NOISE

Kevin McMurtrie wrote:

In article ,
Stacey wrote:

Patrick Boch wrote:

Digital needs to get past
this-plastic like-sheen.

I agree. I've shot with several of the dSLR's and IMHO some just look
"waxy" even though they are void of most noise. I'm more interested in
the overall lood of the images/prints that looking at 200% crops
=looking= for noise. These waxy looking images aren't my cup of tea
either, YMMV.

I think you're seeing noise filtering. If a small detail doesn't appear
to be part of a larger pattern, it gets blurred out. Trees, shrubs,
grass blades, and roof shingles are the usual victims. It gives photos
a VHS video look. CMOS sensors from a few years ago needed big fat gobs
of filtering. CCD sensors need lots of filtering when they heat up from
use. The good news is that today's CMOS sensors need almost no
filtering with an ISO in the 50 to 200 range. Photos look a lot more
realistic.

It was a 10D that had this "waxy" look I didn't care for. I think it's
more what type of sensor and how they process it that does this. I process
the Images now with NO noise reduction and only apply it selectively is
it's really needed. I'd rather see some noise than this waxy look.
--

Stacey

Marli wrote:

Must have been a Canon CMOS then.

It was, it was a Canon 10D. Everyone brags about how clean they are at
1600ISO and they BASH any camera that doesn't as being garbage. I didn't
care for the look this "smooth" sensor gives the overal image, lots of
people must? For me some noise is a lot less of an issue than this waxy
look is.


A CCD is not "Waxy", not mine anyway.
Not sure what the D2X looks like, its a CMOS.


No idea either.
--

Stacey

In article ,
Stacey wrote:

Kevin McMurtrie wrote:

In article ,
Stacey wrote:

Patrick Boch wrote:

Digital needs to get past
this-plastic like-sheen.

I agree. I've shot with several of the dSLR's and IMHO some just look
"waxy" even though they are void of most noise. I'm more interested in
the overall lood of the images/prints that looking at 200% crops
=looking= for noise. These waxy looking images aren't my cup of tea
either, YMMV.

I think you're seeing noise filtering. If a small detail doesn't appear
to be part of a larger pattern, it gets blurred out. Trees, shrubs,
grass blades, and roof shingles are the usual victims. It gives photos
a VHS video look. CMOS sensors from a few years ago needed big fat gobs
of filtering. CCD sensors need lots of filtering when they heat up from
use. The good news is that today's CMOS sensors need almost no
filtering with an ISO in the 50 to 200 range. Photos look a lot more
realistic.

It was a 10D that had this "waxy" look I didn't care for. I think it's
more what type of sensor and how they process it that does this. I process
the Images now with NO noise reduction and only apply it selectively is
it's really needed. I'd rather see some noise than this waxy look.

Strange. I haven't had that problem on a Canon 300D -

100 - Perfection
200 - Very clean
400 - Minor noise, faintest hint of noise reduction
800 - Noise and noise reduction are noticeable
1600 - Noise and noise reduction badly damages fine details

I'm shooting in JPEG mode except when major post-processing is expected.
Were you using a RAW converter that might have applied filtering when it
shouldn't?

"Kevin McMurtrie" wrote:
Stacey wrote:

It was a 10D that had this "waxy" look I didn't care for. I think it's
more what type of sensor and how they process it that does this. I
process
the Images now with NO noise reduction and only apply it selectively is
it's really needed. I'd rather see some noise than this waxy look.

Strange. I haven't had that problem on a Canon 300D -

Neither has anyone else. Nor with any other Canon dSLR. Stacey's simply
wrong on this.

David J. Littleboy
Tokyo, Japan

On Sun, 13 Feb 2005 17:31:05 +0900, "David J. Littleboy"
wrote:


"Kevin McMurtrie" wrote:
Stacey wrote:

It was a 10D that had this "waxy" look I didn't care for. I think it's
more what type of sensor and how they process it that does this. I
process
the Images now with NO noise reduction and only apply it selectively is
it's really needed. I'd rather see some noise than this waxy look.

Strange. I haven't had that problem on a Canon 300D -

Neither has anyone else. Nor with any other Canon dSLR. Stacey's simply
wrong on this.


Mostly I'm wondering how a 13.0 x 17.3 mm sensor
with 8 million pixels is going to deliver a better
quality image than the 10D's 15.1 x 22.7 mm sensor
which only supplies 6 million pixels.

The unit sensor area works out to almost exactly
half in the Oly (2.8E-5 mm^2 vs. 5.7E-5.)

But Stacey is careful to describe the Oly's
advantages in terms that defy measurement or
verification.


rafe b.
http://www.terrapinphoto.com

"rafe bustin" wrote:
"David J. Littleboy" wrote:
"Kevin McMurtrie" wrote:
Stacey wrote:

It was a 10D that had this "waxy" look I didn't care for. I think
it's
more what type of sensor and how they process it that does this. I
process
the Images now with NO noise reduction and only apply it selectively
is
it's really needed. I'd rather see some noise than this waxy look.

Strange. I haven't had that problem on a Canon 300D -

Neither has anyone else. Nor with any other Canon dSLR. Stacey's simply
wrong on this.

Mostly I'm wondering how a 13.0 x 17.3 mm sensor
with 8 million pixels is going to deliver a better
quality image than the 10D's 15.1 x 22.7 mm sensor
which only supplies 6 million pixels.

But you've misunderstood Stacey's point: more noise makes a better looking,
higher dynamic range image than those plasticy low-noise images that Canon
produces. ROFL.

The unit sensor area works out to almost exactly
half in the Oly (2.8E-5 mm^2 vs. 5.7E-5.)

That sounds off. Canon = 15 x 22.5 = 337.5 vs Oly = 13 x 17.3 = 224.9 =? Oly
= 67% of the Canon. Although the _effective_ area of the pixel also depends
on the microlenses.

But Stacey is careful to describe the Oly's
advantages in terms that defy measurement or
verification.

Right. "As good as medium format". Not.

David J. Littleboy
Tokyo, Japan

Stacey wrote:
Marli wrote:

Must have been a Canon CMOS then.

It was, it was a Canon 10D. Everyone brags about how clean they are
at
1600ISO and they BASH any camera that doesn't as being garbage. I
didn't
care for the look this "smooth" sensor gives the overal image, lots
of
people must? For me some noise is a lot less of an issue than this
waxy
look is.


A CCD is not "Waxy", not mine anyway.
Not sure what the D2X looks like, its a CMOS.


No idea either.
--

Stacey
Could you post a link to one of the images off of the 10D that look
waxy to you?

Scott

rafe bustin wrote:
Now as to Stacey's complaint... in fairness,
I too have seen weirdness examining the color
channels in certain of my 10D images. We all
know that the blue channel often sucks, and
it seems Canon's "Digic" chip does odd things
to restore the blue channel image. OTOH, I have
no evidence at all that Oly knows how to deal
with the problem any better than Canon.

But this should not be a problem at all if shooting in raw, in that
case the Digic chip should not be doing any pocessing to the saved file
all all.

I have not seem much of this problem with my 20D but see a lot of it
with my F828. In the case of the F828 if I shoot a blue ocean with
waves the camera tends to remove the waves apparently thinking it is
looking at a noisy sky. But when I shoot this in raw there is no
problem.

Again I have seen very little problems with the jpgs that the 20D
produces.

Scott

On Sun, 13 Feb 2005 23:06:38 +0900, "David J. Littleboy"
wrote:


"rafe bustin" wrote:


The unit sensor area works out to almost exactly
half in the Oly (2.8E-5 mm^2 vs. 5.7E-5.)

That sounds off. Canon = 15 x 22.5 = 337.5 vs Oly = 13 x 17.3 = 224.9 =? Oly
= 67% of the Canon. Although the _effective_ area of the pixel also depends
on the microlenses.


Right. Now:

337.5 / 6M = 5.6E-5
224.9 / 8M = 2.8E-5

Now as to Stacey's complaint... in fairness,
I too have seen weirdness examining the color
channels in certain of my 10D images. We all
know that the blue channel often sucks, and
it seems Canon's "Digic" chip does odd things
to restore the blue channel image. OTOH, I have
no evidence at all that Oly knows how to deal
with the problem any better than Canon.

In particular, have a look at the images he

http://www.terrapinphoto.com/tenD/



rafe b.
http://www.terrapinphoto.com

On 13 Feb 2005 07:12:49 -0800, "Scott W" wrote:


rafe bustin wrote:
Now as to Stacey's complaint... in fairness,
I too have seen weirdness examining the color
channels in certain of my 10D images. We all
know that the blue channel often sucks, and
it seems Canon's "Digic" chip does odd things
to restore the blue channel image. OTOH, I have
no evidence at all that Oly knows how to deal
with the problem any better than Canon.

But this should not be a problem at all if shooting in raw, in that
case the Digic chip should not be doing any pocessing to the saved file
all all.

I have not seem much of this problem with my 20D but see a lot of it
with my F828. In the case of the F828 if I shoot a blue ocean with
waves the camera tends to remove the waves apparently thinking it is
looking at a noisy sky. But when I shoot this in raw there is no
problem.

Again I have seen very little problems with the jpgs that the 20D
produces.


Did you look at the images?
Download them and look at them
in Photoshop. Look in particular at
the blue channel.

All I know is this: I have never seen
anything like this in a film scan.

I believe the Digic chip plays some
role even in RAW format images, if
only as the basic controller and
data-gatherer to/from the sensor.


rafe b.
http://www.terrapinphoto.com