95-98% accurate color grading machines.

[diagem]

I am starting a new thread on the issue which I keep reading on the fact that there are colorimeters that can grade Diamonds colors to an accuracy of 95-98%.

I have never yet met this kind of accuracy not to mention consistency!

David Atlas..., can you please elaborate more on your writing from another thread:

https://www.pricescope.com/community/threads/round-table-best-light-conditions-for-diamond-subjective-performance-comparison.115561/page-3 (I started a new thread to avoid a threadjack on the former).

re> "All technology improves over time, but it must enter the market a bit imperfect and then users along with technicians and inventors improve its performance levels. A machine to grade color currently cannot identify problem diamonds with 100 or 99 percent accuracy. Not yet. I think 95% to 98% is possible right now."

We would love to hear some more.....

 

[strmrdr]

sas 2000 and imagem are supposedly able to get in that range or better.

 

[Garry H (Cut Nut)]

I think the comparison from front face up which is how these devices work, to side pavilion grading is the main issue.

Throw away the later.

But even then - do you measure dull body or intense flashes?
How do you totally avoid diespersion? Does that lower the color of some cuts?

Yuri is THE MAN on this topic

 

[diagem]

Date: 5/23/2009 6:14:23 AM
Author: Garry H (Cut Nut)
I think the comparison from front face up which is how these devices work, to side pavilion grading is the main issue.

Throw away the later.

But even then - do you measure dull body or intense flashes?
How do you totally avoid diespersion? Does that lower the color of some cuts?

Yuri is THE MAN on this topic

Garry,
How can we throw away the later? It must be an integral part of the color grade equation.

If not..., what does the consumer pay for a J-K profile grade which translates to a face up H color grade (with strong fluorescence?)

Do I understand from you there is a problem with dispersion and intense flashes in the faceup grading via machines?

 

[diagem]

Date: 5/23/2009 4:34:12 AM
Author: strmrdr
sas 2000 and imagem are supposedly able to get in that range or better.

Like I said..., in all my years in the industry..., I have not met a machine which is even close to this accuracy!
I am not talking face-up grading as it is a non-issue as of yet...., definitely not a issue if profile color grade is not taking into account!

 

[Garry H (Cut Nut)]

Date: 5/23/2009 6:47:39 AM
Author: DiaGem

Date: 5/23/2009 6:14:23 AM
Author: Garry H (Cut Nut)
I think the comparison from front face up which is how these devices work, to side pavilion grading is the main issue.

Throw away the later.

But even then - do you measure dull body or intense flashes?
How do you totally avoid diespersion? Does that lower the color of some cuts?

Yuri is THE MAN on this topic

Garry,
How can we throw away the later? It must be an integral part of the color grade equation.

If not..., what does the consumer pay for a J-K profile grade which translates to a face up H color grade (with strong fluorescence?)

Do I understand from you there is a problem with dispersion and intense flashes in the faceup grading via machines?

You could stop this DG:
http://images.google.com.au/images?sourceid=navclient&ie=UTF-8&rlz=1T4GGLL_en&q=dinosaurs&ei=zt0XSsPIA8qGkAWyqdyDDQ&gbv=2

And start to think about consumer confidence.

Right now, anywhere on earth, walk up to a group of people and say "Hi, I am in the diamond business, trust me!"

after 5 minutes they stop laughing

 

[diagem]

Date: 5/23/2009 7:30:29 AM
Author: Garry H (Cut Nut)

Date: 5/23/2009 6:47:39 AM
Author: DiaGem

Date: 5/23/2009 6:14:23 AM
Author: Garry H (Cut Nut)
I think the comparison from front face up which is how these devices work, to side pavilion grading is the main issue.

Throw away the later.

But even then - do you measure dull body or intense flashes?
How do you totally avoid diespersion? Does that lower the color of some cuts?

Yuri is THE MAN on this topic

Garry,
How can we throw away the later? It must be an integral part of the color grade equation.

If not..., what does the consumer pay for a J-K profile grade which translates to a face up H color grade (with strong fluorescence?)

Do I understand from you there is a problem with dispersion and intense flashes in the faceup grading via machines?

You could stop this DG:
http://images.google.com.au/images?sourceid=navclient&ie=UTF-8&rlz=1T4GGLL_en&q=dinosaurs&ei=zt0XSsPIA8qGkAWyqdyDDQ&gbv=2

And start to think about consumer confidence.

Right now, anywhere on earth, walk up to a group of people and say 'Hi, I am in the diamond business, trust me!'

after 5 minutes they stop laughing

Maybe in your neck of the woods...., I dont know what you are talking about....

Show me a realistic way it can be done!
Body color is root!

ETA: You are sure not going to gain consumer confidence by selling them "Cape's" as "Fancy Colored Diamonds"!

 

[strmrdr]

The issue with face up grading is talked about here...
Serg gets it, Garry read it,,,,

https://www.pricescope.com/community/threads/round-table-best-light-conditions-for-diamond-subjective-performance-comparison.115561/page-4

Lab color(non-fancy) is not consumer color, lab color is used by the trade to push the agenda of rarity not the color that the consumer will see until it gets to fancy colored diamonds where their is a huge advantage to the trade to all the sudden switch to face up.

 

[diagem]

Date: 5/23/2009 7:48:20 AM
Author: strmrdr
The issue with face up grading is talked about here...
Serg gets it, Garry read it,,,,

https://www.pricescope.com/community/threads/round-table-best-light-conditions-for-diamond-subjective-performance-comparison.115561/page-4

Lab color(non-fancy) is not consumer color, lab color is used by the trade to push the agenda of rarity not the color that the consumer will see until it gets to fancy colored diamonds where their is a huge advantage to the trade to all the sudden switch to face up.

I am not too certain face-up grading is such an advantage for the consumer either....

But am ready to listen to another tune...

 

[strmrdr]

Date: 5/23/2009 8:06:11 AM
Author: DiaGem

Date: 5/23/2009 7:48:20 AM

Author: strmrdr

The issue with face up grading is talked about here...

Serg gets it, Garry read it,,,,


https://www.pricescope.com/community/threads/round-table-best-light-conditions-for-diamond-subjective-performance-comparison.115561/page-4


Lab color(non-fancy) is not consumer color, lab color is used by the trade to push the agenda of rarity not the color that the consumer will see until it gets to fancy colored diamonds where their is a huge advantage to the trade to all the sudden switch to face up.

I am not too certain face-up grading is such an advantage for the consumer either....


But am ready to listen to another tune...

It would require a totally different grading scale and pricing model.

 

[Garry H (Cut Nut)]

Date: 5/23/2009 8:06:11 AM
Author: DiaGem

Date: 5/23/2009 7:48:20 AM
Author: strmrdr
The issue with face up grading is talked about here...
Serg gets it, Garry read it,,,,

https://www.pricescope.com/community/threads/round-table-best-light-conditions-for-diamond-subjective-performance-comparison.115561/page-4

Lab color(non-fancy) is not consumer color, lab color is used by the trade to push the agenda of rarity not the color that the consumer will see until it gets to fancy colored diamonds where their is a huge advantage to the trade to all the sudden switch to face up.

I am not too certain face-up grading is such an advantage for the consumer either....

But am ready to listen to another tune...

Why on earth would a first time buyer care about the color of a diamonds pavilion?

 

[strmrdr]

Date: 5/23/2009 8:08:56 AM
Author: Garry H (Cut Nut)

Date: 5/23/2009 8:06:11 AM

Author: DiaGem

Date: 5/23/2009 7:48:20 AM

Author: strmrdr

The issue with face up grading is talked about here...

Serg gets it, Garry read it,,,,


https://www.pricescope.com/community/threads/round-table-best-light-conditions-for-diamond-subjective-performance-comparison.115561/page-4


Lab color(non-fancy) is not consumer color, lab color is used by the trade to push the agenda of rarity not the color that the consumer will see until it gets to fancy colored diamonds where their is a huge advantage to the trade to all the sudden switch to face up.

I am not too certain face-up grading is such an advantage for the consumer either....


But am ready to listen to another tune...

Why on earth would a first time buyer care about the color of a diamonds pavilion?

because the trade insists it is a sign of rarity and bases the price on it.

 

[diagem]

Date: 5/23/2009 8:08:56 AM
Author: Garry H (Cut Nut)

Date: 5/23/2009 8:06:11 AM
Author: DiaGem

Date: 5/23/2009 7:48:20 AM
Author: strmrdr
The issue with face up grading is talked about here...
Serg gets it, Garry read it,,,,

https://www.pricescope.com/community/threads/round-table-best-light-conditions-for-diamond-subjective-performance-comparison.115561/page-4

Lab color(non-fancy) is not consumer color, lab color is used by the trade to push the agenda of rarity not the color that the consumer will see until it gets to fancy colored diamonds where their is a huge advantage to the trade to all the sudden switch to face up.

I am not too certain face-up grading is such an advantage for the consumer either....

But am ready to listen to another tune...

Why on earth would a first time buyer care about the color of a diamonds pavilion?

Because a Diamond is three dimensional..., you call it the pavilion (only)..., another might see it as part of the design..., (cut or jewelry design).
Cant dismiss this fact.

 

[Garry H (Cut Nut)]

Famous quote from Australia

"tell''em thar dreamin"

 

[diagem]

Date: 5/23/2009 8:23:06 AM
Author: Garry H (Cut Nut)
Famous quote from Australia

''tell''em thar dreamin''

Just dont tell them you are a Diamond dealer from Australia

....

 

[diagem]

Date: 5/23/2009 8:23:06 AM
Author: Garry H (Cut Nut)
Famous quote from Australia

''tell''em thar dreamin''

I did not hear or see anything that can make me see your vision as realistic...

I have no problem joining your vision if you can show me how?

I have touched this point numerous times already and never got past this point.
I am hoping this thread will break the barrier...

 

[Richard Sherwood]

If you're talking loose stones in the cape series, I find the SAS2000 to be in 95 to 98% agreement with GIA-GTL grading. The 2-5% it differs it usually grades tougher than GIA-GTL.

If you're talking mounted stones in the cape series, the accuracy drops down to 85%-95%, in my opinion. This varies according to the type of setting. The metal influence from a full yellow gold bezel or heavy yellow gold channel setting "flattens" the spectrum quite a bit more than that of a four prong platinum setting. The SAS2000 has a correcting feature that compensates for this pretty good, but in the end the eye makes the final call. All in all, it grades both loose and mounted stones better than any other spectrophotometer I've had experience with. Many spectrophotometers don't even grade mounted stones.

When you get into the brown and gray series spectrums, it requires a bit of experience to interpret the results correctly. Also stones with yellow fluorescence. These show different results than the cape series stones, and just require plain old familiarity to grade correctly.

It's great for grading fancy yellows as well, once you've built up a comparison database to work from.

Best $20,000 I ever spent. Diamond color grading is just one of it's many features. The value from colored stone analysis (aid in determing country of origin in particular) is hard to put a price tag on.

 

[Regular Guy]

Rich,

Of course I remember your comments previously about the SAS2000.

I see previous discussion goes to face up vs pavilion, and I don''t see you addressing this, although maybe you are; I also don''t know what "cape series" refers to.

Since you mention your investment in the SAS2000, it caused me to have a different question that may deserve a new thread, but I wonder what motivates the variance in spending for appraisers. For example, you mention thinking that was money well spent. To me it seems enormous, since it duplicates the cost you also had for your master set, and I read where other appraisers slowly fill in just alternate masters over a period of time.

Relatedly, separately, some might decide to invest in various technology to measure angles, in addition, or instead. I''m aware OGI and Sarin come in at different price points. As far as valuation, to me, on the one hand, this might be of greater value than a color definer, which duplicates what a master set does, suggesting maybe Helium would be good to invest in, if you''re going to put out some cash, as right now, only a dealer like Jonathan probably even has it in the US. On the other hand, I also read where only maybe 10% of an appraiser''s clients will arrive with an unmounted diamond anyway, making any such apparatus sitting like a lump mostly. And...maybe with things like ASETs, contemporary discussion aside, some of these costs could be reconsidered for their value as well.

Anyway, while this discussion is current, I thought I''d threadjack with apology for now.

Regards,

 

[diagem]

Thank you Richard for your info...

Date: 5/23/2009 3:10:05 PM
Author: Richard Sherwood
If you''re talking loose stones in the cape series, I find the SAS2000 to be in 95 to 98% agreement with GIA-GTL grading. The 2-5% it differs it usually grades tougher than GIA-GTL.

Cape series you mean off-colored Diamonds (R-Z range?)
What loose stones are you referring to? Rough/semi-polished or just polished?
Shapes? RB''s or are Fancy shapes and novelty shapes also part of your personal experience?
The SAS2000 grades face-up color appearance as well as true body color?

If you''re talking mounted stones in the cape series, the accuracy drops down to 85%-95%, in my opinion. This varies according to the type of setting. The metal influence from a full yellow gold bezel or heavy yellow gold channel setting ''flattens'' the spectrum quite a bit more than that of a four prong platinum setting. The SAS2000 has a correcting feature that compensates for this pretty good, but in the end the eye makes the final call. All in all, it grades both loose and mounted stones better than any other spectrophotometer I''ve had experience with. Many spectrophotometers don''t even grade mounted stones.

When you get into the brown and gray series spectrums, it requires a bit of experience to interpret the results correctly. Also stones with yellow fluorescence. These show different results than the cape series stones, and just require plain old familiarity to grade correctly.

Because it doesnt call the correct hue? Or has it something to do with with the quantity of color?
Does brown come up as a colorless D color result?
Only problem with yellow fluo, does it grade accurately stones with higher than med blue fluo.?
In any case I take it as human intervention is a required here.

It''s great for grading fancy yellows as well, once you''ve built up a comparison database to work from.

Grading Fancy yellow based on Body color or Face up look/appearance?
Does it take cut into consideration?

Best $20,000 I ever spent. Diamond color grading is just one of it''s many features. The value from colored stone analysis (aid in determing country of origin in particular) is hard to put a price tag on.

True!

 

[strmrdr]

Diagem see here:
http://www.adamasgem.org/sas2000.html

 

[Richard Sherwood]

Date: 5/23/2009 5:03:04 PM
Author: Regular Guy
Rich,

Since you mention your investment in the SAS2000, it caused me to have a different question that may deserve a new thread, but I wonder what motivates the variance in spending for appraisers. For example, you mention thinking that was money well spent. To me it seems enormous, since it duplicates the cost you also had for your master set, and I read where other appraisers slowly fill in just alternate masters over a period of time.

One of the major motivations for me was that I had both dealer and private clients bringing ln large, high color diamonds for grading in which the difference of a single color grade made a huge difference in price. Take for example the difference between a 5 ct. J/VS2 versus I/VS2. You''re talking about $10,000 dollars difference between the two.

The responsibility of accurate color grading in these situations weighed heavy upon me. My clients were looking to me to make sure they didn''t overpay, and in those kind of situations it''s comforting to have an "expert" second opinion. At GIA the graders can always pull in another grader or two to get their opinion on an important stone. As a one man show, I don''t have that luxury. Hence my interest in the SAS.

And don''t forget, diamond color grading is only one of the many applications the SAS has. Hence, you''re not just paying for a diamond color grading instrument, you''re also paying for a robust UV-VIS NIR (ultra-violet, near infra-red) spectrophotometer that can help tell you whether you''re looking at a Kashmir sapphire, Burma sapphire, a Sri Lankan sapphire or a Thai sapphire. The difference in price between these can be enormous.

Relatedly, separately, some might decide to invest in various technology to measure angles, in addition, or instead. I''m aware OGI and Sarin come in at different price points. As far as valuation, to me, on the one hand, this might be of greater value than a color definer, which duplicates what a master set does...

If I only had one choice of an instrument, the SAS, the Sarin and the OGI, I''d choose the SAS. It has helped me in far more critical situations than my OGI has.

suggesting maybe Helium would be good to invest in, if you''re going to put out some cash, as right now, only a dealer like Jonathan probably even has it in the US.

I took a serious look at the Helium when I was deciding which scanner to purchase. It''s accuracy was appealing, but at the time it only measured up to 2 carat stones (I''m not sure if it does larger yet). And it was a much larger instrument than I was interested in, as I do a lot of mobile work.

On the other hand, I also read where only maybe 10% of an appraiser''s clients will arrive with an unmounted diamond anyway, making any such apparatus sitting like a lump mostly. And...maybe with things like ASETs, contemporary discussion aside, some of these costs could be reconsidered for their value as well.

Agreed. You need a lot of loose diamond work to make shelling out $15 to $18,000 on an OGI/Sarin cost effective.

 

[Richard Sherwood]

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

Cape series you mean off-colored Diamonds (R-Z range?)

Cape (yellow) series refers to all diamonds whose tint of body color is determined by the trace element nitrogen (~ 95% of all diamonds). The SAS can detect this trace element from E color (possibly even D, I'll ask Marty) down through Fancy Vivid. The more nitrogen, the more color the diamond shows and the farther down the scale the spectrum falls.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

What loose stones are you referring to? Rough/semi-polished or just polished?

If the rough crystal has a clear and transparent crystal face, that'll work. Semi-polished with a nice window polished will work as well. The main reason I stipulated "loose" was to differentiate from "mounted". The contact of metal will "pull the spectrum down", at first glance giving you the impression the stone is darker than it really is. The SAS has a correcting feature for this, where it swivels the spectrum upon its fulcrum point to bring it up to where it would be without the metal's influence.

This feature works better (is more accurate) with minimal metal contact (such as four prongs) than with heavy metal contact (such as a full bezel).

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

Shapes? RB's or are Fancy shapes and novelty shapes also part of your personal experience?

It grades equally well and impartially for their "true" body color.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

The SAS2000 grades face-up color appearance as well as true body color?

No. It grades "true" body color only. That is, it grades the true color of the crystal, that which you would see if you were grading the stone face down, looking through the pavilion against a white background, under color corrected lighting.

It cannot detect an increase or decrease in color caused by the cut.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

When you get into the brown and gray series spectrums, it requires a bit of experience to interpret the results correctly. Also stones with yellow fluorescence. These show different results than the cape series stones, and just require plain old familiarity to grade correctly.

Because it doesnt call the correct hue? Or has it something to do with with the quantity of color?

No, it's because the GIA master spectrums input into the stones are all nitrogen based spectrums. In the case of browns you're talking about either different trace elements or deformation of the crystal (type IIa).

So in the case of brown or gray stones you'd be comparing the spectrums of oranges to those of apples.

That's one of the cool features of this spectrophotometer as opposed to the others available on the market. You can input a library of spectrums into the database to use for future comparison. Everytime I get a GIA certified stone in the brown series for example, I save that spectrum in my database with its corresponding color.

Same with fancy yellows.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

Does brown come up as a colorless D color result?

No, except for a type IIA that's been HPHT treated to a D color.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

Only problem with yellow fluo, does it grade accurately stones with higher than med blue fluo.?
In any case I take it as human intervention is a required here.

It grades fine with blue fluorescence (and even yellow if you take the time to build up a database). In fact, you have the option of grading color with or without ultraviolet light by way of a removable filter. With the filter in it eliminates the UV and gives a color grade the way the GIA originally trained us to grade. Or you can eliminate the filter and color grade with moderate UV influence. It's quite interesting to see the different results sometimes. Not usually a deal breaker, but then again, if you were buying a 5ct I VS2 would you rather have that be a non-UV influenced color grade or a "J" that happens to go "I" because of strong blue fluorescence? (Not that the "J" that goes "I" would be bad if it were priced appropriately, and the buyer had full knowledge to make an informed buying decision.)

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

It's great for grading fancy yellows as well, once you've built up a comparison database to work from.

Grading Fancy yellow based on Body color or Face up look/appearance?
Does it take cut into consideration?

It grades based on true body color, without taking cut or face-up appearance into consideration. Still, it's helpful with the right database built up.

 

[diagem]

Date: 5/23/2009 10:26:48 PM
Author: Richard Sherwood

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

Cape series you mean off-colored Diamonds (R-Z range?)

Cape (yellow) series refers to all diamonds whose tint of body color is determined by the trace element nitrogen (~ 95% of all diamonds). The SAS can detect this trace element from E color (possibly even D, I''ll ask Marty) down through Fancy Vivid. The more nitrogen, the more color the diamond shows and the farther down the scale the spectrum falls.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

What loose stones are you referring to? Rough/semi-polished or just polished?

If the rough crystal has a clear and transparent crystal face, that''ll work. Semi-polished with a nice window polished will work as well. The main reason I stipulated ''loose'' was to differentiate from ''mounted''. The contact of metal will ''pull the spectrum down'', at first glance giving you the impression the stone is darker than it really is. The SAS has a correcting feature for this, where it swivels the spectrum upon its fulcrum point to bring it up to where it would be without the metal''s influence.

This feature works better (is more accurate) with minimal metal contact (such as four prongs) than with heavy metal contact (such as a full bezel).

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

Shapes? RB''s or are Fancy shapes and novelty shapes also part of your personal experience?

It grades equally well and impartially for their ''true'' body color.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

The SAS2000 grades face-up color appearance as well as true body color?

No. It grades ''true'' body color only. That is, it grades the true color of the crystal, that which you would see if you were grading the stone face down, looking through the pavilion against a white background, under color corrected lighting.

It cannot detect an increase or decrease in color caused by the cut.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

When you get into the brown and gray series spectrums, it requires a bit of experience to interpret the results correctly. Also stones with yellow fluorescence. These show different results than the cape series stones, and just require plain old familiarity to grade correctly.

Because it doesnt call the correct hue? Or has it something to do with with the quantity of color?

No, it''s because the GIA master spectrums input into the stones are all nitrogen based spectrums. In the case of browns you''re talking about either different trace elements or deformation of the crystal (type IIa).

So in the case of brown or gray stones you''d be comparing the spectrums of oranges to those of apples.

That''s one of the cool features of this spectrophotometer as opposed to the others available on the market. You can input a library of spectrums into the database to use for future comparison. Everytime I get a GIA certified stone in the brown series for example, I save that spectrum in my database with its corresponding color.

Same with fancy yellows.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

Does brown come up as a colorless D color result?

No, except for a type IIA that''s been HPHT treated to a D color.

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

Only problem with yellow fluo, does it grade accurately stones with higher than med blue fluo.?
In any case I take it as human intervention is a required here.

It grades fine with blue fluorescence (and even yellow if you take the time to build up a database). In fact, you have the option of grading color with or without ultraviolet light by way of a removable filter. With the filter in it eliminates the UV and gives a color grade the way the GIA originally trained us to grade. Or you can eliminate the filter and color grade with moderate UV influence. It''s quite interesting to see the different results sometimes. Not usually a deal breaker, but then again, if you were buying a 5ct I VS2 would you rather have that be a non-UV influenced color grade or a ''J'' that happens to go ''I'' because of strong blue fluorescence? (Not that the ''J'' that goes ''I'' would be bad if it were priced appropriately, and the buyer had full knowledge to make an informed buying decision.)

Date: 5/23/2009 5:36:30 PM
Author: DiaGem

It''s great for grading fancy yellows as well, once you''ve built up a comparison database to work from.

Grading Fancy yellow based on Body color or Face up look/appearance?
Does it take cut into consideration?

It grades based on true body color, without taking cut or face-up appearance into consideration. Still, it''s helpful with the right database built up.

Richard,

Does the SAS2000 work in the same fashion as the Yehuda colorimeter?: http://www.diamond-color.net/Home.aspx
It sounds similar except a few....
How does the SAS2000 compare with IDL''s technology? (Peter Meuss claims its 99% accurate!)

I remember Garry mentioning Marc Brauner said: https://www.pricescope.com/community/threads/new-kid-laboratory-in-town.67744/page-2 (about half way down the thread page).

"...There was the Eickhorst colorimeter 30 years ago, not bad at all but they don''t make them any longer, then came the Yehuda colorimeter which had 2 versions (one very accurate with test a & test b, the other smaller with one test only, not bad but less accurate); the Yehuda colorimeter was especially used for rough diamonds, it could see through skin (coated) of any color, and was accurate for brownies and fluor.

To our knowledge, the Yehuda colorimeter is the most precise colorimeter ever built.

Then the Gran-I (not so accurate and influenced by dirt or ink on the stone, and totally inaccurate when fluorescence or brown-green-grey nuances), and the Gran-II which later became Sarin''s Colorimeter, which is not bad, but then it tends to make inpredictable errors making this toy inconsistent. But overall I must say it''s not bad.

IDL''s technology: Peter Meuss (ex CEO of HRD and now head of IDL and the Dubai diamond center) claims it is 99% accurate, however, during IDL''s presentation in India we were told that once fluorescence is Medium or more, or when the stone is a browny, then the said instrument cannot give a result (or perhaps it could mean that its results are inconsistent, inaccurate when there is a certain level of fluor or a brown nuance, just like Gran''s colorimeter)...."[end quote]

Me also..., based on my experience..., I have not encountered a machine which beats the Yehuda colorimeter in accuracy..., BUT I find it far from being accurate (eg 95-98%)!
My personal experience with the Yehuda colorimeter is:

D-G accuracy in the 90''s %
H-I accuracy is less than the D-G.
J-K-L seems way off (depending greatly on the final shape).
And the most important fact, even Yehuda himself admits that his colorimeter is NOT accurate when it comes to off colored (cape series) Diamonds (can be 2 shades + or - according to him).
Another issue according to Yehuda..., Diamond material from some specific geographic area''s are not grade-able. (period)

So here I am asking again about shapes...., the Yehuda Colorimeter results are based on round brilliant polished shapes (eg, fancy shapes will usually loose a grade or even two depending on the model), doesnt the SAS2000 take shapes into calculation?
Since you say the SAS2000 can''t detect an increase or decrease in color caused by cut..., how can someone match the ''Fancy Color Grade'' issued to the specific cut of the stone? After all, Fancy Colored Diamonds are graded face-up by GIA. How can one built a database on this subject with so many variable Fancy colored cuts existing?

Same with fluorescence, the Yehuda colorimeter cant issue a color grade on Diamonds which possess medium and higher (any colored) fluorescence..., I understand from you the SAS2000 accurately issue a color grade to Diamonds with fluorescence..., that seems as a groundbreaking option that is non-existent in other machines...

 

[John P]

Hi guys - just back from a long hiatus so forgive if I'm not totally up to speed.

For the record, in the big (consumer) picture, I'm with Strm here.

That said, in the gemologist-picture, Rich has given a great description of the versatility of the SAS 2000... Colorimetry is just a feature. It was developed for total spectral analysis, ID and treatment detection of both diamonds and colored stones where possible in the visible NIR range. Marty Haske nicknamed it the “third gemologist who doesn’t leave early, take off weekends or come in with a hangover.”

I'm aware of the SAS 2000 being used at several major labs (not sure about the Yehuda or others?). Branko Deljanin, someone I consider at the front of our trade's detection of man-mades and treatments, has several at his disposal. If you’re interested in the technical aspects there are many prior threads between Serg and Marty which show its capabilities; in particular, those involving cross-checks of GIA’s published forward Monte Carlo brilliance & fire analyses with Marty’s published comparisons.

I was speaking with Marty and mentioned this thread to him. I hope he’ll pop in to address specifics.

 

[adamasgem]

Date: 5/24/2009 8:09:21 AM
Author: DiaGem


Richard,

Does the SAS2000 work in the same fashion as the Yehuda colorimeter?: http://www.diamond-color.net/Home.aspx
It sounds similar except a few....
How does the SAS2000 compare with IDL''s technology? (Peter Meuss claims its 99% accurate!)

I remember Garry mentioning Marc Brauner said: https://www.pricescope.com/community/threads/new-kid-laboratory-in-town.67744/page-2 (about half way down the thread page).

''...There was the Eickhorst colorimeter 30 years ago, not bad at all but they don''t make them any longer, then came the Yehuda colorimeter which had 2 versions (one very accurate with test a & test b, the other smaller with one test only, not bad but less accurate); the Yehuda colorimeter was especially used for rough diamonds, it could see through skin (coated) of any color, and was accurate for brownies and fluor.

To our knowledge, the Yehuda colorimeter is the most precise colorimeter ever built.

Then the Gran-I (not so accurate and influenced by dirt or ink on the stone, and totally inaccurate when fluorescence or brown-green-grey nuances), and the Gran-II which later became Sarin''s Colorimeter, which is not bad, but then it tends to make inpredictable errors making this toy inconsistent. But overall I must say it''s not bad.

IDL''s technology: Peter Meuss (ex CEO of HRD and now head of IDL and the Dubai diamond center) claims it is 99% accurate, however, during IDL''s presentation in India we were told that once fluorescence is Medium or more, or when the stone is a browny, then the said instrument cannot give a result (or perhaps it could mean that its results are inconsistent, inaccurate when there is a certain level of fluor or a brown nuance, just like Gran''s colorimeter)....''[end quote]

Me also..., based on my experience..., I have not encountered a machine which beats the Yehuda colorimeter in accuracy..., BUT I find it far from being accurate (eg 95-98%)!
My personal experience with the Yehuda colorimeter is:

D-G accuracy in the 90''s %
H-I accuracy is less than the D-G.
J-K-L seems way off (depending greatly on the final shape).
And the most important fact, even Yehuda himself admits that his colorimeter is NOT accurate when it comes to off colored (cape series) Diamonds (can be 2 shades + or - according to him).
Another issue according to Yehuda..., Diamond material from some specific geographic area''s are not grade-able. (period)

So here I am asking again about shapes...., the Yehuda Colorimeter results are based on round brilliant polished shapes (eg, fancy shapes will usually loose a grade or even two depending on the model), doesnt the SAS2000 take shapes into calculation?
Since you say the SAS2000 can''t detect an increase or decrease in color caused by cut..., how can someone match the ''Fancy Color Grade'' issued to the specific cut of the stone? After all, Fancy Colored Diamonds are graded face-up by GIA. How can one built a database on this subject with so many variable Fancy colored cuts existing?

Same with fluorescence, the Yehuda colorimeter cant issue a color grade on Diamonds which possess medium and higher (any colored) fluorescence..., I understand from you the SAS2000 accurately issue a color grade to Diamonds with fluorescence..., that seems as a groundbreaking option that is non-existent in other machines...

1) When GIA graded masterstones for AGS , they would only quote to 10% of a AGS color grade range (i.e. 0.05)
2) The 1 sigma standard deviation between the nominal grade of 17 masters and the GIA best estimate on a 17 stone sample I owned, based entirely on GIA''s numberes , was about 16% or so of a color grade range... Those are the yardsticks we deal with in doing any sort of calibration

3) I AM NOT SURE, but i believe the Yehuda is either a bi or tri stimulous colorimeter, and may not be spectrophotometer based, like the SAS20000 with 1.5nm optical resolution and 0.35nm pixel quantization, using this data, I use color science to estimate a color grade, obviously fancy shapes require some adjustment and accuracy will be effected..

4) I have software based solutions for my clients, which included the valuators for the South African government in 2000, to help value the DeBeers stockpile, which involve statistical analysis of rough, and my client list included General Electric, many labs and producers of synthetic and HPHT processed stones, cut and polished.

5) I would say, scientifically, that a quote of 99% accuracy is PURE B***S**T. You may get approaching that in repeatability, but not accuracy. If I were them,I would tone done the hyperbole, rememeber they are in a part of the world that doesn''t like it (Islamic Law) PS (I have a unit in Dubai at the central muncipality government lab)

6) I built rough to polished algorithms for fancy colored diamond dealers , who seem to be happy, they make money knowing what they have

7) I developed a crogenic technique to help sharpen lines in the Vis-Nir spectra for determination of treatments, including Raman to detect HPHT and irradiated (most irradiated detection is easily) at room temperature.

I know diamond dealers want a black box that spits out a price when you put any stone in, it just doesn''t exist.

Rough to polished requires using both the knowledge of the rough shape and projected cut shape, and diamonds are not nessarily homogeneous, so it is no more than an educated guess, or based on statistics (that my fancy colored clients are using based on integral software)

As to fluorescence, you filter it out like it should be, read my website (Is your G a J?) http://www.adamasgem.org , as wellas AGA presentations on their website regarding this issue of overgrading of fluorescent stones.

ANY DEVICE NEEDS a human in the loop, I wave a lot of read flags when I detect anomolies, and try to educate people on a very complex problem. The software also works in identification of colored stones, CIE correlations among
various color systems used in the trade, Foreward Monte-Carlo ray tracing to compare with GIA''s published results on their Brilliance study and Fire study ( threads showing data comps are on PS), gemological database, radiation deposition (electron andxray and gamma), dynamic optics tutorials.. etc, etc, etc..

 

[adamasgem]

Date: 5/23/2009 2:54:23 AM
Author:DiaGem
I am starting a new thread on the issue which I keep reading on the fact that there are colorimeters that can grade Diamonds colors to an accuracy of 95-98%.

I have never yet met this kind of accuracy not to mention consistency!

David Atlas..., can you please elaborate more on your writing from another thread:

https://www.pricescope.com/community/threads/round-table-best-light-conditions-for-diamond-subjective-performance-comparison.115561/page-3 (I started a new thread to avoid a threadjack on the former).

re> ''All technology improves over time, but it must enter the market a bit imperfect and then users along with technicians and inventors improve its performance levels. A machine to grade color currently cannot identify problem diamonds with 100 or 99 percent accuracy. Not yet. I think 95% to 98% is possible right now.''

We would love to hear some more.....

There are confidence bounds on any absolute color estimate because of the variance (uncertainty) of the "blessed" yardsticks used (masters) to calibrate any machine, let alone the cut shape issue or the issue in 30+% of stone on fluorescence, which historically should be eliminated, but G-d In America, in Winter 2008 G&G , has now told us what they had been saying for 50+ years is wrong, yadda, yadda, yadda, as it is all about money Even a faint fluorescent stone could trigger a reaction on a borderline grade to make it appear better. Hell, GIA even used to TEACH in their course material to give the benefit of the doubt to the stone, just as they taught that "diamonds should be graded at their POORER color, in artificial light devoid of ultraviolet". Now they have changed their tune, actually beginning in 1997 when they said that flurescence didn''t effect the color grade and in 2002 (about the time when DeBeers went private) where they showed the grader holding the diamond right up to the light to pump in as much UV as the fluorescent bulb could put out.

My stuff from the late 1990s
http://www.adamasgem.org/giafluor.html

AGA Lighting task force files and presentation 2009
http://accreditedgemologists.org/currevent.php


See also comments above on my last post on the statistics of masters.

 

[Richard Sherwood]

Date: 5/24/2009 8:09:21 AM
Author: DiaGem
Richard,
Does the SAS2000 work in the same fashion as the Yehuda colorimeter?: http://www.diamond-color.net/Home.aspx
It sounds similar except a few....

I''m not familiar with the inner workings of the Yehuda colorimeter. I''d be interested in more technical data regarding it than the simplistic description they give on the URL address you supplied.

My gut instinct though is no, they do not work in the same fashion. The Yehuda colorimeter seems to be more of a "black box" type of instrument symplifying the results for the average user, "dumbing down", if you will.

The SAS is a much more interactive instrument, which involves an interaction between the user and the instrument. In other words, a highly trained gemologist with experience in spectroscopy is going to understand the results of the SAS much better than a non-gemologist, or a gemologist who is not proficient in spectroscopy.

I myself like this, because I like to know the "why" of things rather than just accept a digital reading. I want to see the spectrum with my own eyes, and interpret its readings with my knowledge and experience, just as I would when color grading a diamond with my own eyes.

Date: 5/24/2009 8:09:21 AM
Author: DiaGem
Richard,
How does the SAS2000 compare with IDL''s technology? (Peter Meuss claims its 99% accurate!)

IDL''s technology: Peter Meuss (ex CEO of HRD and now head of IDL and the Dubai diamond center) claims it is 99% accurate, however, during IDL''s presentation in India we were told that once fluorescence is Medium or more, or when the stone is a browny, then the said instrument cannot give a result (or perhaps it could mean that its results are inconsistent, inaccurate when there is a certain level of fluor or a brown nuance, just like Gran''s colorimeter)....''[end quote]

It sounds like Peter Meuss''s claim comes with a qualification (ie, when the stone is a "browny", or has fluorescence of Medium or more.

I can understand these qualifications. A "browny" definitely requires a different analysis and interpretation than a cape (yellow) series stone. And it seems he is acknowledging how fluorescence can affect modern day grading (it didn''t use to).

That''s what I love about the SAS (yes, I admit it, the SAS and I are having an affair). Since the SAS is not a "black box", the user can interact with it to solve these dilemnas. In the case of "brownies", the user can build up a database of the "brown series" of stone (especially certed stones) to help him in future grading. In the case of fluoresence, the user can grade the stone without UV (using the filter), and with UV (minus the filter), and then arrive at his or her determination.

It is a much more informed decision.

Date: 5/24/2009 8:09:21 AM
Author: DiaGem
Richard,
I remember Garry mentioning Marc Brauner said: https://www.pricescope.com/community/threads/new-kid-laboratory-in-town.67744/page-2 (about half way down the thread page).

''...There was the Eickhorst colorimeter 30 years ago, not bad at all but they don''t make them any longer, then came the Yehuda colorimeter which had 2 versions (one very accurate with test a & test b, the other smaller with one test only, not bad but less accurate); the Yehuda colorimeter was especially used for rough diamonds, it could see through skin (coated) of any color, and was accurate for brownies and fluor.

To our knowledge, the Yehuda colorimeter is the most precise colorimeter ever built.

..., based on my experience..., I have not encountered a machine which beats the Yehuda colorimeter in accuracy..., BUT I find it far from being accurate (eg 95-98%)!
My personal experience with the Yehuda colorimeter is:
D-G accuracy in the 90''s %
H-I accuracy is less than the D-G.
J-K-L seems way off (depending greatly on the final shape).
And the most important fact, even Yehuda himself admits that his colorimeter is NOT accurate when it comes to off colored (cape series) Diamonds (can be 2 shades + or - according to him).
Another issue according to Yehuda..., Diamond material from some specific geographic area''s are not grade-able. (period)

Well, from your own experience I would say the Yehuda colorimeter doesn not measure up to the accuracy of the SAS2000.

D-G accuracy in the 95-98% range.
H-I equally as accurate.
J-K-L equally as accurate.
Off-colored cape (yellow) series equally as accurate, bearing in mind that even the GIA resorts to a two grade designation on these stones (ie R-S coloas opposed to the single grade of R, or S)
Regarding specific geographic areas, I would have to agree with this. Some areas yield difficult to grade stones, either with the eye or with a machine. Mostly my experience with these stones is relegated to South American stones, often brown stones with yellow fluorescence. These stones are just plain weird to grade, either with a machine or with the naked eye. They require expert human analysis.

Date: 5/24/2009 8:09:21 AM
Author: DiaGem
Richard,
Then the Gran-I (not so accurate and influenced by dirt or ink on the stone, and totally inaccurate when fluorescence or brown-green-grey nuances), and the Gran-II which later became Sarin''s Colorimeter, which is not bad, but then it tends to make inpredictable errors making this toy inconsistent. But overall I must say it''s not bad.

I refer to the Gran Colorimeter as the "EGL" machine. It grades like the EGL, in my opinion. Usually off a color grade. Way off on brown series stones. One of my good friends is a Belgium diamond dealer here in Sarasota, who owns the Gran. When he has a critical sale pending, he brings the stone to me to grade instead of trusting his Gran colorimeter.

Date: 5/24/2009 8:09:21 AM
Author: DiaGem
So here I am asking again about shapes...., the Yehuda Colorimeter results are based on round brilliant polished shapes (eg, fancy shapes will usually loose a grade or even two depending on the model), doesnt the SAS2000 take shapes into calculation?

The SAS grades the true body color of the crystal. It''s up to the user to determine the final grade.

When I was taught by the GIA, they said to grade a fancy by assessing the color through the "middle" and "ends" of the stone. In an oval, for example, the color tends to be whiter in the center and more concentrated in the ends. They said to take an average of these two colors when assigning the final grade.

The SAS doesn''t differentiate between these distinctions. It just tells you what the body color of the crystal is, based on its nitrogen content. It''s up to you to make the final call, just as you would visually color grading the stone. This is another instance of where a (knowledgeable) human needs to interact with the instrument.

Date: 5/24/2009 8:09:21 AM
Author: DiaGem
Richard,
Since you say the SAS2000 can''t detect an increase or decrease in color caused by cut..., how can someone match the ''Fancy Color Grade'' issued to the specific cut of the stone? After all, Fancy Colored Diamonds are graded face-up by GIA. How can one built a database on this subject with so many variable Fancy colored cuts existing?

It''s a little trickier. I do it by building a database according to the cut of the stone. Everytime I get a "certified" radiant cut fancy light yellow or better stone, I input its spectrum into that particular database. Then, to further qualify it, I input the dimensions to get a feel for the "face-up" look of the stone (deep versus shallow, etc.)

It''s not a perfect system yet, but it''s very helpful. Marty doesn''t advertise this as one of the features of the SAS yet. I''ve been working on him to build up an extensive fancy database, but I don''t think he has the time, and I suspect he isn''t interested in getting involved in the "politics" of fancy colored diamond grading at this time.

Date: 5/24/2009 8:09:21 AM
Author: DiaGem
Richard,
Same with fluorescence, the Yehuda colorimeter cant issue a color grade on Diamonds which possess medium and higher (any colored) fluorescence..., I understand from you the SAS2000 accurately issue a color grade to Diamonds with fluorescence..., that seems as a groundbreaking option that is non-existent in other machines...

It isn''t exactly ground breaking. The SAS technology has been out for decades with this capability.

Another note... Before Marty came out with this technology, you had to spend 100k plus for a machine with these capabilities. Marty basically blew the game out of the water when he came out with his technology.

But, because Marty is another "one man show", his technology often gets eclipsed by the slick marketing of the larger companies like Yehuda, OGI, and Gran/Sarin.

In truth they don''t come close to his work, in my opinion.

 

[oldminer]

The colorimeters which are for sale grade from the face-up position, face down in reality, but it is the same. ImaGem uses face up, face down, too but in a far more sophisticated way. There are multiple measures on a revolving stone. It is not a process done in only one position over a sensor. Many, far from a few, GIA graded diamonds were assessed so that the actual range of scientific color for each color grade was established. This covers the cape series of colors, but the theory can be extended to brown, pink, blue, whatever color might exist. The idea was not to re-make the GIA''s result, but to mimic it with scientific precision.

Of course, if the trained observer disagrees with a result, it is hard to prove a machine is more correct than the trained observer. The human eye is not the very best tool for color grading, but we must live with our physical limitations. The machine can define the exact color space of any color it is programmed for. There will be issues with elongated fancy shapes which intensify face-up color in small zones. In time, machines can do what our eyes can do. It is programming and money dependent, not impossible.

Machines such as ImaGem''s GL3100 can grade with scientific repeatability unlike human graders. The grades are objective and the machines do not tire near the end of the work day. Borderline calls will continue to exist, but the borderline gray zones will get far narrower. ImaGem currently grades round and princess cut diamonds with its GL3100. A qualified GG supervises and can over-ride a grade on a diamond which for some reason, such as a central black inclusion, interferes with the readings. This tightens up the repeatability of grading while allowing for traditional approaches to be further fine tuned. Within a few minutes a diamond can be color graded, fluorescence graded, full measured, and light performance analyzed. This takes 1 tweezer placement to put it one the device and one tweezer placement to remove it. The efficiency this brings in a lab environment is highly important although admittedly not fully understood by a large majority of the trade as yet.

Dr. Aggarwal, the ImaGem principal and inventor, will be on a panel at the JCK Las Vegas Show next week along with Garry H and Peter Yantzer, discussing his processes. I will be in the audience where he will be introducing our mutually developed program, DFS, for grading the craft aspects of a diamond.

The members of the trade may initially be distressed about the likelihood of consistently correct and scientific color grading, but I suspect that in time, they will grow to accept it and even demand it. Humans have a tendency to favor order and discipline. Categorization of the aspects of the world around us into neat packets applies to most everything including diamond color grading. It may take time, but I am looking into my color grading crystal ball and making an informed guess.

 

[diagem]

Thanks you (all) guys,

So I think we are pretty much in understanding at present that human intervention is needed when grading the color [and clarity] of Diamonds...
Machines can definitely help (as they do with most everything else when used efficiently).

*John P..., can you elaborate a bit more on you mentioning:

re> "For the record, in the big (consumer) picture, I''m with Strm here."

Any suggestions how to adapt a new grading model that will be better for consumers?

*Dave...,

re> "The members of the trade may initially be distressed about the likelihood of consistently correct and scientific color grading, but I suspect that in time, they will grow to accept it and even demand it. Humans have a tendency to favor order and discipline. Categorization of the aspects of the world around us into neat packets applies to most everything including diamond color grading. It may take time, but I am looking into my color grading crystal ball and making an informed guess."

I dont think its about being distressed...., and I agree that if it can work the trade will embrace it as it does other issues...
I just dont see how the consumer will gain from it at the present...

But am willing to listen to reason..., not dreams...

 

[Serg]

Date: 5/25/2009 3:26:29 AM
Author: DiaGem
Thanks you (all) guys,


So I think we are pretty much in understanding at present that human intervention is needed when grading the color [and clarity] of Diamonds...

Machines can definitely help (as they do with most everything else when used efficiently).


*John P..., can you elaborate a bit more on you mentioning:


re> ''For the record, in the big (consumer) picture, I''m with Strm here.''


Any suggestions how to adapt a new grading model that will be better for consumers?


*Dave...,


re> ''The members of the trade may initially be distressed about the likelihood of consistently correct and scientific color grading, but I suspect that in time, they will grow to accept it and even demand it. Humans have a tendency to favor order and discipline. Categorization of the aspects of the world around us into neat packets applies to most everything including diamond color grading. It may take time, but I am looking into my color grading crystal ball and making an informed guess.''


I dont think its about being distressed...., and I agree that if it can work the trade will embrace it as it does other issues...

I just dont see how the consumer will gain from it at the present...


But am willing to listen to reason..., not dreams...

re:So I think we are pretty much in understanding at present that human intervention is needed when grading the color [and clarity] of Diamonds...

Diagem,

We near the point where human and machine accuracies for grading color are similar. ( Fluorescence is problem for both approach, because clear rules for grading is absent)
Evolution line for color grading : "Bi-line or Tri-line colorimeters", "transmission absorption spectrum", "Absorption spectrum and exact 3d models".

First two technology based on Statistic and can not account color pattern in diamond, accuaracy depends from positions sample and difference between theoretical or real 3d models had been used for Statistic data base and 3D model object what you want grade .
Third color grading technology less depends from Statistic databased but need much more complex model to model human vision.
Third technology could be ready for industry use in few years.