AVS Forum banner

1 - 20 of 856 Posts

·
Registered
Joined
·
8,928 Posts
Discussion Starter #1


This thread is dedicated for discussion about the upcoming ColourSpace CMS software, which will take display calibration, colour management, and colour workflows to the next level.

ColourSpace expands on LightSpace CMS as the de-facto standard for advanced colour management within the global film, TV, and post-production industries where colour control is not just critical, but requires facility-wide management from a simple to use system that is display format and workflow agnostic.

The background to ColourSpace is some 5 years old, with the equivalent of an unprecedented 15+ years man-hours development time dedicated to re-imagining the needs and requirements for next generation display calibration, colour management and colour workflows.

While there are similarities in some areas of the look and feel of ColourSpace when compared to the existing LightSpace CMS, only the Gen3 Colour engine is the same, with the rest of the system being totally new, built from the ground-up, with no code similarity.



3D Graphs: Unique 3D interactive graphs, enable accuracy assessment in ways previously impossible, with dE tangent lines defining individual point errors, while all graphs can effectively be infinitely zoomed and scrolled.



Interactivity: Speed, combined with advanced 3D graphic charts, unique error tangent lines, and data point colour coding results in a level of user interactivity unmatched in any alternative calibration system.



Building on the unmatched capabilities of Light Illusion's existing calibration system - LightSpace CMS - ColourSpace is a ground-up, and totally new product development that far exceeds the capabilities of any previous or present calibration systems.

ColourSpace enables unprecedented and previously unachievable levels of accuracy and data reporting for display calibration, colour management, and colour workflows within the professional Film, Post-Production, & Broadcast industries, as well as for Display Manufacturers and Home Cinema enthusiasts.

ColourSpace CMS will become the new de-facto standard for advanced colour management, guaranteeing the highest possible level of display calibration and colour control, regardless of final application.



The most visually impressive capabilities within ColourSpace CMS revolve around the unique graphics capabilities, combined with interactive data manipulation, with 3D CIE, XYZ, and normalised Colour Space Graphs, including Error Tangent lines and colour coded measure points.

New user defined patch colour sequences have been added for manual TV CMS adjustments, allowing any patch sequences to be imported for either manual or automated measurements.

Release time-scales have not been fixed, and although LightIllusion is hoping for sooner rather than later, the last 10% of product development always seems to take an inordinately long time, compared to the previous 90% development...

During this finalisation time LightIllusion is happy to answer any question!
 

·
Premium Member
Joined
·
1,177 Posts
Looks like LightSpace on visual steroids! Bring it on! Looking at the development page, it adds several features not available in LS, at least the version level I have. Maybe you can expand on the new features?
 

·
aka jfinnie
Joined
·
4,252 Posts
3D Graphs: Unique 3D interactive graphs, enable accuracy assessment in ways previously impossible, with dE tangent lines defining individual point errors, while all graphs can effectively be infinitely zoomed and scrolled.
Really looking forward to playing with this when it becomes available. I've been trying to get my head around the performance of my JVC projector's internal processing; its calibration tools allow you to create custom gamuts for example for DCIP3 or REC709. I can see there is something mildly entertaining going on judging by what happens if you create a 3DLUT for one of these profiles, but I've currently got no good way to explore the profile, can only do it implicitly by first generating a LUT. At first look in the "traditional" 2D charts you'd look at them and think it looks pretty OK...
 

·
Registered
Joined
·
2,734 Posts
Will it support the Quantum Data 804 and 780 family of PGs? I use an 804a and have a 780 as a backup. CalMAN triggers HDR output on them, but HDR can’t be triggered manually with their normal touchscreen settings.


Sent from my iPhone using Tapatalk
 

·
Premium Member
Joined
·
9,840 Posts
Will there be an upgrade path for Lightspace users? I was hoping this more modern interface was coming in a new main release of the software, not in a new product. I'm not very happy if I have to buy another software just to get the interface that Lightspace was missing and see outstanding issues resolved (assuming the issues outstanding in Lightspace are indeed solved in this new product).
 

·
Registered
Joined
·
2,201 Posts
All information on ColourSpace, including the reasons for its development, and future options, should be within our website ColourSpace page.

https://www.lightillusion.com/colourspace.html

As we finalise development we will keep the page updated.
And if info is missing, shout, and we'll add it.

The initial release will likely just have existing hardware support.
But, future updates will add more and more capabilities.

Edited to add:
This most definitely is not about a more modern interface, although that is very different in ColourSpace.
The LightSpace code base just can't be changed like this.
We really had to start again - from scratch.
And we have been at this for some years!

Steve
 

·
Registered
Joined
·
976 Posts
Wow! It looks promising! Is there any integration with TVs, monitors?
 

·
Registered
Joined
·
2,201 Posts
One of the new developments for ColourSpace is actually also being retro-fitted to LightSpace.
This is a new 'Remote Control' capability, with a 'Secondary Execution' function too.

With these functions anyone can connect a third party program to LightSpace/ColourSpace to enable features/capabilities that are not natively within LightSpace, but would require a lot of the capabilities of LightSpace/ColourSpace to be developed within the third party program if it were not for the remote control option.

A good example is where a third party program needs probe and patch generation capability, which LightSpace/ColourSpace can provide, while the third party program offers DDC type connections/adjustments to different TVs.

We will be releasing a Beta version of LightSpace with this 'Remote Control' and 'Secondary Execution' capability to third party program developers some time next week - hopefully!

Steve
 

·
aka jfinnie
Joined
·
4,252 Posts
My hopes for a few tweaks in this new SW which slowly bug me about LS, but I live with and accept:

"Traditional" absolute gamma chart with standard scaling showing reference gamma and measured.
Ability to show gamut sweep targets on the traditional CIE charts as well as zooming these charts
Ability to "fold in" the drift compensation so that you can look at the charts with the benefit of the drift compensation applied to the profile (on a drifting display most of the profile charts get messy)
More intuitive ways to control pattern generation when driving manually (eg type W25 for 25% white, R10 for red 10%, Y50 for red 50% G 50%, etc).
Less forms that block other forms from even being moved when open (eg manage colourspace prevents interaction with display calibration)
Cancel option for LUT generation and maybe just some kind of indication of progress (not even completion, just so we can get an idea of what is going on)
When exploring a measured profile, show the target xyY for the selected patch in the selected colourspace (at the moment we can just see the result reported in various different ways). Maybe add option to show the targets on the CIE chart.

I'm sure there are others...
 

·
Premium Member
Joined
·
9,840 Posts
My hopes for a few tweaks in this new SW which slowly bug me about LS, but I live with and accept:

"Traditional" absolute gamma chart with standard scaling showing reference gamma and measured.
Ability to show gamut sweep targets on the traditional CIE charts as well as zooming these charts
Ability to "fold in" the drift compensation so that you can look at the charts with the benefit of the drift compensation applied to the profile (on a drifting display most of the profile charts get messy)
More intuitive ways to control pattern generation when driving manually (eg type W25 for 25% white, R10 for red 10%, Y50 for red 50% G 50%, etc).
Less forms that block other forms from even being moved when open (eg manage colourspace prevents interaction with display calibration)
Cancel option for LUT generation and maybe just some kind of indication of progress (not even completion, just so we can get an idea of what is going on)
When exploring a measured profile, show the target xyY for the selected patch in the selected colourspace (at the moment we can just see the result reported in various different ways). Maybe add option to show the targets on the CIE chart.

I'm sure there are others...
There *are* others :):

- Fix the 3D LUT engine so that when the native gamut of the display is undersaturated, or when the target is wider than the display capability, there is no need to manually create a custom gamut and use the LUT Concatenation feature to avoid posterization. It's more and more common to have to create LUTs for a target wider than the display, for example BT2020 or even P3 on some displays, and LS can't handle this well. How difficult is it to automatically create a custom gamut based on the measured primaries and concatenate the LUT if any of the primaries are undersaturated, or at least provide the option to do so?
- Greyscale/RGB balance at 10% and below is worse after the 3D LUT than before, at least here.

Better support for MadVR, which includes:
- replace the old .dll from 2015 that causes an invalid header when a P3 3D LUT is uploaded to MadVR using MadTPG (this was fixed months ago in MadVR but inexplicably is still not fixed in LS, yet it's a simple .dll swap in the installation package. At the moment the *only* way to get a correct DCI-P3 3D LUT with LS is to manually copy the fixed .dll after installation.
- allow to enable/disable the selected 3D LUT in MadVR without having to disconnect the patterns in LS or uploading a NULL LUT, which isn't always desirable if we want to verify, not replace an existing LUT. This is not only clumsy, it's also error-prone as there is no way to know if the 3D LUT is enabled or not when taking measurements.
- support properly 3D LUT levels in MadVR. MadVR always needs video levels in 3D LUTs, and LS is the only software that makes it necessary to manually apply a videoscale filter after the creation of each single LUT, which is bonkers. Failing this, offer an option to apply automatically this filter to each LUT created, to minimize the risk of error and to improve workflow. After I reported this issue, Steve added a note on the website to warn users that they have to manually apply the correct filter after each LUT creation otherwise the LUT created is invalid, but still no way for the user to deal with this automatically.
- offer tabs so that we can save different measurements of the same set (gamut, greyscale, etc) for example to measure before/after or different presets. This is really missing in LS compared to Calman.

All the above has been reported to Steve in July (and earlier for the interface improvements), and still no fix in LS. I understand that they are busy working on the new software, but as a user of LS I find it hard to suffer all the issues with the MadVR implementation. Apparently I'm the only one suffering from these issues, which is supposed to explain why they are not getting fixed... Is there no one else using LS to create DCI-P3 LUTs with MadVR? Or checking their results when the wrong levels are applied? Or wanting an easy way to know if the 3D LUT is enabled, and have a simple way to enable/disable it? Or check for posterization when targeting a gamut wider than the native gamut of the display? :confused:

I'm certainly hoping that all these issues (and others) will be fixed in Colourspace, given that the two software share the same LUT engine, and that the trade up offer will be good, otherwise I don't see any reason to upgrade as a MadVR user.
 

·
Registered
Joined
·
976 Posts
Will LS still receive updates or will it be discontinued?
 

·
aka jfinnie
Joined
·
4,252 Posts
There *are* others :):

- Fix the 3D LUT engine so that when the native gamut of the display is undersaturated, or when the target is wider than the display capability, there is no need to manually create a custom gamut and use the LUT Concatenation feature to avoid posterization. It's more and more common to have to create LUTs for a target wider than the display, for example BT2020 or even P3 on some displays, and LS can't handle this well. How difficult is it to automatically create a custom gamut based on the measured primaries and concatenate the LUT if any of the primaries are undersaturated, or at least provide the option to do so?
- Greyscale/RGB balance at 10% and below is worse after the 3D LUT than before, at least here.

Better support for MadVR, which includes:
- replace the old .dll from 2015 that causes an invalid header when a P3 3D LUT is uploaded to MadVR using MadTPG (this was fixed months ago in MadVR but inexplicably is still not fixed in LS, yet it's a simple .dll swap in the installation package. At the moment the *only* way to get a correct DCI-P3 3D LUT with LS is to manually copy the fixed .dll after installation.
- allow to enable/disable the selected 3D LUT in MadVR without having to disconnect the patterns in LS or uploading a NULL LUT, which isn't always desirable if we want to verify, not replace an existing LUT. This is not only clumsy, it's also error-prone as there is no way to know if the 3D LUT is enabled or not when taking measurements.
- support properly 3D LUT levels in MadVR. MadVR always needs video levels in 3D LUTs, and LS is the only software that makes it necessary to manually apply a videoscale filter after the creation of each single LUT, which is bonkers. Failing this, offer an option to apply automatically this filter to each LUT created, to minimize the risk of error and to improve workflow. After I reported this issue, Steve added a note on the website to warn users that they have to manually apply the correct filter after each LUT creation otherwise the LUT created is invalid, but still no way for the user to deal with this automatically.
- offer tabs so that we can save different measurements of the same set (gamut, greyscale, etc) for example to measure before/after or different presets. This is really missing in LS compared to Calman.

All the above has been reported to Steve in July (and earlier for the interface improvements), and still no fix in LS. I understand that they are busy working on the new software, but as a user of LS I find it hard to suffer all the issues with the MadVR implementation. Apparently I'm the only one suffering from these issues, which is supposed to explain why they are not getting fixed... Is there no one else using LS to create DCI-P3 LUTs with MadVR? Or checking their results when the wrong levels are applied? Or wanting an easy way to know if the 3D LUT is enabled, and have a simple way to enable/disable it? Or check for posterization when targeting a gamut wider than the native gamut of the display? :confused:

I'm certainly hoping that all these issues (and others) will be fixed in Colourspace, given that the two software share the same LUT engine, and that the trade up offer will be good, otherwise I don't see any reason to upgrade as a MadVR user.
I think many of those are "just" bugs(!) to be prioritised. I was thinking of stuff unlikely to ever change in LS and perhaps could perhaps sneak in now a new product is happening.

