For many different and very sensible reasons we have always had a maximum LUT Size within ColourSpace of 33^3.
In the real world of calibration there is rarely any benefit for anything larger.
And it can actually be detrimental if the underlying display is particularly unstable.

However, there is actually no fixed LUT Size limit in ColourSpace
And with the relatively recent release of Multi-Threading for the higher license levels the option of having larger LUT Sizes makes a bit more sense.

The below 3D LUT is a native 128^3 LUT directly generated within ColourSpace.
With Gamut Mapping disabled the processing time on an 8 core PC was a matter of 10s of seconds.
With Gamut Mapping enabled the processing time took minutes, but in single digits.

Less powerful PCs will obviously be significantly slower...

Additionally, the profile was a default 21^3 profile, and larger profiles will have an impact on the processing speed.
But, the same profile also generated a 256^3 LUT without too great an increase in processing times with Gamut Mapping disabled.

So, we have decided we will release the ability for the users of higher ColourSpace license levels to define greater LUT sizes directly within ColourSpace.
Lookout for this in an upcoming ColourSpace Update release.

Steve

Native ColourSpace 128^3 LUT

And here is a native 256^3 LUT, generated from the same profile, just to show that working too.
With such large LUTs there will be memory restrictions, and working with more than one LUT at a time will likely cause issues.
You have been warned!

I also generated the same LUT with Gamut Mapping enabled - it took AGES.
And I mean ages!
I gave up timing the process after a couple of hours.
But the LUT was generated, as can be seen here!

Steve

Native ColourSpace 256^3 LUT

256^3 with Gamut Mapping

I have just tested on a Windows Surface 3 laptop with just 4 CPU cores, and the same 128^3 LUT also processed very quickly, taking only about 50% longer.
With Gamut Mapping enabled the time taken was significantly longer, which is not unexpected as Gamut Mapping is a single core process.
But a result was indeed generated after about 45 mins.
It will be very interesting to see how this ColourSpace development works in the real world.

Steve

In response to multiple email questions we have been sent, we have added more information to the website on the new LUT Generation Size option being added to ColourSpace.

The bulk of the new information can be found in the Advanced Operation page, in a new LUT Size section.

One of the main questions we have been asked is what impact this has on those displays/devices that have the ability to have a 1D LUT separate from a 3D, with the 1D having more steps than the 3D.
That info has been added to the above Advanced Operation link.

The short answer is this is potentially a HUGE change to ColourSpace.

Steve

We now have an Alpha release of ColourSpace with the new LUT Size capabilities available for early testing .
Send me and email to request early access.

Steve

I had the chance to test the new Alpha build of ColourSpace with the expanded LUT Size feature — and I have to say, the results are excellent.

I recently received my Valerion VisionMaster Max from the Kickstarter campaign, so this was the perfect opportunity to test the new LUT functionality together with my madVR Envy Extreme.

For my tests, I used a 65³ LUT size, since anything larger doesn't really make sense in practical terms — and indeed, the 1D LUT benefits noticeably from the increased grid resolution.

I used one of my custom patch sets, which includes a much denser grayscale sampling, and the overall results look fantastic.
No crashes or bugs encountered during my session.

System setup:
· Windows 11 running inside Parallels Desktop on a MacBook Pro (M4 Pro chip)
· LUT generation at 65³ took less than 10 seconds

I've attached a few screenshots from today's session for those interested.

What was the pre-cal. profile like?
I ask as I would have expected far better calibration grey scale based on our testing.

Steve

The pre profile looked better in terms of the RGB balance see attached. I guess I have to do some more tests.

Seeing just the Balance is not really a help.
Need to see the actual profile to see what is going on.
Including Drift, etc.

Steve

Hi all,
With small profiles (for example, simple 13p), is there theoretically an advantage to increasing the LUT size above the standard?

If not.... in your opinion, from what profile size does it make sense to also increase the LUT size (33^3)?

Thanks

It is not that simple.
Issues such as display/probe stability, as well as probe low-light capabilities, all impact the profile patch count needed.
You really will need to test and evaluate different profiles to see what results you get.

Steve

Sorry, maybe I expressed myself poorly.
My question isn't about "real-world" conditions and issues.
I wanted to know if, in a theoretically perfect condition (therefore free of the problems that Steve has highlighted) it is mathematically (or technically) legitimate to expect a superior result.

thank you