clear, ask for clarity and specifics. Otherwise, your post adds nothing to a conversation that so far, only you appear to have difficulty understanding.


http://digitaldog.net/files/LinearityandGamma.pdf
« Last Edit: May 02, 2016, 10:45:57 am by digitaldog »
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 18090
  • When everyone thinks the same, nobody thinks
    • My website
Re: The terms "linearization" vs "calibration"
« Reply #7 on: May 02, 2016, 11:59:07 am »

That question begs another; do you understand English?...

If I don't, that would be understandable.

Before Mark posted a response in simple English (thanks, Mark):

"...it would output 25%, 50%, 75% Black if the Input (the reference values) were each 25%, 50%, 75% Black (i.e. a straight line relationship between the value required and the value printed."


I heard the following "definitions":

- "...uniform intervals (in the context of numerical, or perceptual, or ...)... a modification in the weighting parameters, especially when interpolating between known reference values."

- "...a hand-waving term for 'making the device be better behaved'."

- "...Some devices can be linearized, some can't."
- Wow, deep! Never mind that it fails to define the term.


digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The terms "linearization" vs "calibration"
« Reply #8 on: May 02, 2016, 12:06:46 pm »

If I don't, that would be understandable.
So your post was to troll?
It is pretty obvious by your original post, you don't understand this topic, fine.
It is pretty obvious by your original post, you could have asked for clarification from any of the posts previous; you didn't.
It is pretty obvious by your original post, you were not posting to attempt your typical OT (and sometimes) comical side.
So what isn't obvious is why you posted what you did. But I guess we can thank Mark for clearing up your confusions and can move on.
Quote
I heard the following "definitions":


- "...uniform intervals (in the context of numerical, or perceptual, or ...)... a modification in the weighting parameters, especially when interpolating between known reference values."


- "...a hand-waving term for 'making the device be better behaved'."


- "...Some devices can be linearized, some can't." - Wow, deep! Never mind that it fails to define the term.
You 'heard' but apparently didn't understand which is fine, worse, you didn't ask for clarification. That would have been more sensible to ask about.


At this point, a quote from this film seems appropriate:

"What is your major malfunction, numbnuts? Didn't Mommy and Daddy show you enough attention when you were a child?"   ;D
-Gunnery Sergeant Hartman, Full Metal Jacket (1987)
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 18090
  • When everyone thinks the same, nobody thinks
    • My website
Re: The terms "linearization" vs "calibration"
« Reply #9 on: May 02, 2016, 12:13:38 pm »

...you didn't ask for clarification...

If those were "definitions," I shudder to think what "clarifications" might sound like.

On a related note, I had to learn English. You, however, were born with it, but seem to squander that gift.

Since you brought mommy and daddy into discussion, I'll mention grandma too: "If you can't explain something to your grandma, you do not understand it fully yourself."

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The terms "linearization" vs "calibration"
« Reply #10 on: May 02, 2016, 12:17:07 pm »

If those were "definitions," I shudder to think what "clarifications" might sound like.
Since for some reason, you were unable to ask for any, let alone specific clarifications, that's now moot.
But you got your 15 seconds of attention today so hopefully we can move on, or you can attempt to further aid the OP as everyone else here attempted but you.

Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 18090
  • When everyone thinks the same, nobody thinks
    • My website
Re: The terms "linearization" vs "calibration"
« Reply #11 on: May 02, 2016, 12:28:17 pm »

... or you can attempt to further aid the OP as everyone else here attempted but you.

Posts in a forum are not there just as a dialog between the OP and experts, they serve to educate others as well. I was simply trying to learn something about the topic myself. The first step in learning is a proper definition of terms used.

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The terms "linearization" vs "calibration"
« Reply #12 on: May 02, 2016, 12:38:10 pm »

Posts in a forum are not there just as a dialog between the OP and experts, they serve to educate others as well. I was simply trying to learn something about the topic myself. The first step in learning is a proper definition of terms used.
IF this were a forum on languages, your original post might have some weight to it. It isn't such a forum. Your original post was simply a call for attention. If you wish to be educated, ask a question that has some relevance to the topic. You didn't.


He who asks a question is a fool for five minutes; he who does not ask a question remains a fool forever. -Chinese Proverb
So what's your question about THIS topic?
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 18090
  • When everyone thinks the same, nobody thinks
    • My website
Re: The terms "linearization" vs "calibration"
« Reply #13 on: May 02, 2016, 12:54:10 pm »

... So what's your question about THIS topic?

My question is the same as the OP's:

Quote
terms “calibration” and “linearization”... are they differentiated somehow?

That differentiation can only be achieved if each term is defined separately and then the difference pointed out. Plain English helps the rest of us understand experts.

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: The terms "linearization" vs "calibration"
« Reply #14 on: May 02, 2016, 01:08:24 pm »

Calibration might contain a step that one would call linearization and it might not. Some devices can be linearized, some can't (due to the driver sending data to the printer, an Epson with the Epson driver comes to mind and yes, it could use better linearization). Calibration is an attempt to put the device in a desirable condition so it's a broad term where linearization is far more specific.

From Wikipedia:

The formal definition of calibration by the International Bureau of Weights and Measures is the following: "Operation that, under specified conditions, in a first step, establishes a relation between the quantity values with measurement uncertainties provided by measurement standards and corresponding indications with associated measurement uncertainties (of the calibrated instrument or secondary standard) and, in a second step, uses this information to establish a relation for obtaining a measurement result from an indication."

