Olympus-OM
[Top] [All Lists]

Re: [OM] (OM) Rendition of red colours by OM digital cameras

Subject: Re: [OM] (OM) Rendition of red colours by OM digital cameras
From: Chuck Norcutt <chucknorcutt@xxxxxxxxxxxxxxxx>
Date: Thu, 15 May 2014 15:32:19 -0400
Silvernose nattering.  Send real data.

Chuck Norcutt


On 5/15/2014 10:56 AM, Ken Norton wrote:
Repeat when you have a side-by-side of the same subject taken at the same
time in the same light.

To a certain extent I agree with you. But in actual practice, the
difference is quite real. If a person photographs test charts you'll
see zero significant difference. However, depending on the light, the
E-1 responds is a VERY unique manner. This is related to three primary
contributors:

1. The red sensel sensitivity is shifted more to the IR. The IR cut
filter is not quite as strong. Some flowers actually extend into the
near IR which the eye sees, but sensors usually don't.

2. The red sensel sensitivity has a secondary bump in violet.This
means that the E-1 is actually able to reproduce violet. (purple
flowers may be purple (red+blue) or violet. Most cameras will only
properly capture the purple flowers and will shift the violet ones to
blue. The E-1 fakes this by having the red sensels see the near UV
colors and then gives you a red+blue to mimic violet. This can be
tested with a prism.

3. The red sensel sensitivity doesn't extend as deeply into the
yellows. Most cameras capture yellows better than the E-1 at the
expense of the reds. This narrower spectral response (actually, it's
about the same width as the others, but is just shifted more towards
IR) means that the red channel is less likely to blow out with more
yellow content flowers. This maintains more tonal gradiation in the
highlights of the red sensels themselves.

On the flip side of all this is the green sensels have their issues.
Unlike normal sensors, the E-1 sensor MUST use a 4-pixel mix. Most raw
converters can work with the sensor data on a 3-pixel mix. No, in most
cameras, it does NOT combine two greens with one red and one blue. It
combines one green, one red and one blue. Everybody assumes 4 pixel
mix and comes up with all sorts of crazy explanations as to why there
are two greens, but if you look at the matrix and see how you can
combine data at a single pixel location, you do so with just a triad.
If you go with a quad, you lose a lot of detail--which is what has
happened with the E-1. There are many versions of raw converters, but
nearly all ones (especially the Adobe ones calibrated for use with
Nikon and Canon cameras), they actually use a triad.

However, due to the 4-pixel mix, the way the two greens have different
color/sensitivity responses, AND the added noise dithering, the
effective dynamic range of the E-1 is massive. The E-1 is about the
only camera that you can recover a 4-stop underexposed raw file and
still end up with a usable image. The anti-grain/noise freaks say that
all E-1 images are unusable, so keep this comment in mind as you read
my claim. But I've certainly used 3-stop underexposed images in
wedding albums without grief. This happens when you get a flash
misfire.

The E-1 also has two other traits worth noting:

1. Skin tones. The extended near-IR response of the red sensels tends
to lighten skin tones a little bit, however, not so much that the skin
turns waxy. This has the great benefit of smoothing the skin out a
little bit AND it also helps make zits and blemishes fade.

2. The narrower red response in the visible spectrum combined with the
two green sensels of varying response makes for what appears to be a
greater midtone tonal contrast. So, there is a natural "S-Curve"
applied to the image without an actual "S-Curve" being applied to the
image. There is a lot of mathematical explanation that I could go into
on this, but you can't argue with the results. While you can always
adjust in post to mimic a lot of this, that requires bit-bending and
any time you are bit-bending you effectively end up with a lower
bit-depth per color channel. There are a number of medium-format
digital backs that have this trait as well as the new Fujifilm X-tran
sensor. But it certainly is something that I have not seen with almost
any normal CMOS DSLR. The E-1 files, even SOOC jpegs, have better
midtone contrast than nearly any other camera.

Yes, I do miss mine.


--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz