AVS Forum banner

1 - 20 of 44 Posts

·
Registered
Joined
·
79 Posts
Discussion Starter #1
I was following the Viewing Angle Article.pdf of Spectracal and doing Meter Distance calculation using the formula D = 2L tan(theta/2), as given in the article. I believe a correction is required in the formula to account for the distance between the point of intersection of FOV lines and the surface of the Meter, as a Meter is not a point object. I thought perhaps I should discuss it here.

Mathematically, the foregoing formula provides the Distance from the Display surface, for a virtual point that is formed by the intersection of the field-of-view lines of the Meter (like shown in the attached drawing). That virtual point may be somewhere along the length of the meter or behind the meter. In order to find the exact positioning of that point from the front end of the meter (the foam pad of i1D3 for example), it will be required to know the diameter of the virtual circle that the Meter measures in Contact Mode. Once we have the Contact Mode virtual circle diameter, trigonometric calculations will give the distance between the point of intersection of FOV lines and the front of the meter. I have i1D3 and i1Pro 2 Meters, and am not sure about the corresponding diameters for these. It would be really helpful if the Viewing Angle Article mentioned that too.

I think that Spectracal should revise the formula D = 2L tan(theta/2) with the correction for the point of intersection of FOV lines of the Meter as mentioned above.
 

Attachments

·
Registered
Joined
·
1,007 Posts
Good point omarank. I guess they put the virtual point you are talking about at the center of the "touching" surface of the meters. In contact mode that distance will never be a true zero because i1d3 has the lens 1/2 mm inside the body and i1 Pro 2 has the contact mode suit. But what I'm more curious about is: have you checked if those calcs are accurate? Does i1D3 really read a virtual circle with a diameter of 25 mm at about 136 mm of distance? Is it true that distance is directly proportional to FOV (eg. D: 136 mm, FOV: 25 mm; D= 272 mm, FOV=50 mm)?

Maybe, you should contact directly Spectracal to have a technically deep answer.
 

·
Registered
Joined
·
5,001 Posts
Good point omarank. I guess they put the virtual point you are talking about at the center of the "touching" surface of the meters. In contact mode that distance will never be a true zero because i1d3 has the lens 1/2 mm inside the body and i1 Pro 2 has the contact mode suit. But what I'm more curious about is: have you checked if those calcs are accurate? Does i1D3 really read a virtual circle with a diameter of 25 mm at about 136 mm of distance? Is it true that distance is directly proportional to FOV (eg. D: 136 mm, FOV: 25 mm; D= 272 mm, FOV=50 mm)?

Maybe, you should contact directly Spectracal to have a technically deep answer.
Not Spectracal, but X-Rite. #ConnecTEDDD also has this type of info on his website...
 

·
Registered
Joined
·
9,498 Posts

·
Registered
Joined
·
5,001 Posts
The info available here, are only the info X-Rite has provided for each instrument specifications.

Before some years, Darryl Bird, the NIST lab director of SpectraCAL performed a FOV comparison test of many meters, and we are based to that info to match different meter's FOV: http://www.spectracal.com/Documents/Viewing Angle Article.pdf
My point was, Ted, that they don't need to reinvent the wheel. This work has already been done, and at the distances actually involved, it's good enough. Now, if you were somehow trying to use this to calculate distances to other planets or even to our moon, you'd need more precision. But not for this...
 

·
Registered
Joined
·
9,498 Posts
My point was, Ted, that they don't need to reinvent the wheel. This work has already been done, and at the distances actually involved, it's good enough. Now, if you were somehow trying to use this to calculate distances to other planets or even to our moon, you'd need more precision. But not for this...
The point is to find out if the test performed accurately, for that reason the user is asking some more details.
 

·
Registered
Joined
·
5,001 Posts
How much of a difference do they really think it will make, Ted? Given the other, larger variables we have when calibrating - meter and display repeatability, drift, etc. - this is small potatoes. It's akin to the debate about how many angels can dance on the head of a pin, IMHO.
 

·
Registered
Joined
·
9,498 Posts
How much of a difference do they really think it will make, Ted? Given the other, larger variables we have when calibrating - meter and display repeatability, drift, etc. - this is small potatoes. It's akin to the debate about how many angels can dance on the head of a pin, IMHO.
If you are not interest you can skip that thread, your posts are not helping to any of the user questions here.
 

·
Registered
Joined
·
1,007 Posts
Precision is all, this is science. But everyone is free to do it or interpret it as they like, with easy references and contact mode. It would just be really appreciated if those people won't ruin others' thread or curiosities. Thanks.
 
  • Like
Reactions: omarank

·
Registered
Joined
·
79 Posts
Discussion Starter #11 (Edited)
Good point omarank. I guess they put the virtual point you are talking about at the center of the "touching" surface of the meters. In contact mode that distance will never be a true zero because i1d3 has the lens 1/2 mm inside the body and i1 Pro 2 has the contact mode suit. But what I'm more curious about is: have you checked if those calcs are accurate? Does i1D3 really read a virtual circle with a diameter of 25 mm at about 136 mm of distance? Is it true that distance is directly proportional to FOV (eg. D: 136 mm, FOV: 25 mm; D= 272 mm, FOV=50 mm)?

Maybe, you should contact directly Spectracal to have a technically deep answer.
It is not about where they put the virtual point, but about the use of a formula that gives the distance of the probe to the screen from a virtual point. The distance vs diameter table that is there on the Viewing Angle Article can be easily computed using the formula that is provided in the article. And that formula can be derived through simple trigonometric calculations. If you want to measure from the surface of the meter (which is obviously the convenient way), you will have to add a correction to the formula or alternatively adjust the distance given by the formula with an offset (corresponding to the distance between the surface of the meter and the virtual point).

In order to measure a diameter of 25 mm, you will have to place i1D3 such that the virutal point is at 136 mm from the display. If I assume that the contact mode diameter of i1D3 is 24 mm (as per XRite data) and do the calculations for the virtual point, it should be about 8.6 cm behind the surface (foam pad of i1D3). And so, you should place i1D3 (foam pad) at 13.6 - 8.6 = 5 cm from the display. [Note that 8.6 cm has been calculated using the TLT FOV angle and it will remain the same for FWHM measurements. 25 mm diameter at 13.6 cm virtual point distance is for FWHM measurements.]

Distance is not directly proportional to FOV angle. It is inversely proportional as given by: L = D/(2*tan(theta/2)) where D is the diameter of the area measured. Distance is directly proportional to Diameter though, which you are referring to.

I did contact SpectraCAL before posting here. I wrote email to Darrell Bird about a month ago, then to Tyler (@WiFi-Spy) in a few days. I also tried contacting Tyler through PM on avsforum. But I haven't yet got any reply.

How much of a difference do they really think it will make, Ted? Given the other, larger variables we have when calibrating - meter and display repeatability, drift, etc. - this is small potatoes. It's akin to the debate about how many angels can dance on the head of a pin, IMHO.
Let's take an example. If we know that the Diameter of the area that we want the meter to measure is 15 cm, then the meter distance (FWHM of i1D3) as per the formula will be 81.6 cm. And we need to subtract 8.6 cm from it to calculate the distance (81.6 - 8.6 = 73 cm) between the display and the i1D3 foam pad (which we can easily measure though a tape). So, if your test pattern had a height of 15 cm (or close) and you placed the meter at the formula given distance of 81.6 cm (instead of 73 cm), the meter's field-of-view will be covering some of the black bar as well outside the test pattern. And this has nothing to do with the stability of the display.

Of course, for projectors, when you place the probe at several feet from the screen, it won't matter much.

Whether or not you care about that precision, having this correction in-built in the formula will only decrease the likelihood of incorrect placement of the meter.
 

·
Registered
Joined
·
1,007 Posts
@omarank I perfectly got your point at first read but what I'm asking is "have you verified if at 136mm i1D3 reads a virtual circle with a diameter of 25mm?" or you just made a formula check. I could be wrong (and I really hope so), but if they placed that (wrong) formula I'm pretty sure that those distances are inaccurate. I don't even know if @WiFi-Spy was already working for SpectraCal, but since he's there now he will surely dispel our doubts.
 

·
Registered
Joined
·
79 Posts
Discussion Starter #13
I have not done any experiments to see what area i1D3 measures at a certain distance. As should be clear from my posts, I have just been talking about the formula provided and the correction required. If just the contact mode virtual circle diameters of different probes are provided/confirmed, one can do the calculations to make the correction.

Just to clarify, the formula is not wrong. Only a correction is required, which will refine the distance value that is outputted by the formula. I should also mention that FOV angles (TLT and FWHM) of the different probes have experimentally been observed by SpectraCAL. We have no other reference to confirm if all those angle values are correct. So, if you do an experiment to find what area a probe measures at some distance and see that it does not tally with the results of the formula even after formula correction, it could be that the FOV angle value used with the formula is not in alignment with the actual FOV angle of the probe.

I hope to see Tyler or anyone from SpectraCAL comment here.
 

·
Registered
Joined
·
4 Posts
Hi Omar,
Apologies for not getting your email. I think I wrote that article more than 6 years ago but I’ll try to clear things up as best I can.

For some background, although I stand by the data, the article was never intended to be academically rigorous. It was written with the intention to help a subset of people who were needing to use certain “in contact” meters on something like a projector screen. Some manufacturers may not have originally intended them to be used this way and their data about Field of View was spotty at best but sometimes altogether nonexistent. So the article was written as a general guideline to help mitigate the possibility of someone reading a grossly different read area than expected.

You’re absolutely correct to point out that to make the formula mentioned an exact representation of the meter’s field of view, we would need to adjust it based on the meter’s effective focal point. However to find the effective focal length of the meter’s optical array, we would need to know many details about the meter’s lenses that are not usually published by manufacturers. Some meters have multiple lenses in their optical stack while some meters have none (simply putting a color filter in front of a sensor).

To get around all that ambiguity, I decided to just test the read area response at various distances for all the meters mentioned. I did this by painstakingly mapping out the read areas with a narrow aperture laser at all the listed distances. As mentioned in the article, some meters had a hard cutoff between response and no-response, while some showed a very Gaussian curve with the light response tapering off gradually.

All the read area data in the article comes from actually measuring it, not from the formula. However, within the error margins of my actual measurements, the formula prediction matches the data very well at all distances above 12 inches (~30.5cm) when using the meter’s front point of contact with the screen as the distance (as you appropriately called it “the convenient way”). But you, are correct, the formula is not exact and it does break down at distances below 12 inches. Example: The diameter of the i1D3’s read area is 24mm in contact with the screen, not zero as the formula would imply. So perhaps a better way to have framed the problem would have been to show the read area at a distance in relation to its minimum read area when in contact.

For the example you described (paraphrasing here), measuring a 15 cm target but reading some of the black bar outside the pattern, the i1D3 is absolutely not the right tool for that level of precision. As mentioned in the article, the response shape is Gaussian so there is no hard line between where light is read and is not read. There are certainly valid use cases where a precise area measurement is needed, however, for that type of read area accuracy, one would need something like a Konica Minolta or Photo Research that has the read area explicitly defined and observable through an SLR-like viewfinder. Of course those are prohibitively expensive which gets back to the point of the original article which was to provide a relatively good estimate of the empirically measured read area of more affordable meters.

Hope that helps.
 

·
Registered
Joined
·
1,007 Posts
Hi Omar,

To get around all that ambiguity, I decided to just test the read area response at various distances for all the meters mentioned. I did this by painstakingly mapping out the read areas with a narrow aperture laser at all the listed distances. As mentioned in the article, some meters had a hard cutoff between response and no-response, while some showed a very Gaussian curve with the light response tapering off gradually.

