[argyllcms] Re: Culling outliers from .ti3 input data before camera profiling
- From: Kamil Tresnak <kamil.tresnak@xxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Tue, 10 Jul 2012 23:41:27 +0200
On Tue, Jul 10, 2012 at 11:10 PM, Ben Goren <ben@xxxxxxxxxxxxxxxx> wrote:
> On 2012-07-10, at 6:27 AM, edmund ronald wrote:
>
>> If you really want a good camera profile, Image Engineering sell some
>> devices which can quickly and precisely do the job with the help of a bunch
>> of narrowband filters.
>
> ...at a cost of €9750, which is $12,000 at today's exchange rates. I'm sure
> it's an awesome tool, but you don't need to spend that kind of money to get a
> good camera profile.
Without doubt; this was interesting but a bit impractical advice .. :)
>
> Hint: your local home improvement store should be able to custom match all
> the (classic) ColorChecker colors. And that'll be a spectral match, too --
> the resulting paint will have the same SPD curve as the ColorChecker patches.
Really? Have you measured?
But even if "yes", spectral match is IMO not sufficient, surface
properties are important too. Your idea with custom colours in chart
is interesting, but we have experimented with some custom charts, and
we have big problems particulary with black patch, which give us very
weak black in resulting profile - far worse than we expect according
to measurement. So, i think, making our own - quality - chart is not
as easy.
All the best
Kamil
Other related posts:
