Re: New Image
- From: Mark <lurchl@xxxxxxxxxxxxx>
- To: leica@xxxxxxxxxxxxx
- Date: Mon, 14 Jan 2008 23:11:52 -0800
Jim:
Thanks for the heads up on David Pogue's presentation. I've enjoyed
his columns in Mac World and elsewhere for years.
I agree on real data being better than interpolated. That may be more
of an issue in 'available darkness', with a 12-bit camera. I like the
low-light results with my EOS 1D Mark II and its 8.2uM pixel size,
the same as the 5D's.
The 1D Mark III has 10.1MP, 7.2uM pixels, and 14-bit resolution per
color. But DIGIC III post-processing would be the only reason I'd
consider one. I'm more likely to sell off my 20D and pick up a 40D to
replace it. I have good results from the 20D in adequate light, and
that's how I'd use the 40D.
That said, for wedding and other event photography, I'll always want
an M-camera with 35mm f/2, 50mm f/1, and 90mm lenses. There's nothing
like 'em when I need to be unobtrusive.
Mark
Mark Bohrer
Wildlife Photography on the Urban Edge
www.mountain-and-desert.com
At 07:01 PM 1/14/2008, you wrote:
Mark,
You might be interested in this since it is local. I belong to ASMP
which is why I get these notices.
http://asmpnorcal.org/drupal/?q=node/405
And I personally refuse to fall into the interpolation trap. Making
smaller pixels means fewer electrons stored per pixel means less
data for that point. The only way for small pixel sensors to compete
with large pixel sensors is to interpolate (create) the missing
data. Some algorithms are good but it is still not better than
having the actual data from the scene to use. This is why MF digital
backs produce such stellar images. Nine microns per pixel. An
enormous amount of image data per pixel.
:-)
Jim
At 04:18 PM 1/14/2008 -0800, Mark wrote:
Some numbers:
EOS 5D pixel count: 13.3MP
EOS 5D effective pixels: 12.7MP
EOS 5D sensor size: 35.8 x 23.9mm
EOS 5D pixel size: 8.2um (square)
EOS 5D post-processing firmware: DIGIC II
EOS 40D pixel count: 10.5MP
EOS 40D effective pixels: 10.1MP
EOS 40D sensor size: 22.2 x 14.8mm
EOS 40D pixel size: 5.7um (square)
EOS 40D post-processing firmware: DIGIC III
In formal and informal reviews, DIGIC III has shown noise at least
one ISO-stop better than DIGIC II. That is, a DIGIC III camera
gives you the same noise at twice ISO1 as a DIGIC II camera does at
ISO1. That's a significant advantage. Some of that may be due to
the 40D's improved pixel microlens design over older cameras like the 5D.
The 40D offers 14-bits for each R, G and B color channel, versus
the 5D's 12 bits. That gives finer color gradation in RAW files,
which give you nice adjustment options in Photoshop CS-3's 16-bit commands.
Going by area, an APS-sized chunk of the 5D's sensor gives you 38.4
% of the total. That comes out to 4.877 effective megapixels,
versus 10.1effective MP using all of the 40D's APS-sized sensor.
I used the first EOS 1D several years ago. It had a 4.1MP sensor.
The resolution difference between it and even a 10D's 6MP made me
start using the 10D all the time instead. 4-ish MP just wasn't enough.
The difference between Minox frames and 35mm frames is pretty
extreme. I don't think that's a very good comparison. And 1/2 frame
35mm cameras gave you half the frame of the same piece of film,
with that film's inherent grain and resolution limitations. So
that's not quite a valid comparison either when the smaller sensor
has more resolution.
The biggest difference is in DIGIC III's noise reduction versus
DIGIC II. It's like getting a stop of push-processing with no
increase in grain.
So it comes down to horses for courses. If you shoot wildlife and
compose tightly in the finder, the 40D makes sense. If you use
wides a lot, crop your shots, and/or shoot in good light, the 5D makes sense.
Using an APS-sensored camera on a regular basis accustoms you to
'seeing' with one. So you can still have all the habitat around the
animal to choose from. Cropping an APS-sized frame may make picture
resolution suffer, if you don't have enough pixels to start with
and want a 16 X 20 or larger print. An EOS 1Ds Mark III's 21.1 MP
may be the best $olution.
Mark Bohrer
=========================================================
To Unsubscribe: Send email to leica-request@xxxxxxxxxxxxx with
'unsubscribe' in the Subject field. The acknowledgment that you then
receive MUST be replied to per instructions. You may also log in to
the Web interface to unsubscribe.
=========================================================
To Unsubscribe: Send email to leica-request@xxxxxxxxxxxxx with 'unsubscribe' in
the Subject field. The acknowledgment that you then receive MUST be replied to
per instructions. You may also log in to the Web interface to unsubscribe.
Other related posts:
