So I got my VideoEQ on Mondayactually a day after the beta finished it seems.
The hardware has a lot of potential, but the software is very limited. Right now we just have a command-line tool for uploading/selecting LUTs and a web-based calculator, which has not been producing good results for most people. That said, they have announced new tools, so hopefully that will change very soon.
So on Monday, I set about creating a custom LUT for my CRT. I set the CRT into its highest contrast mode (in its more accurate mode, the black level is raised and the screen is dimmed considerably)
I set it up to provide the peak white level I wanted (I go with reference levels, so 80 cd/m² for video white, 100 cd/m² for peak white) and the absolute darkest black it can do. That is to say, no light output at all.
The pre-calibration results look pretty dire, as there's no control over the greyscale other than choosing a preset (6500K) and obviously gamma is awful as a result of the brightness/contrast settings.
(note: the charts are Luminance RGB, Gamma tracking, Absolute RGB tracking and dEuv)
Spending a bit of time hand-crafting a LUT (which is nothing more than a text-file) and these are the post-calibration results:
Quite an improvement, and things look drastically better than they ever did on the CRT as it now has the ability to do pure black and bright whites while the greyscale/gamma tracking is as flat as can be.
My only concern is that right now I'm seeing posterisation with the VideoEQ in the chain, which could be the result of one or both of these things:
- While the VideoEQ processing is apparently 16-bit, and the LUT is 10-bit, if you send it an 8-bit input you only get 8-bit out. With the LUT being 10-bit, it should always output 10-bit in my opinion.
Around 60% and under, you could definitely notice a lack of precision from the 8-bit output. (as you can see, that's where the dE results start getting worse)
- Because there are no proper tools for LUT creation just now, I'm having to use a spreadsheet calculator I made to go from 12 measured points (0-100% in 10% steps and 109%) to a full LUT.
I don't think I have enough data points near black and have to work on how the low-level points are interpolated. (initially there were significant amounts of posterisation, but changing how things were calculated helped a lot)
It may simply be that I'm asking too much of the LUT to correct the CRT when it's like this as a starting point thoughespecially if it's only outputting 8-bit.
As for the hardware itself, it's a lot smaller than I was expecting, which is great.
I feel that it should have some sort of rubber feet on the bottom though so it doesn't damage the surface it's on and I would also really like it if there were holes in the bottom so that it can be wall mounted or fixed to the underside of a shelf. (like most routers have these days)
The LEDs on it are very bright, though it shouldn't be too hard to cover them up. Would have been nice if there was a hi/lo/off switch on the box though.
The buttons are awkward to press because they're so flat on the face of the unit, but hopefully you shouldn't have to change things often. (it's really designed to be hidden out of the way)
It's also difficult to know which LUT you have selected. There is only a preset/custom LED on the box, even though you have multiple presets and custom LUTs available. One LED for each LUT would be preferable so that you know exactly what is selected.
And of course more precision (if the processing is 16-bit, why isn't the LUT?) and multiple inputs would be nice
Overall I'm pretty happy with it so far, and I haven't seen what the CMS is capable of yet, but this is assuming that proper tools are going to arrive shortly. Certainly though, I can think of plenty of changes I'd want made if there were a second versionit's not perfect by any means.
Considering its price and capabilities, and how limited many displays are, it seems like it's probably a good buy for most people. I doubt the lack of precision (8-bit out with 8-bit in) will be too problematic on most displays. (because they're imprecise already)