Colour calibration - is this possible? And if so - how?

[Cheng]

EDIT:

To simplify the object of what I'm trying to say again:

I have a macbeth colour chart in front of me. How do I make my monitor display a photo I have just taken of this chart as close as possible to the chart I have infront of me.

Is there a way I can use my spyder to 'read' from the chart then 'read' from my LCD and calibrate it until every colour is accurate?


====

I understand the basics of monitor calibration - however - I am yet to try it. I'm not sure if calibration on it's own will be able to solve my problem but will be delighted if it could. I'd appreciate all your input!

I am a dentist, and use a ring flashed camera to take photos of mouths. When we need to make crowns, we have our own shade guides, that are standardised by crown manufacturers. The problem is, when we have taken the photos, they are not accurate to the physical shade guide we have. Colleagues have told me that even when they calibrate their monitors, the photo they have taken is not accurate to the shade guide they have in their hands. We need the photos to be accurate so that we can send photos to the dental lab who also would have a matched calibrated monitor, so they can make a crown's shade accurately to the photo we have.

Using calibration hardware or software, would it be possible to calibrate monitors so that what I see is what appears in the monitor?

Here's a photo of a shade guide we might use:

How might I approach this?

Thanks in advance

 

[Cheng]

One approach was to either calibrate monitors to the colour guides directly (if this is possible?) or to calibrate conventionally then use some sort of custom curve in photoshop to calibrate to the colour guide? (again, how would this be done?)

 

[snappingsam]

would the simple answer not be to photograph the teeth with the appropriate shade guide beside them? - or at least the closest matches?

like this?

 

[andrew gilbert]

I think your problem is in getting the White balance right in the photo. This is probably why the photo looks different from the shade even on a calibrated screen.

You would need to shoot an image of a grey card (you can get them at most camera shops) with the ring flash / camera setup and then use the photo to create a custom White balance setting that you can apply to your other images, as long as the flash dominates the lighting this should give you the right reference point for other shots taken the same way.

Then, if you have a calibrated monitor and you open an image of the shades and adjust it using the settings from the grey card shot you should get a decent match. More importantly because you've corrected to a grey card which is a known reference shade the lab should see the same thing on their screen.

Of course if you print it out your printer needs to be calibrated too...

My digital backs, screens, printers and scanners are all calibrated to each other and I can honestly say that I get what I see on screen and in print - but the grey card is the key to getting the colours right in the image.

Hope this helps!

 

[Cheng]

would the simple answer not be to photograph the teeth with the appropriate shade guide beside them? - or at least the closest matches?

like this?

Yes - that is an option - but if what you see on the screen is different from what you see in your hand, then you're still working 'relatively' - that is, that tooth is a bit redder, so I'll make the crown a bit redder - but not as accurate as comparing it directly.

Would like it so I could perfectly calibrate the screen to a known reference that I have in front of me if that makes sense.

 

[Cheng]

I think your problem is in getting the White balance right in the photo. This is probably why the photo looks different from the shade even on a calibrated screen.

You would need to shoot an image of a grey card (you can get them at most camera shops) with the ring flash / camera setup and then use the photo to create a custom White balance setting that you can apply to your other images, as long as the flash dominates the lighting this should give you the right reference point for other shots taken the same way.

Then, if you have a calibrated monitor and you open an image of the shades and adjust it using the settings from the grey card shot you should get a decent match. More importantly because you've corrected to a grey card which is a known reference shade the lab should see the same thing on their screen.

Of course if you print it out your printer needs to be calibrated too...

My digital backs, screens, printers and scanners are all calibrated to each other and I can honestly say that I get what I see on screen and in print - but the grey card is the key to getting the colours right in the image.

Hope this helps!

I thought of using a whibal - I think I'll give it a go. Unfortunately I don't have a calibrated monitor and the person I talked to does not have a whibal...

Surely in the same way a spyder can calibrate a monitor to true black, it could calibrate a certain hue/shade to a separate reference shade?

 

[u8myufo]

Is the patient not able to make the decision which shade suites best by you holding the templates up against their teeth whilst they look in a mirror? Why do the lab need a picture :shrug: surely each template has a referrence #?

 

[Cheng]

Is the patient not able to make the decision which shade suites best by you holding the templates up against their teeth whilst they look in a mirror? Why do the lab need a picture :shrug: surely each template has a referrence #?

In a word, no. A photo is much better in explaining the subtle colour changes and characteristics of a tooth.

Anyone have any other ideas - surely there must be a lot of applications that require calibrating against a set colour chart?

 

[u8myufo]

In a word, no. A photo is much better in explaining the subtle colour changes and characteristics of a tooth.

Just seems a bit odd to me as I have had three done and they all look fine to me. What I am trying to say is, if you look at the template yourself against the patients teeth can you not match it up :shrug: Going about it the way you are doing will surely complicate matters more, but then again knowing Dentists taking pictures is good enough to bung another fifty sobs on the bill

 

[andrew gilbert]

I thought of using a whibal - I think I'll give it a go. Unfortunately I don't have a calibrated monitor and the person I talked to does not have a whibal...

Surely in the same way a spyder can calibrate a monitor to true black, it could calibrate a certain hue/shade to a separate reference shade?

The spyder will calibrate the monitor but the original colour in the image will still be off - you only need to correct the colours by using the grey card at your end - then the photo has the right shade and adjusting the white balance to correct the colours, even if your monitor is off it will still look right on somebody else's monitor if theirs has been calibrated. Obviously it would be better if your screen is calibrated as well - for your own peace of mind if nothing else...

 

[Cheng]

The spyder will calibrate the monitor but the original colour in the image will still be off - you only need to correct the colours by using the grey card at your end - then the photo has the right shade and adjusting the white balance to correct the colours, even if your monitor is off it will still look right on somebody else's monitor if theirs has been calibrated. Obviously it would be better if your screen is calibrated as well - for your own peace of mind if nothing else...

but will the grey card actually match up all colours, or just shades of grey/black/white?

 

[Cheng]

Just seems a bit odd to me as I have had three done and they all look fine to me. What I am trying to say is, if you look at the template yourself against the patients teeth can you not match it up :shrug: Going about it the way you are doing will surely complicate matters more, but then again knowing Dentists taking pictures is good enough to bung another fifty sobs on the bill

If it looks fine, then great. But think of it like buying a burger at mcdonalds vs getting Gordon Ramsey to cook you a burger...etc.

 

[Ploddles]

If it looks fine, then great. But think of it like buying a burger at mcdonalds vs getting Gordon Ramsey to cook you a burger...etc.

That doesn't make any sense at all. What has cooking a burger got to do with the colour of anything?

As stated above, all the dentists I have been to also do the colour match themselves by holding the card against the teeth that they are matching to. That is going to be far more reliable than trying to colour match from a photo.

 

[Cheng]

That doesn't make any sense at all. What has cooking a burger got to do with the colour of anything?

As stated above, all the dentists I have been to also do the colour match themselves by holding the card against the teeth that they are matching to. That is going to be far more reliable than trying to colour match from a photo.

What I was trying to say is that there are different grades of treatment provided. You can have your basic one shade crown, which may look suitable, but still like a crown, or you can have your crown that has many different layers, translucencies, opacities, characterisations, cracks, stains, fluorosis marks, mamillons, exposed dentine, worn enamel, etc. This crown will then look like a tooth and hopefully be almost indistinguishable from the other teeth.

As writing from the perspective of someone IN the business, it'd be far more convenient if you had a photographic record of the actual teeth, which you can then analyse with more time after the appointment, then you can do when the patient is in the chair. You can also see far greater detail in a photograph than you can remember. It helps the technicians to construct better looking crowns.

Yes, you can eyeball for a few seconds and get an approximate shade, but if there is a method where you could consistently achieve a perfect shade then I'm all for it - you can get colour fatigue and colour shifts by looking at things too long, and if you can eliminate any inconsistencies then thats always better. I hope this explains a bit more.

 

[Cheng]

TO PUT THIS BACK ON TOPIC:

I have a green pencil in front of me, and a blue pencil in front of me. How do I calibrate my monitor so that the green and blue shown on the monitor is as close as possible to the items in front of me.

 

[coldpenguin]

TO PUT THIS BACK ON TOPIC:

I have a green pencil in front of me, and a blue pencil in front of me. How do I calibrate my monitor so that the green and blue shown on the monitor is as close as possible to the items in front of me.

have a whibal card in the photo as you take it.
Then set the white balance of the image depenant on the black, grey and white registered in the card.
Akthough the white is 'white', if there is some strange colour reflection, such as the red area of the inside of the mouth reflecting onto the card, then the calibration of that photo will reduce the redness in the white back to the white on the card being white. The same amount of reduction will be made to the green and the blue, but it will be less noticable. However, the colour registered in the photo now will be closer to the real colour of the blue if shone under a pure white light.

Then to make the monitor look the same, you need to calibrate the monitor. At this point, you have calibrted the image to a know 'range' and calibrated the monitor to a known 'range', the images should 'match'

However, to do this with your patients, you should include a whibal card somewhere in their shot. I suggest one of the smaller cards, the larger ones might be too big to get their teeth around (parts of the above statements might have been tonque in cheek)(of course, having the toungue in the shot might also effect the balance).

 

[Cheng]

have a whibal card in the photo as you take it.
Then set the white balance of the image depenant on the black, grey and white registered in the card.
Akthough the white is 'white', if there is some strange colour reflection, such as the red area of the inside of the mouth reflecting onto the card, then the calibration of that photo will reduce the redness in the white back to the white on the card being white. The same amount of reduction will be made to the green and the blue, but it will be less noticable. However, the colour registered in the photo now will be closer to the real colour of the blue if shone under a pure white light.

Then to make the monitor look the same, you need to calibrate the monitor. At this point, you have calibrted the image to a know 'range' and calibrated the monitor to a known 'range', the images should 'match'

However, to do this with your patients, you should include a whibal card somewhere in their shot. I suggest one of the smaller cards, the larger ones might be too big to get their teeth around (parts of the above statements might have been tonque in cheek)(of course, having the toungue in the shot might also effect the balance).

Cheers. Have you used a whibal card yourself with a calibrated monitor? I have a whibal card, but not calibrated my monitor yet. Does the whibal card actually look identical to the one that's in your hand after adjusting for white balance? And have you tried the scenario I mentioned above - coloured pencils or similar - and then viewed them on the monitor after calibration - is the colour identical? Cheers.

 

[coldpenguin]

I have not personally used a whibal card, which contains white, black and grey. However I have in the past used a grey card. I do not need the accuracy that three points of comparison would give.

The picture of the pencils on the screen might not look the same as the two pencils on your desk.
However, if you took the pencils outside and took an image of them with the whibal card, or took them into a room painted with dark green walls with the whibal card, after calibration the three images should look pretty much identical to each other, although prior to calibration they might not.
IF the manufacturer is viewing their product charts under correct lighting, then the image shown on the screen should be as close as it possible to what they are seeing. It will never be possible to give the manufacturer an idea of what the teeth look like, because it depends a lot on the colour of the light which is falling on the teeth.However, by using a balance card, you can give them a known calibrated image to a 'standard'

Consider how different a car might look under a sodium street light, or a camera flash. At least with the whibal calibrated image, you should have a standard reference point.

Do you have a lowepro bag? Supposedly all of their inners are 18% grey, so could be used as a 'demonstration' test, you could take your teeth shade guide and have some quick practices in different rooms. Take the images raw, copy them and use one in auto white balance (or flash/preset) and one white whitebalance set on the grey card.

 

[snappingsam]

When photographing artwork with high end scanning backs etc - the norm is to photograph a McBeth colour chart - and you then build a profile to get the most accurite colour.

There is a new system from HP that uses a colourimiter to read areas on the original artwork to give more accurate colours.

Neither of these approaches would work for dental photography - however, there is a far more limited range of colours that teeth can be... from very white to.....

The problem with trying to create a screen based system is that anyone looking at the images has to be at the same standard..

If you could make a "step wedge" of all the colours available that could be placed between the teeth that would be near perfect - the other thing is the reflectance of the step wedge would need to be the same as teeth....

The act of getting colour "right" for this is quite different to getting the colours right in general photography - as you don't normally need to have absolute colour matching - in general its just down to colour balance - so a whitebalance is all thats really needed. The dental photography has a fixed colour balance - as you are using the same lightsource so whitebalance is a given.

 

[u8myufo]

TO PUT THIS BACK ON TOPIC:

I have a green pencil in front of me, and a blue pencil in front of me. How do I calibrate my monitor so that the green and blue shown on the monitor is as close as possible to the items in front of me.

Sorry Cheng if you thought it was going off topic, but in your opening thread you state " We need the photos to be accurate " And now you are saying you want to set up your monitor so your pencils are as close as possible to what you see That is two different things imo. You will also have to take into consideration that no two monitors will be the same, the ambient lighting cast on the monitor either end will never be the same and what your eyes see and what the bloke at the lab see`s will never be the same in a month of Sunday`s, therefore you will never get an exact match

 

[Cheng]

Sorry Cheng if you thought it was going off topic, but in your opening thread you state " We need the photos to be accurate " And now you are saying you want to set up your monitor so your pencils are as close as possible to what you see That is two different things imo. You will also have to take into consideration that no two monitors will be the same, the ambient lighting cast on the monitor either end will never be the same and what your eyes see and what the bloke at the lab see`s will never be the same in a month of Sunday`s, therefore you will never get an exact match

Not different at all (IMO) - What I basically want is my monitor to resemble as close as possible to what I have infront of me - be it teeth,pencils, or whatever. And thus if my monitor is calibrated as such, then I can calibrate a monitor near enough for someone else to see 'what I have in front of me' - I hope this makes sense!

I've recently got a spyder 3 - and the colours are coming pretty close! Not perfectly the same mind you - but then again, colours change radically depending on what angle of the LCD i'm looking at. I'm wondering to think if a print out might be better for consistency...which adds another colour management profile into the equation argh....!

I think that calibrating my camera itself against a macbeth might help too and get it profiled with a DNG profiler. Then I will have to calibrate my printer against my monitor too to ensure it's producing the right colours too. This is getting pretty complicated ....!

 

[Cheng]

Or to simplify the object of what I'm trying to say again:

I have a macbeth colour chart in front of me. How do I make my monitor display a photo I have just taken of this chart as close as possible to the chart I have infront of me.

Is there a way I can use my spyder to 'read' from the chart then 'read' from my LCD and calibrate it until every colour is accurate?

 

[snappingsam]

Most calibration software only work in "relative" mode which is perfect for normal photographs as you are not measuring the exact colours - nor have to match to them.

A specialist type of work like this need "absolute" which will match colours within gamut and ignore those outside...

There are actions that will allow you to match gretag charts - but unless you have the full version (which costs over £200) you have just a few "memory" colours...