An interesting (to me) ASET comparison -- JA vs. DiamCalc

[drk14]

As part of the appraisal of a marquise diamond I recently purchased from JA, Dave Atlas provided me with an ASET generated from Sarin scan data in DiamCalc. Given the notorious difficulty of photographing ASETs in general, and JA's past tribulations with their AGS ASET setup in particular, as well as periodically recurring discussions on PS about the usefulness of ASETs in evaluating fancies, I thought it would be interesting to compare the JA-supplied ASET photograph against the DiamCalc version and my own amateur observations of the diamond IRL.

JA ASET Photograph:


Sarin/DiamCalc ASET:


JA High-Definition Photograph:


Overall, I think the agreement between the JA and DiamCalc ASETs is remarkable, which is a testament both to JA's efforts to fix their past ASET issues, and to the accuracy of the ray tracing software. Here are the differences that I see:

• Some regions that are blue in the DiamCalc ASET are black in the JA ASET (most notably, the virtual facets reflected in the vertically oriented pavilion mains, at the inside edge of the table -- i.e., just above and below the red 'mini-bowtie' shape in the center).
• There is more red in the DiamCalc ASET, whereas the some of these facets are green in the JA version.
• The JA ASET suggests a fair amount of leakage under the table, whereas the DiamCalc version only shows a handful of small black (leakage) spots under the table.
• The overall light return (red+green) is fairly uniform in the DiamCalc ASET, but in the JA ASET, the belly region seems much brighter than the points of the MQ.
• The calculated ASET image has crisp virtual facets throughout, whereas the JA ASET has more blur in the facet edges, especially away from the MQ belly.

So, which gives a more accurate representation of my IRL impressions? I think the DiamCalc version is in slightly better agreement with my observations.

In particular, to check for leakage, I placed the diamond on a colored background and examined under a loupe in diffuse lighting. Under these conditions, I could see only a few small spots of the background color under the table, so this observation matched the DiamCalc prediction better than JA's ASET. In separate tests, I also placed a small dot of color against the girdle (i.e., near 0°), and noticed that a faint hue of that color would appear throughout large swaths of the table. To me, this suggests that maybe the JA ASET setup has the girdle (and maybe even a sliver of the crown adjacent to the girdle) below the horizon (thus black), and/or that DiamCalc places all or part of the girdle above the horizon (thus green).

Also, in my (heavily biased) opinion, the uniformity of brightness throughout the MQ body and the sharpness of the virtual facets in the DiamCalc ASET is more representative of the IRL behavior of this particular diamond than is the JA ASET.

I offer this post primarily as a data point to assist those of us who use JA ASETs make purchasing decisions (or help others make purchasing decisions) about fancy cut diamonds. Nonetheless, I would welcome corrections, comments, expert explanations, or discussion!



ETA: I know that the accuracy of ASET photography can depend on diamond size. In case it helps anybody interpret the images above, this particular marquise is 1.05ct, has a total depth 3.34mm, crown height 0.90mm, girdle 0.22mm, and pavilion depth 2.27mm, length 11.08mm, and width 5.00mm (Sarin scan data).

 

[Texas Leaguer]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Very interesting post DRK. Looks like you got yourself a nice marquise there!

I think your observations are very insightful. It is virtually impossible to set up an ASET photo system that faithfully replicates the exacting parameters of the ASET hemisphere. Even the best setups will introduce one or more small variables (as noted in your comments). The beauty of the computer generated approach is that it is repeatable and consistent and true to the confines of ASET theory. (provided you have an accurate scan of course).

It is a little ironic that many people seem to prefer the actual ASET photos over computer generated ones.

 

[Karl_K]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Very interesting post DRK. Looks like you got yourself a nice marquise there!

I think your observations are very insightful. It is virtually impossible to set up an ASET photo system that faithfully replicates the exacting parameters of the ASET hemisphere. Even the best setups will introduce one or more small variables (as noted in your comments). The beauty of the computer generated approach is that it is repeatable and consistent and true to the confines of ASET theory. (provided you have an accurate scan of course).

It is a little ironic that many people seem to prefer the actual ASET photos over computer generated ones.

scanner error is often times bigger than the problems in the aset setup from the better suppliers of real ASET photos particularly with fancies.
The computer generated images on AGS reports are actually some of the worst ones at times.
The bottom line is both computer generated and real ASET photos rely on the care and attention to detail of the person doing them as well as the quality of the setup and that is sometimes a problem with both.
Neither one is perfect but I prefer real images over all.

 

[drk14]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Very interesting post DRK. Looks like you got yourself a nice marquise there!

Thanks, Bryan!

 

[Karl_K]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Very interesting post DRK. Looks like you got yourself a nice marquise there!

Thanks, Bryan!

I agree it is very nice, sorry for not saying so earlier.
Congrates!!!

 

[John P]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

First thought: Very strong correlations between the actual ASET image and computer-generated ASET. Awesome.

Neither one is perfect but I prefer real images over all.

Second thought: I typically agree with Karl here.

That said, we're coming ever-closer to parity in sims (with top equipment, software and regular calibration) and actual images (with attention to the girdle plane, precise angular spectrum and leveling).

 

[John P]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Some regions that are blue in the DiamCalc ASET are black in the JA ASET (most notably, the virtual facets reflected in the vertically oriented pavilion mains, at the inside edge of the table -- i.e., just above and below the red 'mini-bowtie' shape in the center).

This has to do with strength of lighting and some girdle-plane discrepancy, which is difficult to avoid IRL.

There is more red in the DiamCalc ASET, whereas the some of these facets are green in the JA version.

Again, this has to do with the girdle-plane. Perhaps a 1 or 2 degree difference. Not an issue in any case.

The JA ASET suggests a fair amount of leakage under the table, whereas the DiamCalc version only shows a handful of small black (leakage) spots under the table.... The overall light return (red+green) is fairly uniform in the DiamCalc ASET, but in the JA ASET, the belly region seems much brighter than the points of the MQ.

JA lighting situation may be a bit over-harsh - not unexpected and will actually vary from setup to setup with great aplomb.

The calculated ASET image has crisp virtual facets throughout, whereas the JA ASET has more blur in the facet edges, especially away from the MQ belly.

This is focal-depth limitation. The camera is unable to keep the table and culet both in crisp focus. The software adapts.

So, which gives a more accurate representation of my IRL impressions? I think the DiamCalc version is in slightly better agreement with my observations.

To-may-to. To-mah-to.

This is like asking "Which is a more accurate version of Peyton Manning? The video of him, or the simulation I see in Madden 2014?" The compulsory answer is Actual Imagery. Full stop. But, with that said, the simulation might be a strong generic week to week rendering of consistent performance, in a particular standardized environment.

At the end of the day, the most salient version of performance will be the one you see and spend time with in person, through multiple environments, day to day and year to year.

 

[Rockdiamond]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Awesome thread drk- thanks for posting it!!

I would say the CG aset is more like a "glamour shot" compared to a polaroid.
Of course the CG ASET skips the imperfection, easily visible on the photograph.


It all highlights the difficulties of trying to use aset to make buying decisions.
Just like photos, comparing one vendor to another is practically useless.
But the photo provides me with more into than either ASET.
I'd also have a far easier time figuring how just how "real" a photo is compared to an aset.

Having said that- I like the way the CG aset looks much better, from a marketing standpoint- it just looks cleaner.

 

[drk14]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

I agree it is very nice, sorry for not saying so earlier.
Congrates!!!

Thanks, Karl. I have to give credit to Niel for spotting it in JA's catalog, and to Gypsy and others for convincing me to look past the "birthmark" (which turned out to be eye-clean in the end).