All the read area data in the article comes from actually measuring it, not from the formula. However, within the error margins of my actual measurements, the formula prediction matches the data very well at all distances above 12 inches (~30.5cm) when using the meter’s front point of contact with the screen as the distance (as you appropriately called it “the convenient way”). But you, are correct, the formula is not exact and it does break down at distances below 12 inches. Example: The diameter of the i1D3’s read area is 24mm in contact with the screen, not zero as the formula would imply. So perhaps a better way to have framed the problem would have been to show the read area at a distance in relation to its minimum read area when in contact.
Thank you from the deep of my heart for your answer @DarrellB as I am projecting and hopefully producing some accessories for xrite meters basing ALL on your article and results. For that reason I'd like to have a yes or no answer to these questions (I'm sorry if that makes you feel uncomfortable):
1) Let alone the formula, are those data below 12 inches correct? eg. i1D3 at ~136 mm reads a virtual circle with a diameter of 25 mm, while i1 Pro 2 reads the same circle at ~178 mm (obviously with meters set right and perfectly in perpendicular axis with the panel being measured). I don't care if actually they read a slightly bigger circle, All I need to know if you are sure that they read that and not maybe less.

2) Assuming that, as you said, your data comes from real measurements, distances and area measured have a really linear relation (i1D3: circle diameter 25 mm, distance 136 mm; circle diameter 50 mm, distance 272 mm), do you confirm that?

Once again thank you very much.
Miki
 

·
Registered
Joined
·
79 Posts
Discussion Starter #16
Hi Omar,


You’re absolutely correct to point out that to make the formula mentioned an exact representation of the meter’s field of view, we would need to adjust it based on the meter’s effective focal point. However to find the effective focal length of the meter’s optical array, we would need to know many details about the meter’s lenses that are not usually published by manufacturers. Some meters have multiple lenses in their optical stack while some meters have none (simply putting a color filter in front of a sensor).

To get around all that ambiguity, I decided to just test the read area response at various distances for all the meters mentioned. I did this by painstakingly mapping out the read areas with a narrow aperture laser at all the listed distances. As mentioned in the article, some meters had a hard cutoff between response and no-response, while some showed a very Gaussian curve with the light response tapering off gradually.

All the read area data in the article comes from actually measuring it, not from the formula. However, within the error margins of my actual measurements, the formula prediction matches the data very well at all distances above 12 inches (~30.5cm) when using the meter’s front point of contact with the screen as the distance (as you appropriately called it “the convenient way”). But you, are correct, the formula is not exact and it does break down at distances below 12 inches. Example: The diameter of the i1D3’s read area is 24mm in contact with the screen, not zero as the formula would imply. So perhaps a better way to have framed the problem would have been to show the read area at a distance in relation to its minimum read area when in contact.

For the example you described (paraphrasing here), measuring a 15 cm target but reading some of the black bar outside the pattern, the i1D3 is absolutely not the right tool for that level of precision. As mentioned in the article, the response shape is Gaussian so there is no hard line between where light is read and is not read. There are certainly valid use cases where a precise area measurement is needed, however, for that type of read area accuracy, one would need something like a Konica Minolta or Photo Research that has the read area explicitly defined and observable through an SLR-like viewfinder. Of course those are prohibitively expensive which gets back to the point of the original article which was to provide a relatively good estimate of the empirically measured read area of more affordable meters.

Hope that helps.
Hi Darrell

Thanks for your reply. The details that you provided regarding how the measurements were taken are actually helpful.

So, my understanding is that the read areas of the different probes were found experimentally first at different distances. As at larger distances the error in the formula without any correction would be small, the formula could be used to calculate the FOV angles and perhaps that’s how you found the FOV angles. Please confirm if otherwise.

I understand that probes like i1D3 have a gaussian response shape, still a corrected formula will help a user in carefully placing the probe at near distances for desired FWHM read areas (where most of the light is captured). In fact, I have seen people using the formula for 25 mm spot measurement and the formula outputted value of 13.6 cm for inaccurate i1D3 placement.

I am not sure why we need to get the details of lenses etc., when through simple geometric/trigonometric calculations we can find the distance correction for the formula. Don’t we just need to know the contact mode read areas of the different probes to find their respective corrections? For instance, if we do the calculations, the distance correction for i1D3 would be 8.6 cm if we assume the contact mode read area as 24 mm.

If SpectraCAL could confirm the contact mode read areas of the different probes through the same/similar experiments as done previously, that alone would be quite helpful. The aperture size value of the probes may not be the actual read area. Like in case of i1D3, we know its aperture size as 24 mm, but whether the probe actually measures 24 mm area or a smaller area when in contact mode can only be confirmed experimentally.
 

·
Registered
Joined
·
1,150 Posts
How much of a difference do they really think it will make, Ted?
Yep - your right, it's largely irrelevant. Meters are made to measure Luminance (cd/m^2), which is independent of the distance to the display, as long as the display is larger than the instrument FOV, and the pixel size is small enough that multiple pixels are averaged. Most of the effort in calibration and profiling is all relative measurements anyway - absolute brightness accuracy is not that critical, given the adaptive nature of the human eye.

The main effect of putting the instrument distant to the display is letting in ambient light, which may or may not be something that the user intends. For this reason, repeatability is going to be better with direct contact instruments.
 

·
Registered
Joined
·
1,007 Posts
Once again, I would like to invite everyone to respect the curiosity of others, even more if they are based on science and standards. It seems to me that everyone here have their religion and standard of measurement, built on personal experiences and tests. I respect that.

I (and some others) prefer to follow the official standard: IDMS display measurement and metrology standard publication, which does not consider contact mode not even as a possible measurement method for many reasons that you can read by yourself, if you want. Thank you for your time and understanding.
 
  • Like
Reactions: omarank

·
Registered
Joined
·
533 Posts
Been a while since I read the IDMS pub, alas as most of these types of measurements described in the IDMS are done in very control situations and not in consumer homes, offices and the like. Hence light artifacts typically out weigh differential area measurements.

As far as I have concluded with FOV, related to creating matrix's for colorimeters from a spectral device is that is helps being somewhat close with capture intensity's. This is because any conversion still has error, the Y influence can be strong, when the differences get large the errors comes along stronger. They only have to be somewhat close though, I really don't think being exactly mm perfect makes it much better. The secondary issue is when comparing two devices that have differing FOV, they typically have different sensitivity ranges, arguably one should be at the same percentage range than equal FOV. Another discussion.

However, there is an example where area capture can have an influence over anything else. This is the sensor noise floor for low light level measurements.
Keep in mind, as pointed out once a device is in it's sweet spot range, measurements are normalized. Too see differences that the device detects you have to look at raw data readings the sensor outputs.
To overcome the noise floor, just consider sensor basics. Essentially any light sensor is a bucket, some call them "wells", which collect photons. At low photon levels the bucket can't shift the sensor into current output as it has to overcome a whole list of resistances before current is produced. To make life more difficult, the sensor typically has a nm spectral sensitivity curve where they are typically low sensitivity in the blue and red regions of the spectrum. (Just consider that when D65 contains 11% of what we see as white or black).

OK, so how would FOV help? Basically you have two choices, if you have ever worked with telescopes and try to capture images through them you discover this very fast. The choices are, 1. Increase capture time or 2. increase the collection area.

So what happens in sensing displays?, well the software reacts to low light sensing to take longer duration readings. But you find that readings bounce about because they contain a lot of noise. If you increase the capture area you increase photon capture for the same time period, and raise the signal relative to the noise, S/N ratio. So you don't change the normalized output of the reading but you decrease the noise and the reading may be more stable.

The downside is ambient light getting into your sensor from the room, which can be reflected light bouncing off the walls in a darkened room. Hence the light measurement guidelines instruct use of cone frustums to eliminate ambient light noise. Which brings us back to the real world of our reality, for what ever that is(another discussion), where it isn't practical for the average person to control all artifacts and you are left with placing the sensing device at a distance that eliminates any ambient light and allow longer duration readings.
 

·
Registered
Joined
·
1,150 Posts
Once again, I would like to invite everyone to respect the curiosity of others, even more if they are based on science and standards. It seems to me that everyone here have their religion and standard of measurement, built on personal experiences and tests. I respect that.
There's no learning or progress if you regard every opinion as "equally valid", and not open to discussion or criticism.

Many people seem to waste their time pursuing irrelevancies, simply because they don't understand how things actually work. It's all black magic voodoo, and they feel compelled to make cryptic offerings to the gods to ensure success.

Add some science and understanding, and the voodoo can be banished.
 
1 - 20 of 44 Posts
Top