[argyllcms] Re: ArgyllCMS V.2.20 has been released
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Tue, 4 May 2021 18:01:47 +1000
Ben Goren wrote:
Hi,
Spotread works as advertised on an M1 MacBook Air with today’s OS X security
update and an i1 Pro 2.
but the real question is whether dispwin works on a recent 10.15 or 11.XX OS :-)
Any comments on these devices from the perspective of somebody who wants to
use them to measure color?
Overall my impression is that the i1Pro3 is a technical advance
on the previous instruments, and is a well made device. The physical
design is a compromise for the 4mm version though, since it's based
on the 8mm design and is significantly larger than it needs to be
for that reason. That bulk gets in the way sometimes when you are
trying to line things up.
They seem to have fixed the awful USB 5V power problem that
the i1Pro2 often has - the i1Pro3 appears to be more forgiving of
power supply drop though the USB connectors or cables.
The majority of changes are in the area of reflective measurement,
where the incandescent illumination has been replaced by LEDs. There
are 4 LEDs used, a white/phosphor one with a second carefully selected Blue
LED to fill in the gap between the phosphor and the white driver wavelengths.
The UV illumination uses two LEDs at slightly different short wavelengths
(this compares to the i1Pro2 which had only a single UV illumination LED).
I guess this is to finesse the FWA/OBE accuracy.
A drawback of LED illumination is that LED output and wavelength depends
on temperature. In the case of the ColorMunki/Studio they dealt with
this by characterizing the LED droop during calibration. With the i1Pro3
they have taken a different approach, and there is actually a spectral
sensor that is used to measure the illumination in real time (something
like an AMS AS7341), as well as what appear to be NTC temperature sensors
and some other type of LED monitoring. The output from all that is used
to drive a polynomial spectral model of the LED output. Overall
the advantage of LED is that the measurement to measurement repeatability
should be better than an incandescent source, due to the
latter having thermal and tungsten vapor effects that can be
observed in certain circumstances.
A feature of the i1Pro3 is the increased scan rate and the ability
to measure M1/M2 conditions in one pass. This is because it
interleaves UV included and excluded measurements during a scan.
So the scan rate is higher, but the combined UV excluded + UV included
rate is the same as the i1Pro2.
For emissive measurement it appears to be comparable to the i1Pro2,
although the black drift compensation may be a slight improvement.
One puzzle is the way that the difference between the 4.5mm and 8mm
aperture is handled. Everything else being equal, an 8mm aperture
should capture 3.16 times the amount of light, which would be a good
thing for low light sensitivity. But the Plus in fact has the same
underlying sensor sensitivity to a given cd/m^2 as the non-Plus model.
I don't know why that is, given that the integration time can be
more than small enough to compensate for the additional sensitivity
while retaining a very adequate maximum illumination measurement range.
They must have achieved this by either turning down the gain of the
A/D converter, or adding a neutral density filter to the Plus's optical path.
If the former, then the S/N ratio for the Plus version would be better
than the non-Plus on emission measurement, but if the latter, then it
would be about the same, and worse for reflection measurement. Without a
Plus version to play with I don't know which of these is the case.
A follow on from this is the sensitivity and scan speed when
using the polarizing filter with the Plus model. A polarizing
filter adds additional optical losses, and in the case of
this particular filter it appears to be about a factor of 3.
For the case of non-polarizing measurement the Plus appears to
simply put up with raw measurement values of 0.316 what they are
for the non-plus, but for polarized measurement it increases
the integration time by a factor of 10, aiming for full
scale sensor values. This means you have to scan very-slowly if you
want the strip to be recognized. (The ArgyllCMS driver uses
a lower integration time to make it a little more usable while
getting similar sensor values to non-polarized Plus reflection values.)
There's probably other stuff, but that's what comes to mind
at the moment. In summary, if you have an i1Pro1 or i1Pro2 and
are happy with how it works, I wouldn't be rushing out to
buy the i1Pro3 unless you need one its new features (large
aperture, polarized measurement or single scan M1/M2), but
if you are in the market for a new instrument, the the i1Pro3
is a perfectly reasonable successor to the i1Pro2.
Cheers,
Graeme Gill.
Other related posts:
