Drew of ProPhotohome offers a new tool for white reference.

New measurements, or, "Certain Death Averted"

Drew has pointed out a "fatal flaw" in my initial report on the behavior of the Color Parrot, in that (he said) I operated on the assumption that my WhiBal gray card was perfectly neutral.

I didn't say that. I said it was very close to neutral based on credible statements from the manufacturer as to the limits on neutrality error. Thus I would use it as a refernce, but we all needed to be mindful of a small uncertainty introduced by my not knowing its exact reflective chromaticity.

But in the interest of overcoming this death warrant, I have acquired a "pedigreed" WhiBal gray card - one which the manufacturer was kind enough to measure precisely for me

I've also developed a workflow that lets me determine the recorded color of the light through the diffuser, as captured by the camera, not by reading R, G, and B on an 8-bit basis but rather on a 15-bit basis. (That's what the Photoshop "16-bit RGB" reading of the colors represented in a 16-bit TIFF file actually gives!)

Doing this greatly reduces a major source of potential experimental error we had in the prior situation (with RGB quantizing error giving a possible uncertainty of perhaps 0.001 unit of du'v').

So, what's a guy with new toys to do but make some more measurements with them. So I decided to rerun my basic determination of the transmissive chromatic neutrality of my v 0.1 Color Parrot.

And the Easter Bunny also brought me a 2007 model Expo Disc (the one I had was the 2006 model, albeit just recently bought from B&H). So I thought I would include it in this new run for comparison (since we hear a lot about it in discussions of the Color Parrot).

As before, the test were run using a Canon EOS 20D with a Canon EF 24-105 mm f4.0 IS lens. The diffuser shots were taken at a focal length of 24 mm at f/4, with focus at infinity. Exposure was metered, with an EC of +1.0 stop.

The tests were run, as in many prior cases, with "daylight filtered through a white venetian blind" in our breakfast room as the source.

Let me note that, since I have not yet made specific tests on the variation (if any) of the chromatic response of the Color Parrot with angle of incidence, I can't be sure that this particular test lighting setup would give the same result as with some other lighting setup.

In any case, without further ado, here are the test results:



Note that here the chromaticity of the WhiBal gray card is plotted at the values stated by the manufacturer, and the other values are plotted with reference to that. The intent is that their actual response is properly plotted on the chart.

On a related front, we also determined the relative transmission of the two diffusers (as before' based on the camera's metering system readings, in terms of Measured BV, as reported in the makernote, read with ExifTool. It seems that the Color Parrot is about 0.87 stops "hotter" than the Color Parrot for this particular arbitrary angular distribution of the incident light (directivity pattern considerations could well lead to a different comparison for other light source configurations).

This is close to the value noted by Drew ("about twice").

Best regards,

Doug

Alright, Doug.

Enough about the ExpoDisc. In regards to acquiring a different instrument. I will look into it. Until then, I stand by the recommendations of the tool manufacturer and my assertion that the white substrate and the blue gel (which looks to basically be a Rosco 1/8 blue) are both spectrally skewed. They both have too much of the same thing, the color blue. There is NO offset effect that can I can imagine could happen in that circumstance.

I don't see this thread really going anywhere at this point. We have hashed and rehashed our disagreement over this issue over on PPH and now again, here.

It seems that we have a few people who would like to help and collaborate on improving the Color Parrot, or possibly creating a new tool altogether. I think that is cool and I wish I could participate more. But, I'm afraid I've revealed too much about our unreleased v2.0 product already.

I will be happy to discuss the original product further, but I'm afraid I won't be able to help create a new one. I'm sure you're all aware that anything we say or develop in the "open" here can be patented by any reader within a few minutes of its being posted here. It's unfortunate, but that's the way it is. The same goes for me discussing too many ideas for future improvements or products.

The links to purchasing the Color Parrot within your website no longer work! Dead ends to B&H store. The B&H store itself does not have them either! I tried the link in your signature too. That's the same. Am I just not seeing the obvious? What's up?

Click to expand...

Really? What browser are you using? Perhaps the page is cached in your browser.

Please do let me know. I need to get that fixed straight away.

the Easter Bunny also brought me a 2007 model Expo Disc

Click to expand...

What "Easter Bunny" was that?

This is close to the value noted by Drew ("about twice").

Click to expand...

Glad to see that your new measurement reflects my original assertion.

Look, Doug. I have no problem with you writing your own reviews, etc. But, if anything, your measurement tools and methods are more flawed than the ones you are attempting to criticize.

You don't even have a reflective spectrophotometer to independently measure the WhiBal tool. That's not what I would call an honest broker. Although, I have every reason to believe, that you are doing your best to be an honest broker.

How about this, I'll look into getting a transmissive spectrophotometer if you will get a reflective one, and a transmissive one. Please don't take your measurements from a biased party if you are going to present yourself as an independent.

Again, my intent has never been, and is not, to bash other products. It think the WhiBal is a fine product, likewise the ExpoDisc. They both have the capability to deliver quality results. The Color Parrot excels in some areas where these products do not. It is as simple as that.

The primary audience that this tool is marketing to is those who use AWB. I did not develop this tool to try and win over pros who already own other products. Our survey simply asked if they owned other products and what they were. We then asked them what their experience was compared to said other product. Some of those results were published previously. They were all very positive.

Quod scripsit scripsit.

Quod dixit dixit.

(Hey, that's catchy.)

Causa finita est.

Consummatum est.

Doug

Drew,

Still I cannot see purchase info? what is up with the links. It goes the the Pro Store but then no color Parrot, unless I'm blocking it out. I swear I'm using calibrated light with a mean temp of 5000 degrees Kelvin!

Asher

roma locuta est, causa finita est

Ah but Rome no longer rules!

So the case is not closed!

What has been said is still alive and working in our minds. That's the great thing about discussions. Here there is a new product to solve and old problem and do it better/cheaper/ than other options.

At least some effort has been made to examine claims. whereas in Rome one could go to a temple, ask a soothsayer or the Vestal Virgins, we rely on logic.

I personally like to take responsibiility for color management even before the first pictures. I map out all locations with various lights so I don't have to take reading in the middle of something interesting. I do that with the WhiBal. I've thought of an Expodisk but never used one. That could be another easy method. I do not, as yet have the Color Parrot or the Babelcolor target.

This discussion provided the beginning of a look at the issues involved. We still do not see that we have all the information we might want to make a choice for transmissive work with the Parrot device on the camera.

However, it doesn't stop one from trying different solutions.

I know a wedding photographer who shoots mostly in film but also digital. He and his two assistance each have two camera ready at all times and he carries a third, a digital Nikon DSLR.

So naturally I was intersted in how he managed white balance with 7 different cameras all over a mixed and varied lit ballroom and banquet hall. He said "It's pretty straightforward, I never used any such thing. I just send them to my lab in Los Angeles and they know the "look" I'm famous for. My job is just to get hte pictures. The processing? That their job!"

Asher

Further testing - Correction

I recently wrote [color emphasis added]:

Doug_Kerr said:

On a related front, we also determined the relative transmission of the two diffusers (as before' based on the camera's metering system readings, in terms of Measured BV, as reported in the makernote, read with ExifTool. It seems that the Color Parrot is about 0.87 stops "hotter" than the Color Parrot for this particular arbitrary angular distribution of the incident light (directivity pattern considerations could well lead to a different comparison for other light source configurations).

Click to expand...

I had meant to say:

"It seems that the Color Parrot is about 0.87 stops 'hotter' than the ExpoDisc".

My apologies for the error.

Thanks to Asher for catching it.

Doug

Hi Doug,

I am pleased are learning about all the various options. What I realize now is that I'm like an old fashioned physician who just uses a small spread of medicines. I do intend to get the various other WB tools and see how they might work for me.

Does anyone have a current link to purchase the Color Parrot since I must be doing something wrong and can't find the right place!

Asher

Bonsoir Asher

should be there, scroll down click on the "Buy Now" buttons (thy are 3 of them)

For
- The Exciting New Patent Pending White Balance Tool.
- the Five White Balance Filters in One
- the Pro Pack Gel and Attachment Add On for the Color Parrot as well…

http://www.prophotohome.com/index.php?page=prophotostore

That link gives a B&H shop Profoto store but no

- The Exciting New Patent Pending White Balance Tool.
- the Five White Balance Filters in One
- the Pro Pack Gel and Attachment Add On for the Color Parrot as wel

Please recheck URL!

Asher

New data from the L*a*b*o*r*a*t*o*r*y*

Some of my colleagues have commented that they are still not used to appreciating chromaticity difference/error data on the basis of du' and dv' (imagine that!).

To help comprehension, I thought that perhaps I would present some key findings on the basis of the CIE L*a*b color space.

Note that in this color space, the coordinates a* and b* do not (by themselves) denote chromaticity. They only do if accompanied by a value of L*. For comparison purposes, if we are not interested in lightness but only chromaticity, we can state that using L*a*b* notation with a consistent, arbitrary value of L*.

Since the reflective color of reflective targets is often stated by their manufacturers in L*a*b notation, and since L* there is often in the range 65-75, for ease of comparison to the data from manufacturers of reflective targets we will use here an consistent arbitrary L* of 70.

We will also state the Delta E (color difference, in the CIE 1976 form) value from 70,0,0.

From our last tests, then:

ExpoDisc 2007: L*a*b* = 70, +1.27, -0.62 Delta E = 1.27

Color Parrot: L*a*b = 70, -0.15, -2.77 Delta E = 2.78

Note of course that these results are from a completely flawed test method (I read that here on OPF) and used as a reference possibly bogus WhiBal gray card test data provided me by that rascal M. T. (I read that here on OPF, too), with whom I am evidently in cahoots to discredit something or another (I read that here on OPF, too).

Here it's Holy Week and apparently nothing is sacred. Oy vay is mir.

In the interest of full disclosure, I also need to reveal that while I was not Governor of Texas, I had a Republican girl come into the District of Columbia from Virginia to have dinner with me. (Or was it that I had a Republican virgin come into the District . . . well, something like that.)

Ahh! I feel much better now.

Tomorrow, Daisy will help me run the Frisbee aerodynamics tests on several diffusers. (I tricked her into doing this by telling her that there would be a chocolate L*a*b involved. She's part L*a*b herself.)

Well, as I am the shabbat goy in the lab, I'd better get back to my cauldrons. The pureed eye of newt I use to coat my new diffuser prototype has to be watched carefully.

Hang in, guys. This long nightmare will be soon over.

Carla sends her - well, you wouldn't want to know.

Best regards,

Douglas A. Kerr, P.E. [Retired, and bloody glad of it]

And now abideth faith, hope, charity, these three; but the greatest of these is charity. (I Cor. 13:13 KJV)

How much difference in color is noticeable?

Hi Doug,

In layman’s terms, what do you use as a difference between accurate color and deviation from that as a difference that many people can recognize.

Since this involves how the brain evaluates color, I have no doubt that there are confounding issues like, "Would the same slight difference in match between a jacket and a skirt be noticed in a bluer/green room or an a brighter area or under tungsten versus bluer light".

Still, what are the rule of color deviation, delta E, figures we can use to know whether or not a stated difference from true color is going to be noticed, whether for one object that delivered with the slightly wrong color or a picture viewed in the wrong light or wrong surroundings or seen on an a poorly profiled monitor?

Asher

Hi, Asher,

Asher Kelman said:

Hi Doug,

In layman’s terms, what do you use as a difference between accurate color and deviation from that as a difference that many people can recognize.

Click to expand...

Let me put that aside until I discuss some of the matters below, since my thoughts need to be in that context.

Since this involves how the brain evaluates color, I have no doubt that there are confounding issues like, "Would the same slight difference in match between a jacket and a skirt be noticed in a bluer/green room or an a brighter area or under tungsten versus bluer light".

Click to expand...

Indeed, this is a very complex issue (with several different "layers"), and I am not at all expert in most of that area.

We can start with a very "clean" issue: how much difference does there need to be between two colors of light so that a certain fraction of observers will note them as distinct. (And this means, "in the same viewing environment")

I have to state it in that statistical way ("so that a certain fraction . . .") because of course no two observers are the same, nor is a single observer "the same" on two different occasions.

David MacAdam led the pioneering work in this matter. We can simplify the discussion by, for the moment, limiting ourself to color changes that are in chromaticity only (that is, where the two colors being compared have the same luminance).

Macadam found that:

A. The amount of chromaticity change needed to make a certain fraction of the observers notice it depended on what chromaticity we were starting with.

B. For a given "starting chromaticity",. the amount of chromaticity change required for a certain fraction of the observers to notice it depended on the direction of the change (that is, perhaps it was a different amount of change in the blue/yellow direction than in the red/green direction.

As a result, he developed a system of what have come to be called the "MacAdam ellipses", geometric figures (generally elliptical) at different places on the CIE x-y chromaticity plane. Each of these showed, if we start with the chromaticity plotted at the center of the ellipse, how far we would have to go in each direction on the chart before 68.3% of the observers could notice the change (the so-called "one sigma" limit).

Then, for each starting chromaticity, there would be a larger ellipse, which was the limit we could go in each direction before 95.4% of all observers would notice the difference, the so-called "two-sigma" limit. Then there would be an even larger ellipse, which was the limit we could go in each direction before 99.7% of all observers would notice the difference, the so-called "three-sigma" limit. And so forth.

These three ellipses (for example) are said to represent the "one MacAdam step". "two MacAdam step", and "three MacAdam step" chromaticity differences.

The sizes of the MacAdam Ellipses (of any "step size") vary with the starting location on the chromaticity plane, and the inclination of the long axes varies as well.

So if we wanted to specify the chromaticity of, for example, a certain traffic light, with a tolerance, we could say, "It should be within two MacAdam steps of this chromaticity: [chromaticity specification]". That would mean that it should be so close that, in the worst case, only 95.4% of the observers could notice the difference. You might think, "well, that's not so good", but it implies that "if it were 50% worse, then 99.7% of the observers coiuld notice the difference."

Now, if we plot chromaticity on the CIE u'v' chart (you've no doubt heard about that recently), then it turns out that the MacAdam ellipses (of any given "step number" turn out to be very near circles, and very near the same size regardless of starting location on the chart.

Now, you may say, "Great, Kerr - now you can tell us just how big one Macadam step will be on the u'v' plane, so we can assess your results in those terms."

No, I can't tell you, yet, because I haven't figured that out yet. I'm on the way, but there is a lot to it, and a lot of other stuff on my plate.

Now, this is complicated enough, for a very clean concept:discerning the difference between two chromaticities, shown alternately in the same place, in the same viewing environment.

But now you transport us to a more complicated arena of eye chromatic adaptation and the like (which of course is involved at the heart of the concept of color correction. And I'm just in no position to carry us further at this point into that morass.

That's not to say that there is no literature in this field - there is probably a ton (likely numerous doctoral dissertations, in fact). I just haven't read any of it.

Best regards,

Doug

MacAdam ellipses on the CIE u' plane

Hi, Asher and all,

I little research indicates that, as a useful approximation, we can consider the "radius" of a one-step MacAdam ellipse on the CIE u'v' plane in the general area of the "white" illuminates to be about 0.001 unit.

Again, in a very simplified way, this suggests that about 68% of all observers could just discern (under very good viewing conditions) the differences between two chromaticities separated by 0.001 unit on the u'v' plane.

And almost all observers could discern the difference between two chromaticities separated by 0.003 units on the u'v' plane.

Please note that this is a great simplification of a very complicated issue, but may be helpful in allowing us to appreciate what we are looking at in the diffuser accuracy data..

Best regards,

Doug

Hiu, Drew,

Please excuse the delay in responding to your (thought) provoking message. I wanted to be sure I had the benefit of a good night's sleep, two good breakfasts and a lunch, and a chance to see my stepson's 11 new puppies and a Texas sunset before I responded. I'll try and not be too wordy.

Drew Strickland said:

Look, Doug. I have no problem with you writing your own reviews, etc. But, if anything, your measurement tools and methods are more flawed than the ones you are attempting to criticize.

You don't even have a reflective spectrophotometer to independently measure the WhiBal tool. That's not what I would call an honest broker. Although, I have every reason to believe, that you are doing your best to be an honest broker.

<snip>. Please don't take your measurements from a biased party if you are going to present yourself as an independent.

Click to expand...

How dare you.

Douglas A. Kerr, P.E. [Retired]

A serious, but silly, charge

Drew:

On 2008.03.14, you wrote:

Please don't take your measurements from a biased party if you are going to present yourself as an independent.

Click to expand...

You make a serious charge here. Let me make sure I understand it by trying to flesh out the context.

1. I am going to be making measurements of the transmissive chromatic performance ("neutrality") of the Color Parrot. I will, for comparison, be doing the same for the ExpoDisc, a competitive product of the same general type.

2. I will need a reference surface of fairly-precisely known reflective chromaticity to use as a reference. I requested from Michael Tapes a standard Whibal gray card along with with the quality control measurements of its color.

3. You are concerned that Michael could have altered those readings so as to either:

a. Make his product "look more accurate" to me (based on the the readings) than it really was.

and/or

b. Make your product look worse than it is, by choosing or making for me a card whose actual chromaticity substantially departed from the reported, nearly-neutral numbers in just that direction that would make the apparent chromatic error of your product seem worse than it was.

Note that your language (the reference to a "biased party") clearly indicates that your concern was not over the readings accompanying my WhiBal card possibly being erroneous through testing carelessness, but rather that they might have intentionally been altered.

4. You feel that I was irresponsible by accepting those possibly-bogus readings at face value.

Have I got it, bunky?

I don't know what offends me most:

A. The suggestion that Michael Tapes would provide me with a device accompanied by a fraudulent test report.

B. The suggestion that I was irresponsible and a failure as an "honest broker" (whatever that is) by relying on those results as a reference point for my tests.

C. My disappointment in how ill-thought out, and illogical, your charge is.

Since you characterize my interest in these matters as "only of an intellectual nature", you can probably guess which one is the most offensive. But the others run a close second.

Douglas A. Kerr, P.E. [Retired]

OPFrs,

There is no place for personal aspersions against anyone here. I have no doubt that all the tests Doug performed are accurate and at the very least done with care, responsible attitude and an honest intent. I believe very strongly that Michael Tapes provides WhiBal cards meeting published tolerances. If anyone has a card they think is not within tolerance, let me know!

All the devices for measuring color are capable of helping getting the color right. How right is another matter! I have an open mind and welcome the availability of more choices.

We welcome the Color Parrot as an extra item to test in the future for transmissive color along side the ExpoDisc and the BabelColor Target with various gray cards for reflective color. The WhiBal, for the time being is my standard but I will not be biased in that!

Let's not allow a good scientific approach to be spoiled by even an impression of disrespect. I think this would be an error. One of the special things about OPF is the feeling that one will be treated well. Let's continue to remember this.

I strongly advocate support for photographers coming out with anything to help in getting to a great picture. We therefore welcome this critical look at the new Color Parrot. Any other information people have will be welcomed.

I have no doubt, even without any measurements, that anything that diffuses light pretty well can work. If it is small and convenient why not? So starting out of the gate, I like the device even though I have not seen it! However, for critical work it does matter how a device matches claims and how it compares to other devices available to do the same job. So it's very worthwhile for anyone who can take objective transmissive colorimetric measurements to do so and let's see the data.

We must be able to welcome new devices but also to accept that the market will demand proof. This may be stressful, but it’s the free honest market of ideas at work. I hope we do it right!

Asher

Asher Kelman said:

That link gives a B&H shop Profoto store but no

- The Exciting New Patent Pending White Balance Tool.
- the Five White Balance Filters in One
- the Pro Pack Gel and Attachment Add On for the Color Parrot as wel

Please recheck URL!

Asher

Click to expand...

Brings you to Drews Paypal account to confirm (and pay) your order…

Drew Strickland said:

Thanks, Doug and Bart.

The BabelColor target certainly looks nice and smooth.

Click to expand...

I've yet to measure a target that is smoother spectrally or measures with a higher L-star value. Plus, you can sand it down if it gets dirty/scratched/stained. Its small and fits on the wrist. The only downside is, its not inexpensive (but less so then the others).

Andrew Rodney said:

I've yet to measure a target that is smoother spectrally or measures with a higher L-star value. Plus, you can sand it down if it gets dirty/scratched/stained. Its small and fits on the wrist. The only downside is, its not inexpensive (but less so then the others).

Click to expand...

Yes, and it's a very useful tool for getting the exposure 'right', especially in a studio setting. Only specular reflections can be brighter and cause clipping, which is usually (for 'white' light) not a real problem in such a setting because the lights can be controlled.

However, it is a reflection target and therefore less suited for White Balancing of distant scenes. I do use it for that if in an open space, by holding it at arms length in front of the camera while pointing the camera at the scene.

Bart

Spectralon?

Hi, Andrew,

Andrew Rodney said:

I've yet to measure a target that is smoother spectrally or measures with a higher L-star value. Plus, you can sand it down if it gets dirty/scratched/stained. Its small and fits on the wrist. The only downside is, its not inexpensive (but less so then the others).

Click to expand...

Do we know of the BabelColor target is made of Spectralon? It sounds as if it might be.

Here is a reference:

http://www.labsphere.com/productdetail.aspx?id=329

Although I can't seem to put my hand on the reference right now, I recall that one of its benefits was said to be that, if the surface gets corrupted in use, it can easily be "resurfaced", restoring its original reflectance properties.

Best regards,

Doug

Doug_Kerr said:

Do we know of the BabelColor target is made of Spectralon?

Click to expand...

Hi Doug,

I don't know for sure, but from the specs it surely looks like Spectralon® White Reflectance material. In fact this looks just like it, also in appearance.
Good catch.

Bart

Hi, Bart,

Bart_van_der_Wolf said:

Good catch.

Click to expand...

Thanks.

Better lucky than good!

It sounds like a good thing to have (the BabelColor target). I might get one.

We would make it into a "pendant" which can be used to adorn models' cleavage.

At the moment, we use a pendant of Silesian Amber (diffuse whitish in color as a result of millions of tiny entrained gas bubbles, the result of volcanic activity), whose reflective chromaticity we know (but it is not handily close to neutral). It was bought from the Crow Collection of Asian Art museum here in Dallas.

(It was acquired, on a buying trip, by Margaret Crow herself. She is the wife of famed Dallas real estate mogul and developer Trammel Crow, and with him founder of the museum.)

Best regards,

Doug

Nicolas Claris said:

Brings you to Drews Paypal account to confirm (and pay) your order…

Click to expand...

Thanks Nicolas,

That's the identical url I have been using for days now and this morning. Then the "Buy Now" reappeared! Voila, Color Parrot $79, Shipping $12, $91 total!

Asher

Hi Asher,

Sorry you had so much trouble.

For some reason a few people have trouble seeing those links. I can't seem to locate the root cause.

I have gone ahead and added back some text links in addition to the graphic ones.

A new site should be up before too long on a separate site. Thanks for the order.

Well, Drew,

I'm a bulldog, I didn't give up. With all the discussion of different solutions, the least I can do is put one of each in my bag to try out.

Asher

Now maybe can we all agree that Doug did not get and would never have used a specially unfair/ bad/even "selected" card from M.T.?

Doug_Kerr said:

Hi, Andrew,

Do we know of the BabelColor target is made of Spectralon? It sounds as if it might be.

Click to expand...

I have no idea. You'd have to ask Danny at BableColor.

Hi Asher,

I never said, nor was I trying to suggest, that Doug was trying to be unfair. Nor Michael Tapes, or the folks at ExpoDisc.

As I already mentioned to Doug privately, to be truly "neutral," pun intended, when writing a review it is best practice to use regular items purchased through regular channels. The samples that are sent out by manufacturers, like the ExpoDisc the Easter Bunny brought, I must guess, is most likely going to be tested to ensure the best face is being put forward. The same goes for the WhiBal being sent out. In no way am I suggesting that Doug is in cahoots with anyone. Neither are the manufacturers in question. They are simply doing their best to provide the best product example they can. This is absolutely natural.

Even though Doug has no way of knowing this, I made sure I had no idea which unit he was to receive. As Doug is very particular about accurate expression, I am probably overly obsessed with experimental issues of bias, given my experimenter background.

When consumer reports tests their new car models and reports their gas mileage, they don't take the numbers on the window at face value. They drive the car and measure what they actually get using their own independent "instruments."

It is a very basic and simple point.

I am sure Andrew will remind me of my recent lack of separation of church and state as a "reviewer." Point taken. However, I made it clear that we manufacture the product. It was an article with historical context, not technically a true review.

I have no wish to make anyone look bad. Doug is a still an awesome guy in my book!

I have also been instructed, by those who know better than I, that I should not be spending my time "defending" the product. My time is much better spent selling it and promoting it, etc. Andrew also pointed this out previously. Sage advice, I will henceforth attempt to follow.