Does that clarify the definition?  :)
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The terms "linearization" vs "calibration"
« Reply #15 on: May 02, 2016, 01:08:44 pm »

My question is the same as the OP's:
That differentiation can only be achieved if each term is defined separately and then the difference pointed out. Plain English helps the rest of us understand experts.


Geez, like pulling teeth to get you to ask a simple question.... As for the rest of us, you mean (thus far) you! Best speak for yourself.



Calibration is a process whereby a device is placed into some predetermined condition or behavior. An example is when a user calibrates a display. We want to set some parameters that can be controlled, such as the display’s white point, brightness contrast, and TRC gamma. This calibration of the device creates a condition that can be standardized and is repeatable. This allows similar devices in multiple locations to behave the same way. Since a device such as a display is in a state of flux over time, it is necessary to calibrate the device. This calibration returns the device to the original aim point. Calibration is something we need to do on a regular basis. Particularly with such devices as a display, which varies considerably over time. In conventional photographic processes, this would be similar to running control strips and adjusting chemistry to achieve target densities. If we understand that a profile describes the behavior of a device, we should be aware that if the device changes its behavior, the profile is no longer valid. Calibration returns the device back to the original condition, maintaining the integrity of the profile. If the device can no longer reach the original aim point, a new aim point within the capability of the device needs to be created.


Linearization
Some profile-building products offer an optional step they usually call prelinearization. The idea is to output a target with a small subset of patches, usually CMY and K in various steps from light to very dark as seen in Fig. 6-9. This linearization target is measured and the software uses that information to produce an optimal target for profiling based upon the information gathered from the linearization step. This means that the profiling process becomes a two-step procedure. Some devices are quite nonlinear in how they reproduce color. The linearization step can aid in producing quality profiles from such devices. At the very least, linearization allows a good profile to be generated with an initially smaller number of patches. This is possible because the secondary patches generated from the linearization data is better optimized for the printer.
Some products support this prelinearization process and some do not. Products that do support prelinearization usually ask the user if they wish to use this option. If you know the printing is very nonlinear, it’s worth testing. On the accompanying CD is a TIFF file called InkDensityTest.tif, which can be useful for visually evaluating if the output device is nonlinear. If most of the steps block up in color and don’t show good tonal separa- tion, the printer is exhibiting this nonlinear behavior. You can try a different driver setting if available. However, this can often alter the resulting color gamut of the printer. Alternatively, you can try conduct- ing a prelinearization process, but even this can’t produce miracles. A profile can do only so much to overcome the limitations of poor printer behavior (see the sidebar, “Printer Drivers and Their Effect on Quality and Color Gamut”).
Some packages also offer a postlinearization process. This can be quite useful for devices that change their behavior. After an initial profile is built, the user prints a special postlinearization target. This target is measured and the profile can be updated to account for some changes in the output device. This postlinearization process can’t work miracles on output devices whose behavior has greatly shifted. In such a case, building a new profile from a standard target is the only viable solution. However, for some devices where the drift of a printer isn’t large, this postlinearization process is a useful feature. Some users running commercial labs print out a postlinearization target every day and update their existing profile to account for slight device drift.



And DO try to read the materials provided already!
http://digitaldog.net/files/LinearityandGamma.pdf
Paragraph 2 for example.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: The terms "linearization" vs "calibration"
« Reply #16 on: May 02, 2016, 02:58:52 pm »

It's probably worth noting for those with a technical background that "linearization" in printing, or monitors for that matter, has nothing to do with "linear" as used in math/engineering. Visual perception is highly non-linear so linearization in photography related topics typically refers to perceptually linear. This is particularly true for printer profiles where the lookup table conversions are based on L*a*b* (Lab hereafter). Just taking the neutral scale (L) this is actually a power of 3 slope tacked onto a small linear (power of 1) deep shadow front end. For instance L=50 has 18% of the reflected photon density of a L=100 patch. It would be 12.5% from a pure power of 3 but is 18% because of the uplift from the linear front end of the L curve.

It turns out that, for B&W, the closer a printer driver matches the L response curve the more accurate the 3D LUTs that are inside profiles can be made. So achieving perceptual linearization coincides with producing good profiles.
« Last Edit: May 02, 2016, 10:30:20 pm by Doug Gray »
Logged

Lundberg02

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 379
Re: The terms "linearization" vs "calibration"
« Reply #17 on: May 02, 2016, 07:17:56 pm »

What the hell is wrong with the top grey scale in the bottom image. I have never seen "fluting" that bad. Are the boundaries being differentiated by a high pass?
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: The terms "linearization" vs "calibration"
« Reply #18 on: May 02, 2016, 07:20:08 pm »

What the hell is wrong with the top grey scale in the bottom image. I have never seen "fluting" that bad. Are the boundaries being differentiated by a high pass?
Not at all visible on this end; NEC SpectraView PA272W, fully high bit display path.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 608
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: The terms "linearization" vs "calibration"
« Reply #19 on: May 02, 2016, 08:08:41 pm »

A perfect alignment would exactly print the same ink density as that required. For example, it would output 25%, 50%, 75% Black if the Input (the reference values) were each 25%, 50%, 75% Black (i.e. a straight line relationship between the value required and the value printed).
A direct relationship between input value and resulting reflectance value is never likely to be desirable, simply because it's perceptually non-linear. A more linear relationship between input values and perceptually linear results is typically a better goal. This translates to something like linear density or L* response.
Logged
Pages: [1] 2 3 ... 5   Go Up