or Connect
AVS › AVS Forum › Display Devices › Display Calibration › ee/ColorBox, LightSpace, and 3D LUT Calibration
New Posts  All Forums:Forum Nav:

ee/ColorBox, LightSpace, and 3D LUT Calibration - Page 7

post #181 of 253
Quote:
Originally Posted by visca blaugrana View Post

that is what I suspect, but would LOVE to see some stone cold facts about this.
I seems to me that calibration is turning in to why I stopped participating in HIFI threads in general.

Blame the people who make claims and refuse to provide data to back them up. It's also very telling that the whole idea of objective assessment via color science principals in the form of deltaE is discounted on the LS website, very convenient.
post #182 of 253
Hi Ted, just to make my self more clear smile.gif

what im looking for is a un-calibrated display
then do a LS calibration then do a Calman 24 point checker
then do a Calman calibration then do a Calman 24 point checker
to make it interesting a ArgyllCMS calibration with a 24 color checker would be nice smile.gif
post #183 of 253
If LS or CalMAN would like an independent, unbiased observer to compare work product from the various solutions on my plasma I'd be happy to do so. I'm not trying to sell anything. smile.gif
post #184 of 253
It's not easy to fully answer some of the recent questions raised in recent posts, as the algorithms used within LightSpace when generating a calibration LUT from the profile data is key here, but that is not something we are going to explain in-depth, as this is what makes LightSpace work as it does... to use a recent description used by Zoyd, it is LightSpace's secret sauce!

But, Zoyd's answer above is not correct. LightSpace most definitely does NOT us a nearest neighbour local modification, and does NOT use "well-known color space transformations" at all.
ArgyllCMS is simply using standard ICC profile transforms to go from one colour space to another.
LightSpace is very different in the colour engine it uses, an no amount of inaccurate statements will change that fact.

When the LUT is made all points in the profile data are assessed and will influence the correcting for any given point - not just the reading from that point's measurement.
This is key, and is very, very unique, and requires a very high-level colour engine, which is at the heart of LightSpace.

It is the implementation of the final calibration LUT in any LUT box or DI program is the same for each and every LUT.
But it is the generation of the LUT that can be (and is, in the case of LightSpace) very different.

The colour engine within LightSpace is unique - and has no common operation with any other approach used in any other calibration system - simple fact.

A bit more info was posted here: http://www.avsforum.com/t/1461363/ee-colorbox-lightspace-and-3d-lut-calibration/120#post_23146650

Steve
Edited by Light Illusion - 4/1/13 at 6:46am
post #185 of 253
Quote:
Originally Posted by zoyd View Post

If LS or CalMAN would like an independent, unbiased observer to compare work product from the various solutions on my plasma I'd be happy to do so.

Feel free to contact Steve for a trial version of LS smile.gif Oh, I am also only a priavte user who bought LS. I won't sell anything too.
post #186 of 253
Oh - and we do not dismiss the use of Delta-E, we are just pointing out the limitations.
Though some people do seem to want to over exaggerate certain points.

The use of skin tones perfectly displays the problem with 'relying' on Delta-E to define accurate calibration.

If you can 'read' cube images that is the bets way to assess calibration accuracy.
Perform a calibration, and then perform a re-profile on the calibrated display.
The closer the re-profile generated cube is the better the 'total calibration'.
A perfect calibration will generate a perfect cube.

*Please do not expect me to post answers to each any any questions though.
I also find the post and comments made by 'the usual suspects' boring and unnecessary.

If people really want to communicate with me a personal e-mail will always solicit a reply.

Steve
post #187 of 253
Quote:
Originally Posted by visca blaugrana View Post

Hi Ted, just to make my self more clear smile.gif

what im looking for is a un-calibrated display
then do a LS calibration then do a Calman 24 point checker
then do a Calman calibration then do a Calman 24 point checker
to make it interesting a ArgyllCMS calibration with a 24 color checker would be nice smile.gif

Hi, I haven't Stored any Uncalibrated Report of my display.
Sorry but i don't have so many time to do that Software Comparison, Single Charts can't tell you the whole story or show you the difference I'm experiencing.

LightSpace is coming from profesional market, At December 2012 when we had a small group of 5 testers of LightSpace using Consumer Displays, the fact is 5/5 testers bought the LightSpace because they saw the difference... Charts were Good but the difference is beyond the Charts...
post #188 of 253
Quote:
Originally Posted by ConnecTEDDD View Post

Charts were Good but the difference is beyond the Charts...

... the differences can be seen when looking at the skin tones smile.gif

Let me quote a statement of a good friend of mine who also bought LS:
Quote:
Just one last comment - the picture improvement due to the full display characterization, compared with the quick I've done before, was really remarkable! Not least the now absolutely natural looking skin tones are impressive ...
post #189 of 253
Quote:
Originally Posted by Nudgiator View Post

Quote:
Originally Posted by ConnecTEDDD View Post

Charts were Good but the difference is beyond the Charts...

... the differences can be seen when looking at the skin tones smile.gif

Anyone here has his opinion for a software that he hasn't ever used.... I can't comment for ArgyllCMS because i haven't used it, or I can't comment for a JVC Projector because i'm not familiar with.

Check my Signature for Hardware/Software/VideEQ i own/extensive tested... My opinion comes for a lot or experiencing using all these, after toons of hours of my free time trying to get the best possible Image Quaility of my Setup.

Anyone can check how many DI/Color Related Companies are using/trusting LightSpace here

ASSIMILATE Scratch (21.000$ Software with 5.000$ annual fees)......is 'Hollywood' Industry-leading digital cinema Software....They are trusting and selling LightSpace at their webstore for Color Management Solution.

SGO Mistika industry leading high-end colour grading DI Systems are trusting LightSpace too.

Quantel world-leading content creation systems are using LightSpace also.

If all these companies from the professional market are using LightSpace for Blu-Ray Granding, then Profiling at your Home Display with the same Software will bring your image even closer to the ideal.

Professionals are not using any dE Reports/Color Checker as none of the industry standard Software (THX CineCube HD / FilmLight TruLight / LightSpace CMS) has any dE Report Tool or dE Chart.....

Here at consumer world we are hunting the dE numbers so thirsty......
post #190 of 253
Quote:
Originally Posted by Nudgiator View Post

Feel free to contact Steve for a trial version of LS smile.gif Oh, I am also only a priavte user who bought LS. I won't sell anything too.

I believe the trial version watermarks the LUT such that it would be impossible to compare the results to other software. At any rate I have already published data on ArgyllCMS showing that I can drive the mean residual errors significantly down below the limits of human perception (close to the limits of the probe itself) in a fraction of the time. I don't think there is much if any room for improvement in those results given an 8-bit source, and I just don't want people buying into this secret sauce BS. LS will give you excellent results and may have other benefits regarding flexibility of output, ease of use, etc. but improved accuracy has not been demonstrated by anyone and is highly unlikely given the quantization limitations of the source and the fact that it's just mathematical transformations we are talking about. It's the same argument some used to make in comparing the different GUI based calibration programs when they first started, and in the end if you do the math right everybody gets the same answer within the tolerances of the probe being used and the abilities of the user.
post #191 of 253
Quote:
Originally Posted by visca blaugrana View Post

Hi Ted, just to make my self more clear smile.gif

what im looking for is a un-calibrated display
then do a LS calibration then do a Calman 24 point checker
then do a Calman calibration then do a Calman 24 point checker
to make it interesting a ArgyllCMS calibration with a 24 color checker would be nice smile.gif

nobody has done a 17^3 using CM... at least I haven't seen results posted (yet)... from what I understand it would take around 20 hours, one reason why I'm not doing it...

a LS 17^3 (4913 pts) w/ i1D3 profiled off an i1Pro on a VT50 plasma WITH (an optional) 1sec delay (for pattern stabilization on plasma) and special handling for low light readings takes me exactly 5 hours... that is in Closed Loop, ergo the sw controls the pattern generator...

use a Klein and it will take just a little over 1 hour...

and again, myself and others have posted this before:

dE evaluation is very important, but do NOT (!) dismiss visual evaluation as there are factors such as image depth, clarity, "color popping" that no test chart currently can measure... and these are exactly the deciding factors why peeps and Pros choose one solution over the other once dE results of solutions such as LS, CM, CP and Argyll are all in the same range (i.e.: dE differences between results of the various solutions are < 3 and visually not noticeable) ...
post #192 of 253
AFAIK it's not available yet in CalMan. I've done the "standard" which is per CalMan 64pts, 16x6, 286 pts, with great success. Took 1hr32min with 1043 total reads. I've also done the highest setting available (16x9, 778pts, 3hr4min) but so far haven't updated my eecolor box to see the results.
I tend to be very skeptical of results that are not fact-based....
post #193 of 253
Quote:
Originally Posted by Citation4444 View Post

I tend to be very skeptical of results that are not fact-based....

Ted and also Buzz have posted charts from their final LS calibrations here in the thread. You only have to look at it wink.gif
post #194 of 253
Thread Starter 
Quote:
Originally Posted by zoyd View Post

I believe the trial version watermarks the LUT such that it would be impossible to compare the results to other software. At any rate I have already published data on ArgyllCMS showing that I can drive the mean residual errors significantly down below the limits of human perception (close to the limits of the probe itself) in a fraction of the time. I don't think there is much if any room for improvement in those results given an 8-bit source, and I just don't want people buying into this secret sauce BS. LS will give you excellent results and may have other benefits regarding flexibility of output, ease of use, etc. but improved accuracy has not been demonstrated by anyone and is highly unlikely given the quantization limitations of the source and the fact that it's just mathematical transformations we are talking about. It's the same argument some used to make in comparing the different GUI based calibration programs when they first started, and in the end if you do the math right everybody gets the same answer within the tolerances of the probe being used and the abilities of the user.

There is a bottom line that you are apparently not getting, skepticism aside. There are five LightSpace beta testers, four for the previous several months and me for the last year. We are bright people, objective, and are looking for the the best video fidelity available. We are not fan boys. We have used various software packages to produce LUTs and we unanimously agree that the picture we see is better with LightSpace. The difference is small, but it is there. As a result, we individually purchased LightSpace and are convinced it was the correct decision. You act like Steve Shaw should be exposing his own trade secrets or be condemned as a bull shi$$er. Zoyd, it's business. Furthermore, if you had any idea of some of things that have gone on behind the scenes this past year you'd be scampering off with your tail between your legs. You obviously don't, and you've never used LightSpace, and you've never compared LightSpace LUTs side by side with yours or anyone else's. The final step in any professional calibrators job is to view real content. We have. You haven't. Your dE expertise is non admissible evidence.
post #195 of 253
I think the thought here is that you should be able to express why your solution is better by some sort of objective metric.

Otherwise you tend to sound exactly like speaker cable salesman. Which I'm not saying is the case. The average user likes oversaturated colors and vibrant pop. When you use the same buzz words to describe a product professing to create accuracy instead of industry standard metrics for measuring accuracy, you will breed skepticism.
post #196 of 253
Quote:
Originally Posted by sotti View Post

The average user likes oversaturated colors and vibrant pop.

... and such users wil NEVER spend money on a Radiance, eeColorBox or only a probe tongue.gif
post #197 of 253
Beta testers buying the product would only impress me if they paid full retail price. Beta testers generally get rewarded for their efforts. Did you beta testers pay the full retail price?
post #198 of 253
That's a very easy thing: I made use of the following offer: CLICK ! ... and paid the amount of money after 10 days .... and I have never regret it smile.gif

And that offer was available for EVERY person ! I also bought SpaceMan and 2 licenses of SpaceMatch. You can see it in my signature. I am also owner of CalMAN Business Professional.
post #199 of 253
Quote:
Originally Posted by Citation4444 View Post

Beta testers buying the product would only impress me if they paid full retail price. Beta testers generally get rewarded for their efforts. Did you beta testers pay the full retail price?

He were not Beta Testers, we were final product testers to see how LightSpace performed versus other calibration systems we were using. I paid the HT version and I'm private user, which are a lot of money for one display.... but i solved all my calibration problems I had in my mind before, now I have more time To Watch Movies....
Edited by ConnecTEDDD - 4/1/13 at 5:20pm
post #200 of 253
Thread Starter 
Quote:
Originally Posted by Citation4444 View Post

Beta testers buying the product would only impress me if they paid full retail price. Beta testers generally get rewarded for their efforts. Did you beta testers pay the full retail price?

Seriously? The beta testers aren't trying to impress you, rather, inform you. What has price got to do with performance?

The only thing I pay full retail for is a beer. And just to jab your ribs, learn how to bargain. If you knew what I paid for Ultimate you'd puke.
post #201 of 253
Quote:
Originally Posted by buzzard767 View Post

Quote:
Originally Posted by Citation4444 View Post

Beta testers buying the product would only impress me if they paid full retail price. Beta testers generally get rewarded for their efforts. Did you beta testers pay the full retail price?

Seriously? The beta testers aren't trying to impress you, rather, inform you. What has price got to do with performance?

The only thing I pay full retail for is a beer. And just to jab your ribs, learn how to bargain. If you knew what I paid for Ultimate you'd puke.
A perfect example of a post that doesn't impress me, but certainly does insult me. Learn how to bargain? Seriously? I'm looking for objective data, not insults.
post #202 of 253
Thread Starter 
Quote:
Originally Posted by Citation4444 View Post

A perfect example of a post that doesn't impress me, but certainly does insult me. Learn how to bargain? Seriously? I'm looking for objective data, not insults.

The true objective data is to see with your eyes. We have. You haven't. Zoyd hasn't. We have. And we opened our wallets.
post #203 of 253
Another thing that I believe that you aware off.....

When you are fully profilling a Display using 17-Point Cube, Measuring 4913 Color Points, LightSpace creates a file with these measurement Data's.

Let's say that after you have perform this Profiling (1-5 Hours, has to do with your meter/display combination) and you are noticing after 2-3 days that a New LightSpace Version is available for Download (New version, let's say with improved color engine, for example)......

You don't have to run again that full Display Profiling again! You have to open the saved Measured Data's using the latest version you have installed and then convert to your target color space / gamma value you want in less than 10 sec.

That saves a ton of hours every time! compared with other calibration systems that you have to run again new measurements from the start with every new software version....
post #204 of 253
Quote:
Originally Posted by Citation4444 View Post

I tend to be very skeptical of results that are not fact-based....

this thread is getting useless, when people keep asking fort "facts" when there are no measurement tools for what we are experiencing.

I hope you do understand in general that color is PERCEIVED by your eyes. dE is a function that TRIES to express color error in a number, but it is far from perfect, that is why there are a gazillion dE formulas... and new come out every few years... but it is a great tool and the best we can do evaluating results based on numbers... THEN comes the visual verification, since the numbers only tell one side of the story.... you know how many times I thought I had a great greyscale with low dE (< 1) but then confirming by looking at a greyscale step pattern I very easily spotted mild or heavy contamination... and green contamination is way easier noticeable than blue, so the number could be off for blue in the GS dE chart and you still wouldn't actually SEE it that much... dE is a work-in-progress formula that is a good evaluation starting point. That's it.

READ and UNDERSTAND previous posts... if you don't believe what real USERS post here - we are not manufacturers of the solution - then this entire forum might not be for you... the only reason I waste my time posting things here on this forum so other users (such as myself) who are looking for UNBIASED information can benefit, like I have myself from the many of helpful users here...
Quote:
Originally Posted by Citation4444 View Post

Beta testers buying the product would only impress me if they paid full retail price. Beta testers generally get rewarded for their efforts. Did you beta testers pay the full retail price?

now to your point, most of the folks here that have purchased LS owned CM already (make sure you understand that)... I own the Enthusiast version, which I think is around US$299 bucks... take a wild guess why I decided to pay X (i.e.: MULTIPLE) times the amount of that for LS, if the results were not better ?

it doesn't matter "if we paid full price", what matters is that we decided to pay AGAIN, MUCH MORE for another solution... all of this has been stated in the thread already.

whatever you do is your personal decision, as with everything in life: make a decision and live with it.

And please don't come crying afterwards, because... wait for it... WE TOLD YOU SO. biggrin.gif
Edited by Iron Mike - 4/1/13 at 7:24pm
post #205 of 253
Quote:
Originally Posted by buzzard767 View Post

The true objective data is to see with your eyes. We have. You haven't. Zoyd hasn't. We have. And we opened our wallets.

Hi Buzz - Let's just say that given my technical expertise on how this processing works that I am skeptical of the "seeing is believing" argument when used to claim a solution is more accurate in the absence of hard data. But, I have an open mind so could you perform the same A/B comparison using an Argyll generated LUT with a profiled D3 w/1500 OFPS patches and give us your impression of it's performance?
post #206 of 253
Quote:
Originally Posted by Light Illusion View Post

ArgyllCMS is simply using standard ICC profile transforms to go from one colour space to another.

Hmm. There is little that is simple about creating high quality color profiles from real world measurement data, and then linking them to create accurate color transformations.
Quote:
When the LUT is made all points in the profile data are assessed and will influence the correcting for any given point - not just the reading from that point's measurement.
This is key, and is very, very unique, and requires a very high-level colour engine, which is at the heart of LightSpace.
The same can be said for many commonly used color device models, from the simple matrix shaper models all the way up to more locally influenced but globally aware algorithms that ArgyllCMS uses to create cLUT profiles, so it is hard to understand the claim that such a property is unique. I have my doubts that it is even desirable.
post #207 of 253
Thread Starter 
Quote:
Originally Posted by zoyd View Post

Hi Buzz - Let's just say that given my technical expertise on how this processing works that I am skeptical of the "seeing is believing" argument when used to claim a solution is more accurate in the absence of hard data. But, I have an open mind so could you perform the same A/B comparison using an Argyll generated LUT with a profiled D3 w/1500 OFPS patches and give us your impression of it's performance?

