RGB-values -> f-stops?

Started Apr 23, 2006 | Discussions thread
erwinv
Regular MemberPosts: 329
Like?
Re: RGB-values -> f-stops?
In reply to Jonas B, May 6, 2006

Jonas B wrote:

Situation: I have a picture opened in Photoshop, in RGB-mode. Let's
say I convert it to greyscale so all the RGB vaslues are the same.
If I measure a dark area to say 32 and a very light area to 186...
is there a way to interpret these figures to f-stops then? In other
words: what relation do the 0-255 figures have to eachother,
expressed as f-stops?

A note to clarify: In my example above I have taken a scenic
picture and for the moment it is living inside photoshop only. It
is not printed (and I guess things get changed then as white isn't
more white than the paper is).

I'm asking since i once saw a table with RGB values related to the
zones in Adam's Zone system starting with Zone 1 at 90% and RGB
value 30 and ending in Zone X at 0% and RGB at 256. That would
cover 10 stops which make me wonder. I also read somewhere (got no
link, sorry) that one f-stop should equal to approx a jump of 50 in
RGB values. That sounds more plausible but leaves me uncertain as I
know nothing about the discussion behind the figure.

It's a logarithmic scale, with the set gamma (2.2 normally) as exponent.

the 1 stop "factor" is 1.37 (because 1.37^^2.2=2) for the next bigger stop, or 0,73 for the next smaller stop.

A difference of half a stop is the square root of these numbers, or 1.17 and 0.85

So a brighness table in f-stops, epressed in RGB values looks like
255= 100% Zone 10
186=50% 9
136=25% 8
99=12.5% 7
72 6
53 5
38 4
28 3
20 2
15 1
11 0

Don' know the details about the zone system, but though including some numbers would be usefull.

If you make blank rectangles with these RGB values, and measure them on your screen, in a dark room, with a light meter, you will find out that it's correct. (It's also a good test of monitor settings and quality)

For actual pictures however, this is only valid after a straight conversion from RAW. Of the 4 RAW converters I have tried sofar, only PWP4 does this. PPL, RSE, and picasa, and also the in camera convertor apply a contrast curve to the values while converting. The purpose of these curves is to show more contrast in the midtones then there was in reality.

The PWP4 conversion actually seems a bit less contrasty then reality. Fstop diffences created by making an exposure series are more like 1.34 factors then 1.37 factors.

The usefull dynamic range found in this test with my DS at ISO 200 was indeed atleast 10 stops, a bit more if you include slight differences in noise levels as sensitivity.

The reason I was testing this is to make batch conversion with a self made slide film like curve. I know how a slide film curve looks like, but to copy it to digital I needed to know the precise position of the "Zones"
Works pretty well, think I will buy the PWP program when the trial expires.

Erwin

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow