AVS Forum banner

2281 - 2300 of 2655 Posts

·
Registered
Joined
·
2,200 Posts
No, Hint is not performing probe matching, as the profiles are normalised.
That means the profile that Hint is being used on will set the primary RGB values.
The master/guide profile fills in the missing Volumetric info.

Steve
 

·
Registered
Joined
·
976 Posts
Let me do an example just to see if I understand how the hint feature works: Fabio and I got the same TV. Fabio did a 21^3 profile with a Klein K10A on his TV. I could use his profile as a hint for my TV by only run a QP primaries+secondaries. Would my results be as near as possible to his?
Also, I feel like this hint feature will be very useful to adjust old calibration. For example, I could use the profile I ran 18 months ago on my Kuro and by just doing a QP I would adjust the LUT to the actual panel “situation”. Right?
 

·
Registered
Joined
·
2,200 Posts
Yep, that is basically the concept behind Hint.
But, your calibration will use your measures data to set the overall colour settings.
Fabio's 12^3 profile will 'direct' the volumetric definition/shape of your LUT, not the overall colours.
To have a 'closer' match to Fabio's TV calibration you would have to match your probe to his.

Steve
 

·
Registered
Joined
·
2,200 Posts
Thanks Steve.
Also, does this mean you could effectively avoid the meter profiling step? I mean, could you generate the large hint profile without having corrected the meter, and then apply the meter correction just by measuring RGBW in the hint process?
We've been thinking about this, and with LG's concept of a 3-Matrix Calibration for WOLEDs we do indeed think that Hint (with some adjustment) could be used for an advanced Probe Matching process, and provide a far more accurate method for probe matching than the existing FCCM or LG's 3-Matrix idea, especially for any non-additive displays and volumetrically non-linear displays, which WOLEDs obviously are.
(That does ignore the requirement for the Spectro to not suffer Metameric issues with WOLEDs, which is a whole other issue...)

We're also working on a new method to effectively extract different profile data from a large volumetric profile to enable different ares of the profile to be treated differently - which will be another real help with difficult displays, and when excessive changes are needed, such as with HDR.

Steve
 

·
Registered
Joined
·
8,913 Posts
LightSpace CMS 10.0.0.2968 (25 July 2019) Beta is available.




Release Notes

This Beta has a new 'Hybrid' mode in the Convert Colour Space options (Peak Chroma, etc).

This process will potentially take a longer processing time than Peak Chroma, but will do two things with a profile that is volumetric (cube data, not a Quick Profile).

It will replace all the LUT data below 5% of the LUT range with data extracted via a process that re-interpolates bad profile data.

This should deal with bad probe readings below 5%, as well as improve the final grey scale response.

(We are using 5% or LUT range, rather than profile range as that changes with the display gamma.)

It will also replace the grey scale in the final LUT with a grey scale processed in isolation of the rest of the profile data.

This means the Grey Scale will not be influences by volumetric color data, that may be ‘bad’.

With the normal Peak Chroma/Peak Luma processes, all points in a LUT are treated equally.

With volumetric profile data, that means the grey scale is calibrated to be as accurate as any other point in the LUT.

However, if there are bad colour readings within the profile, they can also affect the Grey Scale calibration.

With the new Hybrid process Grey Scale is treated separately to the volumetric colours.

LightSpace has always been more accurate than any other calibration system with respect to Volumetric calibration.

But, the grey scale can sometimes be slightly less accurate.

This new Hybrid approach should provide the best of both.

Download

As is usual, for that beta download link, please contact me here only (no PM's): http://www.displaycalibrations.com/contact_us.html
 

·
Registered
Joined
·
588 Posts
Great new feature! Is the hybrid mode also useful for users with a K10 or any other reference colorimeters? This could replace the 1D LUT substraction and I would like to know how this works when there are no bad data points under 5% does LightSpace do nothing then and just separate the grayscale values from the volumetric data and treats them separately?
 

·
Registered
Joined
·
2,200 Posts
The new Hybrid mode will work with any profile that has volumetric info.
The probe used is not really relevant.
It will also replace the profile data regardless of the 'accuracy', or otherwise, of the data.

Steve
 

·
Registered
Joined
·
588 Posts
LG released the final firmware 5.10.03 for the 2018 models so I decided to give the hybrid mode a chance and I can say that im pretty amazed. Here is the setup I used:

LG 65 C8 with 5.10.03 FW
Klein K10-A
DVDO TPG @ 4:2:2 12bit
LightSpace 10.0.0.2969 BETA
21^3 Cube with Finseq Patchset

The first step was to install the latest firmware for the TV if you didn't receive it OTA the you can download the file from the LG official website put it on the thumb drive and update manually.

http://gscs-b2c.lge.com/downloadFile?fileId=3pIgx3J18zpknk9q6eAMg

After a firmware update I always use IN-STOP of the service menu to be sure that everything is erased and no custom LUT is active anymore so you can start from the scratch but be careful because the UTT counter will be erased also!

Then you create a QP I used Gray oy large and also Aug Data for the 1D LUT. At this step you will use peak chroma for colorspace conversion.

After LUT upload via device control you will create the profile with volumetric data. In the colorspace conversion tab you will choose hybrid.



In the past I substracted the 1D LUT from the 3D LUT to improve the overall grayscale performance but that's not necessary anymore.

Here are my results with the approach of substraction. As you can see the average CCT is with ~6600k a little bit to cool because of the blue is dominating the RGB balance.





And here is the result of the new hybrid mode nearly perfect 6500k and the almost perfect RGB balance the colorcheker performance improves also. This is a huge step for more accurate grayscale performance well done!
Would like to use the results from other user also especially with a id3.



 

·
Registered
Joined
·
6,624 Posts
LG released the final firmware 5.10.03 for the 2018 models so I decided to give the hybrid mode a chance and I can say that im pretty amazed. Here is the setup I used:

LG 65 C8 with 5.10.03 FW
Klein K10-A
DVDO TPG @ 4:2:2 12bit
LightSpace 10.0.0.2969 BETA
21^3 Cube with Finseq Patchset

The first step was to install the latest firmware for the TV if you didn't receive it OTA the you can download the file from the LG official website put it on the thumb drive and update manually.

http://gscs-b2c.lge.com/downloadFile?fileId=3pIgx3J18zpknk9q6eAMg

After a firmware update I always use IN-STOP of the service menu to be sure that everything is erased and no custom LUT is active anymore so you can start from the scratch but be careful because the UTT counter will be erased also!

Then you create a QP I used Gray oy large and also Aug Data for the 1D LUT. At this step you will use peak chroma for colorspace conversion.

After LUT upload via device control you will create the profile with volumetric data. In the colorspace conversion tab you will choose hybrid.



In the past I substracted the 1D LUT from the 3D LUT to improve the overall grayscale performance but that's not necessary anymore.

Here are my results with the approach of substraction. As you can see the average CCT is with ~6600k a little bit to cool because of the blue is dominating the RGB balance.





And here is the result of the new hybrid mode nearly perfect 6500k and the almost perfect RGB balance the colorcheker performance improves also. This is a huge step for more accurate grayscale performance well done!
Would like to use the results from other user also especially with a id3.



What I find interesting is even with this new method you still have that elevated brown to deltaE of around 2 in the CMS scan. I've noticed this and elevated orange on both the Sony and the LGs.
 

·
Registered
Joined
·
38 Posts
Hi guys,
I have a question for people who have C8 from LG and update to latest software. Is it worth to do it or should we stay on .20?
Thx
 

·
Registered
Joined
·
397 Posts
@ BJ did you have a look at near black ire 0-5 with this new 05.10.03 software,havent had a chance to re-cal myself after update.
 

·
Registered
Joined
·
588 Posts
@ BJ did you have a look at near black ire 0-5 with this new 05.10.03 software,havent had a chance to re-cal myself after update.


They fixed the black crush near black performance is is comparable to the old 4.10.31 but with all benefits of the new dethtering.

Unfortunately I have no scan of the 6 point near black workflow.
 
  • Like
Reactions: T( )( )L

·
Registered
Joined
·
976 Posts
What I find interesting is even with this new method you still have that elevated brown to deltaE of around 2 in the CMS scan. I've noticed this and elevated orange on both the Sony and the LGs.
Maybe because the hybrid function treats differently only GS?
@BlackJoker since the 3DLUT is before the 1DLUT in the processor video pipeline and new hybrid function seems to don’t require the previous workflow to achieve excellent GS, have you tried to create a directly 3DLUT after resetting 1D/3DLUT and then put a unity lut as 1DLUT? Just curious...
 

·
Registered
Joined
·
8,913 Posts
since the 3DLUT is before the 1DLUT in the processor video pipeline

LG 2018 OLED image processing chain seems to be:



HDMI Input Signal -> Contrast / Brightness (Color + Tint also) -> 3D LUT -> 1D LUT -> 3x3 Matrix - 1D LUT -> 1D LUT -> OLED Light -> Display Panel.

There 3x1D LUTs (1024-Point) one for each Red, Green and Blue channel. (Red 1D LUT Box of the video chain diagram)

After the 3D LUT (we can upload), there is a De-Gamma 1D LUT (un-does the default gamma by applying reverse gamma values) followed by a 3x3 Matrix (for color transform) followed by a Re-Gamma 1D LUT (it will re-apply the default gamma) followed by the 1D LUT (we can upload).

The 3x3 Matrix can be used for linear scale color transformation between different colorspaces. Its a simple mathematical formula that can reposition, rotate, and scale, so describing the global size/position of the gamut, but not manage the internal (volumetric) content in a non-linear fashion.

De-Gamma LUT, 3x3 Matrix and Re-Gamma LUT are disabled by loading UNITY values during the Reset LUT procedure of LG Template with DeviceControl.

Also OLED Light is an OLED Panel gain control and is separate from the video signal path.

According to that way the 1D LUT is implemented in the LG OLED TV, you must do the 1D LUT first, and then the 3D LUT profiling. If you change the 1D LUT you must re-do the 3D LUT. But if you change the 3D LUT only, you don't need to re-do the 1D LUT.
 

·
Registered
Joined
·
976 Posts
Why?
 

·
Registered
Joined
·
8,913 Posts
Because if you perform 3D LUT first (which is including '1D LUT inside' of 33-Point Grayscale), and you will perform after loading the 3D LUT the 1D LUT profiling (which is 1024-Point), the adjustments of 1D LUT will un-align some thousand colors which are already been corrected from 3D LUT table.

For the same reason we say to don't adjust any display calibration control (grayscale for example) when you are using an external LUT Box like eeColor, when you have loaded a correction 3D LUT table, as the internal display controls will work like a 'Output 1D LUT', the same as its happening with LG, where you use that Output 1D LUT for White Balance/Gamma.
 

·
Registered
Joined
·
976 Posts
Yes Master, but if you read more accurately my post, I suggested to leave a unity LUT as 1DLUT.
 

·
Registered
Joined
·
2,200 Posts
What I find interesting is even with this new method you still have that elevated brown to deltaE of around 2 in the CMS scan. I've noticed this and elevated orange on both the Sony and the LGs.
Where is that colour in volumetric space?
It looks close to the gamut edge, so is likely a limitation in the display's gamut?

Steve
 

·
Registered
Joined
·
6,624 Posts
Where is that colour in volumetric space?
It looks close to the gamut edge, so is likely a limitation in the display's gamut?

Steve
I don't know. Just pointing this out. Maybe someone can test and see where it is, etc..
 
2281 - 2300 of 2655 Posts
Top