RE: Do I need to upgrade to i1Profiler?
RE: Do I need to upgrade to i1Profiler?
- Subject: RE: Do I need to upgrade to i1Profiler?
- From: Ethan Hansen <email@hidden>
- Date: Thu, 07 Apr 2011 16:51:29 -0700
Andrew Rodney wrote:
> > But that is the problem: those swatches were meant for cameras and
> not printers.
>
> Not really. Yes, the ColorChecker can be used for making a DNG profile.
> But the target, its colors and manufacturer go back way farther than
> even ICC color management! Its been a visual reference for film for
> decades.
The ColorChecker dates from the mid-70's. It was designed to mimic how a set
of colors were reproduced on film. The patches exhibit low levels of
metamerism, making the chart useful under a variety of lighting conditions.
The chart included with the shipping version of i1Profiler has circular
holes punched in the middle of each patch. It offers an OK means of visually
comparing profiled output to theoretical input.
> > Also, for the "smoothness" slider: I feel that you may get a
> different result if you test printed a black-to-white gradation
> instead of solid swatches. I would imagine the purpose of that slider
> is to improve the reproduction of smooth gradients: seeing any delta E
> change when messing with that slider would defeat the whole purpose!
>
> So I can build two profiles using both options, make a print using
> each and you see no difference on the output. There is a spectral
> gradient in the test image and it t looks identical with both
> profiles. The differences with the two settings and profiles is simply
invisible!
[SNIP]
> The bottom line was visually and numerically, the slider does next to
> anything.
Our in-house testing confirms Andrew's findings. Starting with a reasonably
linearized printer, be it Epson inkjet, Silver-Halide, or a Vutek, the
measured difference in a patch set printed through a profile made with
"smoothness" set to 0 or 100 was negligible.
I experimented with this slider using different datasets. Thankfully,
i1Profiler can be coerced into reading old, MeasureTool-generated CGATS
files. I built profiles using measurements from a long-gone Epson 1280 in
NCA mode - one of the least-linear printers around. Softproofs showed the
smoothness slider making some difference. I then intentionally screwed up
the K-linearization on an Epson 7890 and repeated Andrew's experiment. This
time the smoothness slider did something. There was more of a visual
discontinuity in prints made with "smoothness" set to 0 than 100. The
measured data reflected this; I saw a maximum DE-2000 of 5. Given that
difference, X-Rite's choice to make an adjustment slider with 100 increments
is a curious one.
Regarding the original question of whether i1Profiler is worth the
upgrade... well... I'm not sure. The profile quality is closer to Profiler
than PMP, but you gain the ability to use arbitrary measurement charts to
better handle problem areas. Also, the full spectral data are used, rather
than just LAB values in Profiler. Profiler's scum-dot problems are also a
thing of the past. All this leads to profile performance that is, visually,
usually an improvement on what PMP or Profiler produced. The improvements
are not huge, and given the time that X-Rite has spent on the project, are
smaller than one might have hoped for.
Only the i1Pro and i1iSis are supported for measurement. You can import
spectral data from MeasureTool without a problem. If you used ColorPort,
pray that you had data scaling set to 0-1 rather than 0-100. i1Profiler
can't read 0-100 data. The software is also optimized for the new XRGA
calibration standard. Legacy data won't be scaled appropriately, and profile
accuracy suffers both measurably and visually. You can convert data using
ColorPort, but this is a painful process. We have also seen weird results
with measurements made with a polarizing filter. Some of the profiles made
by i1Profiler have nasty discontinuities, while the same dataset processed
through PMP or Profiler comes out fine.
My take on i1Profiler is that it is targeted at people wanting a more
accurate and powerful product than the ColorMunki. This is likely fine for
hobbyists and people with an inkjet or two to profile. The guided workflow
will be helpful if you only use the software occasionally. For the rest of
us, i1Profiler falls woefully short in usability and features. Gone are
batch processing capability, the ability to easily specify where to save
data and profiles, any semblance of profile editing, and all the data
analysis and processing functions of MeasureTool.
i1Profiler defaults to a wholly new set of file formats, making use of your
existing targets problematic at best. With enough fiddling with paper and
patch sizes, it is possible to come close. The txf format chart definition
files used by i1Profiler appear to assume fixed values for some parameters,
unlike the older definition files. You can manually edit a few settings, but
others are not present. On the plus side, i1Profiler uses a new set of
instrument drivers that, on the iSis at least, produce shorter measurement
times. The downside of this is that we see fairly frequent chart read
errors, where previous X-Rite software reads the same charts without issues.
All to many of the settings in i1Profiler revert back to default values each
time the software is launched. Most - but not all - of the settings can be
saved, but it requires loading them manually on each page of the workflow.
Settings are saved in obscure locations in the file system, making it
difficult to synchronize between computers. Simple items such as a "Browse"
button to choose a location and file name to save a profile in are not
present. The software helpfully offers to save any profiles in the system
and user folders. Fine if you are a one-man shop with a single computer, not
so helpful if your outfit uses multiple machines.
After all the development time, i1Profiler still comes across as a Beta
product. The overall profile quality is an improvement on both PMP5 and
Profiler, but the user interface and workflow are disasters. Omitting a
profile editor, the ability to manually specify K-Gen curves, no data
analysis, ColorPicker, etc. all make i1Profiler appear to be something less
than an upgrade to those of us who paid all that money for PMP and Profiler.
Ethan Hansen
Dry Creek Photo
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden