Round Table: Best light conditions for Diamond Subjective performance comparison

[diagem]

Date: 5/19/2009 9:38:25 AM
Author: Serg

Date: 5/19/2009 9:33:40 AM
Author: oldminer
Sergey, I am talking about ''grading'' not a final, personal judgment of ''beauty''. Please don''t confuse what I have said.


On the issue of Subjective Grading, it makes no sense to speak of it when we can grade objsctively while still retaining personal taste and budget in the final selection. We grade D color, but not everyone wants a D color. If we would accept objective color grading, there would be no change in the ability to choose, but we could choose with way more confidence. I see no difference in the argument and am not trying to take the thread off course. I am trying to remind people that subjective grading has been successful while leading to a lot of problems which objective grading could help solve.

Dave,

I did not understand your last post

It will always remain subjective...., even if the name changes to objective

.

 

[Regular Guy]

Date: 5/19/2009 8:19:45 AM
Author: Serg


re:To GRADE things we always pick a standardized environment, not multiple environments.

1) Sport car, Land Cruiser, Urban Cruiser,.. Could you Grade these car in one environment
2) Should we GRADE BEAUTY by single number? Is it possible? Do you know such example from other industry, life ..?
3)Should consumer, professionals test Brilliancy, Fire, Scintillation in ONE environment

Biggest disadvantage of current gemological grade, what gemologists grade what they can and how they can, instead what consumers need and want.

Yes, Serg great post.

I wonder if "camera" logic helps?

Modern camera do come with different assumptive backgrounds you pick...at least point & shoot modern models. Canon & others have you make a choice: portrait, night, kids & running. But, there is also an automatic option. What does it assume.

And then, assume either prior to point & shoot, or higher level cameras where it''s understood the picture taker manipulates all the settings. Is there a more helpful "alpha" position.

Relatedly, does it make any sense to speak about the extent a diamond''s cut has more inherent "bounce-abiltiy." Somewhat? I can see, for example, that it would be different to talk about both inclusions and color, since they are both "with" the diamond." The light performance...it''s not "in" the diamond. It''s how the diamond will allow and encourage a bounce from something not in the diamond.

 

[strmrdr]

RG,
light performance is not a physical property of the diamond.
It is a reaction of the diamond to its environment.

color
clarity
polish
symmetry

Are all physical and should be facts not subjective like they are now.

Light performance on the other hand will always be subjective.
But you can use information/facts to measure performance under one condition.
If you do that for enough conditions you start getting close to stating the performance as a fact but you will never get there.

 

[Regular Guy]

Karl

Date: 5/19/2009 10:35:01 AM
Author: strmrdr
RG,
light performance is not a physical property of the diamond.
It is a reaction of the diamond to its environment.

color
clarity
polish
symmetry

Are all physical and should be facts not subjective like they are now.

Light performance on the other hand will always be subjective.
But you can use information/facts to measure performance under one condition.
If you do that for enough conditions you start getting close to stating the performance as a fact but you will never get there.

I thought I said most of what you did, too. But, I thought I was stretching in a positive way by introducing the logic of what a camera does. Take your last line. It ls worth approximating. They made calculus from that. And it may help us?

 

[Moh 10]

Today we have:

1. Grading - Labs/appraisers grade with what they consider to be standardized lighting.
2. Selling - Sellers will always light (live and in photos) diamonds to maximize sales.
3. Consumers - They live with their diamonds moving in an out of a zillion lighting conditions.

What could possibly change?
I''m confused, Strm, what is broken that you are trying to fix with this thread?

 

[oldminer]

When grading, one grades by standardized and accepted conventions. Grading of most everything is based on accepted rules. Grading is not a perfect conversion to absolute truths nor does it necessarily directly correlate to personal opinion. A grade is only as important as we make it. We can choose to not go for a top grade because we have a preference for some lower grade. In this way, the final choice may be subjectively chosen in spite of objective measures and grades which may be available.

When one grades color for dying cloth or printing, one uses standardized lighting, standard color swatches and not many different lighting scenarios. You may want to check for metamerism(sp) based on different lighting color temperatures with specific dyes and inks, but color grading of things other than diamond is a one lighting scenario. Light performance needs only one lighting scenario to create a meaningful grade. You always have the freedom to use other lighting to make a final choice and I am for broad choices to be available, but grading needs to be done with as few variables as possible, made highly consistent and repeatable and must correlate tightly with human perception.

I do not equate "beauty" directly with "grading" Successful grading strategies must be developed for any shape graded which makes human perception generally correlate to the grade given. I am certain there are diamonds which look their best when not returning the maximum possible light. Asscher cut is one which comes immediately to mind. How to construct a way to properly grade such diamonds for their unique light behavior properties is a big task, but it is being tackled. When done correctly, the grade and the subjective comparison approach ought to be quite similar. If you are an engineer and simply want to know which diamond returns more light, that is an easy task which can be done now. However, lhigh amounts of ight return alone do not prove beauty or quality. It is a physical fact which engineers like to know, but it stops right there.

If all you want is subjective grading, we don't need to advance at all. We have our eyes and minds now and can grade subjectively without any problem. Honesty is the worst problem when it comes to credibility of subjective grading provided to consumers. Give me a perfectly truthful seller who is unbiased and experienced. That would be a very big and difficult thing to come by and is the real reason to get to objective grading.

You can compare subjectively in any lighting conditions you prefer, but grading should be on a higher plane.

 

[Judah Gutwein]

It seems like I''m not the only one who was/is confused by the crux/premise of this round table discussion.

My initial understanding based on the title of this thread was that Storm was initiating a discussion on the best way to capture realistic performance/output of a diamond under standardized and authentic lighting conditions. Thus, my earlier response was to point to two companies who manufacture lighting instruments for supplying a uniform and balanced source of pure white light (using diffuseres), as well as mimicking natural daylight (daylight-LEDs (5800° Kelvin). I pointed out that proper viewing conditions would not automatically extend to proper technique for actually capturing these conditions once the external variable of the camera is introduced into the equation.

Storm, came back to clarify that he is talking here about optimal conditions for "subjective" comparison of loose diamond LIGHT performance, specifically.
He further clarified for Moh 10 on this thread, that his intentions here are to cover (optimal light performance conditions for) both, viewing and photographing loose diamonds.

What has actually evolved here, seems to be a cross between a focus on light performance and a focus on a standardized environment for viewing/assessing loose diamonds and with no special emphasis on light performance. These are two VERY different tangents as Ira points out earlier.

Optimal viewing conditions can be independent from optimal light performance conditions. ''Viewing conditions'' is actually referencing and alluding to the inherent physical qualities and characteristics from within the diamond. Whereas, optimal light performance is a reaction of the diamond to its environment (as pointed out earlier by Storm). A loose diamond on the finger is never static and is always reflecting and refracting available light back to the viewer in the form of ''light performance''.

Some have come on to post with emphasis relating to the (standardized) actual assessment and grading of loose diamonds (color/clarity). What does this have to do with light performance?? It seems as though not everyone is on board with the crux of this discussion. Given my understanding from reading these postings, things could get pretty convoluted here and very quickly.

Therefore, I think it would be beneficial to re-calibrate here and to distinguish between these 2 (very different) variables and to re-emphasise Storms intentions here.

The discusssion here is on finding an optimal way to view and capture (using photography for online viewers) loose diamond light performance.

Within this context, Storm uses the word "subjective" in his title. This is important, since once you essentially realize that this stuff is subjective, it is then extremely difficult, by definition, to develop an objective standard condition(s) for deriving a subjective result??

Therefore, the best we can do is to discover and create the BEST or OPTIMAL conditions to view and capture the above.

We looking at Consumers? Grading labs? Sellers?

This is not going to be a simple ''fix'' that''s for sure!

Judah

 

[Serg]

Date: 5/19/2009 11:37:29 AM
Author: oldminer
When grading, one grades by standardized and accepted conventions. Grading of most everything is based on accepted rules. Grading is not a perfect conversion to absolute truths nor does it necessarily directly correlate to personal opinion. A grade is only as important as we make it. We can choose to not go for a top grade because we have a preference for some lower grade. In this way, the final choice may be subjectively chosen in spite of objective measures and grades which may be available.


When one grades color for dying cloth or printing, one uses standardized lighting, standard color swatches and not many different lighting scenarios. You may want to check for metamerism(sp) based on different lighting color temperatures with specific dyes and inks, but color grading of things other than diamond is a one lighting scenario. Light performance needs only one lighting scenario to create a meaningful grade. You always have the freedom to use other lighting to make a final choice and I am for broad choices to be available, but grading needs to be done with as few variables as possible, made highly consistent and repeatable and must correlate tightly with human perception.


I do not equate ''beauty'' directly with ''grading'' Successful grading strategies must be developed for any shape graded which makes human perception generally correlate to the grade given. I am certain there are diamonds which look their best when not returning the maximum possible light. Asscher cut is one which comes immediately to mind. How to construct a way to properly grade such diamonds for their unique light behavior properties is a big task, but it is being tackled. When done correctly, the grade and the subjective comparison approach ought to be quite similar. If you are an engineer and simply want to know which diamond returns more light, that is an easy task which can be done now. However, lhigh amounts of ight return alone do not prove beauty or quality. It is a physical fact which engineers like to know, but it stops right there.


If all you want is subjective grading, we don''t need to advance at all. We have our eyes and minds now and can grade subjectively without any problem. Honesty is the worst problem when it comes to credibility of subjective grading provided to consumers. Give me a perfectly truthful seller who is unbiased and experienced. That would be a very big and difficult thing to come by and is the real reason to get to objective grading.


You can compare subjectively in any lighting conditions you prefer, but grading should be on a higher plane.

re:When one grades color for dying cloth or printing, one uses standardized lighting, standard color swatches and not many different lighting scenarios. You may want to check for metamerism(sp) based on different lighting color temperatures with specific dyes and inks, but color grading of things other than diamond is a one lighting scenario.

Dave,

There are different standardized lightings for grading dying cloth or printing or paint. For example A, B, C, D55, D65, D75,..

It is reasonable and very helpful to use for grading paint the color temperature becoming environment using this paint( for example indoor or outdoor)

 

[oldminer]

Sorry for pulling away from the topic of this thread.

I think the right lighting for displaying diamonds to consumers are several both singly used and in combination. Pibnpont lighting brings out the sparkly effect and I have seen this used by some sellers with excellent success in comparing diamonds side by side while tilting them in a video presentation. Lighting scenarios which tend to exaggerate sparkle, fire and contrast are the way to show and display diamonds. Especially when diamonds are compared, side by side, in the same lighting, we can readily see which one may look better to us.

 

[Rockdiamond]

No one has described what is meant by "backlighting"- I'm curious about that one.

To some extent this is a moot conversation regarding the lights as they relate to photography.
For one thing, we all agree that grading a diamond based on photos does not work- and I believe we can all agree there's no way it could.
Therefore photos are relegated to representations based on the photographer's - or editor's viewpoint.

This is no different than anything else one might take a photo of- like a person. Francisco Scavullo doesn't get six figures for nothing- there's a lot of skill involved. I believe that diamonds also require specialized photo skills- and will be colored by the vision of the photographer.

Another aspect is the practical one.
Diamonds are small things- that means close up photography is indicated.
That , in turn, necessitates a lot of light to obtain focus. For these reasons, the photographer of a diamond has limitations which must be overcome.
The nature of the beast is that different photographers will come up with different solutions.
I don't believe we could ever agree on which is the "right' one.

 

[strmrdr]

Date: 5/19/2009 11:45:33 AM
Author: Judah Gutwein

This is not going to be a simple ''fix'' that''s for sure!


Judah

ditto that

Welcome to the world of PS round tables... the conversation runs all over the place.
When it gets to the point no one can figure out what is being discussed they dry up.

But really these are all different sides of the same dice.

 

[strmrdr]

Date: 5/19/2009 2:28:38 PM
Author: Rockdiamond

For one thing, we all agree that grading a diamond based on photos does not work- and I believe we can all agree there''s no way it could.

it is done and done well dozens of times a day on PS.
1000s of times a year. Many 10s of thousands of happy diamond owners.

There is a huge difference between online product photography and art photography.
In art there is no fraud.
When you post an image of an item for sale you are claiming it represents that item.
If it does not and its your art interpretation then it is deceptive.

 

[Rockdiamond]

Date: 5/19/2009 2:45:26 PM
Author: strmrdr

Date: 5/19/2009 2:28:38 PM
Author: Rockdiamond

For one thing, we all agree that grading a diamond based on photos does not work- and I believe we can all agree there''s no way it could.

it is done and done well dozens of times a day on PS.
1000s of times a year. Many 10s of thousands of happy diamond owners.

There is a huge difference between online product photography and art photography.
In art there is no fraud.
When you post an image of an item for sale you are claiming it represents that item.
If it does not and its your art interpretation then it is deceptive.

I was speaking of color and clarity grading.

I suppose a strong case for ASET/IS is the use to grade what is termed "performance"- sorry, i was not thinking of that aspect......
For that part to work, you have to be someone looking for a stone that performs well on ASET/IS


I don''t feel that using terms like "art photography" mean much in this conversation. We could make the case that diamond cutting itself is an art.

Storm- just so I know if I''m being insulted here- were you speaking to me directly when you used the word you?

 

[diagem]

Date: 5/19/2009 2:32:44 PM
Author: strmrdr

Date: 5/19/2009 11:45:33 AM
Author: Judah Gutwein

This is not going to be a simple ''fix'' that''s for sure!


Judah

ditto that

Welcome to the world of PS round tables... the conversation runs all over the place.
When it gets to the point no one can figure out what is being discussed they dry up.

But really these are all different sides of the same dice.

Any other performances Diamonds offer other than "light" that I should know about?

 

[strmrdr]

Date: 5/19/2009 2:56:09 PM
Author: Rockdiamond
Storm- just so I know if I''m being insulted here- were you speaking to me directly when you used the word you?

no I meant it as anyone in general who posts a product image...
"one" would have been a better word to use in hindsight.

 

[Rockdiamond]

Date: 5/19/2009 2:57:31 PM
Author: DiaGem

Date: 5/19/2009 2:32:44 PM
Author: strmrdr

Date: 5/19/2009 11:45:33 AM
Author: Judah Gutwein

This is not going to be a simple 'fix' that's for sure!


Judah

ditto that

Welcome to the world of PS round tables... the conversation runs all over the place.
When it gets to the point no one can figure out what is being discussed they dry up.

But really these are all different sides of the same dice.

Any other performances Diamonds offer other than 'light' that I should know about?

ABSO-FREAKIN- LUTELY


Thanks Storm

 

[strmrdr]

Date: 5/19/2009 2:57:31 PM
Author: DiaGem
Any other performances Diamonds offer other than ''light'' that I should know about?

Well I heard that they can do good things for your sex life lol

 

[Stephan]

Date: 5/19/2009 3:30:21 PM
Author: strmrdr
Well I heard that they can do good things for your sex life lol

I avoid to have a sex life while wearing my diamond.
There are some things that are hard to clean.

 

[diagem]

Date: 5/19/2009 3:30:21 PM
Author: strmrdr

Date: 5/19/2009 2:57:31 PM
Author: DiaGem
Any other performances Diamonds offer other than ''light'' that I should know about?

Well I heard that they can do good things for your sex life lol

"Storm- just so I know if I''m being insulted here- were you speaking to me directly when you used the word you?"

 

[Serg]

Date: 5/19/2009 2:56:09 PM
Author: Rockdiamond

Date: 5/19/2009 2:45:26 PM

Author: strmrdr

Date: 5/19/2009 2:28:38 PM

Author: Rockdiamond


For one thing, we all agree that grading a diamond based on photos does not work- and I believe we can all agree there's no way it could.

it is done and done well dozens of times a day on PS.

1000s of times a year. Many 10s of thousands of happy diamond owners.


There is a huge difference between online product photography and art photography.

In art there is no fraud.

When you post an image of an item for sale you are claiming it represents that item.

If it does not and its your art interpretation then it is deceptive.

I was speaking of color and clarity grading.


I suppose a strong case for ASET/IS is the use to grade what is termed 'performance'- sorry, i was not thinking of that aspect......

For that part to work, you have to be someone looking for a stone that performs well on ASET/IS



I don't feel that using terms like 'art photography' mean much in this conversation. We could make the case that diamond cutting itself is an art.


Storm- just so I know if I'm being insulted here- were you speaking to me directly when you used the word you?

re:For one thing, we all agree that grading a diamond based on photos does not work- and I believe we can all agree there's no way it could.

RD,
I am disagree. It could be done. I do not see fundamental limitation if you are speaking about set of photos for one diamond.
One photo is not enough.

 

[diagem]

Date: 5/19/2009 4:19:25 PM
Author: Serg

Date: 5/19/2009 2:56:09 PM
Author: Rockdiamond

Date: 5/19/2009 2:45:26 PM

Author: strmrdr

Date: 5/19/2009 2:28:38 PM

Author: Rockdiamond


For one thing, we all agree that grading a diamond based on photos does not work- and I believe we can all agree there''s no way it could.

it is done and done well dozens of times a day on PS.

1000s of times a year. Many 10s of thousands of happy diamond owners.


There is a huge difference between online product photography and art photography.

In art there is no fraud.

When you post an image of an item for sale you are claiming it represents that item.

If it does not and its your art interpretation then it is deceptive.

I was speaking of color and clarity grading.


I suppose a strong case for ASET/IS is the use to grade what is termed ''performance''- sorry, i was not thinking of that aspect......

For that part to work, you have to be someone looking for a stone that performs well on ASET/IS



I don''t feel that using terms like ''art photography'' mean much in this conversation. We could make the case that diamond cutting itself is an art.


Storm- just so I know if I''m being insulted here- were you speaking to me directly when you used the word you?

re:For one thing, we all agree that grading a diamond based on photos does not work- and I believe we can all agree there''s no way it could.

RD,
I am disagree. It could be done. I do not see fundamental limitation if you are speaking about set of photos for one diamond.
One photo is not enough.

It might be done..., but with a significant margin of error...

I believe as long as you cant grade the clarity on the realistic size of the Diamond at subject, the grade will always be approximate.
Especially since "eye-clean became a "grade".

Same with color..., pictures (on-screen or off) vary to much...

 

[strmrdr]

Date: 5/19/2009 3:54:44 PM
Author: DiaGem

Date: 5/19/2009 3:30:21 PM

Author: strmrdr

Date: 5/19/2009 2:57:31 PM

Author: DiaGem

Any other performances Diamonds offer other than ''light'' that I should know about?

Well I heard that they can do good things for your sex life lol

''Storm- just so I know if I''m being insulted here- were you speaking to me directly when you used the word you?''

Of course I was.
Just offering advise to a friend.
Start with taking home a 10ct and if that don''t work try a 20ct :}

 

[Rockdiamond]

Yes, 10 or 20 carat diamonds REALLY perform in that way.....but if you put them in the wrong positon it can be VERY painful....hehehe

 

[Serg]

RD,
re:It might be done..., but with a significant margin of error...

I believe as long as you cant grade the clarity on the realistic size of the Diamond at subject, the grade will always be approximate.
Especially since "eye-clean became a "grade".

Same with color..., pictures (on-screen or off) vary to much...

----
I did not say anything about grade by expert.

I am speaking about grade without any expert(Human ). Lets discuss it later. It is not important for current thread .

 

[Regular Guy]

Thanks to Storm for initiating this thread, Serg for continuing it, and Judah for trying to four square realign the question.

Reviewing here...

Date: 5/19/2009 11:45:33 AM
Author: Judah Gutwein

What has actually evolved here, seems to be a cross between a focus on light performance and a focus on a standardized environment for viewing/assessing loose diamonds and with no special emphasis on light performance. These are two VERY different tangents as Ira points out earlier.

Optimal viewing conditions can be independent from optimal light performance conditions. 'Viewing conditions' is actually referencing and alluding to the inherent physical qualities and characteristics from within the diamond. Whereas, optimal light performance is a reaction of the diamond to its environment (as pointed out earlier by Storm). A loose diamond on the finger is never static and is always reflecting and refracting available light back to the viewer in the form of 'light performance'.

Some have come on to post with emphasis relating to the (standardized) actual assessment and grading of loose diamonds (color/clarity). What does this have to do with light performance?? It seems as though not everyone is on board with the crux of this discussion. Given my understanding from reading these postings, things could get pretty convoluted here and very quickly.

Therefore, I think it would be beneficial to re-calibrate here and to distinguish between these 2 (very different) variables and to re-emphasise Storms intentions here.

The discussion here is on finding an optimal way to view and capture (using photography for online viewers) loose diamond light performance.

Within this context, Storm uses the word 'subjective' in his title. This is important, since once you essentially realize that this stuff is subjective, it is then extremely difficult, by definition, to develop an objective standard condition(s) for deriving a subjective result??

Judah, I didn't eliminate much of your text...I kept most of it, and maybe I should have kept it all. I did read it more than once. Really, I think what I highlighted above is the main point, and the concept of subjective can either come out, or put into the wash with the whole of the discussion.

A couple of things.

On the one hand, maybe there are two types of data. But...rather then two formally separated types...I think it is relative...with some data being more "fixed," and some being more flexible.

Think about becoming a teacher. Your good with data. You could choose to teach math or teach computers. Both are quantitative. The guy who chooses to teach math will create a set of notes, and not have to revise them for a long time, if ever. The computer guy...he'll have to revise them every 6 months, maybe. But both are teachers of quantitative oriented data.

Some environments for looking at diamonds are probably similar, and their variance may be incidental. Probably, somewhat regardless of the environment, inclusions can be seen with specificity and the results will be similar, regardless too much of the environment.

Color...the environment is probably a lot more important.

But light...yesterday, I tried to look up the definition of tautology to see how it describes the confound, and I didn't find it too satisfying. Fundamentally, though, if we are measuring "light performance," and wanting to defining the "light" environment in which we do this...the challenge increases. You have to be flexible in understanding the variables.

I don't know a lot about databases, but intuit that the difference between fixed and relational databases may apply a similar logic. Likewise, I think it in substance modeling that Serg does, and such a study I think demands certain specific methodology that we need to be sensitive to with this, that may be specific for something like light, versus something like inclusions...on the other side of the spectrum.

I think in statistics, the idea of resampling may help.

Generally, if when you change one variable (the fixed light environment) the other variable you study changes significantly, understanding this needs to be part of how you deal with the definition of the environment to begin with, I think.

Talking out loud, I wonder if modelling as a science does offer help. I think this is not Kansas anymore, as David A would have hoped.

 

[Judah Gutwein]

Ira,

All interesting points and much food for thought and analysis.

Your analogy re: tautologies (..hated that stuff in school..'if it is raining, you use an umbrella and such...

') to illustrate somewhat of a deficiency with respect to defining fundamentals of light performance, can also be extended to color/clarity definitions as well.

I didn't wish to bring this up in my earlier posting, since I did not want to throw a 'monkey wrench' into an already complicated task of ascribing divergent (finite) principles to properly assessing light performance versus the standard already in place for color/clarity grading.

However, consider the following:

It is true that there is a "fixed" standard for assessing color/clarity. Indeed, the standard bearers of the industry (GIA/AGS) have clearly deliniated conditions for proper quanitification of color/clarity. But the bottom line is, there are lesser labs of dubious distinction

, who will routinely misgrade color/clarity (better or worse than what the stone might actually deserve and usually be$$er than....

) for the diamonds they review.

If there is a standard, why does this happen?

Moreover, why are there many instances where GIA/AGS graders look at the same exact diamond twice and give it a different color or clarity grade each time?

The truth of the matter (and the answer to the inconsistencies above) is, that (unlike carat weight, where there is no way to fudge or misread the data....and where, essentially, 'it is what it is'..) when it comes to grading color and clarity, there may be a standard in place, but that standard is subject to a subjective interpretation of the individual results to the viewer/grader, since they are using their eyes to determine these grades.

In this respect, there will always be an inherent similarity between affixing a standard for subjective evaluation of diamond light performance and the subjective interpretation of standardized metrics already in place for color/clarity evaluation.

In this respect, your analogy re: the math vs. computer teacher is not entirely applicable, since the evolving science in the area of computers, is a result of new quantifiable, hard data which needs to be incorporated into the curriculum. Whereas, the potential for divergent interpretation of the color/clarity standard definitions, will not cease with the incorporation of new variables into that standard (imo), since the interpretation itself will always be reliant on the human eye.

 

[strmrdr]

Date: 5/20/2009 12:43:32 PM
Author: Judah Gutwein



In this respect, there will always be an inherent similarity between affixing a standard for subjective evaluation of diamond light performance and the subjective interpretation of standardized metrics already in place for color/clarity evaluation.

since the interpretation itself will always be reliant on the human eye.

Good post and this actually brings us full circle sort of...
The grading/classification of color by machine is a mature science in everything but the diamond trade.
The hard part is not the machines but 2 other things.
Identifying when the machine is having problems and raising an error message rather than letting a bad grade be assigned.