I'm curious about your comment that scanner errors may be a bigger problem the ASET photography setups. It would seem to me that the latter has more variables and moving parts (and therefore harder to get right). What are some typical sources of error in scanning?

 

[Texas Leaguer]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Awesome thread drk- thanks for posting it!!

I would say the CG aset is more like a "glamour shot" compared to a polaroid.
Of course the CG ASET skips the imperfection, easily visible on the photograph.


It all highlights the difficulties of trying to use aset to make buying decisions.
Just like photos, comparing one vendor to another is practically useless.
But the photo provides me with more into than either ASET.
I'd also have a far easier time figuring how just how "real" a photo is compared to an aset.

Having said that- I like the way the CG aset looks much better, from a marketing standpoint- it just looks cleaner.

Im not sure how, but you seem to have missed the point of the original post. DRK indicated that both ASET images were very close, and that the computer generated ASET was most accurate to real life observation.

How is it that you use this as a pretext to disparage ASET or even photos? The incongruity of your argument calls into question your intent.

 

[Rockdiamond]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

hi Bryan,
What is your intent?
I am not looking to argue with anyone- but it is an open discussion which, if held respectfully, can be informative.

Scanner error, placing the stones on the tray in the exact correct orientation- these issues plague all of us who would like to use ASET in representing diamonds. Are you disputing the fact that there's no "benchmark" on how to take an ASET picture, or what it should look like?


DRK- the stone does look amazing.

 

[drk14]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Some regions that are blue in the DiamCalc ASET are black in the JA ASET (most notably, the virtual facets reflected in the vertically oriented pavilion mains, at the inside edge of the table -- i.e., just above and below the red 'mini-bowtie' shape in the center).

This has to do with strength of lighting and some girdle-plane discrepancy, which is difficult to avoid IRL.

There is more red in the DiamCalc ASET, whereas the some of these facets are green in the JA version.

Again, this has to do with the girdle-plane. Perhaps a 1 or 2 degree difference. Not an issue in any case.

The JA ASET suggests a fair amount of leakage under the table, whereas the DiamCalc version only shows a handful of small black (leakage) spots under the table.... The overall light return (red+green) is fairly uniform in the DiamCalc ASET, but in the JA ASET, the belly region seems much brighter than the points of the MQ.

JA lighting situation may be a bit over-harsh - not unexpected and will actually vary from setup to setup with great aplomb.

The calculated ASET image has crisp virtual facets throughout, whereas the JA ASET has more blur in the facet edges, especially away from the MQ belly.

This is focal-depth limitation. The camera is unable to keep the table and culet both in crisp focus. The software adapts.

John,
Thanks for the insights, I was hoping to get some information of this nature.

So, which gives a more accurate representation of my IRL impressions? I think the DiamCalc version is in slightly better agreement with my observations.

To-may-to. To-mah-to.

This is like asking "Which is a more accurate version of Peyton Manning? The video of him, or the simulation I see in Madden 2014?" The compulsory answer is Actual Imagery. Full stop. But, with that said, the simulation might be a strong generic week to week rendering of consistent performance, in a particular standardized environment.

Usually I agree with your analogies, but this one I'm not sure about...

Note that I was not trying to draw a comparison between the ASETs and some universal assessment of the diamond's actual performance, but rather, just stating which ASET image better corresponded to my impressions in real life. Since I'm using my own subjective experience as a reference, I get to be right!

The most significant difference (to me) was the difference in the predicted leakage under the table. My impressions based on playing with the diamond was that the dark swaths in the JA are not actually "leaky" in the sense of drawing light from the pavilion, but that the light returned in those regions come from the girdle and crown (and probably more so from the shallower angles of the crown). I think this observation is actually consistent with both ASETs, if one assumes that in the JA setup, the girdle (and possible parts of the crown) were on the "wrong side" of the horizon plane.

However, with only the JA ASET as a reference, I would have no way of knowing that those dark "leakage" areas pull light from near 0°. Those regions would look equally dark if the light was pulled from negative 90° (like a window)!

Given the difficulty in aligning the girdle plane exactly, perhaps it would be useful to add a fourth band of color in the ASET, to represent light coming in just below the horizon (say -10° to 0°, or even better, -10° to +10°)? In my opinion, that would make the system more robust, and allow one to estimate the severity of leakage.

 

[John P]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

I'm curious about your comment that scanner errors may be a bigger problem the ASET photography setups. It would seem to me that the latter has more variables and moving parts (and therefore harder to get right). What are some typical sources of error in scanning?

Non-contact scanners have limitations. Diamond facet surfaces are extremely small at normal sizes, more-so as you move to low carat weights and extra-faceted cuts. It's also very hard to tease out facet-arrangements which graduate with subtle changes. This is why the commonly acknowledged error for popular HD scanners is ± 10 microns linear and ± 0.1 degree angular. Less expensive and portable models have greater error. If not regularly calibrated, the error can increase. Human operation also comes into play. Cleanliness of the stage and the diamond are key in every scan.

In my experience the accuracy of the scan relies on both the equipment and the operator. Big mistakes are usually attributable to inattention or inexperienced users using out of date software and a dirty machines. A careful analyst running a clean operation with good equipment can produce scans which are reliably spooky-accurate.

With that said, we occasionally see subtle ASET imprint errors on AGSL grading reports coming back to us, and they are careful operators. The lab is fast to correct them when we point them out. Many would say we're grousing in the macro since such errors are nowhere near enough to influence the (current) overall performance grade, but we're pedantic about this stuff so I'll mention it here in the context of a micro-discussion.

FWIW, experienced users of diamond design software can spot a wonky scan pretty fast, especially if we have the source file to open.

 

[Rockdiamond]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

drk- do you think the presence of leakage is "bad" in a marquise?
In fact, leakage is an essential element in many super well cut marquise diamonds.

 

[John P]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Usually I agree with your analogies, but this one I'm not sure about...

No worries. For the record, I thought it especially apt (and still do), and since I received a private email from a well-known CG/ASET expert giving me a high five for that analogy I'm going to stand behind it.

(PS: Hey CG/ASET expert - you know who you are - add your 2cents here )

Note that I was not trying to draw a comparison between the ASETs and some universal assessment of the diamond's actual performance, but rather, just stating which ASET image better corresponded to my impressions in real life. Since I'm using my own subjective experience as a reference, I get to be right!

100%. You have Peyton in your living room. Whether he's more like the game or the video from MNF - you can tell us.

The most significant difference (to me) was the difference in the predicted leakage under the table. My impressions based on playing with the diamond was that the dark swaths in the JA are not actually "leaky" in the sense of drawing light from the pavilion, but that the light returned in those regions come from the girdle and crown (and probably more so from the shallower angles of the crown). I think this observation is actually consistent with both ASETs, if one assumes that in the JA setup, the girdle (and possible parts of the crown) were on the "wrong side" of the horizon plane... ...However, with only the JA ASET as a reference, I would have no way of knowing that those dark "leakage" areas pull light from near 0°. Those regions would look equally dark if the light was pulled from negative 90° (like a window)!

Can we see the CG ASET with a backlit background to really shake things up? To your comments, do remember the CG is using completely absent light under the girdle plane. But even the best-controlled actual setups have some level of ambient light sneaking around underneath since the diamond is transparent. Add to that; you're seeing things IRL with stereo vision - mitigating some of the low-angled light and leakage - and the VFs we perceive in this mag view may or may not create fans or events large enough for you to perceive.

Say all this to say: ASET for me is a great indicator of a diamond's brightness, contrast distribution and light-loss in certain conditions. I don't attempt to go too far beyond that.

Given the difficulty in aligning the girdle plane exactly, perhaps it would be useful to add a fourth band of color in the ASET, to represent light coming in just below the horizon (say -10° to 0°, or even better, -10° to +10°)? In my opinion, that would make the system more robust, and allow one to estimate the severity of leakage.

That's GilbertonScope (searchable in past discussions). AGS purchased the patent from him, from which ASET was derived.

 

[Karl_K]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

I agree it is very nice, sorry for not saying so earlier.
Congrates!!!

Thanks, Karl. I have to give credit to Niel for spotting it in JA's catalog, and to Gypsy and others for convincing me to look past the "birthmark" (which turned out to be eye-clean in the end).

I'm curious about your comment that scanner errors may be a bigger problem the ASET photography setups. It would seem to me that the latter has more variables and moving parts (and therefore harder to get right). What are some typical sources of error in scanning?

If you look at the wire frame output from a diamond scan you will often see facets that don't exist on the real diamond or facets that don't line up the way they do on the real diamond.
When a facet is yawed(tilted) or the scanner thinks it is yawed from the additive errors the scanned image will often split the facets.
Basically the less the angle difference the harder it is for the scanner to tell them apart and fancies often have angles that are closer to each other than a RB.

Dust and even oil film under the table of the diamond is a huge source of scanning errors because it is not sitting level on the scanner. Dust or oil on any facet is a source of many other errors.
Then you have the issue of calibration and holding calibration.
A scanner that looks calibrated on a simple design may be out of calibration on a more complex design with close angles.

Basically the scanners are optical devices they take a bunch of pictures of the diamond as it rotates on a platter. Then using shadows and edge finding tools it analyzes those photos and finds facet edges and compares the angles to the platter the diamond is sitting on. There are other complex steps but that is the basics.
Different scanners and different settings on the scanners use different numbers of photos.
On complex stones different firmware levels can give different results as can different tuning where the software adapts to the hardware based on parameters entered.
What it boils down to is that complex designs require much more care in setup, calibration, settings and attention to detail to get the max accuracy the scanner is capable of and overall that is not that high.
Often all that does not come together well on even a RB which it was designed around and is relatively easy.

 

[drk14]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

David,
Thanks for your response.

Of course the CG ASET skips the imperfection, easily visible on the photograph.

No argument there. One would be silly to make a purchasing decision based on an ASET alone, whether from a photographic ASET setup or a computer simulation. In this thread, I'm only discussing assessment light return qualities.

It all highlights the difficulties of trying to use aset to make buying decisions.
Just like photos, comparing one vendor to another is practically useless.

I agree that there are can be challenges in making ASETs reproducible between vendors. However, that's a far stretch from making the technology "useless". On the contrary, the comparison of the two ASETs in the original post above primarily shows that the two versions are remarkably similar, which is evidence that these two imaging techniques, although very different in their methods, are both quite accurate. I'm confident that technology will improve further to make ASET imaging even more reproducible.

Of course, some vendors may not bother to make sure they get the ASETs right, but in a forum like PS, it will quickly become known which vendors are able to generate accurate ASETs.

Having said that- I like the way the CG aset looks much better, from a marketing standpoint- it just looks cleaner.

It is cleaner, probably as a combination of tracing only a finite number of light reflections, and having an infinite depth-of-focus. In the case of my diamond, I think this works to the benefit of the DiamCalc ASET, because my general IRL impression of the diamond is that is has "crisp" facets. However, if I had a diamond with mushy/blurry virtual facets (which is often seen in marquise), the I think the computer-generated ASET would be misleading: it would probably still look like it had sharp facets.

 

[Karl_K]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

lol I see John and I were posting at the same time.
I like John's explanation it fits pretty well and I was the one that sent him a Bravo.

Talking about ASET white and black I find white overly harsh and black not right either.
I think ASET 18% grey or ASET neutral would be more real world but impossible to implement in real life vs computer simulation.

 

[Rockdiamond]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Thank you drk

About increases in technology....I too am waiting to see what is around the corner.
By simply enlarging the hole on top, it's now possible to very easily take an aset picture with an iphone.

But photos taken that way look like carp- especially compared with the scan based ASET- which as a person who wants their site to look nice- that matters a lot.
If there's general agreement that scan based ASET is acceptable, that will make it easier for sellers to provide a consistent type of picture. But it seems there's no consensus on that.
It's not that I'm arguing against the technology - more the interpretation of the info it provides.

For example, your questions about leakage and how it affects light performance.

 

[Texas Leaguer]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

I agree it is very nice, sorry for not saying so earlier.
Congrates!!!

Thanks, Karl. I have to give credit to Niel for spotting it in JA's catalog, and to Gypsy and others for convincing me to look past the "birthmark" (which turned out to be eye-clean in the end).

I'm curious about your comment that scanner errors may be a bigger problem the ASET photography setups. It would seem to me that the latter has more variables and moving parts (and therefore harder to get right). What are some typical sources of error in scanning?

If you look at the wire frame output from a diamond scan you will often see facets that don't exist on the real diamond or facets that don't line up the way they do on the real diamond.
When a facet is yawed(tilted) or the scanner thinks it is yawed from the additive errors the scanned image will often split the facets.
Basically the less the angle difference the harder it is for the scanner to tell them apart and fancies often have angles that are closer to each other than a RB.

Dust and even oil film under the table of the diamond is a huge source of scanning errors because it is not sitting level on the scanner. Dust or oil on any facet is a source of many other errors.
Then you have the issue of calibration and holding calibration.
A scanner that looks calibrated on a simple design may be out of calibration on a more complex design with close angles.

Basically the scanners are optical devices they take a bunch of pictures of the diamond as it rotates on a platter. Then using shadows and edge finding tools it analyzes those photos and finds facet edges and compares the angles to the platter the diamond is sitting on. There are other complex steps but that is the basics.
Different scanners and different settings on the scanners use different numbers of photos.
On complex stones different firmware levels can give different results as can different tuning where the software adapts to the hardware based on parameters entered.
What it boils down to is that complex designs require much more care in setup, calibration, settings and attention to detail to get the max accuracy the scanner is capable of and overall that is not that high.
Often all that does not come together well on even a RB which it was designed around and is relatively easy.

This is an excellent overview Karl. You should turn it into an article. In your spare time.
Of course you would also want to hit on the strengths as well as weaknesses. It is a pretty incredible technology.

 

[Texas Leaguer]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

In the case of my diamond, I think this works to the benefit of the DiamCalc ASET, because my general IRL impression of the diamond is that is has "crisp" facets. However, if I had a diamond with mushy/blurry virtual facets (which is often seen in marquise), the I think the computer-generated ASET would be misleading: it would probably still look like it had sharp facets.

This is an interesting idea. I would love to hear commentary on this.

 

[Karl_K]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

In the case of my diamond, I think this works to the benefit of the DiamCalc ASET, because my general IRL impression of the diamond is that is has "crisp" facets. However, if I had a diamond with mushy/blurry virtual facets (which is often seen in marquise), the I think the computer-generated ASET would be misleading: it would probably still look like it had sharp facets.

This is an interesting idea. I would love to hear commentary on this.

It is incredibly complex and I am hesitant to tackle it because it can lead to reading things into an ASET image that are not there.
Focus stacking real ASET images can lead to a more crisp VF representation.
Basically what you do is take a 100+ photos of the diamond without changing anything but the focus point of the camera then combine the best focused parts of each image into one image using software.
But the interpretation of such an image can lead you down a rabbit hole.

 

[Karl_K]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Part of the issue is the crispness of a virtual facet needs motion to be seen.
A crisp virtual facet is like a light bulb on a switch on and of on and off.
A less crisp virtual facet is like the same light bulb on a dimmer where the knob is slowly turned.
Both can look the same in a static ASET image both virtual and real.
That is also incredibly complex to account for all the factors that affect it.

 

[Texas Leaguer]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Not looking to take this down a rabbit hole. I'm claustrophobic and a little afraid of the dark. And rabbits.

But, what would you say to drk's theory that a CG ASET (from a good scan) could be misleading vs a well captured real ASET photo in terms of crispness of the facet patterns. Is this a weakness of CG ASET?

 

[Serg]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

David,
Thanks for your response.

Of course the CG ASET skips the imperfection, easily visible on the photograph.

No argument there. One would be silly to make a purchasing decision based on an ASET alone, whether from a photographic ASET setup or a computer simulation. In this thread, I'm only discussing assessment light return qualities.

It all highlights the difficulties of trying to use aset to make buying decisions.
Just like photos, comparing one vendor to another is practically useless.

I agree that there are can be challenges in making ASETs reproducible between vendors. However, that's a far stretch from making the technology "useless". On the contrary, the comparison of the two ASETs in the original post above primarily shows that the two versions are remarkably similar, which is evidence that these two imaging techniques, although very different in their methods, are both quite accurate. I'm confident that technology will improve further to make ASET imaging even more reproducible.

Of course, some vendors may not bother to make sure they get the ASETs right, but in a forum like PS, it will quickly become known which vendors are able to generate accurate ASETs.

Having said that- I like the way the CG aset looks much better, from a marketing standpoint- it just looks cleaner.

It is cleaner, probably as a combination of tracing only a finite number of light reflections, and having an infinite depth-of-focus. In the case of my diamond, I think this works to the benefit of the DiamCalc ASET, because my general IRL impression of the diamond is that is has "crisp" facets. However, if I had a diamond with mushy/blurry virtual facets (which is often seen in marquise), the I think the computer-generated ASET would be misleading: it would probably still look like it had sharp facets.

If we speak about Crispy images/DoF( depth of field) then most macro photos are even more misleading.
DoF is much bigger for Human observation conditions than in a typical macro photo.

 

[drk14]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

drk- do you think the presence of leakage is "bad" in a marquise?
In fact, leakage is an essential element in many super well cut marquise diamonds.

I'd prefer for this thread not to stray too far from it's original topic, but to answer your question, I do not automatically assign any negative connotation to the word "leakage", either in marquise or in any other cut. To me it's just a shorthand for describing regions in which the returned light was originally drawn from the pavilion.

If admiring loose stones, "leakage" can enhance the appearance of a diamond. However, in my case, the marquise will be set in a basket head of a ring, which will block some of the pavilion light. Therefore, I expect that any large areas of "leakage" would be somewhat darker or less lively after setting the diamond in this manner. In this specific context, then, large leakage zones would not be desirable, in my opinion.

 

[Rockdiamond]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

David,
Thanks for your response.

Of course the CG ASET skips the imperfection, easily visible on the photograph.

No argument there. One would be silly to make a purchasing decision based on an ASET alone, whether from a photographic ASET setup or a computer simulation. In this thread, I'm only discussing assessment light return qualities.

It all highlights the difficulties of trying to use aset to make buying decisions.
Just like photos, comparing one vendor to another is practically useless.

I agree that there are can be challenges in making ASETs reproducible between vendors. However, that's a far stretch from making the technology "useless". On the contrary, the comparison of the two ASETs in the original post above primarily shows that the two versions are remarkably similar, which is evidence that these two imaging techniques, although very different in their methods, are both quite accurate. I'm confident that technology will improve further to make ASET imaging even more reproducible.

Of course, some vendors may not bother to make sure they get the ASETs right, but in a forum like PS, it will quickly become known which vendors are able to generate accurate ASETs.

Having said that- I like the way the CG aset looks much better, from a marketing standpoint- it just looks cleaner.

