Re: Photographer needs monitor recommendations(Eye-One/iCColor)
Re: Photographer needs monitor recommendations(Eye-One/iCColor)
- Subject: Re: Photographer needs monitor recommendations(Eye-One/iCColor)
- From: Graeme Gill <email@hidden>
- Date: Mon, 24 Nov 2003 11:22:16 +1100
Terry Wyse wrote:
>
So, YOU'RE saying more samples per patch is no better than a single
>
measurement?
What I was saying was what I said. Number of samples per patch, in itself,
is meaningless. There are a multitude of factors that will determine the
overall accuracy with which you can take readings of a devices color
characteristics. A "More readings of each patch is better" approach,
without taking into consideration all the other factors, could lead you astray.
One primary factor is likely to be the total time taken to make a reading.
Random factors such as noise in the detector, and photon noise
will tend to average out over a long reading, while the measurement itself
integrates. Sampling a greater area of printed patch seems like a good
idea too, although I can't quite figure out how this would work, given the
practical constraints. One of the main constraints is usually the time
taken to print and read samples. Driven by this, the patches are usually
made as small as possible, to fit as many as possible on a single sheet.
The size is then determined by the optical area sampled by the instrument
with a single reading. If the patch is minimal size, there is no scope to
make multiple readings of the same patch at different locations, because there
is only one location that has a clear "view" of the patch, and doesn't start
to be "contaminated" by adjacent patches.
Strip reading instruments that make multiple readings, if they are used
with charts that pack as many patches as possible into a single sheet, may
have to throw away more that 50% of their readings, since only a few will
not intersect the boundaries between patches. It is for this reason
that strip reading instruments often use elongated sample patches
in the scanning direction, and may also have an asymmetric sampling
area for an individual reading.
For certain profiling algorithms, I'd tend to favour measuring
more sample patches, slightly less accurately, because one of the limitations
of profiling accuracy currently seems to be the sampling density in
the CMYK space. 1000, or 3000 sample points doesn't go very far,
when you're searching in 4 dimensions. A profiling system that applies some
level of smoothing, will tend to reduce random sample noise in individual
readings, while improving the detail with which the device is characterized.
>
I will go on my merry way comforted by the fact that I'm getting several
>
samples per patch for free as opposed to the 1-3 samples per patch I used to
>
get. Whether that's BETTER data or not is for someone else to decide. For
>
me, ignorance is bliss!
Why endanger your peace of mind reading the colorsync list then ? :-) :-)
Graeme Gill.
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.