Are you asking me to learn another software program?

Both LS and CM have gotten considerably better over the past few months. I'm running a CM Cube Generator at its maximum capability presently and will immediately follow that with a LS 17 point display characterization. Both meter profiles (i1Pro and K10-A) are less than a week old and should be accurate enough. The reason I purchased LS in the first place was that I could indeed see a difference, and as I have stated many times, what is easy to detect in still images is next to impossible in moving video.

The Cube Generator will produce a 65x LUT so I'll be able to load both of them into an eeBox and switch back and forth. Even doing that produces an A/B difficulty because when the LUTs are switched the progression is LUTx, Unity, LUTy and viewing the Unity LUT in between makes it much more difficult to "see". It would be so much easier if the two LUTs could be viewed simultaneously on a split screen. I'll let you know how it goes. Honesty is one of my strongest attributes. smile.gif
post #208 of 253
Quote:
Originally Posted by buzzard767 View Post

Are you asking me to learn another software program?

yes. I know you enjoy chasing down every option and tweak available. We just figured out how to implement an Argyll cLUT on the ee box a week ago and although the image looks stunning to me and measures just as well, I don't have the other software to do my own comparisons and see what visual differences there might be.
Quote:

Both LS and CM have gotten considerably better over the past few months. I'm running a CM Cube Generator at its maximum capability presently and will immediately follow that with a LS 17 point display characterization. Both meter profiles (i1Pro and K10-A) are less than a week old and should be accurate enough. The reason I purchased LS in the first place was that I could indeed see a difference, and as I have stated many times, what is easy to detect in still images is next to impossible in moving video.

So what is the benefit (other than psychological) if one can not detect a difference between the two LUTs in moving images?
Quote:

The Cube Generator will produce a 65x LUT so I'll be able to load both of them into an eeBox and switch back and forth. Even doing that produces an A/B difficulty because when the LUTs are switched the progression is LUTx, Unity, LUTy and viewing the Unity LUT in between makes it much more difficult to "see". It would be so much easier if the two LUTs could be viewed simultaneously on a split screen. I'll let you know how it goes.

I know, I have the same problem with the intervening native image when doing these comparisons. I try to focus on one particular small portion of the image with well defined gradients and "hold" that in memory while the box cycles through the native image. It's another reason why I tend to favor meter assessment, especially when the differences are this small. But I do not agree with your statement that your eyes are "the true objective measurement" for this type of calibration, a well calibrated meter is much better at this if your goal is to reproduce a standard.

Your eyes are very good at picking out relative anomalies and pretty good at matching memory colors, which is why they are important in a final calibration assessment when you have to make sure that a control adjustment using test patterns has not compromised the display performance when viewing real video or when making trade-offs with relatively large errors on a display with limited controls. But generating cLUTs is not a typical calibration exercise and at the level of residual errors these boxes produce with any of the software, I have a hard time believing that your eyes are of any use what-so-ever, again speaking strictly from the point of view of reproducing a standard.

If the argument is that reproducing a standard, with measurable tolerances below the level of human detection, is somehow "not accurate enough", then we have a fundamental difference on what calibration means and I don't think any amount of discussion would resolve that.
Quote:

Honesty is one of my strongest attributes. smile.gif

That's why I asked. smile.gif
post #209 of 253
Quote:
Originally Posted by zoyd View Post

We just figured out how to implement an Argyll cLUT on the ee box a week ago and although the image looks stunning to me and measures just as well, I don't have the other software to do my own comparisons and see what visual differences there might be.

Let me know how to do a 17^3 with Argyll and export the LUT in the correct ee format, and I'll do some comparison... been using Argyll and DispCal for some of our Notebooks and results were better than basICColor and a few other solutions, but - to be fair - all these other solutions (incl. basICColor) do not support meter profiling (Argyll does), so they were somewhat limited in accuracy...

- M
post #210 of 253
Thread Starter 
Quote:
Originally Posted by Iron Mike View Post

Let me know how to do a 17^3 with Argyll and export the LUT in the correct ee format, and I'll do some comparison... been using Argyll and DispCal for some of our Notebooks and results were better than basICColor and a few other solutions, but - to be fair - all these other solutions (incl. basICColor) do not support meter profiling (Argyll does), so they were somewhat limited in accuracy...

- M

zoyd & I are talking. It looks like I can do an i1Pro only profile with both A and LS but I can't get on it until the weekend.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › ee/ColorBox, LightSpace, and 3D LUT Calibration