Re: Averaging profiles?
Re: Averaging profiles?
- Subject: Re: Averaging profiles?
- From: John MacDonald <email@hidden>
- Date: Wed, 26 Jan 2005 10:51:57 -0500
- Organization: Dodge Color, Inc.
Hi Douglas,
I'm not sure if I understand how averaging profiles would be beneficial.
If you've taken the effort to get good profiles for, say, three rolls of
media, you have a good profile for each one of the rolls. Averaging them
would create one mediocre profile for all three, and would require an
additional step. So it wouldn't improve your quality, and it wouldn't
save time.
Maybe I don't understand what you're planning to do. Are you planning to
profile a sample of rolls, and then average them with the idea that
you won't reprofile for a while? For example, you'd profile three of a
batch of ten rolls, average them, and then use that profile for all ten?
I don't know if that would work or not. (Actually, it would probably be
up to you whether it worked or not because it depends on how much color
shift you're willing to put up with.)
John MacDonald
Dodge Color, Inc.
Douglas Rhiner wrote:
Would it make any sense to average profiles for media?
Is this possible?
For example:
We print on quite a bit of canvas.
We have a profile for our printer-media(canvas) combination.
By nature the manufacturing process of the canvas we use results in what
I call "qualitative drift" in each roll provided to us. Not a noticeable
difference to the every day person, several others in the industry
concurr with my observations.
Regardless of time & effort invilved, would it be beneficial to print
test/profiling patterns at the beginning of each roll, create a profile
and then average this with the previous profile thus creating a running
"average" profile for this media?
The "qualitative drift" is real from production run to production run.
Each manufacturer has it's own QC (quality control) percentage of drift
allowed as in any industry manufacturing industry.
We (I) would print measure targets on two (three if time allowed) of
each lot number, measure and then average the data using ProfileMaker's
MeasureTool averaging function, averaging two into one data set and then
averaging one against the first averaged two. Over two years this
allowed us to target legacy files and current files to within the same
parameters.
Profilemaker's compare function allowed us to see the drift from any lot
number to current lot numbers. It also allowed us to show our supplier
that certain lot numbers (in particular the substrates) were
unacceptable because occasionally OEM'd products change suppliers and
don't change the packaging or tell the end-users. It is obvious to
notice some substrate differences but when the dE steps way out of the
previous range you can usually assume something is up.
Quality media providers generally provide better QC, just as quality
suppliers will accept demonstrated returns based on data.
I should note that quite often for canvas and art paper the first two
wraps of any roll are often not suitable for measuring. Check the
beginning of your roll against the center for each brand and that will
help ensure stability in your profile averaging. Ageing (and
discoloration) seems to strike the beginning and ends of each roll of
some manufacturers more than others. After some time you begin to see
the differences between premium and third party economy brands.
It is well worth your time to average data from the previous lot and
compare this to a profile averaged from earlier lots. (Methods of
measure, software and hardware do play a role here too. Bear in mind
their is a certain amount of subjectivity and, for some, methodology
which go into organizing and determining which data sets to average.
Good luck...
;0)
-- joel johnstone Color Canuck (A lesser-known of the Bangkok joels)
_______________________________________________
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