The New Lumagen 125-Color Calibration Technology

[realzven]

Je n'ai pas trouver de topic qui parlait de la révolution qui se prépare dans le monde de la calibration :

NEWS DE TOM HUFFMAN SUR AVS FORUM

Tom Huffman a testé le lumagen avec la nouvelle technologie "125 color calibration"

Description and Review of Lumagen's New 125-Color Calibration Technology

3 hours, 16 minutes ago

I have spent quite a bit of time with the new Lumagen color correction technology that address a 5x5x5 matrix for a total of 125 colors. This is a really big deal for affordable consumer video calibration. For a description and review, see the ChromaPure New Gear page.

Tom Huffman

NEWS DU CEDIA 2012

SpectraCal Shows Radiance Color Cube Technology at CEDIA 2012

(Indianapolis, IN) – September 5, 2012 – Lumagen, Inc., designers of the Radiance series of video processors, announced the first color cube processing for home theater at the CEDIA Expo today. The processor is being shown in the SpectraCal booth, booth 1237 at the Indiana Convention Center in Indianapolis.

The color cube will allow more precise calibration to ensure the most accurate video image.

Color cubes are also known as 3D Look-Up Tables (LUTs), although the technology is unrelated to stereoscopic (aka 3D) video. Cube technology is currently used for characterizing high-end reference monitors in professional video production and post-production. Color cube correction has never been available to the viewer at home.

“The new color cube processing in the Radiance is the first time the home user has been able to apply the same color correction technology as professionals who adjust color for a living,” said Derek Smith, founder and CTO of Seattle-based SpectraCal, Inc.

Simultaneous with the availability of color cube functionality in the Radiance series, SpectraCal is releasing software to make use of the added levels of control. SpectraCal’s CalMAN 5 automatically characterizes any video display and sets the Radiance appropriately to correct the display’s deviation from the standard.

The Radiance color cube technology allows settings at 5 points of luminance, saturation, and hue, for a 5 x 5 x 5 cube.

“The 125 points of control ensure that nonlinearities in display technologies which could not be corrected by previous video processors can now be completely addressed by the Radiance,” said Jim Peterson, Lumagen’s President.

“It’s amazing the difference in the quality of the video image before and after calibration with the Radiance 3D LUT,” said Joel Barsotti, SpectraCal’s Director of Software Development, who personally developed the Radiance implementation. “After calibration, every color is visually indisguishable from the reference standard,” Barsotti said.

“Even though the 125-point calibration is much more precise, the calibration itself is actually much faster and easier,” Smith said.

The color cube technology, still in pre-release testing, is freely available to current Radiance owners.

Une petite explication de Jim Peterson, le président de LUMAGEN: http://www.convergent-av.co.uk/forum/in ... #msg_19273

To my knowledge, all consumer level displays/projectors use the HLS color space for CMS. The HLS (or HSL if you prefer) is a non-linear color space and so is not mathematically correct for this calibration -- it's just easier to implement, which is why it is used I suspect. This means that even with a display/projector with linear response, using HLS color space means only the calibrated primary/secondary points are correct. Points in between are not correct.

The Radiance 8-point Gamut CMS uses a linear-Gamma RGB color palette. This means that if the display/projector has a linear response it can do an excellent job of calibrating CMS. However, if the display/projector has a non-linear response, such as red-push or green-push, interior points in the color cube need to be adjusted for correct response. With the 8-point Gamut, one can use Color Hue and their offset to mitigate red and green push, but this must be done by eye looking at real material since there is no easy way to automate it. It is tough to get this just right.

The new 125 point (5x5x5) linear-Gamma RGB color palette can correct for this and other non-linear responses, and should certainly give better performance than anything available in a consumer projector or display.

Note that the Radiance 11/21-point Grayscale is calibrated first. Then the 125 point CMS palette is calibrated. They are logically in series with the CMS first followed by grayscale

Il est là (pour le XD ici: http://www.lumagen.com/testindex.php?mo ... ce_updates )

Beta 090412- NEW 125 point CMS added! That's a 5x5x5 matrix for color correction and calibration. A big upgrade from our previous 8 point matrix which allowed calibration from only the corners of the color cube. The new 125 point gamut has it's calibration points located at 0, 25, 50, 75,100% of video min/maximums. The older 8 point gamut was located at digital min/maximums which did not give quite as much precision. Previous calibration values in the CMS gamut menu are interpolated to the new gamut locations so the numbers displayed in the menu will be different. This doesn't change your previously performed calibration results. Instead, the changed values represent the same calibration but from the new gamut locations. You now have the option of calibrating with 8 or 125 points. If you perform an 8 point calibration you can change to the 125 point mode and keep the previously set 8 points. The other 117 points are interpolated from the original 8 when changing from 8 to 125 point mode. With this many points to adjust, auto calibration software from ChromaPure and SpectraCal really helps. They have both added support for our new 125 point gamut with excellent results. Note: We are in the process of updating the Configuration Utility to support backing up the new 125 point gamut data to a PC via rs232. This should be available by tomorrow.
Gamma correction: Range has been extended and can now be adjusted in steps of .01.
Test Patterns: Added 2.35 letterbox geometry overscan pattern.
Bugfixes,etc: Problems with Top/Btm 3D formats to Frame Packed and Frame Sequential output was fixed. An issue of using test patterns with certain settings not restoring the input aspect was fixed. Improvement to output setup menu allows multiple changes within the command to be performed more easily. Added sending 601/709 color information needed by Darbee Darblets with firmware before 2.8.2214. Added a direct command to globally disable video on output 1, "menu 0985". Other small fixes and improvements were also made.

Quelques détails sur l'implémentation de la mise à jour , source: http://www.convergent-av.co.uk/forum/in ... #msg_19282

You do not need to wait until your next calibration to update to the 5x5x5 software.

Patrick spent a lot of time making sure old 8-point, and grayscale, calibrations yield the same calibration points (RGBCMYW, and 11/21 grayscale calibration points) using the new 5x5x5 software. Note: If you did find any issues, you could always go back to older software and your old data will still be intact.

The new 5x5x5 CMS release uses the 5x5x5 FPGA-hardware for the 8-point CMS calibration. So, the software does need to interpret the old 8-point data and apply it to the new hardware. While this means there might be a very small chance of a significant measurable difference, we do not think there will be.

That said, there is an advantage to using the new 5x5x5 software with your old 8-point calibration. The new hardware has higher precision, and the blend of Grayscale and CMS has been improved. So while you should get the same measured calibration points, the image color accuracy will improve with the new 5x5x5 software using your old 8-point calibrations.

=====

Credit where it is due: On the Lumagen side this improvement was due to Patrick's efforts for both software and hardware. While I did have some input on architecture, Patrick spent nearly a month (possibly more. I am not sure) on optimizations to allow this to fit in our FPGA, and on the actual implementation, and deserves the lion's share of your praise for the new 5x5x5 CMS.

And again thanks to both Spectracal and Chromapure on there input, software development, testing and feedback on this exciting new feature.

La méthode pour calibrer, source: http://www.convergent-av.co.uk/forum/in ... #msg_19284

Because the new 5x5x5 FPGA hardware works differently than the old 8-point calibration, there is a new order of steps for calibration.

Here is the new best practices order:

1. Set display controls to neutral and turn off all special features in the display (including dynamic-iris).
2. Set black level display controls. Use Radiance Dark-PLUGe for black (go to contrast pattern and press number 4 on remote).
3. Set white level in display being careful no color channel is crushing. Use the 2-white/2-black box contrast pattern and others.
4. Set 100% white color-of-gray in display using appropriate sized window pattern, if available.
5. Check Gamma and use Gamma factor for coarse Gamma adjust.
6. Recheck black and white and again use display controls to adjust them if Gamma factor changed them.
7. Calibrate Grayscale/Gamma with 11/21 point. Note, once any CMS entry is changed, do not tweak any grayscale entries!
8. Use CMS white point to reduce 100% white Luma only (R=G=B), if needed, such that RGBCMY have enough Luma.
9. Calibrate other CMS points.

This sequence works for both 8-point and 5x5x5 in the new software. The old procedure does NOT work properly with the new 5x5x5 software. Old 8-point CMS data does get interpreted correctly with new software as mentioned in previous post.

[EDIT] Interestingly, this new calibration sequence works with old software.

We will be updating our Tech-Tip 2 on calibration with the information soon.

source: http://www.convergent-av.co.uk/forum/in ... #msg_19285

The new Beta Software has been posted!

When you update your software, the old calibration will be accurately mapped into the new color cube. This feature has been tested by a group of beta testers.

Randy Freeman
Lumagen

MERCI A LA LUMAGEN TEAM

(reste plus qu'à tester maintenant)

les topics sur AVS forum :

- Description and Review of Lumagen's New 125-Color Calibration Technology
- Radiance 3D LUT (5x5x5 Cube) Calibration with CalMAN 5

[realzven]

SPECTRACAL

NEWS DU CEDIA 2012

SpectraCal Shows Radiance Color Cube Technology at CEDIA 2012

(Indianapolis, IN) – September 5, 2012 – Lumagen, Inc., designers of the Radiance series of video processors, announced the first color cube processing for home theater at the CEDIA Expo today. The processor is being shown in the SpectraCal booth, booth 1237 at the Indiana Convention Center in Indianapolis.

The color cube will allow more precise calibration to ensure the most accurate video image.

Color cubes are also known as 3D Look-Up Tables (LUTs), although the technology is unrelated to stereoscopic (aka 3D) video. Cube technology is currently used for characterizing high-end reference monitors in professional video production and post-production. Color cube correction has never been available to the viewer at home.

“The new color cube processing in the Radiance is the first time the home user has been able to apply the same color correction technology as professionals who adjust color for a living,” said Derek Smith, founder and CTO of Seattle-based SpectraCal, Inc.

Simultaneous with the availability of color cube functionality in the Radiance series, SpectraCal is releasing software to make use of the added levels of control. SpectraCal’s CalMAN 5 automatically characterizes any video display and sets the Radiance appropriately to correct the display’s deviation from the standard.

The Radiance color cube technology allows settings at 5 points of luminance, saturation, and hue, for a 5 x 5 x 5 cube.

“The 125 points of control ensure that nonlinearities in display technologies which could not be corrected by previous video processors can now be completely addressed by the Radiance,” said Jim Peterson, Lumagen’s President.

“It’s amazing the difference in the quality of the video image before and after calibration with the Radiance 3D LUT,” said Joel Barsotti, SpectraCal’s Director of Software Development, who personally developed the Radiance implementation. “After calibration, every color is visually indisguishable from the reference standard,” Barsotti said.

“Even though the 125-point calibration is much more precise, the calibration itself is actually much faster and easier,” Smith said.

The color cube technology, still in pre-release testing, is freely available to current Radiance owners.

NEWS FROM AVS

With CalMAN 5 the Radiance LUT calibration is done in two steps. First we run the grayscale/gamma AutoCal at 21 points to deal with gamma and RGB coupling. Second we run the Cube AutoCal at 125 points minus 16,16,16 black. The 21 point grayscale takes about 10-15 minutes and the 125 point Cube takes about 20-25 minutes. So your total time including setting up the display/projector should not be more than an hour to get the best calibration available with consumer gear.

This processes is the same for CalMAN 5 for all the 3D LUT devices we support only difference is the number of points the devices has in it's cube taking longer.

If at max luminance a primary or secondary is under luminance we reduce the maximum luminance of white independently from the gamut luminance to correct for this error.

When the cube is done, everything should be correct for luminance.

Our whole automated process takes about 30-40 minutes, once you got the basic calibration on the display (brightness, contrast, 2 point white balance).
A 21-point grayscale calibration with a C6 takes 10-15 minutes and the cube portion takes between 15-25 minutes.
The result is 125 points with an AVERAGE dE2000 <1.

Custom Gamma and targets

Yes you can use any exponent or any of our supported formulas.
We also include a custom gamma target editor, so you can create any kind of gamma target you'd like.
We can use measured black levels to compensate for high black levels.

Infact you can even target alternate gamuts, so if you wanted to calibrated for DCI/P3 gamut, or if you wanted to do AdobeRGB you could do that as well (assuming your display has a wide enough native gamut).

Our process is extremely flexible to let you calibrate it the way you want to. It's important to have this flexibility baked in, saturation changes significantly when you change gamma targets and once you've calibrated the cube, the levels may not line up well enough with the grayscale to be able to successfully modify gamma. Then even if you could, you may or may not have correct saturations.

The other side of that coin is that by default CalMAN 5 uses gamma 2.2 power curve as its AutoCal target so for those that don't need or want to learn about the complexities of gamma it’s easy just press the button. But now that CalMAN 5 is being used in many industries including post and production we had to adapt all our AutoCal code to support any type of gamma, white balance, gamut including full user created custom. The benefit of that is some of these high end features trickle down to our consumer products.

You can manually use DDC to touch it up.

But our cube autocal completes in less than 20 minutes on the radiance (in most situations). I can't imagine you'll be able to find and adjust a handful of points in less time. Managing 125 points manually is extremely cumbersome, and I wouldn't recommend it, but yeah we'll let you do it.

The look up points are discrete, our DDC controls are not. We interpolate our DDC controls using the same algorithms we use during auto cal to allow you to adjust a few points without having to adjust every point. We probably need to add a toggle to allow users to lock down all the points so you can manually tweak just a value or two. After an AutoCal session, all the points will be locked down. But even just flipping through all the points and updating the patterns and DDC control to get to the correct points is likely going to be more work than just running AutoCal.

I'm not sure what you mean by 2nd run. Creating a good cube isn't a 1-way, multplass procedure. So most of the cube is fully calibrated by the time we get to the last few points.

Also you CAN NOT do grayscale with a cube in place, the levels no longer line up, you need to do grayscale with the cube in unity, then calibrate the cube. Since the cube may be using a white limiter it may turn 100% into 97%, so now your 100% window doesn't line up with either grayscale 95% or 100%, so how do you adjust that afterwards? You don't.

Like I said above, the manual controls are there, but you're going to spend more time and likely shoot yourself in the foot if you try it. Our results with AutoCal
are very good.

des screens du 3D LUT sur Calman V5

Some examples of the DDC 3D LUT pop-out editor in CalMAN 5. The pop-out editor can be placed anywhere on the screen and used just like a remote control.

les 1er retours sur AVS Forum

As we are now released from NDA, I just wanted to post a copy of my last (as in most recent, not the very last I hope:)) post in the Calman 5 beta forum. I'm sure results will vary depending on each display and user, but as far as I'm concerned, I couldn't be happier. And those who know me know I'm not easy to please... Here we go:


Just wanted to say thanks again to Spectracal and Lumagen for this breakthrough. I've just done a new calibration with a 2.3 target (my usual target in my bat cave) and the results are nothing but spectacular on screen.

It is by far the best my rs45 has ever looked, both regarding color accuracy and dimensionality/shadow detail.

Having both the right chromaticity and gamma at every level is just spectacular, especially in scenes with low light.

I went through a series of clips I know really well, and the level of detail is simply incredible. Some colors which used to suffer from the necessary compromises in a manual calibration are fully restored.

Well done again!


Ps: I'm using a Lumagen radiance mini with the latest beta f/w, an i1pro2 as a reference meter and an i1display Pro as a field meter. I do a basic calibration before hand, especially setting the gains at 109%, but all the hard core work is done automatically in 45mn by Calman pushing 2 buttons.

Sent from my iPad using Tapatalk HD

Thanks, Joel. It works like a champ. With the revised workflow I LUTed my 65VT50 and the results are excellent. After first watching some reference content I added a Darblet between the Radiance and the plasma. 40% Darblet is now doing what 55% used to do. I'm presently watching today's Sprint Cup race I recorded and am pleased to report I'm watching the best picture I've ever seen. Kudos, guys.

[realzven]

CHROMAPURE

NEWS DE TOM HUFFMAN SUR AVS FORUM

Tom Huffman a testé le lumagen avec la nouvelle technologie "125 color calibration"

Description and Review of Lumagen's New 125-Color Calibration Technology

3 hours, 16 minutes ago

I have spent quite a bit of time with the new Lumagen color correction technology that address a 5x5x5 matrix for a total of 125 colors. This is a really big deal for affordable consumer video calibration. For a description and review, see the ChromaPure New Gear page.

Tom Huffman

HOT NEWS FROM AVS FORUM

Selon tom Huffman him-self :

ChromaPure 2.3 Announced

We will be releasing a new version of ChromaPure in a few days when the production release of the Lumagen firmware that supports the new and revolutionary 125-color calibration capability is ready.

See our News page for release notes for 2.3.

We have also released a firmware update to the AccuPel DVG-5000.

For a write-up of the new Lumagen technology and what it means for consumer color correction, see our latest New Gear article.

Précisions concernant les sondes conseillées

BTW, I should have mentioned, anyone opting for this should be using a fast and accurate meter with good low-light sensitivity, such as the Display 3 PRO, Hubble, or K-10. If you try and do this with a Chroma 5, Display 2 or i1Pro you either won't get the desired results or you may get them after the 2 hours it takes for the process to complete. I would also recommend not using a reference spectroradiometer either. Each calibration may take 600-700 readings. That's the best way I know of to wear out a shutter.

Précisions concernant le Gamma à 2.22

The rest of auto-cal is the same.

Actually, it is not advisable to set a higher standard gamma for auto-cal (and certainly not a lower). If you run a standard power law gamma much higher than 2.22, then you will get poor shadow detail. Running a higher gamma above 40% and a lower one below is fine, but that sort of fine tuning should be done independently of auto-cal. We may implement BT.1886 in the future builds, but my inclination is to leave auto-cal at 2.22, as that is a good compromise for most systems.

[realzven]

réservé

[realzven]

réservé

[realzven]

réservé

[realzven]

Mise à jour du post 1 avec création des 2 sections SpectraCal et Chromapure

[jacko]

Depuis le temps qu'on le demandait !

[realzven]

Jacko tu va tester la nouvelle calibration avec chromapure 2.3 ?

[jacko]

Certainement !

Mais avant ça mes sondes doivent repartir chez Chromapure pour un ré-étalonnage, ajout de profile ....

Comme je vais tester un 1000ES bientôt (en tous cas j'espère) les sondes ne partiront pas dans l’immédiat.
De toute façon je préfère attendre que la 2.3 soit dispo avec le firmware du Radiance.
Pour être certain qu'il n'y est pas de fâcheuse incidence de licence et de profile ...

[realzven]

Certainement !

Mais avant ça mes sondes doivent repartir chez Chromapure pour un ré-étalonnage, ajout de profile ....

Comme je vais tester un 1000ES bientôt (en tous cas j'espère) les sondes ne partiront pas dans l’immédiat.
De toute façon je préfère attendre que la 2.3 soit dispo avec le firmware du Radiance.
Pour être certain qu'il n'y est pas de fâcheuse incidence de licence et de profile ...

Vaut mieux être prudent

[realzven]

Post section Calman mis à jour :


NEWS FROM AVS

With CalMAN 5 the Radiance LUT calibration is done in two steps. First we run the grayscale/gamma AutoCal at 21 points to deal with gamma and RGB coupling. Second we run the Cube AutoCal at 125 points minus 16,16,16 black. The 21 point grayscale takes about 10-15 minutes and the 125 point Cube takes about 20-25 minutes. So your total time including setting up the display/projector should not be more than an hour to get the best calibration available with consumer gear.

This processes is the same for CalMAN 5 for all the 3D LUT devices we support only difference is the number of points the devices has in it's cube taking longer.

If at max luminance a primary or secondary is under luminance we reduce the maximum luminance of white independently from the gamut luminance to correct for this error.

When the cube is done, everything should be correct for luminance.

Our whole automated process takes about 30-40 minutes, once you got the basic calibration on the display (brightness, contrast, 2 point white balance).
A 21-point grayscale calibration with a C6 takes 10-15 minutes and the cube portion takes between 15-25 minutes.
The result is 125 points with an AVERAGE dE2000 <1.

[realzven]

Calman : Custom Gamma and targets

Yes you can use any exponent or any of our supported formulas.
We also include a custom gamma target editor, so you can create any kind of gamma target you'd like.
We can use measured black levels to compensate for high black levels.

Infact you can even target alternate gamuts, so if you wanted to calibrated for DCI/P3 gamut, or if you wanted to do AdobeRGB you could do that as well (assuming your display has a wide enough native gamut).

Our process is extremely flexible to let you calibrate it the way you want to. It's important to have this flexibility baked in, saturation changes significantly when you change gamma targets and once you've calibrated the cube, the levels may not line up well enough with the grayscale to be able to successfully modify gamma. Then even if you could, you may or may not have correct saturations.

The other side of that coin is that by default CalMAN 5 uses gamma 2.2 power curve as its AutoCal target so for those that don't need or want to learn about the complexities of gamma it’s easy just press the button. But now that CalMAN 5 is being used in many industries including post and production we had to adapt all our AutoCal code to support any type of gamma, white balance, gamut including full user created custom. The benefit of that is some of these high end features trickle down to our consumer products.

[realzven]

Calman : suite

You can manually use DDC to touch it up.

But our cube autocal completes in less than 20 minutes on the radiance (in most situations). I can't imagine you'll be able to find and adjust a handful of points in less time. Managing 125 points manually is extremely cumbersome, and I wouldn't recommend it, but yeah we'll let you do it.

The look up points are discrete, our DDC controls are not. We interpolate our DDC controls using the same algorithms we use during auto cal to allow you to adjust a few points without having to adjust every point. We probably need to add a toggle to allow users to lock down all the points so you can manually tweak just a value or two. After an AutoCal session, all the points will be locked down. But even just flipping through all the points and updating the patterns and DDC control to get to the correct points is likely going to be more work than just running AutoCal.

I'm not sure what you mean by 2nd run. Creating a good cube isn't a 1-way, multplass procedure. So most of the cube is fully calibrated by the time we get to the last few points.

Also you CAN NOT do grayscale with a cube in place, the levels no longer line up, you need to do grayscale with the cube in unity, then calibrate the cube. Since the cube may be using a white limiter it may turn 100% into 97%, so now your 100% window doesn't line up with either grayscale 95% or 100%, so how do you adjust that afterwards? You don't.

Like I said above, the manual controls are there, but you're going to spend more time and likely shoot yourself in the foot if you try it. Our results with AutoCal
are very good.

[realzven]

des screens du 3D LUT sur Calman V5

Some examples of the DDC 3D LUT pop-out editor in CalMAN 5. The pop-out editor can be placed anywhere on the screen and used just like a remote control.