Re: So how do we know if the colour you see is correct?
- From: Chris Malcolm <cam@xxxxxxxxxxxxxxxxx>
- Date: 10 Mar 2008 09:38:13 GMT
Colin_D <nospam@xxxxxxxxxxx> wrote:
Chris Malcolm wrote:
Wolfgang Weisselberg <ozcvgtt02@xxxxxxxxxxxxxx> wrote:
____ <internetphobic@xxxxxxxxxxxxxxx> wrote:
The problem with using white is that you don't know how white-white is,
in terms of brightness.
Which is completely irrelevant concerning colour accuracy.
It would only be of interest if you wanted to calculate a correct
ISO/exposure/aperture triple based on the amount of light reflected
from the card.
The grey card always reflects the same 18%.
Which isn't the correct setting for middle gray in photography,
anyway.
http://www.bythom.com/graycards.htm
http://www.richardhess.com/photo/18no.htm
Very interesting references! Anyone who cares to experiment with a
light meter can easily discover that 18% is a tad too dark for a
mid-reflectance gray, but I never knew why. As Hess points out in your
reference, the true figure of 12-13% is a simple consequence of the
geometry of Lambertian hemispherical reflectance.
There is a very simple explanation for using 18% grey, and why 12% is
better *for modern cameras and sensitive materials*.
It used to be that an exposure range of five stops was accepted as the
useful dynamic range (DR) of the films of the day. Another way of
saying this is to say the DR was 2^5, or 32:1, which is equivalent to
five stops. The half-way point, or more correctly the geometric mean,
is therefore 2^2.5, which is 5.65:1. Another way of doing this is to
take the square root of 32 - which is 5.65. What this says is that the
ratio of 5.65:1 is the same as the ratio of 32:5.65 so 5.65 is the
geometric mean between 1 and 32.
Now, 5.65/32*100 is - wait for it - 17.68%, rounded to 18%.
So, 18% grey is the geometric mean between 'black' and 'white' *for a DR
of five stops*, or in other words, 'white' is 32 times brighter than
'black'.
What happens if we take a DR of six stops as being a more reasonable
approach to real-life DR of the average scene, assuming that modern
sensors can handle that range (they can with ease, for the doubters!)?
As before, six stops is a DR of 2^6, or 64:1. 2^3, or sqrt(64) is 8, so
the g.mean is 8/64*100 which is - right again! - 12.5%.
If you want to go further, seven stops is 128:1, g.mean is 11.31, and
11.31/128*100 is 8.8%
So, if you are into wide dynamic range shots, you might well use an 8.8%
grey card.
I hope this clarifies why grey cards are 18% and why they should be
12.5%, and why meters in cameras are generally calibrated to 12.5%.
That's a very simple and persuasive argument! So much so that I'm
embarrassed to have previously accepted a more complex one :-)
It makes it even odder that so many modern photographers with
sophisticated accurate meters built into their cameras remain
convinced of the magic of the 18% gray card.
--
Chris Malcolm cam@xxxxxxxxxxxxxxxxxxxx DoD #205
IPAB, Informatics, JCMB, King's Buildings, Edinburgh, EH9 3JZ, UK
[http://www.dai.ed.ac.uk/homes/cam/]
.
- Follow-Ups:
- Re: So how do we know if the colour you see is correct?
- From: Floyd L. Davidson
- Re: So how do we know if the colour you see is correct?
- References:
- So how do we know if the colour you see is correct?
- From: Pete D
- Re: So how do we know if the colour you see is correct?
- From: ____
- Re: So how do we know if the colour you see is correct?
- From: Chris Malcolm
- Re: So how do we know if the colour you see is correct?
- From: ____
- Re: So how do we know if the colour you see is correct?
- From: Chris Malcolm
- Re: So how do we know if the colour you see is correct?
- From: ____
- Re: So how do we know if the colour you see is correct?
- From: Wolfgang Weisselberg
- Re: So how do we know if the colour you see is correct?
- From: Chris Malcolm
- Re: So how do we know if the colour you see is correct?
- From: Colin_D
- So how do we know if the colour you see is correct?
- Prev by Date: Sony α200 colour oddities
- Next by Date: Re: Sony a200 colour oddities
- Previous by thread: Re: So how do we know if the colour you see is correct?
- Next by thread: Re: So how do we know if the colour you see is correct?
- Index(es):
Relevant Pages
|