Fast measuring of large numbers of patches. Important new tech direction, finally.

[Deleted member 16349]

A very interesting technology has been introduced. It measures large numbers of colour patches quickly, .

It is called the Rapid Spectro Cube or RSC from ColorGate and is mentioned in this WTT article:

http://whattheythink.com/news/76388-colorgate-solutions-digital-industrial-printing-inprint-2015/

Here is the web site for this measuring device.

https://www.colorgate.com/rapid-spectro-cube-rsc/

There is a nice video showing how it works.

This is interesting for me since I have been waiting for such a technology since 2001, when I started to think about what was needed for the future direction for simple and effective prepress.

This particular product is claimed to be able to measure up to 10,000 small patches, as small as 1x1 mm, at one time. This then can be used for ICC profiles or other profile methods.

Why is this important. Well for many reasons.

With this capability, there will be no need for plate adjusting curves for dot gain.
There will be no need for G7 and its wasteful press runs or consultants.
The results of the offset press and digital printers will be more predictable and can use the same methods, because this leads to non device specific approach.
There will be little or no colour management required.
More accurate separations will be possible due to the higher patch count and the lack of need for dot modelling and curve adjusting.
Existing Standards become irrelevant. Aiming at some printing condition will be irrelevant since this technology can greatly help the real goal which is the reproduction of colour matching the printed image to the initial target image at all points in the image.
More capability to easily operate with other ink sets, with extended gamut, than the conventional CMYK.
Simple method to find the closest build for Pantone colours, plus the ability to know which Pantone colours are obtainable or not even within the gamut space.
Huge simplification to the operation for manufacturing print which will save time and money.
And maybe most importantly, it will make Frank Romano happy to see that simplification of obtaining colour is possible.

I am sure there are many other important reasons why this is a good direction but the ones I have given so far hopefully is enough to get you thinking.

I do not know if the RCS technology performs well enough but what it is trying to do is without question, the right direction. I am sure there are other ways to do the same thing.

It looks promising.

 

[gordo]

A very interesting technology has been introduced. It measures large numbers of colour patches quickly, .

It is called the Rapid Spectro Cube or RSC from ColorGate and is mentioned in this WTT article:

http://whattheythink.com/news/76388-colorgate-solutions-digital-industrial-printing-inprint-2015/

Here is the web site for this measuring device.

https://www.colorgate.com/rapid-spectro-cube-rsc/

There is a nice video showing how it works.

This is interesting for me since I have been waiting for such a technology since 2001, when I started to think about what was needed for the future direction for simple and effective prepress.

This particular product is claimed to be able to measure up to 10,000 small patches, as small as 1x1 mm, at one time. This then can be used for ICC profiles or other profile methods.

Why is this important. Well for many reasons.

With this capability, there will be no need for plate adjusting curves for dot gain.
There will be no need for G7 and its wasteful press runs or consultants.
The results of the offset press and digital printers will be more predictable and can use the same methods, because this leads to non device specific approach.
There will be little or no colour management required.
More accurate separations will be possible due to the higher patch count and the lack of need for dot modelling and curve adjusting.
Existing Standards become irrelevant. Aiming at some printing condition will be irrelevant since this technology can greatly help the real goal which is the reproduction of colour matching the printed image to the initial target image at all points in the image.
More capability to easily operate with other ink sets, with extended gamut, than the conventional CMYK.
Simple method to find the closest build for Pantone colours, plus the ability to know which Pantone colours are obtainable or not even within the gamut space.
Huge simplification to the operation for manufacturing print which will save time and money.
And maybe most importantly, it will make Frank Romano happy to see that simplification of obtaining colour is possible.

I am sure there are many other important reasons why this is a good direction but the ones I have given so far hopefully is enough to get you thinking.

I do not know if the RCS technology performs well enough but what it is trying to do is without question, the right direction. I am sure there are other ways to do the same thing.

It looks promising.

Your enthusiasm is noted.
However, the product seems to be designed only to generate ICC profiles (albeit on a wide variety of substrates) so I don't understand how it would:
Eliminate the need for G7 and its wasteful press runs or consultants.
Result in little or no required colour management.
Make existing Standards irrelevant.
Make for a simple method to find the closest build for Pantone colours, plus the ability to know which Pantone colours are obtainable or not even within the gamut space.
Simplify the operation for manufacturing print (other than the variety of substrate capability).

I do agree that maybe most importantly, it will make Frank Romano happy to see that simplification of obtaining colour is possible. ;-)

 

[Deleted member 16349]

Your enthusiasm is noted.
However, the product seems to be designed only to generate ICC profiles (albeit on a wide variety of substrates) so I don't understand how it would:
Eliminate the need for G7 and its wasteful press runs or consultants.
Result in little or no required colour management.
Make existing Standards irrelevant.
Make for a simple method to find the closest build for Pantone colours, plus the ability to know which Pantone colours are obtainable or not even within the gamut space.
Simplify the operation for manufacturing print (other than the variety of substrate capability).

I do agree that maybe most importantly, it will make Frank Romano happy to see that simplification of obtaining colour is possible. ;-)

Hum. I am a bit surprised that you don't see the implications but maybe you need to think about it more.

Yes, it is capable of making ICC profiles or potentially other types. The point is that normally the way ICC profiles, with using LUTs, are made, they do not use enough patches to accurately define how a device will print.

If one talks about an inkjet printer, then one is talking about rgb inputs, where each channel has 256 levels. It is not real rgb but one should think of it as input 1,2 and 3. Anyhow, the number of combinations one can input will be 256 to the 3 power or 256x256x256 = 16.7M combinations.

For CMYK plates, where one would have the range for each input to be 0 to 100%, then the number of combinations will be 101 to the 4th power or 101x101x101x101= 104M combination.

Either device, the number of combinations is huge relative to the normal number of measurements made for normal ICC profiles. With such a large volume in these data spaces, interpolation with only about a thousand patches or so, is way too small to obtain accurate results.

But as I understand, ICC profiles can be made with very many patches which will improve accuracy but for practical reasons, it is not easy and quick to measure many thousands of patches, one by one with the existing methods. Part of the problem is the size of the patch. Presently, fairly large patches are needed to be able to get an accurate reading. The other part of the problem is to have equipment of measure thousands of patches on the large size of paper that would be required.

So the point is that if a more accurate data set describes how you device prints, it tells you everything the device can do or can not do for that printing condition.

If you have an printing device and you know exactly how it will print, then you don't need anyone else to tell you how it will print.

If you have a printing device that does not print consistently enough, then no method will tell you how it will print.

I hope you think about it and eventually get the concept but if you don't, I don't think I will try to persuade you. You would need to understand it on your own.

 

[Stephen Marsh]

I am intrigued by this “high resolution spectro” that appears to capture data similar to the speed of a digital camera in a single shot, without the limitations of a digital camera… Perhaps I am missing how it actually works and it does not “take a single snapshot of the entire chart”?

For comparison, with say i1Profiler the maximum patch count is 6000 (approx. 7mm square patches), this creates around 7 A4 charts which would take less than 45 minutes to measure using an automated i1iO table in strip reading mode (longer for spot patch by patch).

Looking forward to seeing real reviews and not just press releases.


Stephen Marsh

 

[Deleted member 16349]

I am intrigued by this “high resolution spectro” that appears to capture data similar to the speed of a digital camera in a single shot, without the limitations of a digital camera… Perhaps I am missing how it actually works and it does not “take a single snapshot of the entire chart”?

For comparison, with say i1Profiler the maximum patch count is 6000 (approx. 7mm square patches), this creates around 7 A4 charts which would take less than 45 minutes to measure using an automated i1iO table in strip reading mode (longer for spot patch by patch).

Looking forward to seeing real reviews and not just press releases.


Stephen Marsh

Stephen. My guess of how this device works is that it is similar to a camera but that the receptors are not rgb but are just a single channel receptor. The spectral data then is obtained by taking multiple pictures, using a set of filters or other optical method that divides the spectrum into different wave length zones. This similar method has been used in tests to image fine art to get spectral curve data of the image, point by point.

Thanks for providing the information on the i7Profiler's count capability and time. It is a good reference.

Back in 2001, I was working with a group in the UK on their simple proofer product, which was meant to give their customers a quick check on what might be the output. Anyhow, we made an experimental test form that had almost 5000 small patches on a single A4 sheet . The idea was to read them using a table top scanner. At that time, I was just starting to read about colour theory and soon realized that the data obtained by the scanner would not be good enough for managing colour. I always liked the concept of taking a quick image that could obtain colour values, but it has taken a long time for one to come out on the market.

I know there have been some technologies that claim they are able to get spectral data from moving webs. That would of course be very fast capture but I suspect that some of these technologies are using simplified filter sets that would not necessarily provide high enough accuracy for prepress. But it is a nice trend.

The RSC, in the description, seems to be measuring only patches but I see no reason why it could not measure an image and record the spectral data of each pixel. That would bring a very interesting potential since one can use it for confirming that an image meets its intended targets at any point in the image. Also would be helpful in R&D for developing other prepress technologies.

The RSC technology has to prove itself in the market place of course but IMO it is something that is moving in the right direction. I hope it works as advertised.

 

[gordo]

Hum. I am a bit surprised that you don't see the implications but maybe you need to think about it more.

Yes, it is capable of making ICC profiles or potentially other types. The point is that normally the way ICC profiles, with using LUTs, are made, they do not use enough patches to accurately define how a device will print.
[SNIP]
So the point is that if a more accurate data set describes how you device prints, it tells you everything the device can do or can not do for that printing condition.

If you have an printing device and you know exactly how it will print, then you don't need anyone else to tell you how it will print.

If you have a printing device that does not print consistently enough, then no method will tell you how it will print.

I hope you think about it and eventually get the concept but if you don't, I don't think I will try to persuade you. You would need to understand it on your own.

What I understand from their website is that they are gathering more data points in order to build an ICC profile. That is an aspect of the characterization of the printing.
But I don't see how just simply having more data points to build a profile would:
Eliminate the need for G7 and its wasteful press runs or consultants.
Result in little or no required colour management.
Make existing Standards irrelevant.
Make for a simple method to find the closest build for Pantone colours, plus the ability to know which Pantone colours are obtainable or not even within the gamut space.
Simplify the operation for manufacturing print (other than the variety of substrate capability).

 

[Deleted member 16349]

What I understand from their website is that they are gathering more data points in order to build an ICC profile. That is an aspect of the characterization of the printing.
But I don't see how just simply having more data points to build a profile would:
Eliminate the need for G7 and its wasteful press runs or consultants.
Result in little or no required colour management.
Make existing Standards irrelevant.
Make for a simple method to find the closest build for Pantone colours, plus the ability to know which Pantone colours are obtainable or not even within the gamut space.
Simplify the operation for manufacturing print (other than the variety of substrate capability).

Gordon, I don't think I can help you understand. You see it or you don't. Right now you don't. Hopefully you will.

Think about what the data base (LUT) of the profile is. It contains basically ALL the colour information of how the printing device will print for a specific ink set, paper, screen etc. conditions. Ideally the printing device should be very consistent and predictable, which they aren't and the industry has not wanted to correct that for offset. The consistency of the printing device is a different issue but if we are thinking of an ideal printing device then the data base will have the information of the resulting colour due to dot gain, wet trap, inks, paper etc.

Maintaining consistency of things like density and dot gain are a process control problems and need to be controlled but are not necessary to know their values when solving the colour reproduction problem. Only output relative to input is important.

If you have such a data base, you can ask it all kinds of questions. Like where are the gray data values? Where are the combinations that result in the same colour but require the least amount of ink? Where is the combination of input values that result in a specific Lab value that represents a Pantone colour? Does a particular Pantone colour exist in the data base? What is the closest colour in the data base to a Pantone colour that is not in the data base? All these kinds of questions can be automated accurately when you have the highly defined data base.

The data base (profile LUT) is like the traditional Atlas but in more detail and usable for making separations.

The general problem of colour reproduction is much more complicated than the problem of density control on offset presses. For over 15 years on Printplanet, I got basically nowhere with that effort, so I don't expect much to happen with this one on this forum. My aim is not to educate everyone but to possibly reach some of the few who can see the opportunity. Therefore I am not going to make any big effort to explain this in any detail. My suggestion is to think deeper about this and try to forget what you already know because what you know is getting in the way of understanding what is possible.

 

[nirupakah]

Anyone with some real-world experience with this device? It seems it is actually on the market so it was not just a mad experiment...

There is (and probably was not a year ago) a nice video on the ColorGate website explaining how it works: https://www.youtube.com/watch?v=L3IQ7Qpuc1M

Does anyone happen to know what the price of this magic box might be?

As for the requirement "Make for a simple method to find the closest build for Pantone colours, plus the ability to know which Pantone colours are obtainable or not even within the gamut space." mentioned above, I am quite sure there must be a tool or even multiple tool for this on the market. Or isn't there? (well, it definitely is in the Caldera RIP software).

 

[Schnitzel]

Sorry, I'm with Gordon here. I fail to understand how simply measuring faster and slightly more patches than what is usually done today, will suddenly Make Life Great. Measuring more patches is not always better - printing devices can be very noisy across the tone-reproduction curve, and to get smoother color transitions it's sometimes better to measure less patches and interpolate between them, at the cost of color accuracy.

It would be helpful if you could describe a real world situation where such a device will, for example, eliminate color management completely.

 

[arossetti]

But I don't see how just simply having more data points to build a profile would:
Eliminate the need for G7 and its wasteful press runs or consultants.

If you had 1000's of patches of CMY gray up and down the L* axis couldn't you do away with G7 curves?

 

[Deleted member 16349]

If you had 1000's of patches of CMY gray up and down the L* axis couldn't you do away with G7 curves?

Very good. I think YOU are starting to see the potential.

In colour science, there is nothing special about gray. It is just a colour like any other.

So if one can measure thousands of colours, this makes it more accurate to interpolate the colours between these measured ones for all the screen combinations. Then with multi millions of combinations of screens generated from the measured values and the interpolated values, one can have them in a look up table LUT and search for the grays or any other colour you want.

There is so much fuss about gray but CMY does not print a very good gray so we use K to help do that. Well this is not an isolated situation. There are many other colours in the gamut that can not be printed well. That is just the way it is but one could pick out the screen combinations from the LUT that come closest. Also by looking at the LUT, you will know what colours in the gamut that will be hard or impossible to print. One can't find that out any other way except to make a test, which wastes time.

In the LUT there will be many combinations that have basically the same colour. So for a need to have GCR one would just look into the LUT and pick the screen combination that has the desired amount of K which replaces some CMY at the same colour. For some graphic affect this could be different in different parts of the image. It would be flexible and still be accurate.

You want a specific pantone combination. Look it up in the LUT. Done.

Since the LUT represents the reality of how the press should print, it should be predictable.

OK, the offset press is not so consistent but it can be made better but that is a different problem.

I hope now that you can see WHY there is no need for tone curves, or G7. Tone values would be only for process monitoring but not for colour.

I am glad you asked the right kind of question.