Add your own Correlation spectral files to ColourSpace

This thread has been edited as we have moved away from the proprietary EDR format, and now use an open .csv based Correlation file format.
The latest ColourSpace includes a WOLED Cor. File as an example.
Any user can now generate their own correlation files, as per the website User Guide.

Note: Correlation files should not be relied on for calibration accuracy, as they apply a relative correction to the i1D3's underlying factory calibration, not an absolute correction, meaning the underlying accuracy of the i1D3 is not improved. This can be seen specifically with Y measurements.
The use of Probe Matching to a Spectro is always recommended, using ColourSpace's unique probe matching capabilities.
Always use SPD2Matrix to verify the suitability of any Corr. Files before use!

Steve

Not 100% sure about the format:
Manual says (with in bold my correction) : "Correlation files are simple .csv files, with a minimum of 3 lines, with each line being one of 401 floating point values.
The values need to be 380 to 780 nm in 1 nm increments, in Watts per Nm."

So each csv line has nm and Watt, for a max of 401 floating point values. Correct?

You just need three - ideally 'near' RGB primary, and near nominal peak to prevent clipping - sets of such data, but in lines/rows, NOT in columns.
(Your 'correction' in bold is wrong.)
The provided WOLED_Example Cor. file has 4x lines/rows of 4 sets of such data - WRGB.
(The WRGB example values are approx. 100 nits based, and approx. Rec709 gamut to prevent excessive gamut compression due to the additional white pixel...)

Steve

I see .... so basically three (RGB) or four (WRGB) rows, depending on the display type. I guess the row order is important, i.e. RGBW should not work ...

No, the row order is actually not important, but if doing just three RGB sets of values, its is logical.
And, you can have a row for any colour you want...
RGB is just the logical three - with W added for a WOLED, but you would get a result without W.
So you could do RGBCMY, or ANY other combination of ANY colours.
How well the correlation would work if you had three colours that were close together, would be questionable...
Especially as there is absolutely no guarantee regarding the accuracy of using Corr. Files.
(Check SPD2Matrix discussions.)

Steve

Most standard spectral files for the EODIS3 contain one or more RGBW or WRGB measurements, i.e. 4 to 20 lines.
I have just converted all the standard spectral files to CSV format for the new ColourSpace version and Ted will be releasing the package shortly.
Peter