A new debate over diamond color grading is shaping up, and it’s a battle of man vs. machine.
The machine is an electronic grader called the Gran Colorimeter. An Israeli manufacturer introduced the instrument 11/2 years ago, but it’s only now making an impact, thanks to a new full-fledged marketing program to retailers and the diamond trade.
The Colorimeter works like this: the user places a diamond on a small pedestal in the instrument. A light beam is directed through the stone and a series of filters. A receptor then analyzes the beam and measures the absorption of various colors. Results are shown in a Gemological Institute of America color scale reading. The $6,750 instrument is distributed through GIA Gem Instruments, Kassoy Inc. and Gesswein Co.
Proponents say the Colorimeter can help retailers who may not have diamond grading facilities or expertise of their own and those who want to use it as a countertop selling tool.
Detractors say the resulting grade is on the “generous side” because it measures the stone’s color face up (the opposite of what human graders do in gem labs) and because the instrument overlooks grays and includes fluorescence, resulting in higher color readings.
Advantages, limitations: Inventor Paul Gran of Gran Computer Industries Ltd. in Migdal Ha’Emek, Israel, says he’s well aware of the limitations and doesn’t try to promote the Colorimeter as a substitute for grading reports. But he does say that measuring diamonds face up is closer to reality than lab procedure. “Who looks at a diamond from the pavilion?” he asks. “Labs don’t do justice to diamonds because everyone else looks at them through the crown.”
He also says the Colorimeter is at least as accurate as a gem lab grade, possibly more so because it grades on a continuous scale. “If the diamond is a low G color, it will appear as such on the reading,” he says. “With the major gem labs, the diamond will be a G or an H, and that’s it.”
In addition, he says, the Colorimeter is consistent regardless of the shape, cut or weight of the diamond. “[Such factors] are why the same diamond may be graded differently by different laboratories,” he says, “or even graded differently by the same lab at different times.”
Gran says the instrument is accurate to plus or minus one-third of a grade (labs usually claim accuracy to within a half grade). And he’s working to calibrate filters so the accuracy is even better in future models. The repeatability (the same instrument giving the same grade to the same diamond in subsequent tests) is “about two-tenths of a grade with current models,” he says.
The Colorimeter also can be connected to any personal computer, with results shown on the computer’s monitor.
Gran says the Colorimeter can be set to what he calls different standards used by the world’s four major gemological laboratories (GIA, the Diamond High Council of Antwerp, the International Gemological Institute and the European Gemological Laboratory). The labs insist they all grade to the same standard (see JCK, July 1995, pp. 60-66). But Gran says the range between the highest and lowest is about a half grade; he declines to say which labs are highest and lowest.
Labs’ reaction: Several gem labs say they are testing the Colorimeter or using it to check their staff graders. They say it’s the best grading instrument ever produced. But they also stress it’s not accurate enough for everyday use in grading reports.
Tom Tashey, director of European Gem Lab in Los Angeles, Cal., uses the Colorimeter as an “extra opinion” in borderline cases where staff graders disagree. In the meantime, Tashey is creating a database to measure the instrument’s accuracy. “We’ve found it’s reasonably accurate to within a half grade 65%-70% of the time,” says Tashey. “The other times it can be off a grade or two. Still, it’s much better than previous color grading machines.”
Jerry Ehrenwald, president of the International Gemological Institute in New York, N.Y., believes the instrument is still too inaccurate to be considered for general use. “It’s dangerous because there’s still too much margin for error,” he says. Ehrenwald acknowledges that human grading leaves room for subjectivity, but notes that it’s based on a set group of master stones.
Gemologist Martin Haske of Adamas Gemological Lab in Woburn, Mass., worries that too many people may use the Colorimeter as a substitute for learning how to grade diamonds. “And for the cost of one of these machines, one can buy a very good set of master stones. Right now, I think they’re a better investment.”
GIA sells the Colorimeter through its Gem Instruments division. A report published in The Loupe, GIA’s newsletter, says the Colorimeter “has potential to help give approximate visual color readings of near-colorless round diamonds.” GIA research found the instrument most accurate with Cape series (where the color yellows as it deepens) and less accurate with brown and gray stones, fluorescent stones and fancy shapes.
“There are fundamental differences between the way a human eye sees color and the way instruments measure it,” says GIA Vice President Robert Kammerling. It would be difficult &endash; maybe impossible &endash; to eliminate or compensate for these differences, he says.
Concerns: Gran is aware of the instrument’s problems and is working to overcome them. Here are the major concerns and Gran’s response:
· Fluorescence. “When a laboratory grades a fluorescent diamond, it uses a quality light source with no ultraviolet…. The result is the diamond appears less white [lower grade] than it would if it were exposed to ultraviolet,” says Gran. “The white light used in the Colorimeter contains a small amount of ultraviolet. Therefore, the Colorimeter’s grading of a fluorescent diamond may differ from the laboratory’s grade for the same stone.” Gran says slight and moderate fluorescence presents no problems for the Colorimeter and that new filters he plans to introduce should reduce or eliminate false readings caused by strong fluorescence.
· Broken culets or strong color concentrations near the culet give false (lower) readings. Gran says he has no way to overcome this problem yet.
· Gray diamonds cause false readings because the color typically results from very fine black impurities, which the instrument can’t recognize. “The Colorimeter can’t sense the existence of gray and grades the stone as if it were clear,” he says. Labs invariably give such diamonds lower grades, he says.
· Some in the trade worry that different Colorimeters may be calibrated differently so one could not repeat the findings of another, and that anyone with fraud on his or her mind could recalibrate them to give bogus high readings. Gran says he will introduce a series of calibration discs so two Colorimeters can be calibrated to within one-tenth of a grade instead of the current margin of about one-third. He also says it’s impossible to reprogram the Colorimeter to cheat. “All of the calibration codes are encrypted so they’re virtually impossible to access &endash; let alone change,” he says.
Man vs. machine: Even though gem lab executives say the Colorimeter is the best machine grader they’ve seen, they don’t believe it will replace human graders anytime soon.
The main reason: differences between human vision and instrument measurement. Human subjectivity can lead to mistakes, they say. But it also can see obvious things in a diamond (such as gray coloring and broken culets) that the Colorimeter cannot interpret.
Add to this the expense of buying and maintaining grading instruments, they say, and it could be a long time before there’s a “magic box” to accurately color grade all kinds of diamonds.