AGS new cut grade system early 2005

[adamasgem]

----------------
On 10/27/2004 3:08:48 PM Serg wrote:

Marty,
Please explain :why does your plot have 58 different points if you use 29 stones only?

----------------

The first 30 stones were those defined in the original GIA brillance article (Fall 1999) where they defined stones and gave WLR values, the next 28 (RD01-RD29) were from their G&G Fire article (Fall 2001) where they gave both WLR and DCLR values..

 

[adamasgem]

----------------
On 10/27/2004 3:08:48 PM Serg wrote:

Marty,
Please explain :why does your plot have 58 different points if you use 29 stones only?

----------------

Since I only had DCLR data on RD01-RD29, my plot for the Fire comparison did not include the original 30 stones.. The WLR comparison plot did..

 

[Serg]

----------------
On 10/27/2004 4:13:54 PM adamasgem wrote:

----------------
On 10/27/2004 3:08:48 PM Serg wrote:

Marty,
Please explain :why does your plot have 58 different points if you use 29 stones only?

----------------

Since I only had DCLR data on RD01-RD29, my plot for the Fire comparison did not include the original 30 stones.. The WLR comparison plot did..
----------------

Please see the abscissa axis on your plot in your post "Posted: 10/26/2004 10:03:08"
and reread my question.

 

[Diermint]

Hi,

Back to the one of the original postings in this thread - I was thinking that you could describe the new system (the diagonal black oval that Garry drew on the HCA) using reasonably simple maths rather than just making smaller boxes. This is a reasonable linear approximation that could be refined to fit curves later:

# = C x ((6 x P/1.3)-154.53

where, C = Crown angle
P = Pavillion angle
# = Arbitrary quality number

This creates “sloped lines” on the graph rather than relying on stepped lines around multiple boxes.

An example of defining limits could be:

Good Cut: Crown angle 31-37
Pavillion angle 40.2-41.5
# of ~ 1100-1200 is very good.

Excellent Cut: Crown angle 32-35.5
Pavillion angle 40.5-41.0
# of ~ 1140-1160 is excellent

The equation only requires the two variables - crown and pavillion angle. Also, given either a crown or pavillion angle, you can calculate the range of the other to give best cut, i.e. a higher pavillion angle wants a lower crown angle. As a crude first approximation, what do you think? It makes a nice graph too (I like graphs).

 

[adamasgem]

----------------
On 10/27/2004 11:42:00 PM Serg wrote:

----------------
On 10/27/2004 4:13:54 PM adamasgem wrote:

----------------
On 10/27/2004 3:08:48 PM Serg wrote:

Marty,
Please explain :why does your plot have 58 different points if you use 29 stones only?

----------------

Since I only had DCLR data on RD01-RD29, my plot for the Fire comparison did not include the original 30 stones.. The WLR comparison plot did..
----------------

Please see the abscissa axis on your plot in your post 'Posted: 10/26/2004 10:03:08'
and reread my question.

----------------

Sergey, the TABLE I posted yesterday gives the parameters of the 58 stones on the WLR plot you question. I don't understand your question or concern.

My DCLR Chromatic flare comparison plot contains only stones RD01-RD29 (no RD2 since I didn't have GIA DCLR values for the GIA stones in their original Fall 1999 Brillance article

I did manage to make a retro reflector model. Square stone, with as close to 45 degree pavilion angles as I can get given the numerical precision and the way I define depths, currently the pavilion mains are at 45.0000000058758 Deg.
I have to have a non zero culet (0.01%) in my models because of the way I have set up the generic problem. I think it is close enough for government work.

Doing a ray trace on parallel beam inputs, I can account for over 99.5+% of the total light energy.

What scenarios would you like to see with it??

 

[Diermint]

As a crude first approximation, what do you think? It makes a nice graph too (I like graphs).
---------------------------

I haven''t attached a file before so I hope it works - if not, please bear with me...

 

[Serg]

re:What scenarios would you like to see with it??

Could you publish your plot with 59 stones? Stone 59 is retro reflector model.

re:Sergey, the TABLE I posted yesterday gives the parameters of the 58 stones on the WLR plot you question. I don''t understand your question or concern.

I ask because I took date WLR RD45 from your plot and wanted to check is I received correct result:

GIA WLR for RD45( nailhead diamond) is more than 1.

 

[adamasgem]

Date: 10/28/2004 10:22

8 AM
Author: Serg
re:What scenarios would you like to see with it??

Could you publish your plot with 59 stones? Stone 59 is retro reflector model.

re:Sergey, the TABLE I posted yesterday gives the parameters of the 58 stones on the WLR plot you question. I don''t understand your question or concern.

I ask because I took date WLR RD45 from your plot and wanted to check is I received correct result:

GIA WLR for RD45( nailhead diamond) is more than 1.

Sergey, I don''t have the newly defined RD45 on my plot, perhaps that is the confusion. See the Table I published earlier in the thread for the model defs I used..

 

[adamasgem]

Date: 10/28/2004 5:53:35 AM
Author: Diermint
As a crude first approximation, what do you think? It makes a nice graph too (I like graphs).
---------------------------

I haven''t attached a file before so I hope it works - if not, please bear with me...

Diermint..

One can do all the curve fitting one wants to, but the important question is whether the data one is fitting is "correct", in that does it represent performance in a defined and realistic and "common" viewing environment.
Realize, though, that one would need such a "fit" for every "set" of conditions, it is a multi variate problem (many dimensions: table size, crown angle, girdle thickness, pavilion depth, culet size, star length, pavillion break length, color grade are the primary ones)

I personally don''t think that the vast majority of "retail" lighting environments are appropriate because I feel that they blur the true distinction between diamonds. These environments are generally too harsh, with way too much intense lighting, especially in the typical retail "mall" environments of the mass merchandisers.

 

[Serg]

Date: 10/28/2004 12:29:57 PM
Author: adamasgem

Date: 10/28/2004 10:22

8 AM
Author: Serg
re:What scenarios would you like to see with it??

GIA WLR for RD45( nailhead diamond) is more than 1.

Sergey, I don''t have the newly defined RD45 on my plot, perhaps that is the confusion. See the Table I published earlier in the thread for the model defs I used..

Thanks. I understood now. Sorry.
We have not information about GIA( or Marty) WLR RD45 yet

 

[adamasgem]

Date: 10/28/2004 3

9

1 AM
Author: Diermint
Hi,

Back to the one of the original postings in this thread - I was thinking that you could describe the new system (the diagonal black oval that Garry drew on the HCA) using reasonably simple maths rather than just making smaller boxes. This is a reasonable linear approximation that could be refined to fit curves later:

# = C x ((6 x P/1.3)-154.53

where, C = Crown angle
P = Pavillion angle
# = Arbitrary quality number

Technically, your equation is non linear. It seems to be part of the general equation for an ellipse.
0=K1*C*C + K2*C*P +K2*P*P +K3*P +K4*C + K5

 

[adamasgem]

Date: 10/28/2004 12:51:51 PM
Author: Serg

Date: 10/28/2004 12:29:57 PM
Author: adamasgem

Date: 10/28/2004 10:22

8 AM
Author: Serg
re:What scenarios would you like to see with it??

GIA WLR for RD45( nailhead diamond) is more than 1.

Sergey, I don''t have the newly defined RD45 on my plot, perhaps that is the confusion. See the Table I published earlier in the thread for the model defs I used..

Thanks. I understood now. Sorry.
We have not information about GIA( or Marty) WLR RD45 yet

Sergey, I''ll try to put together a list of the NEW stone definitions from GIA and run them in the next week

In the meantime, what do you want run with the retro-reflector model..

 

[adamasgem]

Date: 10/26/2004 2:21

5 PM
Author: Serg


re:''corner cube'' issue within the framework of my software, Table and four pavilion mains is what needs to be done I believe.

Yes. It is right.

What is GIA ''brightness'' model ?

I see this term firstly.

Well, I guess we both screwed up

, the correct model for a retro reflector is a triangular stone with three pavilion mains. I''ll rebuild it over the weekend..

. The pyramidal stone I built worked well for overhehead parallel rays though

 

[Serg]

Marty,
Please publish result( for ABH and other light models) for pyramidal stone too.

 

[adamasgem]

Sergey My pure retroreflector design

Triangualar stone length of each side = 1.00
Length of each side of culet =0.001
Depth=-0.40784004

Had to itterate to get the correct depth, what a pain

Those numbers will give perpendicularity of pavilion main planes to better than one million''th of a degree, good enough for government work I guess

How do you want to define the pyramidal stone.? It can''t be a retroreflector with four mains, I believe

 

[Diermint]

Technically, your equation is non linear. It seems to be part of the general equation for an ellipse.
0=K1*C*C + K2*C*P +K2*P*P +K3*P +K4*C + K5
_____________________________________________


Yes… you''re quite right – I started off trying to think of ways to define it using a simple linear relation but it didn''t turn out to be quite as neat as it seemed it should. I think it ends up cleaner this way.

Anyway, I still think that having only one relation for a given depth etc. is simpler to work with than the number of ranges (boxes) that has been proposed and this can hopefully be made to fit better at the same time killing two birds with one diamond.

Either way you’re going to have to define ranges based on all those variables you mentioned so why not simplify it a bit where possible?

I worry

that unless you fit the desired area a bit better, then there will be a decently high percentage that are "in" that should be "out" and/or vice versa - especially when dealing with such large boxes. Or does it not make much difference? I thought that some of you might like a single defined blob rather than lots of mini-ranges. Of course the boundaries can be modified as desired and we could use much better equations utilising some sexier maths.

Anyway, just a thought...

 

[Garry H (Cut Nut)]

I am afraid the elliptoid idea will not work with AGS

 

[Diermint]

I am afraid the elliptoid idea will not work with AGS
----------------------------------------------------

Correct me if I''m wrong... but I''m assuming that those rectangular boxes are only that shape because they are defined by the two angles - pavillion and crown angles- and this does not actually reflect reality? Is it not true that the boxes are average values that are of high enough value to include, and that they are not actually homogenous, ie. the corners are likely to be considerably worse than the middle in some cases. That blocky-thing (

I think that''s what it is called) on the left is a good example of this - the upper edge is likely to be a smooth curve in reality rather than stepped and the upper right of the steps is likely to be of poorer quality than the rest. The inner corners of the steps that are not included are probably better than the outer corner that is included.

Hence wouldn''t it be more accurate if you could describe them using an oval shape or slopes rather than a rectangles? You might think it does not matter much, but when the blocks are as large as they are the error in approximating the shape also becomes very large.

By approximating in this way I think it should statistically be possible to include a wider range of values than are currently included without skimping on quality - at a guess maybe 10-15% more.

Admittedly the left hand blocky-thing would be a little more difficult to describe, but not impossible. For example, you could use three curves as shown to describe it rather than quite a few boxes.

Anyway, I''d be interested to hear other views on this.

 

[Garry H (Cut Nut)]

Dont shoot me
Tell AGS.

This is blown up a lot - the boxes are 0.2 degree by 0.2 degree, and this is the only way the existing scanners and the data could be presented.

But if AGS would like us to help them design a better solution, we are ready and able

 

[Diermint]

Dont shoot me
Tell AGS.
----------------------------------------


0.2 degree is pretty coarse feature resolution considering the range we''re talking about and it does seem a bit odd that AGS hasn''t come up with something better, especially when you consider they have the advantage of seeing a ready-made first-class system, namely the HCA. The HCA is awesome, it''s orders of magnitude better than what the AGS uses.

Anyway... I guess we''ll see how long they take to catch up in technologically.

 

[Serg]

Date: 10/28/2004 6:41:26 PM
Author: adamasgem
Sergey My pure retroreflector design

Triangualar stone length of each side = 1.00
Length of each side of culet =0.001
Depth=-0.40784004

Had to itterate to get the correct depth, what a pain

Those numbers will give perpendicularity of pavilion main planes to better than one million''th of a degree, good enough for government work I guess

How do you want to define the pyramidal stone.? It can''t be a retroreflector with four mains, I believe

Marty I think to find result for the retroreflector with four mains is very interesting too.

 

[strmrdr]

Date: 10/29/2004 3

3:45 AM
Author: Diermint
Dont shoot me

Tell AGS.

----------------------------------------



0.2 degree is pretty coarse feature resolution considering the range we''re talking about and it does seem a bit odd that AGS hasn''t come up with something better, especially when you consider they have the advantage of seeing a ready-made first-class system, namely the HCA. The HCA is awesome, it''s orders of magnitude better than what the AGS uses.

Anyway... I guess we''ll see how long they take to catch up in technologically.

.2 degrees is the accuracy of the more common diamond scanners in use.
HCA, AGA, AGS or anything else based on the measurements is going to have that margin of error.
There are some better tools available or soon to be available but there will always be a margin of error it is just huge right now.

 

[Serg]

re:There are some better tools available or soon to be available but there will always be a margin of error it is just huge right now.

Strmrdr,

Could you give more information about "some better tools available or soon to be available "

 

[adamasgem]

Date: 10/29/2004 1:56:23 AM
Author: Diermint







I am afraid the elliptoid idea will not work with AGS
----------------------------------------------------

Correct me if I''m wrong... but I''m assuming that those rectangular boxes are only that shape because they are defined by the two angles - pavillion and crown angles- and this does not actually reflect reality? Is it not true that the boxes are average values that are of high enough value to include, and that they are not actually homogenous, ie. the corners are likely to be considerably worse than the middle in some cases. That blocky-thing (

I think that''s what it is called) on the left is a good example of this - the upper edge is likely to be a smooth curve in reality rather than stepped and the upper right of the steps is likely to be of poorer quality than the rest. The inner corners of the steps that are not included are probably better than the outer corner that is included.

Hence wouldn''t it be more accurate if you could describe them using an oval shape or slopes rather than a rectangles? You might think it does not matter much, but when the blocks are as large as they are the error in approximating the shape also becomes very large.

They can generate the "boxes" with any quantization they deem appropriate, commensurate with the accuracy of the measurement on the particular axis.

 

[adamasgem]

Date: 10/29/2004 3:34:56 AM
Author: Serg

Date: 10/28/2004 6:41:26 PM
Author: adamasgem
Sergey My pure retroreflector design

Triangualar stone length of each side = 1.00
Length of each side of culet =0.001
Depth=-0.40784004

Had to itterate to get the correct depth, what a pain

Those numbers will give perpendicularity of pavilion main planes to better than one million''th of a degree, good enough for government work I guess

How do you want to define the pyramidal stone.? It can''t be a retroreflector with four mains, I believe

Marty I think to find result for the retroreflector with four mains is very interesting too.

Sergey: the four mains design (with 45 degree pavilion angles) is NOT a retroreflector ( EXCEPT for a parallel wavefront perpendicular to the table, e.g. GIA''s DCLR illumination). [ I believe that one can make a retroreflector with six pavilion mains also, but I''m not sure]

 

[Serg]

Marty,

re:Sergey: the four mains design (with 45 degree pavilion angles) is NOT a retroreflector ( EXCEPT for a parallel wavefront perpendicular to the table, e.g. GIA''s DCLR illumination).

I know it. But I do not see any reason : Why you are not publishing result for stone with "the four mains design (with 45 degree pavilion angles) is NOT a retroreflector "

We lost a lot of time for discussing what is retroreflector. Much more easy to publish result .

 

[Diermint]

Date: 10/29/2004 9:39:39 AM
Author: strmrdr


.2 degrees is the accuracy of the more common diamond scanners in use.
HCA, AGA, AGS or anything else based on the measurements is going to have that margin of error.
There are some better tools available or soon to be available but there will always be a margin of error it is just huge right now.

That''s interesting Strmrdr, thanks, I had just assumed they would be measured more accurately than that but I hadn''t ever looked into it before. Do you know offhand what the error in measuring the dimensions is - ie. they seem to be often quoted to the nearest 0.1mm - is that a limitation in the measurement or has it to do with the variability in the stone?
Sometimes the measurements vary for the same stone on different certificates.

I''m guessing that carat is easier to understand - although a good analytical balance can provide more significant figures than are usually quoted, a difference of 1.001 and 1.002 carat is unlikely to make any major difference unlike the difference between say 41.2 or 41.4 degrees in pavillion angle.

I reckon that 0.2 seems like a lot if you''re aiming for a box only ~ 1.1 degrees high (~20% error). Conceivably the top 20% and bottom 20% might or might not actually be AGS0, making 40% of the range possible measurement errors that include stones from outside the range (greater if you include the sides too)?

 

[adamasgem]

Date: 10/29/2004 10:29

2 AM
Author: Serg
Marty,

We lost a lot of time for discussing what is retroreflector. Much more easy to publish result .

Sergey.. Just got to dig up the correct Tolkowsky data to normalize results to, but here is an interesting result from the retroreflector..

Plane parallel rays normal to table covering ENTIRE STONE, excluding GLARE, the refracted EXITING rays are confined to the area of an inscribed hexagon, little known neat factoid, huh

I'm going to bed, it is 5am

PS I have to rebuild the pryamid, as I trashed it when I realized it was not a retroreflector.

View attachment Rr_pint.jpg

 

[Garry H (Cut Nut)]

Cool refractoid Marty

 

[strmrdr]

Date: 10/29/2004 9:52:52 AM
Author: Serg
re:There are some better tools available or soon to be available but there will always be a margin of error it is just huge right now.


Strmrdr,


Could you give more information about 'some better tools available or soon to be available '

helium that your involved in.
Dave at aga is working with someone on a new one.
Sarin has or will soon release a better high end scanner. (memery recall is fuzzy on the details of this one.)
OGI has released a better scanner.

Those are just the ones Iv heard mentioned.
Time will tell if they actualy are better.

ps. did I pass your test? :}