The biggest problem is integrating the results with the system in place that is arbitrary and frankly messed up beyond salvation.
A machine could measure the color of a diamond accurately and repeatably down to .01% error 99.99% of the time and not match the lab assigned grade once because diamond color grading as practiced by the labs today is a moving target.

 

[oldminer]

"The hard part is not the machines but 2 other things.
Identifying when the machine is having problems and raising an error message rather than letting a bad grade be assigned.

The biggest problem is integrating the results with the system in place that is arbitrary and frankly messed up beyond salvation.
A machine could measure the color of a diamond accurately and repeatably down to .01% error 99.99% of the time and not match the lab assigned grade once because diamond color grading as practiced by the labs today is a moving target."


I can answer these questions with responses years in the making and highly reviewed.


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.

BUT, if every diamond which goes through a machine for color assessment is reviewed by a skilled GG, the process is greatly speeded up, and the results are far tighter than the current 65% agreement level of graders in major labs. The number of diamonds which can be color graded from D to J color with consistent, accurate and repeatable grades is close to 98%. The remaining 2% will be the room for further improvment, but GG's may always have a role in weeding out the odd, exceptional stones which likely will never be 0%. Diamonds below J and into the fancy colors will take more time to get better sample data, but it is not going to be a big problem other than getting stones and GIA grading to compare them to.

I see the major stumbling block being the diamond trade which always wants the edge on a borderline call. The borderline calls will never go away, but the border will be far tighter and narrow. Guys who have millions tied up in diamonds don't like tight grading. No one likes tight slot machines at the casino, either. We all tend to want more money, faster..... who doesn't? If big diamond clients of the major labs don't want stricter, or what we might call more consistent grading, then which lab would have the courage to introduce it? Especially during these difficult financial times, the trade is limping along looking for hard to find profits. Investing in technology or choosing to give up the benefit of the doubt, does not look like it will make more money with any immediacy. In the long run, we'll have machine grading, but there is resistance at many levels right now.

Machine grading will bring efficiency and cost savings to the labs. They don't think so, just yet, but when they take the time to look at the cost per stone with all human grading and the cost per stone with machine assisted grading, they may see big bucks added to their bottom lines.

None of this sticks to the main topic of the thread, but I am responding to what just came before. For me, subjective grading is the judgment of beauty. Objective grading of the future is going to be for color, clarity, durability, finish, size/weight and light return characteristics. beauty in all types of lighting conditions is what consumers want to see. The more they get, the better they get to know the diamond even from a long distance.

 

[strmrdr]

Date: 5/20/2009 4:01:42 PM
Author: oldminer
For me, subjective grading is the judgment of beauty. Objective grading of the future is going to be for color, clarity, durability, finish, size/weight and light return characteristics. beauty in all types of lighting conditions is what consumers want to see. The more they get, the better they get to know the diamond even from a long distance.

Before you can have objective grading of light return characteristics you have to know what is important.
That changes when light condition changes so in order to do that you need to nail down a range of real world conditions for subjective light performance testing.
Once that is place you can begin to build a system around it.

That is where GIA failed badly and AGS is weak, the initial AGS 3D program was geared to much towards brilliance but to give them credit they are working hard on improving that.

edit: fixed wording

 

[Regular Guy]

David, I think these are good process points, and I realize that was I was guessing about the continuum of relevance and irrelevance of the lighting environment may not work so good after all, as you and Judah say.

Not wanting to take up too much space here, but...

...could it be that a way to look at a difference on the continuum is where the data returned in an investigation is meaningful in the feedback loop to understanding the correctness of the answer?

With math, it is an almost just theoretical system. You set rules, they are reliable, and that 4X4 = 16 is not that interesting, and not relevant to also understanding 3 X9 = 27.

But, maybe with light, it is different?

Take color, as has been discussed. We are not now talking about actual application, where labs focus on human graders, but on best practices. If we can appreciate the great reliability from SAS2000 or other tools, and see this, it becomes reasonable, since the data parsimoniously works with the machine, to use the machine...and consequently, the rules the machine understands, and that are embedded.

With respect to the above example of color, do we understand that lighting will not have been an important variable?

Take light performance. If data changes predictably with a change in environment, and the changing environment is substantive to the pattern of changing scores in light performance, doesn''t the changing light as environment need to become part of the evaluative set?

Sorry, I''ve really not worked this out at all.