Admittedly none of the MarVR ones would make it to my love-list ;) but the behaviour of the LUT engine with smaller than target gamut displays is particularly annoying as it gets quite wonky around any LUT points corresponding to out-of-display-gamut areas; it's really hard to see any logic in what happens (it's almost as if the points bounce back off the cube edge, instead of getting clipped to the outer edge of the RGB cube), leaving the only option to use Fit Space LUT. Are you using the LUT concatenation to grab the better greyscale out of the fit or map LUT, and then concatenating that with the colour portion of the fit LUT? That's the only workflow that seems to work (ish) for these displays for me.
 

·
Registered
Joined
·
2,201 Posts
With a small gamut display the colour engine in LS functions exactly as it is designed to do.
It maintains colour accuracy with the colours that 'can' be corrected accurately.
Any other approach is just wrong.
It will not sacrifice colour accuracy for a larger, inaccurate, gamut.

Steve
 

·
Registered
Joined
·
4,910 Posts
There *are* others :):

- Fix the 3D LUT engine so that when the native gamut of the display is undersaturated, or when the target is wider than the display capability, there is no need to manually create a custom gamut and use the LUT Concatenation feature to avoid posterization. It's more and more common to have to create LUTs for a target wider than the display, for example BT2020 or even P3 on some displays, and LS can't handle this well. How difficult is it to automatically create a custom gamut based on the measured primaries and concatenate the LUT if any of the primaries are undersaturated, or at least provide the option to do so?
- Greyscale/RGB balance at 10% and below is worse after the 3D LUT than before, at least here.
This may be off-topic, but have you tried LS' new Augmented LUT generation? Although I'm not trying to shrink DCI-P3 into a much smaller gamut, one of my TVs, an old CRT rear-projector, has a gamut in which Blue is fine, but Red and Green are both undersaturated and outside the gamut triangle for Rec. 709. Using the Augmentation procedure has resulted in a visual improvement in darker area detail as well as better grayscale chart performance.
 

·
Premium Member
Joined
·
9,840 Posts
With a small gamut display the colour engine in LS functions exactly as it is designed to do.
It maintains colour accuracy with the colours that 'can' be corrected accurately.
Any other approach is just wrong.
It will not sacrifice colour accuracy for a larger, inaccurate, gamut.

Steve
The current LUT engine is causing posterization when the target is wider than the native gamut of the display.

The current workaround to this issue is to manually spot when the primaries are undersaturated, manually create a custom gamut, then manually concatenate the LUTs.

This is not acceptable. We often have to use 3D LUTs with a target wider than the display. BT2020 is the most common example, as no consumer display reaches BT2020, yet it's the target we have to use for all consumer sources. Only MadVR or the Radiance allow to use DCI-P3 as a target. And even then, some displays can't reach 100% of DCI-P3, either because of native gamut limitations, or because they prefer to privilege brightness over color.

The current workaround isn't acceptable. This is why I'm using DisplayCAL at the moment to generate my 3D LUTs. I'm not going to go through a dozen of steps for every LUT generated, given that I already have to manually apply the videoscale filter as well as otherwise the MadVR LUTs generated by Lightspace are invalid.

This isn't the case using DisplayCAL, from the same LS profile, so this is clearly an LS issue, not a profiling or meter issue.

DisplayCAL generates a valid LUT (with video levels) that doesn't have posterization. In one step.

Also I really don't see how resolving posterization is sacrificing color accuracy. You prefer to have LUTs that cause posterization because that's more "accurate"?

You need to fix this, or at least to give the user the option to get something less "accurate" if that solves the posterization issue. :rolleyes:
 

·
Premium Member
Joined
·
9,840 Posts
This may be off-topic, but have you tried LS' new Augmented LUT generation? Although I'm not trying to shrink DCI-P3 into a much smaller gamut, one of my TVs, an old CRT rear-projector, has a gamut in which Blue is fine, but Red and Green are both undersaturated and outside the gamut triangle for Rec. 709. Using the Augmentation procedure has resulted in a visual improvement in darker area detail as well as better grayscale chart performance.
Yes, doesn't work.
 
1 - 20 of 856 Posts
Top