SpyderX Review - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 13Likes
  • 4 Post By TomHuffman
  • 1 Post By TomHuffman
  • 5 Post By nodixe
  • 1 Post By Vicent
  • 1 Post By TomHuffman
  • 1 Post By gwgill
 
Thread Tools
post #1 of 29 Old 06-18-2019, 10:41 PM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
SpyderX Review

See this review of the new DataColor Spyder5 colorimeter.


Hint: It offers a huge improvement over previous Spyder instruments.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
Sponsored Links
Advertisement
 
post #2 of 29 Old 06-19-2019, 07:45 AM
AVS Forum Special Member
 
BigCoolJesus's Avatar
 
Join Date: Feb 2009
Posts: 4,992
Mentioned: 18 Post(s)
Tagged: 0 Thread(s)
Quoted: 2090 Post(s)
Liked: 1702
Quote:
Originally Posted by TomHuffman View Post
See this review of the new DataColor Spyder5 colorimeter.


Hint: It offers a huge improvement over previous Spyder instruments.

Is the peak nits known for use on HDR displays? With the soon to be released Vizio's breaking into the 2000nits territory, looking for a new affordable meter than can handle measurement accuracy that high.

Thanks!
BigCoolJesus is offline  
post #3 of 29 Old 06-20-2019, 07:44 PM
AVS Forum Special Member
 
Killroy's Avatar
 
Join Date: Oct 2008
Location: Ninth Circle of Hell
Posts: 2,329
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 433 Post(s)
Liked: 327
Has anyone tried the new Spyder X for generating 3DLUT corrections using ArgyllCMS and DisplayCAL for use in MadVR video renderer.

I have a Spyder 5 that worked great for monitors but for projectors it seemed to wane in accuracy.
Killroy is offline  
Sponsored Links
Advertisement
 
post #4 of 29 Old 06-21-2019, 01:04 AM
Member
 
Join Date: Nov 2018
Posts: 27
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 2
I find this a bold claim: "The SpyderX's accuracy is comparable to the i1Display Pro" since colorimeter accuracy (raw reading) is limited to how close they are to an standard observer in wavelengths were a display's light peaks are placed.
I mean, testing raw uncorrected accuracy with one backlight type could give you close results for two different colorimeter models (like SpyderX and i1d3) and with another backlight very different ones.
Maybe I missed something but in that text/"review" I only see a WOLED, I see no information if they were raw uncorrected readings, corrected by custom matrix (so they wil match) or corrected by vendor bundled corrections.
As a consumer I would want that such testing compares several backlight types like:
-WLED sRGB-like consumer TV/monitor
-LED P3 devices like those WLED PFS phoshpor screens in LED P3 TVs, multimedia monitors or newer macs
-Widegamut graphic arts displays like GB-LEDs or WLED PFS with AdobeRGB coverage
-consumer WOLED (ok, not "backlight")

Also since CIE xyY is not perceptually uniform I would like to see raw readings & vendor corrected readings tested against a reference device in deltaE00 color distance for those RGB primaries, D65 white and gray scale.
I wrote "I would" because as a consumer I want to see these things before buying but YMMV.


Becasue of that, I find claims made by that review not reliable at all beacuse of the missing data. Maybe a review update filling the gaps could fix those issues so I can consider it a reliable advice for choosing which colorimeter I should buy.



Also most users want that their measurement device could be upgraded to new backlight types and from Graeme Gill's findings it seems that SpyderX does not support spectral corrections in the way i1d3 does. That means no vendor or community support for new backlight types unless you buy/rent/borrow a spectrophotometer.
Thack lack of functionality will negate that claim of being close to an i1displaypro and even make an SpyderX an usleess device for most of the potential buyers of a colorimeter.
Vicent is offline  
post #5 of 29 Old 06-21-2019, 03:22 PM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by BigCoolJesus View Post
Is the peak nits known for use on HDR displays? With the soon to be released Vizio's breaking into the 2000nits territory, looking for a new affordable meter than can handle measurement accuracy that high.
DataColor, for reasons known only to them, does not publish this sort of spec. DataColor told me that the SpyderX would work with HDR, but that's a little vague. I did my own test on a HDR display and the SpyderX measured about 8.5% lower than my reference device reading of 1020 nits. I don't know what to make of this. On the one hand absolute luminance is not a very important performance spec so long as the device has adequate dynamic range. Relative luminance is much more important. Since I don't have a 2000 nit display to test it on I cannot verify that it has that dynamic range. Maybe someone who does have such a display can test it and report here.


Also, the SpyderX does not read in illuminance mode, which might matter to those with front projectors. They will have to read off the screen.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #6 of 29 Old 06-21-2019, 03:47 PM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by Vicent View Post
I find this a bold claim: "The SpyderX's accuracy is comparable to the i1Display Pro" since colorimeter accuracy (raw reading) is limited to how close they are to an standard observer in wavelengths were a display's light peaks are placed.
I mean, testing raw uncorrected accuracy with one backlight type could give you close results for two different colorimeter models (like SpyderX and i1d3) and with another backlight very different ones.
No, it might. Then again it might not. You can't know until you measure them. You have to understand, there is a limited amount of time I am willing to spend on an uncompensated review that I did as a public service. The test I did took 2-3 hours. Perhaps I can add more information later when I have more time. Of course, you are always welcome to perform the test yourself. Also, to be fully comprehensive you would need to test an entire batch of meters rather than just one. The i1Display Pro has a rather wide unit-to-unit variation. The SpyderX may or may not. I have tested only one. Testing 10 meters would add another several hours to the exercise.
Quote:
Originally Posted by Vicent View Post
Maybe I missed something but in that text/"review" I only see a WOLED, I see no information if they were raw uncorrected readings, corrected by custom matrix (so they wil match) or corrected by vendor bundled corrections.
The test data was raw, uncorrected. The reference data was based on a CR-100 colorimeter corrected with a CR-300 spectro.

Quote:
Originally Posted by Vicent View Post
Also since CIE xyY is not perceptually uniform I would like to see raw readings & vendor corrected readings tested against a reference device in deltaE00 color distance for those RGB primaries, D65 white and gray scale. I wrote "I would" because as a consumer I want to see these things before buying but YMMV.
Perceptual uniformity is not really relevant to this test, since the readings for both meters were exact replicas. However large the error was with one meter, it would be exactly the same for the other. Perceptual uniformity would allow us to weigh the importance of the reported errors, but whatever that would be it would be the same for both since the respective test conditions were identical. Also, you are free to convert the xy readings (I didn't concern myself with luminance for color error testing) to ab or uv yourself.

Quote:
Originally Posted by Vicent View Post
Also most users want that their measurement device could be upgraded to new backlight types and from Graeme Gill's findings it seems that SpyderX does not support spectral corrections in the way i1d3 does. That means no vendor or community support for new backlight types unless you buy/rent/borrow a spectrophotometer.
Thack lack of functionality will negate that claim of being close to an i1displaypro and even make an SpyderX an usleess device for most of the potential buyers of a colorimeter.
No. The meter could be pre-corrected for several display types using a reference spectro and those corrections included in the accompanying software. Your easy transition from short of optimal to "useless" is hyperbolic to say the least.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #7 of 29 Old 06-22-2019, 12:26 AM
Advanced Member
 
Paul Butler's Avatar
 
Join Date: Sep 2000
Location: The Old World
Posts: 616
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 19 Post(s)
Liked: 16
There are two models, at least in the uk; the Spyder X Pro and the higher priced Spyder X Elite. No idea what the difference between the two is, might just be the accompanying software.
Paul Butler is offline  
post #8 of 29 Old 06-22-2019, 02:18 AM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by Paul Butler View Post
There are two models, at least in the uk; the Spyder X Pro and the higher priced Spyder X Elite. No idea what the difference between the two is, might just be the accompanying software.
That's right. The only difference is the software. The meters are identical.
Paul Butler likes this.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #9 of 29 Old 06-22-2019, 02:23 AM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by Killroy View Post
Has anyone tried the new Spyder X for generating 3DLUT corrections using ArgyllCMS and DisplayCAL for use in MadVR video renderer.

I have a Spyder 5 that worked great for monitors but for projectors it seemed to wane in accuracy.
That was likely due to the S5's abysmal low-light sensitivity.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #10 of 29 Old 06-22-2019, 03:35 AM
Member
 
Join Date: Nov 2018
Posts: 27
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 2
Quote:
Originally Posted by TomHuffman View Post
No, it might. Then again it might not. You can't know until you measure them. You have to understand, there is a limited amount of time I am willing to spend on an uncompensated review that I did as a public service. The test I did took 2-3 hours. Perhaps I can add more information later when I have more time. Of course, you are always welcome to perform the test yourself. Also, to be fully comprehensive you would need to test an entire batch of meters rather than just one. The i1Display Pro has a rather wide unit-to-unit variation. The SpyderX may or may not. I have tested only one. Testing 10 meters would add another several hours to the exercise.
The test data was raw, uncorrected. The reference data was based on a CR-100 colorimeter corrected with a CR-300 spectro.

Perceptual uniformity is not really relevant to this test, since the readings for both meters were exact replicas. However large the error was with one meter, it would be exactly the same for the other. Perceptual uniformity would allow us to weigh the importance of the reported errors, but whatever that would be it would be the same for both since the respective test conditions were identical. Also, you are free to convert the xy readings (I didn't concern myself with luminance for color error testing) to ab or uv yourself.

No. The meter could be pre-corrected for several display types using a reference spectro and those corrections included in the accompanying software. Your easy transition from short of optimal to "useless" is hyperbolic to say the least.

So your review is almost useless as a buyers advice: 1 backlight type, no vendor corrections and the fact that "mid cost spectro" (xrite) will not measure properly most of those WLED PFS/Triluminos displays with default driver (10nm readings).

I mean, you are free to spend as much time you want testing those things for free, but the data you provided is almost useless for a consumer and I have the right to say it in a polite & non violent way explaing all the faults of your so called "review"
Vicent is offline  
post #11 of 29 Old 06-28-2019, 05:17 PM
Senior Member
 
nodixe's Avatar
 
Join Date: Nov 2009
Posts: 289
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 55
Quote:
Originally Posted by Vicent View Post
So your review is almost useless as a buyers advice: 1 backlight type, no vendor corrections and the fact that "mid cost spectro" (xrite) will not measure properly most of those WLED PFS/Triluminos displays with default driver (10nm readings).

I mean, you are free to spend as much time you want testing those things for free, but the data you provided is almost useless for a consumer and I have the right to say it in a polite & non violent way explaing all the faults of your so called "review"
So you admit it has some use, just not the use you want. Look dont get mad im just gonna try and be real here. Its a basic theory of social discourse/transaction called being polite or acting courteously. Not because its a rule or because I say you have to but because it is proven more likely to elicit a positive response. You came off as dismissive of toms work and demanded, as a consumer, that more testing be done. A more proper response would have been to thank him and make a request, as a forum member, to do the tests you require. This is a supportive community and people are more likely to do more work if their work feels apreciated. That's all im saying.

Sent from my SM-G935V using Tapatalk
Chad B, fhoech, chunon and 2 others like this.
nodixe is offline  
post #12 of 29 Old 06-29-2019, 02:16 PM
Member
 
Join Date: Nov 2018
Posts: 27
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 2
Although I understand your concerns about my message, this... "review" looks exactly like any of those you get if you write "SpyderX review" in google, mostly washing Datacolor's bad name and against consumer interest and without relevant data.
And I think I wrote my message in a constructive way because I wrote what is missing and would be very useful to customners, so that review can be updated.

You can think of other kind of reviews as useless as the article we can read in 1st messsage link, but not linked to Datacolor. Let's use an example.
You or I may write a review about... inter instrument variability in i1d3 family. It's just an example. If you take 5 o 7 devices purchased in different years and measure raw readings and put mean and std dev in an excel... it would be as useless as that SpyderX review.
Why? Because people using a i1d3 do not care about that (unless they were extremely bad). Why? Because EDR/CCSS corrected measurements accuracy rely on a matching between actual colorimeter behavior and behavior stored in firmware (unless colorimeter observer is far off from the standard observer you want to use). Also if you have an spectro then you make a matrix and do not care about that variability.
So if you or I wrote an "review" about i1d3 family inter instrument varability just with raw uncorrected measurements and mean and std dev it will be biased and even made wrong on purpose to put a halo of "unreliable instrument" even if such devices measure in a very accurate way out of the box with free software or vendor software or expensive 3rd party software suites.
In the same way you can dump i1d3's colorimeter behavior (colorimeter observer) from firmware data (using ArgyllCMS for example) and make the same inter instrument variability analysis ... but this time you check firmware observer of those outliers far from mean. And let's say that we have a surprise... you may find that those distant outliers share the same firmware observer or that their observer stored in firmware is so close that cannot explain that instrument variability you see when compared against a reference instrument. THAT will be an extremely useful information for customers based on actual data about that hypothetical issue in i1d3 (it's just an example of a review made in the right way, I'm not sayong that such issue is real)... but if you or I wrote only "I saw wide inter instrument variability fo i1d3" it will be USELESS and maybe somebody that knows a little of what such devices are under the hood may even think that such article is written just to mislead consumers.

I hope that I explanied why I wrote that negative comment on this thread's review, so you are no longer worried about why I wrote it in that way.
Also I'm not a native english speaker so maybe something is lost in translation or should be explained using other synonyms that does look rude to native speakers.

Last edited by Vicent; 06-29-2019 at 02:21 PM.
Vicent is offline  
post #13 of 29 Old 06-30-2019, 04:02 PM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by Vicent View Post
Although I understand your concerns about my message, this... "review" looks exactly like any of those you get if you write "SpyderX review" in google, mostly washing Datacolor's bad name and against consumer interest and without relevant data.
Other than the generally obnoxious tone of your comments, the biggest problem with what you write is this:

I clearly reviewed the SpyderX according to three criteria: color accuracy, speed, and low-light sensitivity, while pointing out that of the three color accuracy was the LEAST important because this can always be corrected by a better instrument. The other two performance criteria are uncorrectable. You then proceed to completely ignore two-thirds of the findings and then obsess about one third, the least important part.

I discovered that the SpyderX offers profoundly improved performance in speed and low-light sensitivity compared to previous versions. That's interesting and useful information. If I have some time this evening, I will post additional color accuracy measurements on different display types. I would have done it already, except I was so annoyed by your hostile reaction that I felt "What's the point?"

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #14 of 29 Old 06-30-2019, 04:09 PM
Advanced Member
 
DIY Guy's Avatar
 
Join Date: Oct 2001
Location: Ann Arbor, Michigan
Posts: 675
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 16 Post(s)
Liked: 29
Does anybody know if the SpyderX will work with the JVC auto-cal feature on their projectors?

Steve
DIY Guy is online now  
post #15 of 29 Old 06-30-2019, 05:48 PM
AVS Forum Special Member
 
Killroy's Avatar
 
Join Date: Oct 2008
Location: Ninth Circle of Hell
Posts: 2,329
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 433 Post(s)
Liked: 327
Quote:
Originally Posted by DIY Guy View Post
Does anybody know if the SpyderX will work with the JVC auto-cal feature on their projectors?
Not yet.
Killroy is offline  
post #16 of 29 Old 06-30-2019, 07:26 PM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
I have updated my original review to include additional color accuracy data and a test of repeatability, which I forgot to include in the original review.

This additional testing revealed nothing to alter my original assessment that the SpyderX is greatly improved over previous Spyder versions and in most ways competes with the i1 Display Pro.

My only reservations are that I would like to see a somewhat better low-light sensitivity and a diffuser for illuminance readings.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #17 of 29 Old 07-01-2019, 03:56 AM
Member
 
Join Date: Nov 2018
Posts: 27
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 2
Quote:
Originally Posted by TomHuffman View Post
I clearly reviewed the SpyderX according to three criteria: color accuracy, speed, and low-light sensitivity, while pointing out that of the three color accuracy was the LEAST important because this can always be corrected by a better instrument. The other two performance criteria are uncorrectable. You then proceed to completely ignore two-thirds of the findings and then obsess about one third, the least important part.
Yes, but you still miss the point. You need to have that reference instrument at your disposal: rent, buy... whatever.
With some other colorimeters you can skip that part and still get reasonable results (even better that if you have a 10nm spectro for some backlight types): you just need a spectral sample of the backlight type you are trying to read.

AFAIK SpyderX does not have that feature and that is a serious drawback for that device.


It is not uncommon to see threads in AVSFroum woth people asking tor spectral sample/corrections for their Samsung Quantum Dot display. Xrite & Datacolor do not bundle corrections for them in their suite. In the same way you can expect the same for foreseeable backlight types in near future.
Actually we see/suffer it in some vendor calibration solutions, some backlights type corrections are not available for all, or just a few apps support them for the same colorimeter model.
So one device type offers inexpensive software upgradeability, others do not.


I think that a review ingoring that fact projects a distorted image of what a customer is going to get from that device.
How do you think that such SpyderX owner is going to feel when he ask "how can I measure XXXXX backlight technology? I saw you guys are sharing some correction for them" an the answer is "Well... it does not work with your device, it's a well known limitation, so you need to rent some i1Studio/i1pro2/3".



I hope that maybe the missing feature is "hidden", that somehow SpyderX firmware stores a table with the spectral sensivities of their filters and that sooner or later Datacolor or 3rd party GNU driver is going to made that information available to calibration/profilling apps: compute a matrix correction for raw measurements on the fly when you provde an spectral sample.
Until then... it's a severe limitation and reviews should include it. If it is not included and it's posted on a forum like this... I'll will point it. Potential customers will be happy to hear about this before paying.
markmon1 likes this.

Last edited by Vicent; 07-01-2019 at 04:00 AM.
Vicent is offline  
post #18 of 29 Old 07-01-2019, 09:29 AM
Member
 
stama's Avatar
 
Join Date: Feb 2013
Posts: 75
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 71 Post(s)
Liked: 43
Tom, thank you for your test.

Your colorimeter and spectro tests are appreciated, and you're one of the few to post such tests and probe comparisons. As far as I can tell, the common recommendation to use the i1D3 for measuring OLEDs is largely due to your test of this probe a while ago.
stama is offline  
post #19 of 29 Old 07-01-2019, 01:35 PM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by Vicent View Post
Yes, but you still miss the point. You need to have that reference instrument at your disposal: rent, buy... whatever.
With some other colorimeters you can skip that part and still get reasonable results (even better that if you have a 10nm spectro for some backlight types): you just need a spectral sample of the backlight type you are trying to read.

AFAIK SpyderX does not have that feature and that is a serious drawback for that device.

It is not uncommon to see threads in AVSFroum woth people asking tor spectral sample/corrections for their Samsung Quantum Dot display. Xrite & Datacolor do not bundle corrections for them in their suite. In the same way you can expect the same for foreseeable backlight types in near future.
Actually we see/suffer it in some vendor calibration solutions, some backlights type corrections are not available for all, or just a few apps support them for the same colorimeter model.
So one device type offers inexpensive software upgradeability, others do not.

I think that a review ingoring that fact projects a distorted image of what a customer is going to get from that device.
How do you think that such SpyderX owner is going to feel when he ask "how can I measure XXXXX backlight technology? I saw you guys are sharing some correction for them" an the answer is "Well... it does not work with your device, it's a well known limitation, so you need to rent some i1Studio/i1pro2/3".

I hope that maybe the missing feature is "hidden", that somehow SpyderX firmware stores a table with the spectral sensivities of their filters and that sooner or later Datacolor or 3rd party GNU driver is going to made that information available to calibration/profilling apps: compute a matrix correction for raw measurements on the fly when you provde an spectral sample.
Until then... it's a severe limitation and reviews should include it. If it is not included and it's posted on a forum like this... I'll will point it. Potential customers will be happy to hear about this before paying.
First, as I explained in one of my previous posts, the meter itself does not have to support corrections internally. Matrix corrections for a variety of display types can be built-in to the supporting software. Our software does. Others may as well. Thus, the problem you emphasize simply does not exist so long as a buyer is willing to invest in third-party software.

Second, even without corrections, the SpyderX already gets REASONABLE results for all of the display types I tested. The highest uncorrected dE I observed was 3.6 and the highest average dE for color and white was 2.0.

Third, my intent was to test the performance of a stock, out-of-the-box meter, and not its capabilities when enhanced by third-party software.

Fourth, the capabilities of a color analyzer are the product of several criteria: color accuracy, speed, low-light sensitivity, and repeatability. Your obsessive focus one one of these criteria to the exclusion of all others as though they don't exist is more than a little mysterious.

Fifth, the description of color analyzers that do "not work with your device" applies to no existing hardware I have ever heard of. All color analyzers operate within prescribed tolerances. Even professional spectroradiometers costing tens of thousands of dollars are imperfect devices that are accurate only within a specified +- range of error. The question is not whether a meter "works" or "doesn't work" with a given display device. That is a meaningless question. The real question is does it perform within reasonable tolerances given its cost? The SpyderX not only meets this standard but represents a profound improvement over previous Spyders and offers the first reasonable competition to the i1 Display Pro at a low cost.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #20 of 29 Old 07-02-2019, 04:52 AM
Member
 
Join Date: Nov 2018
Posts: 27
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 2
Quote:
Originally Posted by TomHuffman View Post
Matrix corrections for a variety of display types can be built-in to the supporting software. Our software does. Others may as well. Thus, the problem you emphasize simply does not exist so long as a buyer is willing to invest in third-party software.
AFAIK a matrix correction for a colorimeter without knowing its spectral sensivities requires that user or a person hired by him to measure with a reference device such display. Otherwise you will suffer from inter instrument variability errors.
So the issue is real and it is not solved by 3rd party software. You need a to rent/buy/borrow a spectro and measure that display, or hire somebody to do it.


*IF* SpyderX stores somewhere in their firmware its spectral sensivities (somehow accurate) the 3rd party software trick works and it is a reasonable way to avoid inter instruent variability issues when applying a "generic correction". That's how i1d3 family works.
AFAIK Spyder X does not store that info from preliminary analysis from Graeme Gill and its driver. I'll be very happy to be wrong in that subject so those floating point spectral sensivity curves are stored in firmware and we will be able to read them in near future.



Quote:
Originally Posted by TomHuffman View Post
Third, my intent was to test the performance of a stock, out-of-the-box meter, and not its capabilities when enhanced by third-party software.
We do not agree in this part, I find necessary the other kind of test (check if spectral sensivities are stored in firmware, and if they can be retrieved by calibraction suite, vendor or 3rd party one).



Maybe other AVSForum users find that data "mandatory" or in a review too, maybe not. I pointed it because as a consumer I find this data extremely useful.
I won't label reviews without that information as useful, in the same way I will do the same with display reviews which do not test color uniformity (splitting color error an brightness error numerically so I can spot color tints across the screen).

YMMV.

Last edited by Vicent; 07-02-2019 at 05:07 AM.
Vicent is offline  
post #21 of 29 Old 07-02-2019, 01:21 PM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by Vicent View Post
AFAIK a matrix correction for a colorimeter without knowing its spectral sensivities requires that user or a person hired by him to measure with a reference device such display. Otherwise you will suffer from inter instrument variability errors.
So the issue is real and it is not solved by 3rd party software. You need a to rent/buy/borrow a spectro and measure that display, or hire somebody to do it.
No, not as long as the matrix correction was created using a display type whose colorimetric response is close to the display being calibrated.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #22 of 29 Old 07-02-2019, 11:54 PM
Member
 
Join Date: Nov 2018
Posts: 27
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 2
Quote:
Originally Posted by TomHuffman View Post
No, not as long as the matrix correction was created using a display type whose colorimetric response is close to the display being calibrated.
That is false.


A matrix corrects a particular instrument measurement (backlight spectral distrubution multiplicated by THAT colorimeter observer, then integrated over visible spectrum = electric response of sensor with some filters applied, then multiplied or scaled to get numers in your desired units) to measure as a reference instruments (which we assume that is almost equal to backlight spectral distribution of that display multiplied by some CIE standard observer integrated over visible light).
That means that such correction stores the particular behavior of THAT colorimeter (let's call it "THAT PARTICULAR colorimeter observer") for some backlight type.


As explained before, if some colorimeter model suffers from some mild to severe inter instrument variability if you apply that correction to another unit on the other side of that model spread of measurements probability distribution you are going to get an error.
Most calibration suites or thir documentation warn user about matrix corrections are not "portable" between instruments because of this.



That is why Xrite apply EDR/CCSS corrections: you store for each colorimeter unit (or ar least for each batch) its spectral sensivities so you can avoid that inter instrument variability error and at the same time use a "generic" correction for a backlight... so I' starting to think that you do not know them or how to use them, but maybe I'm wrong and you are evading this subject.

Anyway, a lot of calibration suites use spectral corrections: ArgyllCMS ecosystem, Calman, Light Illusion, most monitor vendor suites, etc... and it is good for their users that those calibration suites use them. Of course you can build a FCCM for YOUR PARTICULAR colorimeter with those suites if you have a refrenece measurement device at your disposal.


With SpyderX unless it's proven that each instrument (or batch) spectral sensivities are stored somewhere in firmware ... you need that reference instrument at your disposal if you wish to measure an unknown backlight type for every colorimeter that you want to correct. Otherwise inter instrument variability is going to add an unwanted error to corrected measurements.
If they stored such information you'll need just an spectral sample of that backlight technology offered by vendor or by some community contributor.


Spyder X may have improved low light sensivity, speed and filter degradation, whixh is good and I have no doubts your claims about it... but is a big jump backwards in terms of being "future proof" (even it is poorly "present proof" given the extremely limited of bundled corrections) because the lack of spectral corrections support.
This is a strong handicap versus competitors in its price range (i1d3 family).


So again: you are free to do not care about such things, but if you wrote a review without that information in a forum somebody is going to "complete" it with the missing information. And I'm sure that somebody will find that warning useful.
Vicent is offline  
post #23 of 29 Old 07-03-2019, 02:18 AM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by Vicent View Post
That is false.

A matrix corrects a particular instrument measurement (backlight spectral distrubution multiplicated by THAT colorimeter observer, then integrated over visible spectrum = electric response of sensor with some filters applied, then multiplied or scaled to get numers in your desired units) to measure as a reference instruments (which we assume that is almost equal to backlight spectral distribution of that display multiplied by some CIE standard observer integrated over visible light).
That means that such correction stores the particular behavior of THAT colorimeter (let's call it "THAT PARTICULAR colorimeter observer") for some backlight type.

As explained before, if some colorimeter model suffers from some mild to severe inter instrument variability if you apply that correction to another unit on the other side of that model spread of measurements probability distribution you are going to get an error.
Most calibration suites or thir documentation warn user about matrix corrections are not "portable" between instruments because of this.

That is why Xrite apply EDR/CCSS corrections: you store for each colorimeter unit (or ar least for each batch) its spectral sensivities so you can avoid that inter instrument variability error and at the same time use a "generic" correction for a backlight... so I' starting to think that you do not know them or how to use them, but maybe I'm wrong and you are evading this subject.

Anyway, a lot of calibration suites use spectral corrections: ArgyllCMS ecosystem, Calman, Light Illusion, most monitor vendor suites, etc... and it is good for their users that those calibration suites use them. Of course you can build a FCCM for YOUR PARTICULAR colorimeter with those suites if you have a refrenece measurement device at your disposal.

With SpyderX unless it's proven that each instrument (or batch) spectral sensivities are stored somewhere in firmware ... you need that reference instrument at your disposal if you wish to measure an unknown backlight type for every colorimeter that you want to correct. Otherwise inter instrument variability is going to add an unwanted error to corrected measurements.
If they stored such information you'll need just an spectral sample of that backlight technology offered by vendor or by some community contributor.

Spyder X may have improved low light sensivity, speed and filter degradation, whixh is good and I have no doubts your claims about it... but is a big jump backwards in terms of being "future proof" (even it is poorly "present proof" given the extremely limited of bundled corrections) because the lack of spectral corrections support.
This is a strong handicap versus competitors in its price range (i1d3 family).

So again: you are free to do not care about such things, but if you wrote a review without that information in a forum somebody is going to "complete" it with the missing information. And I'm sure that somebody will find that warning useful.
This is getting really irritating. You would think that you would take the trouble to know what you were talking about before speechifying with such authority. Yes, matrix calibrations are useful only if done for each individual meter instead of creating a generic correction for all. That's why we do precisely that. We have been doing for about 10 years.

You have now transitioned this from a discussion of the SpyderX to an evaluation of various software packages, at least one of which you are completely misinformed.

So, no, what I wrote was not "false."

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #24 of 29 Old 07-03-2019, 05:22 AM
Member
 
Join Date: Nov 2018
Posts: 27
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 2
Quote:
Originally Posted by TomHuffman View Post
This is getting really irritating. You would think that you would take the trouble to know what you were talking about before speechifying with such authority. Yes, matrix calibrations are useful only if done for each individual meter instead of creating a generic correction for all. That's why we do precisely that. We have been doing for about 10 years.

You have now transitioned this from a discussion of the SpyderX to an evaluation of various software packages, at least one of which you are completely misinformed.

So, no, what I wrote was not "false."

It is false sir, because I can buy a SpyderX wherever I want: on a graphics art store in my city, on amazon... etc
I'm afraid that it was you in your last message who deviates a calibration suite agnostic discussion about which features offer new Datacolor colorimeter to what you offer in your store which I would label as "offtopic" and I'm very polite naming it.



If I buy a SpyderX in amazon for example (it's just an example!) and SpyderX has not stored its spectral sensivities in firmware (or stored in some Datacolor server by unit/batch so I can download them) the only way I can make an accurate correction that avoids (in a pactical way) inter instrument variability to compute a matrix measuring on my screen with my colorimeter and a refrenece device.
That is not "software", that is a service. Of course offering such service is legitimate and it is reasonable that such service has to be paid. No one has denied it... but it is a service.


Other devices can avoid it and still offer reasonable results using spectral corrections. That is an objective drawback not related to calibration suite.
So this is "vendor agnostic" and it is not related at all with you or with what you do to earn money.



We are talking about what a SpyderX colorimeter can offer right now and in near future because a customer could be in Japan, Shanghai, EU, USA, Israel, India or Argentina and buy whatever colorimeter he wants from whatever retailer he has near him.
SpyderX limitations as a future proof meter are caused by how it was built and what Datacolor chose to (do not) offer with it.

It's the device, not the software suite.

Last edited by Vicent; 07-03-2019 at 05:30 AM.
Vicent is offline  
post #25 of 29 Old 07-06-2019, 05:02 PM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Despite several abandoned initial criticisms, the focus of your complaint in the end boils down to extolling the virtues of the i1d3’s EDR files, which provide enhanced accuracy. The i1d3 has this feature and the SpyderX doesn’t. Notice, these are just assertions. No tests, no data, just an evangelical belief in the essential nature of this technology.

So I decided to spend some time actually testing this. I chose four display types: CCFL LCD, plasma, wide gamut LED (Samsung QLED)*, and OLED. I compared measurements of these displays using the SpyderX in its optimal mode--which I found was wide gamut LED—with the i1d3 in both its default CMF (Color Matching Function) mode and targeted modes using X-Rite’s EDR files for each of these display types.

I have the X-Rite SDK, and while I cannot quote from it directly because of my NDA with X-Rite, I can comment on its features generally. It provides the EDR files, 8 in total, and a test utility that allows you to easily switch back-and-forth between the default mode and the targeted mode using the relevant EDR file.

I first measured white on each display type with the SpyderX in its most accurate mode. Second, I measured white on each display with the i1d3 using the default mode. Next, I measured white on each display with the i1d3 using the targeted EDR mode. Finally, I compared all results with the reference values from a CR-300 spectroradiometer and calculated a dE value for all in CIE94. Here are the results.


The EDR files improved the i1d3’s accuracy in most cases, except in the OLED case where it resulted in a significantly LESS accurate reading. I can only speculate as to why this is so. My guess is that the original EDR file, which as created several years ago, was based on a RGB OLED, which has considerably different characteristics than the White OLEDs used in the consumer world. The accuracy with OLED in its default mode is actually quite good, slightly better than the SpyderX.

The bottom line is that if we include OLED, the EDR files provide essentially no improved accuracy. If we eliminate OLED as an outlier, then the EDR files improve accuracy over the default mode by 1.7 dE, a meaningful but hardly profound increase in accuracy. The only case where that occurs is with Plasma (5.6 dE improvement), and that is largely because the default reading is so poor.

To sum up, the i1 Display Pro is a better meter than the SpyderX. It is also more expensive. Its low light sensitivity is a entire level of magnitude better than the SpyderX. It also provides a built-in diffuser, which in some cases is essential for calibrating front projectors. However, its color accuracy, even using EDR files, is only slightly better than the SpyderX, about 1.0 dE. If you include its OLED readings in EDR mode, it is actually LESS accurate than the SpyderX, by about 0.6 dE.

You may not care about low-light sensitivity, or speed, or repeatability, but if color accuracy is your sole focus, then there’s really not much to distinguish the i1d3 from the SpyderX. My original assessment stands. The color accuracy of these two meters is roughly the same.


* The EDR files include a White LED and a RGB LED. Evidently, the RGB LED is supposed to be used for wide gamut LED, but it didn't work very well, so I used the WLED file instead.
fhoech likes this.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #26 of 29 Old 07-09-2019, 02:06 AM
Member
 
Join Date: Nov 2018
Posts: 27
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 18 Post(s)
Liked: 2
Your table is misisng information, since Xrite provides (at least for some developers*) several EDRs for various flavors of the same panel technology.
The devil is in the details and there are at least 14 EDRs from Xrite available to developers.

Just a few examples:

-OLED: there are spectral samples for RGB LED refernce monitors and some AMOLED displays (similar spectral powers distributons... and also some software vendor have an EDR for WOLED consumer TVs. Depending on which you choose for your test displays results could or could not be good.
OLEDFamily_28Aug18.edr (RGB OLED)
FSI_XM55U_23Jan19.edr (WOLED)

(+community samples that can be obtained for free)

-Widegamut LEDs, right now RGBLED backlight is not used since last RGB Dreamcolor.
Current widegamut LED techologies are GB-LED (in older widegamuts like CS2730), WLED PFS in at least 3 flavors (AdobeRGB coverage in green, full P3 for macs, 95% P3 for multimedia displays) and QLED TVs/monitors (in limited to 8x%-9x% P3 flavor or full/almost full AdobeRGB/P3 flavor)
Also i1d3 support via community CCSS for soem backlights unsupported by Xrite like QLEDs for TVs or AUO QLED monitors, also for P3 macs.
Depending on which you choose for your test displays results could or could not be good... and my bet in L*a*b* units is that error between them is going to be in a* rather than in b* due to differences in spectral power distribution between these backlight types that you grouped in to "widegamut LED" in a extreme simplification (IMHO).
RG_Phosphor_Family_25Jul12.edr (last sample packed here is a GBLED)
NEC_64_690E_PA242W_2013-02-28.edr (a GB-LED with a slightly shorter wavelength peak in blue channel)
PFS_Phosphor_Family_31Jan17.edr
HP_DreamColor_Z24x_NewPanel.edr (full AdobeRGB WLED PFS like those new CG from Eizo)
Panasonic VVX17P051J00.edr (multimedia monitors with 95% P3)
RGBLEDFamily_07Feb11.edr (old RGB HP Dreamcolor, wrong correction used by LG or Benq software)
(+community samples that can be obtained for free for several QLEDs and P3 Mac)


-WLED correction for Xrite is a set of 3 sRGB like displays spectral power distribution packed in a EDR.
Since user can have the good ones in several commercial packages or in free software, there is no need and it is discouraged to use it for Widegmaut LEDs.
WLEDFamily_07Feb11.edr
(+community samples that can be obtained for free for several sRGB-like variants in yellow phosphor + panel)

It's very good to see an update in the good direction seeing the capabilities of SpyderX that a consumer is going use ... but comparison tables are missing important information that distorts results or even question the validity of that test (for i1d3 comparisons).
Also looking for consumer interest and needs, the availability of community corrections (CCSS) for display tecnologies NOT supported by colorimeter vendor cannot be hidden (P3 Macs, QLED TVs or QLED photography monitors with AUOptronic AHVA panels... or whatever foreseeable display tecnology).
AFAIK SpyderX is not going to be able to use them and I'm not going to repeat it again since is explained in this thread. As I wrote previosly I hope that I'm wrong here and colorimeter spectral sensibilities can be obtanied somehow from firmware or form Datacolor... but untill them a SpyderX owner is not going to be able to use them.

IDNK how many EDRs or which ones have you at your disposal, but as we can see here you are far from seeing the whole picture regarding that subject (i1d3 support for different display technologies).

*= I'll skip which calibration suites supports all of them to avoid further offtopic discussion... but if somebody needs clarification about calibration suite an set of correction I can write a detailed list.

Last edited by Vicent; 07-09-2019 at 02:13 AM.
Vicent is offline  
post #27 of 29 Old 07-09-2019, 05:27 PM
AVS Forum Special Member
 
gwgill's Avatar
 
Join Date: Jan 2013
Location: Melbourne, Australia
Posts: 1,071
Mentioned: 23 Post(s)
Tagged: 0 Thread(s)
Quoted: 332 Post(s)
Liked: 244
Thanks for the accuracy comparisons Tom - very interesting. It would also be interesting to know (but a lot of work and expense for someone to check!) how the SpyderX compares to the i1DisplayPro statistically - i.e. how consistent the filter production is between the two manufacturers. It does indicate that the AMS sensor is worthy of respect in its colorimetric credentials. For me, the lack of black thermal stability and lack of low level resolution is the main drawback to the SpyderX, and a reason to pay a little more for a ColorMunki Display or i1DisplayPro.
Rolls-Royce likes this.

Author of ArgyllCMS and ArgyllPRO ColorMeter
gwgill is offline  
post #28 of 29 Old 07-10-2019, 02:02 AM - Thread Starter
AVS Forum Special Member
 
TomHuffman's Avatar
 
Join Date: Jan 2003
Location: Springfield, MO
Posts: 7,876
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 993 Post(s)
Liked: 831
Quote:
Originally Posted by gwgill View Post
Thanks for the accuracy comparisons Tom - very interesting. It would also be interesting to know (but a lot of work and expense for someone to check!) how the SpyderX compares to the i1DisplayPro statistically - i.e. how consistent the filter production is between the two manufacturers. It does indicate that the AMS sensor is worthy of respect in its colorimetric credentials. For me, the lack of black thermal stability and lack of low level resolution is the main drawback to the SpyderX, and a reason to pay a little more for a ColorMunki Display or i1DisplayPro.
Yes, this would be interesting, but the only way to do it would be to test an entire batch of SpyderX's. I only have one on the premises.

Tom Huffman
ChromaPure Software/AccuPel Video Signal Generators
ISF/THX Calibrations
Springfield, MO

TomHuffman is offline  
post #29 of 29 Old 07-10-2019, 04:48 AM
Member
 
Join Date: Jun 2017
Posts: 24
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 25 Post(s)
Liked: 2
Anyone know if this can this be used on the JVC 540 Projector for autocal as the Spyder5 has been discontinued in Aus, and the JVC specifies the Spyder5

My 5.4.2 HT
KEF Q750 Fronts and Q600 Centre
KEF C130 Atmos in-Ceiling & KEF Ci200 In-Wall Surrounds
2 x Speedwoofer 10s
JVC 7900 Projector, Yamaha 2070 Receiver, Panasonic 420 4K Player
Terminate is offline  
Sponsored Links
Advertisement
 
Reply Display Calibration

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off