Consumer advisory: GIA Cut Grade Rounding Problems

[oldminer]

Thanks guys! It truly helps all of us to step back once in a while and make it clear what its all about in terms that one can digest.

It is certainly right that if a diamond, or any object, that we'd hope is also to be sparkly, has too much brilliancy, that the appearance of sparkle will be diminished or non-existent.

Diamonds in jewelry are observed at many differing angles, but when buying or selecting a diamond, one looks at rather more closely and with more intense observational conditions than at any other time. Admirers who later take a close, possibly envious, look at a set diamond also may use nearly the same degree of observational intensity. At these times, the viewer does not look from a wide angle but from a basically face up to a near face-up view. If a diamond is only good face-up, it will be an underperformer otherwise anyway. That would be the sort of diamond that one would hesistate to boldly show off or brag about. There's a lot of those out there and their owner's love them, but they are not the best stones.

The potential for scintillation is what I'd call sparkle. This is a static measure of the standard deviation of the greyscale between all the pixels inside the face-up view. The broader, the larger, the standard deviation, the more sparkle is there and the more that will be there when the stone is moved and scintillation happens. Although the ASET scope shows pretty colors, I believe the measure of the stone's ability to sparkle can just as well be derived from greyscale.

I suppose it is nice to know what the angles of light are coming into the stone. For the life of me, I can't figure out who is in a position to decide what angles are "right" ones. AGS is, I suspect. If common sense prevails, I think we have pretty good answers already. Now, what do we do to go further. How will ALL the shapes be handled? I don't see this tool being of much use with some other shapes. I must admit, the Asscher above, looks super with the ASET images. It took a lot of work to construct them, didn't it? I see a direct digital solution as being easier to achieve, more elegant, more scientific, and more flexible for variation in faceting, outline and shape......

 

[strmrdr]

Date: 2/4/2006 5:28:12 PM
Author: adamasgem

Date: 2/4/2006 5:13:33 PM

Author: strmrdr

here is the gem file for my comparison model.

open both it and the one above side by side and set both to office lighting and hit play on both of them.

This one will dance more because it has more scintillation.

Now you are learning... But you also have to consider if there is no low angle lighting, remember, there are usually light absorbers in a room at low angles... In cluding your body..

yea its not perfect but it gives the idea.
This stuff is old hat and its been posted a dozen times before.

 

[strmrdr]

interesting Dave,
yes gray-scale tilted images could be used for the same thing.
As its tilted the various shades of gray would move and show the same thing.
The colors of the aset make it easier to see and add a little more information.
More than likely AGS has software that they input the data and it spits out all 45 images but Iv got to do it the hard way.
When I get a larger hard drive installed and get vb.net installed I will write a small program to automate DC to produce them and post a bunch of them.

With the diamond in hand and the hand held aset its easy just tilt the diamond slowly and observe the shift.

I think the aset is the best thing to happen to asschers in a long time.
Its near perfect for judging patterns in them.
When using virtual aset to grade them you have to use helium data, sarin has problems scanning them properly.
I cant wait to see the ags cut grade for them and really really hope they don''t mess it up.

 

[adamasgem]

Date: 2/4/2006 5:36:18 PM
Author: oldminer
Thanks guys! It truly helps all of us to step back once in a while and make it clear what its all about in terms that one can digest.

It is certainly right that if a diamond, or any object, that we''d hope is also to be sparkly, has too much brilliancy, that the appearance of sparkle will be diminished or non-existent.

The potential for scintillation is what I''d call sparkle. This is a static measure of the standard deviation of the greyscale between all the pixels inside the face-up view. The broader, the larger, the standard deviation, the more sparkle is there and the more that will be there when the stone is moved and scintillation happens. Although the ASET scope shows pretty colors, I believe the measure of the stone''s ability to sparkle can just as well be derived from greyscale.

Well dave, you also have to consider what you are measuring. The less optical symmetry, the mor or what you term sparkle, BUT the small the areas of each "sparkle". What this does if effect the contrast AS WELL AS effect the FIRE.

The simple analogy is broddflash versus pinfire opal, up close you can see the fire in a pinfire, take a step back and you can''t. Same goes with broken up patterns in diamonds. Lack of optical symmetry also effects the internal absorption of light. It appears, from what I''ve seen empirically, optically symmetric stones face up a better "color" than what the color grade leads you to expect.

Richard Von Sternberg wears a K color Eightstar, and most woud say it is a G or H face up.. I believe it is because of the lower pathlengths the light travels inside an optically symmetric diamond before it is returned to the viewer vis a vie one without optical symmetry.

I have shown that FIRE is effected by optical symmetry, and surmise from the data (photos) that in a non optically symmetric diamond there is so much mixing internally such that the FIRE photos become muddy, earth tones versus purer spectral hues..

 

[He Scores]

Marty...just out of curiosity...does Richard''s K color flouresce (sp)?

I''ll have to take a G with me to Vegas and check this out in an independant observation study.



Bill

 

[adamasgem]

Date: 2/4/2006 7:17:35 PM
Author: He Scores



Marty...just out of curiosity...does Richard''s K color flouresce (sp)?

I''ll have to take a G with me to Vegas and check this out in an independant observation study.



Bill

Do''n honestly know the answer to that question... See you in Vegas

 

[valeria101]

Date: 2/4/2006 5:36:18 PM
Author: oldminer

For the life of me, I can''t figure out who is in a position to decide what angles are ''right'' ones. AGS is, I suspect.

I would pic up this one question too...

Is this even a matter of choice? I remember that at some point was said that in the older Gilbertonscope the boundaries between color zones were assigned by chance. Was there some logic behind the positions chosen for the ASET?

Is there anything else to decide about the construction of the ''scope? (considering the coloring of the resulting image carries a value judgement)

 

[strmrdr]

34.2, 41, 56t,60%star,80% lgf
aset vs jewelery shop DC lighting with black background.
0 to 7 degrees.

 

[strmrdr]

same 7 to 15 degrees

 

[strmrdr]

Dave,

As the above images show, which i didnt change to gray-scale but the are close enough too it I think, that gray-scale images could be used but that the aset is much easier to read and tells you which angle the light is coming from and not just that one area is darker or lighter than another.
An expert could interpret the non-aset images but I think the ASET images are much easier to understand by the average person.

 

[Serg]

Dave,

re:I see a direct digital solution as being easier to achieve, more elegant, more scientific, and more flexible for variation in faceting, outline and shape......

We do not need new mantra. If you want help to us, please give real input( information). Mantra is not information for us.

 

[adamasgem]

Date: 2/5/2006 12:28:16 AM
Author: strmrdr
34.2, 41, 56t,60%star,80% lgf
aset vs jewelery shop DC lighting with black background.
0 to 7 degrees.

Storm.. I think the way to view the tilt concept is to use a symmetrical lighting environment, which I believe the "jewelry" is not. The point source effects are highly azimuth dependent.

Sergey: Can you suggest a symmetrical envirionment in diamondcalc for photoreal modeling representation.
I''d like to see the cloud covered sky and a solar disk (head obscured with facial reflected light) blue sky modeled in DiamondCalc.
My thinking is, is that an asymmetrical model obscures the issue, and is random anyway, relative to the assymmetry in a stone. With a limited number of point sources, you run the risk of hiding or overly influencing the result.

 

[valeria101]

Date: 2/5/2006 2:45:04 AM
Author: adamasgem

Sergey: Can you suggest a symmetrical environment in diamondcalc for photoreal modeling representation.
I''d like to see the cloud covered sky and a solar disk (head obscured with facial reflected light) blue sky modeled in DiamondCalc.

A ''white'' ASEET model?

 

[adamasgem]

Date: 2/5/2006 2:49:17 AM
Author: valeria101

Date: 2/5/2006 2:45:04 AM
Author: adamasgem

Sergey: Can you suggest a symmetrical environment in diamondcalc for photoreal modeling representation.
I'd like to see the cloud covered sky and a solar disk (head obscured with facial reflected light) blue sky modeled in DiamondCalc.

A 'white' ASEET model?

Yes and No Ana.. Think of it as grayscaled..
White light D6500K in the case of the the standard cloud covered sky model in color science and the appropriate (I'll have to check) color temperature distribution for the blue sky model, with a nominal non zero intensity of 20 to 60% spectrally dependent for the head obsuration. Pure science, theoretically physically realizable. GIA's hemisphere is uniformly illuminated from the zenith to the horizon. It is not real world.
Since body or envirionment obscurations are random relative to the stone, you wind up having to consider and average over more azimuths and tilts to avoid introducing a bias into the problem.

 

[valeria101]

Date: 2/5/2006 3:09:36 AM
Author: adamasgem

Date: 2/5/2006 2:49:17 AM
Author: valeria101

Date: 2/5/2006 2:45:04 AM
Author: adamasgem

Sergey: Can you suggest a symmetrical environment in diamondcalc for photoreal modeling representation.
I''d like to see the cloud covered sky and a solar disk (head obscured with facial reflected light) blue sky modeled in DiamondCalc.

A ''white'' ASEET model?

Yes and No Ana.. Think of it as grayscaled..

Hm... ok... no discrete ''steps'' then (as in ASET or any other layered lighting model). I am not going to crowd this thread more with trial an error.


Thanks for the hint

 

[Garry H (Cut Nut)]

Date: 2/4/2006 6:58:03 AM
Author: Serg

Please, see very interesting examples for all cut grade system.

What is your idea about GIA cut grade for this examples?
What is your idea about adequate cut grade for this example?

This diamond has crown and pavilion painted. I will publish 3d model in next post
I will publish full report and my comments in Monday - Tuesday

Can we have your simple answers folks to these questions of Sergey''s (and my new question)

1. What is your idea about GIA cut grade for this examples?

2. What is your idea about adequate cut grade for this example?

3. What do you think AGS would give the stone for LIGHT PERFORMANCE

AGS will down grade the stone +4 grades for having a 6% girdle at the thickest part (but the stone has a very good spread 1% better than Tolkowsky) - but it is the appearnace of the stone that is most interesting.

Here is some of the stats again:

 

[valeria101]

Oups, for got the third question ...

Date: 2/5/2006 5:27:22 AM
Author: Garry H (Cut Nut)


Can we have your simple answers folks to these questions of Sergey''s (and my new question)

1. What is your idea about GIA cut grade for this examples?

''Good'' (FacetWare says ''very good'' for the proportions, but they punish ''non-traditional brilliantring'' so...I went down a step from the FacetWare prediction)



2. What is your idea about adequate cut grade for this example?

One step below ''top'' because of reduced contrast. It is a nice fancy round better suited for smaller stones, IMO. If the diamond was larger, I would have been allot less inclined to like the cut.



3. What do you think AGS would give the stone for LIGHT PERFORMANCE

Ideal

 

[Serg]

Marty,

re:Sergey: Can you suggest a symmetrical envirionment in diamondcalc for photoreal modeling representation.


You can try use "GemCad COS" + observer head blockage( end in light scheme by light button) May be name is wrong, I have not DC at home. It was done 3-4 years ago.

But I do not like all light scheme in DC. its good for study ( understanding cut work) only.
We are adding HDR texture to DC now. If this work will successful anybody can work with any REAL( not synthetic) light scheme.

 

[strmrdr]

Date: 2/5/2006 5:27:22 AM
Author: Garry H (Cut Nut)

Date: 2/4/2006 6:58:03 AM

Can we have your simple answers folks to these questions of Sergey's (and my new question)


1. What is your idea about GIA cut grade for this examples?

gia good

2. What is your idea about adequate cut grade for this example?


vg for pendant, honestly I think for the overall market good is a grade I would agree with on this one


3. What do you think AGS would give the stone for LIGHT PERFORMANCE

lack of contrast would kick LIGHT PERFORMANCE down to a 4 I think, it isnt well balanced

AGS will down grade the stone +4 grades for having a 6% girdle at the thickest part (but the stone has a very good spread 1% better than Tolkowsky) - but it is the appearnace of the stone that is most interesting.



Here is some of the stats again:

 

[adamasgem]

Date: 2/5/2006 6:45:36 AM
Author: Serg

Marty,

re:Sergey: Can you suggest a symmetrical envirionment in diamondcalc for photoreal modeling representation.


You can try use ''GemCad COS'' + observer head blockage( end in light scheme by light button) May be name is wrong, I have not DC at home. It was done 3-4 years ago.



But I do not like all light scheme in DC. its good for study ( understanding cut work) only.
We are adding HDR texture to DC now. If this work will successful anybody can work with any REAL( not synthetic) light scheme.

Thanks Sergey for the hint.. Attached is Sergeys example in a dmc file with head obscuration, GEMCAD COS lighting, multiple reflections and colors added in raytrace options..

Special thanks to Leonid for quickly changing code to allow DMC file extensions for DiamondCalc users.

 

[strmrdr]

bug someplace dmc file not downloading.
says file not found.

 

[pricescope]

Date: 2/5/2006 4:04:54 PM
Author: strmrdr
bug someplace dmc file not downloading.
says file not found.

should be OK now...

 

[strmrdr]

34.2, 41, 56t,60%star,80% lgf
Marty''s lighting 0 to 15 degrees tilt.

 

[adamasgem]

Date: 2/5/2006 9:44:13 PM
Author: strmrdr
34.2, 41, 56t,60%star,80% lgf
Marty''s lighting 0 to 15 degrees tilt.

Please understand that I do not know the distribution characteristics of the DC COS lighting scheme.... Anyone know Sergey? I just pulled it up on Sergeys suggestion as a non uniform distribution, so I don''t necessarily "bless" it, so to speak..

I reallly would like to see an altered Moon and Spencer Cloud covered sky model and/or a Pokrowski blue sky model used..

 

[Garry H (Cut Nut)]

Date: 2/5/2006 10:49:57 PM
Author: adamasgem
Please understand that I do not know the distribution characteristics of the DC COS lighting scheme.... Anyone know Sergey? I just pulled it up on Sergeys suggestion as a non uniform distribution, so I don''t necessarily ''bless'' it, so to speak..

I reallly would like to see an altered Moon and Spencer Cloud covered sky model and/or a Pokrowski blue sky model used..

Does this help Marty?
It shows you where all the pages and #''s are - although it is only 14 degree obstruction.

BTW you have not voted yet?
We only have votes in from Ana and Storm.

It will be more fun if you all play the game.

Remeber you probably have more chance to be right than one of the labs - so there will be no loosers on PS

 

[Garry H (Cut Nut)]

This is the actual lighting from your file Marty - it differs from the one at the bottom of the I just posted because i used the side bar drop down - I never realised they were different (35 degree obstruction)

 

[strmrdr]

Garry,
how do you get too the lighting map screen your showing?
Im not finding it.

never mind found it.......

 

[Garry H (Cut Nut)]

Date: 2/6/2006 12:27:53 AM
Author: strmrdr
Garry,
how do you get too the lighting map screen your showing?
Im not finding it.

never mind found it.......

Options > advanced > Lighting map
Then I usually use spherical and color radio buttons

 

[adamasgem]

Date: 2/6/2006 12:22:11 AM
Author: Garry H (Cut Nut)
This is the actual lighting from your file Marty - it differs from the one at the bottom of the I just posted because i used the side bar drop down - I never realised they were different (35 degree obstruction)

I played with the head obscuration HOWEVER, obscuration is not perfect, as I posted before, the head reflect 20 to 60 % of the incident light, so any obsuration model should relfect that point..

I''m thinking about two realistic lighting models based on the distributions I published before. Lower angle lighting is attenuated in amplitude and color by the local envirionment, trees, buildings, etc

 

[Garry H (Cut Nut)]

Date: 2/6/2006 12:53:13 AM
Author: adamasgem

Date: 2/6/2006 12:22:11 AM
Author: Garry H (Cut Nut)
This is the actual lighting from your file Marty - it differs from the one at the bottom of the I just posted because i used the side bar drop down - I never realised they were different (35 degree obstruction)

I played with the head obscuration HOWEVER, obscuration is not perfect, as I posted before, the head reflect 20 to 60 % of the incident light, so any obsuration model should relfect that point..

I''m thinking about two realistic lighting models based on the distributions I published before. Lower angle lighting is attenuated in amplitude and color by the local envirionment, trees, buildings, etc

File attached should be saved and opened with the lighting Load control.

Marty I made the COS dome 306mm away from the stone (600 diameter) and the dark zone 290 from the white part, which still has a 35 degree relative size.

The Dark Zone is now set at 20 brightness and the white part at 100. The dark part is infront of the white part and is shading it.

Happy?