It is cleaner, probably as a combination of tracing only a finite number of light reflections, and having an infinite depth-of-focus. In the case of my diamond, I think this works to the benefit of the DiamCalc ASET, because my general IRL impression of the diamond is that is has "crisp" facets. However, if I had a diamond with mushy/blurry virtual facets (which is often seen in marquise), the I think the computer-generated ASET would be misleading: it would probably still look like it had sharp facets.

If we speak about Crispy images/DoF( depth of field) then most macro photos are even more misleading.
DoF is much bigger for Human observation conditions than in a typical macro photo.

100% true in my experience Serg.
The camera has to pick a spot to focus on, the human eye has an infinitely more variable capacity to constantly re-focus.
Sometimes I actually prefer a slightly out of focus macro shot for these reasons.

 

[Serg]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

David,
Thanks for your response.

Of course the CG ASET skips the imperfection, easily visible on the photograph.

No argument there. One would be silly to make a purchasing decision based on an ASET alone, whether from a photographic ASET setup or a computer simulation. In this thread, I'm only discussing assessment light return qualities.

It all highlights the difficulties of trying to use aset to make buying decisions.
Just like photos, comparing one vendor to another is practically useless.

I agree that there are can be challenges in making ASETs reproducible between vendors. However, that's a far stretch from making the technology "useless". On the contrary, the comparison of the two ASETs in the original post above primarily shows that the two versions are remarkably similar, which is evidence that these two imaging techniques, although very different in their methods, are both quite accurate. I'm confident that technology will improve further to make ASET imaging even more reproducible.

Of course, some vendors may not bother to make sure they get the ASETs right, but in a forum like PS, it will quickly become known which vendors are able to generate accurate ASETs.

Having said that- I like the way the CG aset looks much better, from a marketing standpoint- it just looks cleaner.

It is cleaner, probably as a combination of tracing only a finite number of light reflections, and having an infinite depth-of-focus. In the case of my diamond, I think this works to the benefit of the DiamCalc ASET, because my general IRL impression of the diamond is that is has "crisp" facets. However, if I had a diamond with mushy/blurry virtual facets (which is often seen in marquise), the I think the computer-generated ASET would be misleading: it would probably still look like it had sharp facets.

If we speak about Crispy images/DoF( depth of field) then most macro photos are even more misleading.
DoF is much bigger for Human observation conditions than in a typical macro photo.

100% true in my experience Serg.
The camera has to pick a spot to focus on, the human eye has an infinitely more variable capacity to constantly re-focus.
Sometimes I actually prefer a slightly out of focus macro shot for these reasons.

David, I would like explain your statement for others.

Human eye has 18mm focal length. It means for typical diamond observation conditions with naked eye DoF is infinite ( diamond and light sources are in focus in same time, and diamond become crispy)
in macro photo DoF is much less than distance between diamond and light sources. so if you want show sharp light sources reflections in diamond you have to do the diamond out focus( you need use focus point between diamond and light sources)

 

[drk14]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

Add to that; you're seeing things IRL with stereo vision - mitigating some of the low-angled light and leakage.

John, thanks for your comments, and for the pointer to the Gilbertonscope. To the above statement, I would just like to clarify that when examining for leakage IRL, I was using a loupe, so mono vision only. Not that I think it will change the discussion any.

 

[Rockdiamond]

Re: An interesting (to me) ASET comparison -- JA vs. DiamCal

DRK- to your point about setting the diamond: My experience is that the leakage will indeed STILL add life and attractive aspects to the diamond, even after setting.
In terms of large areas of leakage, yes, that can be seen naked eye, and might ( but not always) be a negative aspect.
Consider that any large areas of color in ASET, be they red, green or "leakage" are going to give a less sparkly and more glittery appearance. Such as a step cut with broader swaths of color in ASET

I see no problem areas of leakage in the ASET, or pics of your stone.
BTW, you show a very strong understanding of these concepts.
In my experience here on PS, posters are oft times warned about leakage that is no problem whatsoever.