Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: picman on January 31, 2019, 09:40:48 am
-
Out of curiosity how do specters read patches?
For example, I'm using a ColorMunki Print, when reading the patches:
Is it one reading per patch?
A continuous reading that is averaged?
A series of readings that are averaged?
If averaged, would moving slower add more data to the average?
-
Approximately 100 samples per second more or less then averaged.
-
You asked about the ColorMunki, but just to add-in, not all spectros do the same. The i1Pro2, for example, reads 200 samples per second and averages them. So how many readings are registered and averaged for each patch depends on the speed at which you pass a spectro over the row of patches on the target. The target I use most often (X-Rite 2371 patches on 4 pages) has 29 patches per row. If I'm not paying too much attention to the speed I'm working at, I may read such a row over a period of roughly 8 seconds. Therefore each patch is getting 0.276 seconds of read time, resulting in 55 samples per patch (0.276 seconds x 200 patches per second). Judging from the profile quality I'm achieving, this seems to be fine.
-
You asked about the ColorMunki, but just to add-in, not all spectros do the same. The i1Pro2, for example, reads 200 samples per second and averages them. So how many readings are registered and averaged for each patch depends on the speed at which you pass a spectro over the row of patches on the target. The target I use most often (X-Rite 2371 patches on 4 pages) has 29 patches per row. If I'm not paying too much attention to the speed I'm working at, I may read such a row over a period of roughly 8 seconds. Therefore each patch is getting 0.276 seconds of read time, resulting in 55 samples per patch (0.276 seconds x 200 patches per second). Judging from the profile quality I'm achieving, this seems to be fine.
That's sure true. The I1Pro 2 also doesn't depend on color transitions between patches. It uses additional optical sensors that provide quadrature info so it keeps track of where it is in the row to about .5mm accuracy. You don't have to be as careful to smoothly swipe across the patches. Heck, you can even stop or reverse then continue to the end and everything still works. All it needs are the thin, black/white registration lines on both ends and the quadrature keeps track from there. Can't do that with the older I1Pro or ColorMunki.
-
Therefore each patch is getting 0.276 seconds of read time, resulting in 55 samples per patch (0.276 seconds x 200 patches per second).
Not quite. The instruments have a finite aperture size, so measurements that straddle two patches have to be thrown away. Add a margin for uncertainty (higher for heuristic patch edge recognition, smaller using the i1Pro2 zebra stripe), and the number of average patches will be lower than you estimate (maybe 15-25). Longer patches + slower scan speed = larger number of measurements averaged.
-
That's useful Graeme, thanks.
-
The OP is using a ColorMunki, and that scans at 50 times / second. But I imagine that works fine, as larger patches end up giving you a slower scan speed quite naturally.