D65 vs. C

Re: Cream&Sugar - an N9 reflective screen mix.

MississippiMan said:

:gah: :thud:

I hate when this happens. :hissyfit: :duh:

I misspoke earlier. Upon further reflection (...ie: a PM from Todd) I discovered that Todd did a "Gain" measurement, and Gary was in fact the "Tester' who ascertained the neutrality factor.

Back to the question of whether or not Gary's test was accurate....that I cannot validate. But his approach seemed to be no less intensive and well intended than anyone else's, and a whole lot more involved than my own (...which is to say none at all but the "Eye"...)

So I apologize to all for posting what might...or might not be an inaccurate statement.

Click to expand...

I'll field this one...

To say his readings were in error or inaccurate isn't really the case. They are correct, but for the C neutral reference not D65. Does it matter? A little bit yes. An analogy could be is there much difference between 1000 yards and 1000 meters? At a quick glance a meter stick and yard stick look the same length. Hold them together and there is a difference. I know, not exactly the same thing here but the point I am trying to make is there is a difference between them. Is it huge? No not huge, but things do change slightly when you change your reference point. Now, I will say this and I have verified it even though I know this was contested by some. If something is measured neutral for C it will be neutral for D65 too. That is one of the characteristics of a true neutral. Not many things measured neutral for C back then though. Which means when referenced to D65 it would shift accordingly on the CIE Chromaticity graph.

The reason I personally use D65 as my reference point is because it is the standard used thoughout the video industry. Seeing that we have been nailing it consistantly lately, means we can achieve that goal.

Another thing I know that is always stated at some point or another by some is that the screen doesn't matter as much as the projector calibration. They both go hand in hand. Seeing that projector manufacturers strive to meet D65 color calibration, it just makes sense to try to have our screens at the same standard. For many it won't matter as long as the screen is at least close to our reference point, but for the purists and videophiles out there, every bit of improvement does matter to them. (Ironically, for those that have said the screen really doesn't matter that much, they also have made statements of superior screen performance... seems like a bit of a contradiction to me. For one argument it doesn't matter, then for the next debate/argument it does matter :scratch

It's like how some people don't mind, or can't even see any change in image quality with keystone set to max, while others would find the image almost unwatchable. If hitting D65 was difficult, then sure, we would have to accept whatever we have available, but since it is something we can achieve, then why not? Last analogy... I used this one before and happen to like it We all know cars will run on regular gasoline or premium gasoline. Premium burns cleaner and better though, increases performance and horse power too. For some cars, and even the owners, they may not see much of a difference in performance, and only a slight improvement in mileage. For others, they may see a bigger difference. Now it boils down to this... if both regular gas and premium gas are both the same price... which one would people get? Kinda a no brainer if you ask me, and that relates perfectly to D65 screens- if they cost the same and are just as easy for a person to obtain, and one does give you better 'mileage', even if its just slight, then why not? In most cases though people really do see a difference between a neutral screen and one that is not.

Anyway, we reference D65 with all of our readings. Older readings were referencing C. Comparing the two will give some odd numbers since they have different reference points. I just wanted to point that out for people to keep in mind...

Re: Cream&Sugar - an N9 reflective screen mix.

To put some hard numbers to the D65 vs C situation...

My Valspar flat enamel tinted Bermuda Beige. This is a pinkish-peach color to the eye.

"D65" L*a*b* values: 84.28, 6.36, 8.79

"C" L*a*b* values: 84.35, 5.82, 8.84


My Behr 1850 tinted Reference Gray (a very neutral dark gray).

"D65" L*a*b* values: 76.56, 0.01, 0.18

"C" L*a*b* values: 76.56, 0.05, 0.09

:nerd::bigsmile:

Re: Cream&Sugar - an N9 reflective screen mix.

Harpmaker said:

To put some hard numbers to the D65 vs C situation...

My Valspar flat enamel tinted Bermuda Beige. This is a pinkish-peach color to the eye.

"D65" L*a*b* values: 84.28, 6.36, 8.79

"C" L*a*b* values: 84.35, 5.82, 8.84


My Behr 1850 tinted Reference Gray (a very neutral dark gray).

"D65" L*a*b* values: 76.56, 0.01, 0.18

"C" L*a*b* values: 76.56, 0.05, 0.09

:nerd::bigsmile:

Click to expand...

Good example Harp.

So what does all that mean? It's the same color but different numbers. That's because there are two different reference points used.

Now it can be clearly seen why when people talk about color in detail they must use the same reference. Even though the color is the same, if you convert those L*ab values to RGB to get a visual representation of what the color looks like, you will get two different RGB values, meaning two different colors, yet they are from the exact same source. So it really is important to make sure we are all talking the same language and using the same reference point or nothing can be compared.

If you know something was measured referencing a different Illiminant then it's easy to convert that to D65 and then to the RGB value.

Harp this also demonstrates what I said about a true neutral... notice how the neutral gray reference still meets neutral tolerance for both C and D65? (Although if you convert those values to RGB it may not look too neutral based solely on RGB numbers). The neutral reference values based on C come out to be 193 187 197 and based on D65 they are 189 189 189. Bermudia Beige 'C' - 234 204 203, D65- 229 206 194.

I think this does demonstrate very well how important it is to know what reference points are used when discussing colors. If we didn't know that the BB readings above were done in D65 and C, and say these were two different people reporting back... it would be very easy to say that one person got a bad batch when in reality they are perfectly fine. Even the neutral looks off with just the RGB numbers and might get passed over as a true neutral because two people are talking different references and don't know it. If someone showed me those numbers and included the L*ab values I would know right away what was going on, but others may not pick up on that and end up dismissing something when it is actually very good. I do know of one occasion when this exact thing happened, someone nailed a neutral with a mix by eye but because the RGB numbers were based on C, they abandoned that mix and went on to others, and none were as neutral as the one that ended up overlooked.

Like I said, it isn't wrong to use C, but to discuss the color balance everyone would have to be using the same reference, or know what was used and then convert to D65. Sometimes conversions can have user error and throw things off too, so me personally... I'll stick with D65, especially since it's the same language everyone in the film and video industry talk as well.

This is why I am such a stickler for correct data and not scanner or MS Paint values. We tend to throw a lot of things out on the table but unless there is a common reference, everything might as well be swept from the table to the trash because we can't discuss and compare colors unless we know exactly how the readings were done and what they were referenced to.

Harp: Thanks for the information, I am familiar with the Lindbloom calculator. It is in fact what I use for conversion.

For the record, I didn't start this thread. Apparently mech or another mod did, using my ID. My comments originally appeared in this thread, and were in response to this comment by mech:

mechman said:

Hopefully that wasn't tested by prof55. I know that a lot of what he's done previously was in error. I believe everything was measured in C or something like that.

Click to expand...

An interesting statement, to be sure. Please elaborate.

mechman said:

Sure. But I think a better question would be why did you use C when D65 is the standard? And before I answer, prove to me that you have any idea how to do it.

Click to expand...

The question was directed to Harp, and he has answered. If you wish to go further, again, please elaborate. Initially you state "I believe everything was measured in C or something like that", and now you seem sure, to the point of asking me why... You further demand that I "prove" to you my knowledge? I was unaware of any "admission requirement" to this thread, particularly when it appears that I started it.

I'm perfectly willing to discuss the topic of the thread, but I respectfully suggest that you refrain from this confrontational approach, and offer your knowledge freely without requiring "proof" from others that they are worthy to enter the conversation.

Thanks,
Garry

Garry being a moderator yourself elsewhere I'm sure you're aware that when you move a post from a thread and spin it off into a new thread, it takes the starting post as the thread originator.

Nobody did anything funny using your account or anything like that.

As I mentioned in the PM I have to admit your post was a little odd sounding to me. Whether it was your intentions or not it sounded a bit like a test and would be jumped on after it was answered. I personally would have prefered to see you explain why you feel there is a problem doing conversions and provide some examples, and then explain how you do it and show the process. That way it is an educational experience for people and not something that sounded like a test. In that sense 'testing' someone could be construed as just as inappropriate as you felt mech's comment on you proving you knew how to do it yourself. Not trying to be confrontational here, but I personally would like to know if the person testing someone else actually knows it themself. In this case I am well aware that you are versed in all of this, so again... why not have just explained some of the problems you feel you have seen and what people have done that was wrong.

I tried to explain in my post why it is important to have a common reference point when discussing colors and detailed color theory. Without a common reference it is impossible for people to talk about a color and know if they are indeed talking about the same thing or not.

Although it is true that some industries such as the textile industry still use C as their reference, D65 is the newest and accepted standard. There is nothing wrong with using one value over another as long as everyone is using the same reference value. Seeing that everyone in the Film, Television, and video industry all use D65, that is why we use D65. If we used something else, then we could not discuss and relate things to what is the standard within the industry itself.

One thing to note too is that sRGB is directly tied to D65, which is one reason why some calculators seem to be off a bit when using a different Illuminant source other than D65. One sure way to help avoid and eliminate any confusion and potential 'error' is to just use D65 to begin with.

An interesting entry on "Standard illuminant" can be found at http://en.wikipedia.org/wiki/Standard_illuminant.

In the context of this thread, the following quote from this article may be helpful.

"Illuminants B and C

Illuminants B and C are daylight simulators. They are derived from Illuminant A by using a liquid filters. B served as a representative of noon sunlight, with a correlated color temperature (CCT) of 4874 K, while C represented average day light with a CCT of 6774 K. They are poor approximations of any common light source and deprecated in favor of the D series.

"Illuminant series D
See also: D65

Relative spectral power distribution of illuminant D and a black body of the same correlated color temperature, normalized about 560nm.

Derived by Judd, MacAdam, and Wyszecki, the D series of illuminants are constructed to represent natural daylight. They are difficult to produce artificially, but are easy to characterize mathematically."

Harp: No intent to confront in my request, and my apologies if it appeared so. I was merely interested in your conversion method. If I seemed a bit blunt, I expect it was my reaction to being told by mech that a lot of my previous work was "in error", and in his later unwillingness to elaborate on his statement without "proof" of my knowledge.

Regardless, the issue of D65 vs C is an interesting one. As you say, "the D series of illuminants are constructed to represent natural daylight. They are difficult to produce artificially, but are easy to characterize mathematically." It is for this very reason that color measurement instruments seldom use an actual D65 source; rather, they measure at another temperature and convert mathematically to D65, returning results in L*a*b* or other color spaces.

Further conversion is necessary if RGB results are desired, and this is where it gets interesting. Your use of the Lindbloom calculator is a good choice in my opinion, but there are others that will give varied results. Some calculators are not particularly useful for conversion between color spaces; unfortunately, EasyRGB falls into this classification. I arrived at this conclusion after extensive conversation with the author of this program, and perhaps this shortcoming is what mech makes reference to (though he seems to lack any actual facts on the topic). Any error in my previous figures has long since been corrected, and thoroughly explained.

More importantly, one must also bear in mind that RGB represents only three points on a spectral curve, and these points may not coincide with the RGB of the actual projector used. It is quite possible to concoct a mix with a perfect RGB that is unusable as a screen. This is why I have encouraged the use of spectral curves over RGB, a practice I am happy to see is being continued here.

Garry

Color Space

Let's talk about color spaces for a minute.

This actually does tie in with the D65 vs C topic...

It is actually possible for a person to develop their own personal color space and work from there. I have seen some use scanners and then MS Paint or PhotoShop (the better of the two choices) to then sample the scanned image and 'determine' the color value. If the individual setup their own color profile within PhotoShop, calibrated their scanner and monitor, they can develop and use their own personal color space. The key word there though is 'personal'.

They then could work on things and 'develop' things and everything would be relative to their personal color space. However they would not be able to discuss this information with anyone else because it would be meaningless since the other person has no common reference point to compare things too.

Now for most people painting a room or border trim or other items like that, things don't have to be as precise. We are dealing with a more precise environment though, so a common reference really is necessary.

Here is one potential problem with personal color spaces. It is totally possible and probably that say a person's multipurpose printer/scanner/copier/fax machine dies and they get a new one, that it will not match the existing color space profile they were working from. Depending on how long they have been using their own color space, they could actually lose a lot of information that they may have to redo. Seeing that it is a personal colorspace though, usually it isn't that critical and the person just opts to keep all their old work and continue on.

Hopefully it can be seen how this can be used only as a single and private color space and not something that can be used to discuss color values with a person across the globe, on the other coast, or even in the same town. Once it moves to a discussion amongst others, they have no reference to use and cannot be sure what they are looking at is what the person on the other end is talking about.

Again, for many things aside from screens it isn't that big of a deal, but we are working with neutral tolerances and specifications, and without a universal reference and set of standards all this would be impossible to do.

Once we move away from personal color spaces we also start to deal with highly precise, and expensive equipement such as colorimeters and spectrophotometers. Not only are these units a bit pricey, but the expense does not end after the initial purchase. These units require routine calibration certification to ensure they remain within specs. This is another piece of the puzzle that ensures us that when we take a reading here in the US, Georgio in Italy can take a reading and get the same results, or take the numbers from our reading, and be able to know exactly what we are talking about.

Part of the ability to be able to do this is because there is a set of standards used in the color field. The other part is Georgio or anyone else for that matter has to know how the reading was taken and what the reference is that was used. Calibration is also a standard. In this example the person that is say in Italy assumes our equipment is working properly and has a current calibration as well. If when they receive the material being tested and get a different reading, they are sure to question how the sample was analyzed on the other end, and also inquire as to if the spectro has been calibrated and certified.

So as it can be seen, without standards it would be impossible to have any of these discussions, at least in a meaningful way.

Let's put all this into a real world scenario that has nothing to do with screens, but everything to do with color science and standards. Say you own a nice bright red Ferrari. Now say you had a minor fender bender. There is no need to repaint the entire car, just the area that was repaired. Without standards, the likelyhood of having the two colors match is not very good, and I am sure you wouldn't be very happy to see your Ferrari with a noticably different colored fender than the rest of the car. Believe it or not, most body shops actually have a spectrophotometer! Paint actually changes its color slightly as it ages due to Sunlight and other conditions. So even though the body shop could just order a quart of Ferrari Red, most likely there will be a slight difference in the colors. They use the spectro to take a reading of the actual color of the car and then they can color match and even age the paint so nobody would ever know there was any damage to the vehicle.

Again, without standards and specs, it would be very difficult to match things or remain consistant when we discuss color values and neutral specifications.

I know all of this is long and boring stuff to read though, but standards really are important when it comes to being precise and accurate. Again I know that some will argue that we don't have to be this precise and there are a wide variety of things that will work. That is true, but some things really do work and perform better than others. Without standards and common references, nobody would be able to determine why one thing works better than something else. We do have the capability of determing that as well as understanding why one thing is better than another whether it is a slight improvement or a drastic one, but in order to do that... we all have to be 'talking the same language', which in this case would be using the same reference point, or at least knowing the reference point the other person used and then converting to the one that you use.