Types of dust spots in images

There are around a dozen dust spots in this 5D image. I'm wondering
why some are irregular while others appear very round. It's almost
like they are being imaged off different surfaces and that the round
ones are diffraction patterns of spots too small to actually resolve
as actual dust particles.
-Rich

http://dc.watch.impress.co.jp/cda/parts/image_for_link/31744-2346-14-1.html

 

I don't understand Japanese but I understand a lot about digital images
and their faults. I examined this image and another offered from the
same site by someone in digital.slr a few days ago.

This image (and the other one) has very clear evidence of the problems
which might face 5D owners. Chromatic Aberrations and purple fringing
are the bane of digital photographers. The more tightly packed the
sensor, the more likely to get the faults.

Sigma and Tokina both produce "designed for digital" lenses. The Tokina
effort produces pretty poor quality images when used on a 35mm film
camera but good to excellent when used on a DSLR. I haven't used the
Sigmas. But having the light rays fall at 90 degrees to the sensor can
only benefit DSLR images.

Seeing as the world at large is yet to pass judgment on the 5D, I guess
we'll have to wait and see how good (bad?) the sensor design is at
preventing purple fringing. These early photos are certainly not good news.

 

Maybe they're tiny pieces of plastic?

 

Dust motes are tiny pieces of many different substances that flake off by a
fracture failure of the top layers of the (usually) crystalline surface.
There's absolutely no reason to expect any uniformity of size or shape.

That said, there are, as you suggest, many different surfaces inside a
camera from which unwanted images can form ranging from the surface of the
sensor itself to the front, rear, and even internal surfaces of the lens,
diaphragm blades, etc.

Norm

 

A couple of questions about this image.
1)Are you sure it was taken with a 5D, the name of the image is
12-24-Kodak-standard.jpg why?
2) the image was taken with a 12-24 f4.5-5.6 lens at 12mm, this is a
Sigma lens not a Canon.
Could this have been taken with a Kodak DCS-N or C? Name of the file
seems to suggest that, would explain a lot of the aberations.
Check Luminos Landscape on the latest user (not quite)review of the 5D.

Tom

 

The file name is odd but the EXIF data says Canon 5D, conversion using Canon
DPP and pixel count 4368 x 2912, which fits with the 5D sensor.

Mike Bernstein

 

Checked the EXIF, was the Canon 5D. The Luminos Landscape article
mentioned that he had to clean the sensor more than other cameras, and
this sensor sure could have used a cleaning.
What was the thing about CA in this pic, took it up to 300% and we
could hardly find any, and that could have been atributed to jpeg
compression.

 

Well, a REAL (TM) photographer would use T-grain dust motes, so they
would be uniform size and shape.

 

Not even Canon would put much plastic in a $3500 camera body, if they
wanted it to sell, that is.
-Rich

 

Don't count on it!