[argyllcms] Sv: Re: Questions about average
- From: Semund Svelle <semund@xxxxxxxx>
- To: "argyllcms@xxxxxxxxxxxxx" <argyllcms@xxxxxxxxxxxxx>
- Date: Fri, 26 Feb 2016 22:26:28 +0000 (UTC)
Thank you for the answer.
Yes, I know I'm on the edge of what can work.
S
Den Fredag, 26. februar 2016 22.50 skrev Ben Goren <ben@xxxxxxxxxxxxxxxx>:
On Feb 26, 2016, at 2:31 PM, Semund Svelle <semund@xxxxxxxx> wrote:
Is it possible to tell average to filter out values that are located far from
other test values when creating an average of several readings?
You're looking for a median average rather than the mean average that Argyll's
tool does. I don't think the current version supports that. I don't think it'd
be too much for Graeme to implement it, but I wouldn't want to speak for him.
In the mean (heh) time, you can load the .ti3 files into your favorite
spreadsheet application and perform the math there. They're very amenable to
such manipulation...Argyll doesn't care much what sort of space-type delimiter
the file has when it reads it, and it's very consistent when writing the files.
You might also consider _why_ you're getting outliers. Are all the outliers
from a single pass? Are you more likely to get outliers from a particular hue?
And so on.
Also...unless the outlier is waaaay out there, it's unlikely to make all that
much of a difference. With color, if you're within a couple percent, you're
typically doing as well as it gets. The average of 5 readings of 100 and one of
90 is 98.3, which is still a reasonable margin of error even though the lone 90
is way out of whack. It's rare to work in an environment where that'd represent
an actual problem...and even more rare to be using equipment with that much
variation in such an environment.
Cheers,
b&
Other related posts:
- » [argyllcms] Sv: Re: Questions about average - Semund Svelle
