GIA has started providing the cut grade dimensions

[adamasgem]

Date: 11/27/2005 4:15:24 PM
Author: valeria101

Date: 11/27/2005 3:52:31 PM
Author: adamasgem






First they average and then they round.. ,

[...]

But what do you expect when the object was to make most stones VG to EX cut on paper.., hence they sell more paper and get more $$$$$..

This is a bit of rhetorical Q... mosty because I am somewhat lost in details.

Is the respective averaging & rounding mostly to blame for the overly lax grading? Or is it more in the scores associated with those measurements?

In other words, can those numbers be interpreted to obtain a more realistic ''top quality'' cut off or that can only be gotten with other measurements than GIA''s and they have extracted all information there is from the numbers reported.

First, I don''t own either Helium, Sarin or OGI, but from the data I have seen, and recomendations, I would probably buy a Helium

It is interesting that the quantization level of the Helium data is one tenth (0.1) that of Sari or OGI.. Quantization level can make a machine appear much better than it actually is..the finer the quantization the better the device, as in scales..

I have seen repeatbility and machine to machine data, but since I did not take it, can''t attest to its'' accuracy.

I believe that there is a lot of fudging going on out there in various releases of software, and that GIA, kept it too too simple by averaging and then rounding again.. keeps their software data base much smaller. I Guess that it defines a "best" grade, but not the worst grade a particular stone could get, but then I guess that where you have to send it to the lab to get their objective opinion (be sure to enclose a check made out to cash, and you might get a better grade

)..

I had been looking at Saring data on EightStars, and the Sarin MESH points, which define the facets, would suggest that there isn''t meet point faceting on probably the best cutting available.. WHICH says that the SARIN machine is seriously screwed up (but better than a leveridge guage), and it didn''t just happen on EightStar data.. so average, to make a stone appear symmetric when it isn''t, round up or down to the nearest 0.5 degree, and put out FARCEWARE(TM), and forget all about min max about the averages, etc, etc, etc..

 

[Regular Guy]

Date: 11/27/2005 2:28:06 PM
Author: JohnQuixote

This is one thread that might make Ira happy, as I think there is agreement on a few things here.

1. Numbers on a lab report are a start and can indicate zones of 'traditional success.'
2. Using HCA to get a prediction - not treated as ultimate, just as a 'nod' - can help with sight-unseen decisions.
(at this time, for several reasons, I find HCA more useful than Facetware)
3. A reflector image, IS or ASET, provides reliable direct assessment of broad overall light return potential.

You'd be surprised by what may or may not make me happy.

Actually, what may make anyone reading this happy is the knowledge that a longer post I'd constructed just went up in smoke, but I will say that I'm very glad that not only still Ana, but others, too are: "... very fond of the classics and the Renaissance hunt for metrics of beauty...." perhaps allowing us to benefit from some of the standards allowing for the beauty of nice diamonds we have today.

Beyond that, although # 2 & 3 you mention, John, are true enough, I'm afraid that Dave's statement about GIA's implementation of rounding: "The numbers need to be rounded as the equipment used to generate the numbers has fairly wide machine error which has not yet been cured," points to what may be the best of their paternalistic intentions, and where I'm afraid Marty's imputing of what they really have in mind may more so be the case, regardless, and now, having furrowed my brow about what this rounding GIA is doing, seeing Storm's good example, above, concerning consequences of it, and considering implications for a shopper, regarding your #1 above, John:

"1. Numbers on a lab report are a start and can indicate zones of 'traditional success.'"

I'm now more unhappy about the prospects of what GIA will be providing than I had been before. Despite their information capture, and although a broad assessment and screening can perhaps now be done that could not before, for measures of cut quality, it's not clear that GIA will be any better than EGL, since both employ rounding, apparently, and so, although a measure of options can be excluded, I'm afraid not as many can be excluded as I would have hoped...and only for the factor of rounding...and as a result, I'm afraid that where the quick search had held out great hope for me as being a more precise tool for pre-final selection, I now think that the effective "search by cut" tool, where vendor's own separate assessments of measurements will still be needed to peform the level of precision delineation that an average but discriminating shopper will want.

Finally, for me, it's all about implementation. What can a shopper do, based on resources and data available. So, although Dave says it here, it's really implicit in the intended boundaries of help that tools like the HCA can provide, too:

" Now that they are going to provide the numbers, the business has advanced to a point where the numbers are not going to mean much to anyone who has a choice. My own point of view remains, that numbers help to screen stones, but should not serve as substitutes for actually seeing and appreciating how beautiful they are and how they compare to one another. Enough variance is left unmeasured that one cannot use numbers to make purely correct decisions. "

Although this may be so, how is the shopper who uses Pricescope to do their shopping expected to use this information to do their job. Despite the preferred approach of "actually seeing" and comparing, generally, we don't get to do too much of that. We do see numbers, and pictures, ask for not more than 3 to see, and make our choice. That, at least, is typically the case.

With GIA's new presentation of data, I wonder how many diamonds we can really screen out, now, more than we could before?

 

[strmrdr]

Date: 11/27/2005 6:17:01 PM
Author: Regular Guy

Date: 11/27/2005 2:28:06 PM

With GIA''s new presentation of data, I wonder how many diamonds we can really screen out, now, more than we could before?

You hit the nail on the head with that question.

my opinion:
With the rounding there are going to be some exellent diamonds sorted out as my example shows.
Will there be some so-so diamonds that using the rounded data will show up as exellent (under 2). I havent tried it but logic says there will be.
Both ends of the best ring performance zone have some pretty steep drop offs for the leakers and the pendant stones that are going to get hit by the rounding.

 

[Garry H (Cut Nut)]

Date: 11/27/2005 6:17:01 PM
Author: Regular Guy

Date: 11/27/2005 2:28:06 PM
Author: JohnQuixote

This is one thread that might make Ira happy,

You''d be surprised by what may or may not make me happy.

No surprises as to what makes me happy
http://images.google.com/images?sourceid=navclient&ie=UTF-8&rls=GGLG,GGLG:2005-20,GGLG:en&q=women+wine+and+song

 

[Garry H (Cut Nut)]

GIA''s www.Facetware.gia.edu may be a con.

It might really be designed to have steep deep diamonds submitted for grading to get the numbers up - but the stones may actually be graded with stricter numbers.

e.g. - say a stone measures 41.2P and 36C - it gets GIA excellent - but GIA have really allowed this could be a 41P 35.5C (which is still steep deep, but not as bad) - and they use the tolerance factor in their favor.

But say a stone with the same proportions on the clients scanner - say it is measured by GIA as 41.5P and 36.5C - then gIA bumps it down - but in the meantime they got an extrat $100 on an extra 1,000 stones a week that would otherwise go to EGL or IGI

 

[mdx]

Hi Guys
Here is the Scan as discussed on the weekend with some ref to GIA rounding

Hope this helps

Johan

 

[adamasgem]

Date: 11/27/2005 8:36:48 PM
Author: mdx
Hi Guys
Here is the Scan as discussed on the weekend with some ref to GIA rounding

Hope this helps

Johan

Do you have the srn file for that stone?

 

[Garry H (Cut Nut)]

or a Gem adviser - it is easier on my computer than .srn

 

[mdx]

How do you upload the file to the Forum?

Johan

 

[adamasgem]

Date: 11/27/2005 8:36:48 PM
Author: mdx
Hi Guys
Here is the Scan as discussed on the weekend with some ref to GIA rounding

Hope this helps

Johan

That is picture of a report from one scan of a stone.. what is more interesting, and what is needed, are the results from a large number of scans on the same stone... How much do the numbers vary from scan to scan (report to report)..

Look at the picture of stone in the Sarin web viewer versus what the "report" shows for the stone.. I have the SRN web viewer but can't gernerate the type of report you show..

Every step of the process which includes rounding adds more errors to the process.. Sarin used to say that the individual angles were good to +/- 0.4..

We have issues of absolute error, repeatability, and rounding errors, plus the FARCEWARE rounding, andwho knows what we have statistically.. we are back to what we were doing in the 90's, working with a bunch of charts on min maxes for the "average" angles to place the stone in some category..

Of course, GIA's own words "We discovered that personal and regional preferences were an inherent part of a truly functional cut grading system. Although we confirmed that most observers could discern five overall grade categories for diamond cut quality, personal and/or regional tastes often dictated which diamond was finally preferred within each of those categories. This underscores the fact that there can be many different, yet equally pleasing, appearances within a grade category." talks to the system being tailored not to science, but subjective judgements...


Additionally thier statement "Measurement precision and rounding was determined by considering:
1) The precision of measuring devices and other measuring methodologies
2) The practical abilities of most diamond cutters and manufacturers
3) The differences in cut quality that were found to be consistently distinguishable during our extensive observation testing "

should have nothing to do with "practical abilities of most diamond cutters"... most diamond cutters produce junk, that is the biggest part of the market. FARCEWARE doesn't reward the very best, it groups the best with the marginally good., and seems to give a deceptive viewppoint.

One of the things I'll do when I have the time, is look how "cut grades" have expanded and widened at GIA..

Interestingly, GIA could now give a diamond with an extremely thin (i.e. knife edge) girdle a VERY GOOD cut grade, when ALL previous grading systems from GIA and GIA put it into a class 4, the old bottom grade and the NAJA system seemed to say it could be a 3B at best.. GIA CONFIDENTIAL internal lab documentation (circa 199 that I have seen states to the graders:

"Girdle thicknes is called extremely thin BECAUSE there is a risk that the diamond may be chipped in that location"..

"Again, look carefully at the point of pears, marquies and hearts. There may be naturals or extra facets that make the girdle thickness in these areas extremely thin. Again, a knife edge at the point creates a danger of chipping the diamond"

Now, should that ever get a VERY GOOD cut grade.. seems to me there has been a little gtade creep going on in FARCEWARE(TM), in an effort to sell paper, and is a little misleading to consumers..

Did you notice the word "extra facets" above.. they aren't always plotted if they are not visible face up,, they are called "non callable", you see those to improve clarity grades, but the "paper" looks clean..

More later

 

[mdx]

Hi Marte

I think we are missing each other here. I need to know how do I post the file on the forum so you guys can play with it.

The report layout is a custom one we designed

Johan

 

[belle]

johan,
the report layout is great. i like the side by side values for the actual measurements and the gia rounded version.

to post a file:

click reply>
attach file>
browse>
(make sure the file has a unique name. no spaces or unusual charaters in the name)
upload>
submit

 

[mdx]

Thanks Belle, Thanks Gary

Hopefully here is the Gem Adviser file


Johan

 

[Garry H (Cut Nut)]

Dug out upper girdles and slightly dug out lower girdles, which overall helps a slightly shallower stone .

7% less over all light return than a top stone, with slightly better contrast - overall a quite nice stone - one we should wait and see what GIA does with it after 1st January.

It would look much brighter than a 41.4P 36C stone, and as it is - it has a quite good spread, where as the steep deep would not.

 

[JohnQuixote]

Date: 11/27/2005 11:45:05 PM
Author: mdx


The report layout is a custom one we designed

An attractive one too, Johan.

 

[strmrdr]

Date: 11/28/2005 12:27:42 AM
Author: mdx
Thanks Belle, Thanks Gary


Hopefully here is the Gem Adviser file



Johan

something wrong with the gem file when opened in diamcalc it gives a diameter of 1.99mm

can we get the .srn file?

 

[JohnQuixote]

I get 1.99 mm too, though the angles/proportions appear as reported on first glance. Garry, did it work properly for you?

 

[adamasgem]

Date: 11/27/2005 11:45:05 PM
Author: mdx

Hi Marte

I think we are missing each other here. I need to know how do I post the file on the forum so you guys can play with it.

The report layout is a custom one we designed

Johan

Johan Nice report..

You might run the same srone, take it on an off the Sarin and see what resultsyou get.. Try to upload the srn files to the forup or I''l send you my email address and you can send it as and attachment..

 

[adamasgem]

Hi Garry..

I think some of the repeatability tests on machines should be posted here. Ask you know who to maybe publish the work that has been done, but then again, some may argue it is self promotion, but I belieb=ve, data is data, and names can be omitted to protect the guilty.

Someone whe isn''t getting any press here is Bill Bray, a cutter, who developed BrayScore, in US Patent appication 20040051861, which evaluated diamonds from a cutters perspective, using the facet to facet variations found in the data. People should look at how he discounts for the variation among a particular set of facets, which of course is ignored in FARCEWARE(TM) which uses averages.

There is, of course your application 20020052170, but I haven''t seen anything published for FARCEWARE(TM) ..

People can go to http://www.uspto.gov and search for issued patents or published patent applications. Seems FARCEWARE(TM) might step on YOUR toes just a little bit

.

 

[adamasgem]

Date: 11/27/2005 8:36:48 PM
Author: mdx
Hi Guys
Here is the Scan as discussed on the weekend with some ref to GIA rounding

Hope this helps

Johan

Since the crown angle, table size and crown height are interrelated, the rounded figures GIA uses are going to play havoc with any attempt to ray trace the stone, since most software demands mathematical consistency, unless you let some parameter float and not use what GIA publishes. Same goes for the pavilion. We alway knew that most data was like that..

 

[JohnQuixote]

Date: 11/28/2005 1:47:37 AM
Author: adamasgem

Since the crown angle, table size and crown height are interrelated, the rounded figures GIA uses are going to play havoc with any attempt to ray trace the stone, since most software demands mathematical consistency, unless you let some parameter float and not use what GIA publishes. Same goes for the pavilion. We alway knew that most data was like that..

You know, relative to this, there is a compensation for overall depth that must be done when you round CA or PA in Facetware or else it may return a lower grade.

Example: 60.3, 55, 34.3, 40.6, 55, 80, Thin-Med, Ex Ex.

In this example the given CA is 34.3, but Facetware rounds up to 34.5. Without any compensation those numbers result in VG. Consider though, that CA was raised .2 degrees to 34.5 but there was not compensation done for overall depth. If you adjust the 60.3 depth% up to a 60.4 (or 60.5) it results in a grade of EX.

 

[mdx]

Date: 11/28/2005 1:47:37 AM
Author: adamasgem

Date: 11/27/2005 8:36:48 PM
Author: mdx
Hi Guys
Here is the Scan as discussed on the weekend with some ref to GIA rounding

Hope this helps

Johan

Since the crown angle, table size and crown height are interrelated, the rounded figures GIA uses are going to play havoc with any attempt to ray trace the stone, since most software demands mathematical consistency, unless you let some parameter float and not use what GIA publishes. Same goes for the pavilion. We alway knew that most data was like that..

Marty
On this stone, on the single scan, the results look pretty close to the GIA scan.

There is however another stone in the same batch where the tolerance between my Sarin and the GIA machine is stretched to the limit of both machines.
This now adds a dramatic third dimension to the rounding.

These stones have been sent for lazer inscription today and will be back in a few days.

I would happy to scan the second one about 10times and e-mail you the results with the GIA vendors report.
If anyone else is interested I could also post results on the forum.

Just a small problem, since installing the latest software version from Sarin I am getting an error saving to the .srn extension, so save in Gemcad or .dxt
I am busy discussing this with Sarin and should have it rectified by the time the stones are back.
Let me know

johan

 

[Serg]

re:Just a small problem, since installing the latest software version from Sarin I am getting an error saving to the .srn extension, so save in Gemcad or .dxt
I am busy discussing this with Sarin and should have it rectified by the time the stones are back.


It is reason of 1.99 mm in GA. "Our" Gemcad( ASCII) format has not information about real scale.

 

[mdx]

Thanks Serg

I noticed when importing the file that the initial image was very small.

All the ratios and facet data are however in exact proportion and are an accurate rendition of the diamond

Johan

 

[Garry H (Cut Nut)]

Marty I would rather not publish the comparison you have seen because I do not think it can be properly attributed.

But you or someone mentioned about Helium reporting angles to the second decimal place. In a way this makes for bigger variations - a less accurate scanner might get 41.3 degrees for each of 9 scans. So the repeatability is great based on the "least count" being ten times larger. Helium might have a 0.01 repeatability error over 9 tests. In this case the data could look like this.
The result could be that a lessor quality scanner gets a better sigma 3 result (a statistics standard used in these cases).

But in this instance I am referring to the repeatability of a single device. Johan is saying that his repeatability compared to 2 GIA stones is not great. But he is using the same scanner. It is likely GIA used 2 different scanners - so it might be that we are seeing a discrepency between 2 different scanners at GIA.

They could just replace them all with Helium, but perhaps GIA has strong links and bonds with people and companies from Israel?

 

[mdx]

Gary

I would say the chances of the GIA scanners being 2 different ones are very high as one stone was cut in USA and submitted in NY and the other cut in India submitted to Rappaport Mumbai

Its possible they both ended up on the same machine but I doubt it very much.

Johan

 

[strmrdr]

Date: 11/28/2005 2:46:00 AM
Author: mdx

Date: 11/28/2005 1:47:37 AM

Author: adamasgem

Date: 11/27/2005 8:36:48 PM

Author: mdx

Hi Guys

Here is the Scan as discussed on the weekend with some ref to GIA rounding


Hope this helps


Johan

Since the crown angle, table size and crown height are interrelated, the rounded figures GIA uses are going to play havoc with any attempt to ray trace the stone, since most software demands mathematical consistency, unless you let some parameter float and not use what GIA publishes. Same goes for the pavilion. We alway knew that most data was like that..

Marty

On this stone, on the single scan, the results look pretty close to the GIA scan.


There is however another stone in the same batch where the tolerance between my Sarin and the GIA machine is stretched to the limit of both machines.

This now adds a dramatic third dimension to the rounding.


These stones have been sent for lazer inscription today and will be back in a few days.


I would happy to scan the second one about 10times and e-mail you the results with the GIA vendors report.

If anyone else is interested I could also post results on the forum.


Just a small problem, since installing the latest software version from Sarin I am getting an error saving to the .srn extension, so save in Gemcad or .dxt

I am busy discussing this with Sarin and should have it rectified by the time the stones are back.

Let me know


johan

I would love to see the reports posted here.
Thank you for taking the time to do this :}

 

[Garry H (Cut Nut)]

Date: 11/28/2005 5:53:31 AM
Author: mdx
Gary

I would say the chances of the GIA scanners being 2 different ones are very high as one stone was cut in USA and submitted in NY and the other cut in India submitted to Rappaport Mumbai

Its possible they both ended up on the same machine but I doubt it very much.

Johan

I believe the Indian stones sent under bond to GIA by Rapaport may be graded in Carlsbad California Johan. Is that your understanding also?

 

[adamasgem]

Date: 11/28/2005 5:42:44 AM
Author: Garry H (Cut Nut)
Marty I would rather not publish the comparison you have seen because I do not think it can be properly attributed.

But you or someone mentioned about Helium reporting angles to the second decimal place. In a way this makes for bigger variations - a less accurate scanner might get 41.3 degrees for each of 9 scans. So the repeatability is great based on the ''least count'' being ten times larger. Helium might have a 0.01 repeatability error over 9 tests. In this case the data could look like this.
The result could be that a lessor quality scanner gets a better sigma 3 result (a statistics standard used in these cases).

But in this instance I am referring to the repeatability of a single device. Johan is saying that his repeatability compared to 2 GIA stones is not great. But he is using the same scanner. It is likely GIA used 2 different scanners - so it might be that we are seeing a discrepency between 2 different scanners at GIA.

They could just replace them all with Helium, but perhaps GIA has strong links and bonds with people and companies from Israel?

Yup, that data that both you and I have seen is very interesting regarding repeatability on ONE machine is very interesting. One Sarin, for example, to anothe, can be expected to be much worse. Yes, I saw the truncation or rounding, which makes the variance zero at the reporting LSB(least Significant Bit), and I think i commented that it could make one appear more consistent regarding repeatability.

One has to do multiple machines, like 31 Sarins on the same stone to get "absolute" accuracy levels, then again, if they, or anyone, have/has screwed up the calibration software, you can get any number..

Thanks for publishing the data examples pulled from the file we saw..

Marty

 

[adamasgem]

Here is a Sarin Mesh file that a friend of mine who has a $20$ Sarin did on an Eightstar. I would say that the representation of meet pont facetining is a little screwed up, and based on that, I wouldn''t trust the facet angles at all. I decoded the srn file and pulled out the Mesh data on the left and the right hand side was the Sarin webviewer. Makes you think that something is wrong, doesn''t it, and IT ISN''T THE DIAMOND..