Inconsistency in D200's noise performance?

You're leading this thread to a direction I wanted to aviod, because what I want to find out is not if D200 has lower noise than D70, but rather whether previously produced D200 have big variation in noise performance and whether new or current ones have tighter variation.

But just to make a few points clear.

First, the only thing I care is how an image look at full screen and printed with sizes up to 12x18". Pixle was mentioned only for a quantified discussion. And I do have concern because at such viewing conditions I did see my previous D200 showing more noise than D70.

Second, you're simplified things too much. Pixel density is higher on D200, but it doesn't mean the active pixel (actual CCD) area is smaller on D200. There're different architectures of sensor designs and technologies that affect final noise.

Anyway, fine with me if D200's noise is designed to be higher than D70. I just need to know what am I expecting to get when I go for one.

Hans Giersberg said:

Just because you want the D200 noise to be no worse than the D70
isn't going to make it so. It has smaller photosites. Physics is
physics. The D200 yields very low noise images up to ISO 800, and
acceptable noise at higher ISOs. Not as low as the D70, but still
very low. And better than my D2X.

It's a tradeoff. I wanted a small body for times when a D2X was a
bit much. I considered the D70S and the D200. I opted for the D200,
even tho I expected the noise to be a little worse, because I
wanted the higher MP, I wanted the faster shooting rate, I wanted
the much better viewfinder, I wanted the much better build quality,
I wanted weather sealing. To get all those things, I accepted
slightly higher noise. Noise which I don't see in my prints, and
noise I don't see on my downsampled web site images.

If you're going to spend all your time pixel peeping at 100%,
you're never going to be satisfied. But the good thing is that
these things don't actually matter in the real world of prints and
web images, which is how most of us use our photos. I don't know
too many people who show of their photos by pulling them up on
screen and zooming in to 100%

--
D200 - The Best $1600 I Ever Spent On Photo Gear

Web Site - http://www.hgiersberg.com/

Click to expand...

This is very surprising... please show your results soon. I'm very curious.

Using the same or very similar sensors between D50 and D70(s), yet they are at the two extreme of your comparison group?

Could it be the effects of NR (that smears details at the same time)?

Could it be the effects of WB settings? For this I've found a dramatic difference with D70 while shooting under tungsten lights (higher noise with AWB or settings other than Incandescent).

Still wonder if anyone has ever tested a few D200's to check copy-to-copy variation???

Steve Bingham said:

Under identical conditions and lens I tested the D50, D70, D200,
and D2x. Before the D200, the D50 ruled. Now the D200 lies
somewhere between the D50 and D2x and better than the D70 at ISO
1600. (Kelvin 3400 interior flood, Nikon 17-35, f8, tripod)

Simply one test by one person. In summary, D50, D200, D2x, D70. All
small increases. All shot raw. Still surprised how well the D50 did.
--
Steve Bingham
http://www.dustylens.com

Click to expand...

Hans Giersberg said:

Just because you want the D200 noise to be no worse than the D70
isn't going to make it so. It has smaller photosites. Physics is
physics. The D200 yields very low noise images up to ISO 800, and
acceptable noise at higher ISOs. Not as low as the D70, but still
very low. And better than my D2X.

It's a tradeoff. I wanted a small body for times when a D2X was a
bit much. I considered the D70S and the D200. I opted for the D200,
even tho I expected the noise to be a little worse, because I
wanted the higher MP, I wanted the faster shooting rate, I wanted
the much better viewfinder, I wanted the much better build quality,
I wanted weather sealing. To get all those things, I accepted
slightly higher noise. Noise which I don't see in my prints, and
noise I don't see on my downsampled web site images.

If you're going to spend all your time pixel peeping at 100%,
you're never going to be satisfied. But the good thing is that
these things don't actually matter in the real world of prints and
web images, which is how most of us use our photos. I don't know
too many people who show of their photos by pulling them up on
screen and zooming in to 100%

--
D200 - The Best $1600 I Ever Spent On Photo Gear

Web Site - http://www.hgiersberg.com/

Click to expand...

My own tests comparing the noise of the D70 to the D200 are in agreement with Hans' findings. The D200 has more noise as predicted.

http://forums.dpreview.com/forums/read.asp?forum=1034&message=17236791

--
Bill Janes

I'm not oversimplifying. The published photosite size for the D200 is smaller. The size is in the spec sheet.

As to variation, I have an early vintage D200 (if there is such a thing) and I'm not seeing the variation that you report. So I guess my answer to your question is that there is at least one D200 in the world that doesn't exhibit that.

What you may be seeing is variation in the color of the subjects you are shooting, and that can translate into more or less noise at the same ISO. The blue channel tends to have more noise, so if some of your shots are of more blue subjects, I wouldn't be surprised if the noise is more. Just a thought. Also, if you are using AWB, you may see some variation in the colors, and thus in noise, as you change the composition slightly and shift white balance. Try some tests with AWB off so that you get consistent white balance from shot to shot. Again, just a thought. And one final thought. If you've got autosharpening set, that could cause different levels in the noise in your images.

I guess I would explore all these avenues first before concluding that your D200 has an issue.

--
D200 - The Best $1600 I Ever Spent On Photo Gear

Web Site - http://www.hgiersberg.com/

My 1st body was fairly clean at ISO1000-1250

2nd body was very noisy at ISO1000

i have images to prove it also - Actually, the ISO1250 shot from body#1 was taken in lower light and is cleaner than the ISO1000 shot from body #2. D200 has some serious QC issues.

Well, I wouldn't trade my D200's high ISO performance for a thousand d70s!
the high ISO of the d50 or d2HS OTOH, is a different story
I've noticed two things about the D200 vs. D70 in noise:

1) With High ISO NR on LOW, only in certain extreme conditions is there color noise at least through ISO 1000 if not ISO 1250 and 1600. My D70 would have noticeable color noise throughout at ISOs above 500--640, but nothing a bit of PP couldn't clean up OK...
2) with NR on Normal, I basically can't find color noise until the HI levels.

3) midtone Shadows seem a tiny bit more prone to 'graininess' due to luminance noise...however...

4) this is purely subjective, but I think that the noise in D200 shots feels more a 'part of the shot' than D70 noise... it may have to do with processing but I always felt that noise with the D70 if it was there esp. color noise would feel superficial or almost like someone had taken a paintbrush dipped in paint, flipped the bristles back and let the slowly slip forward, flinging paint on to the image, creating a more superficial feel.

5) D200 color noise in my experience is mostly yellows and typically shows up in certain difficult to color manage areas of the shot.

6) I have my camera set with a 1/3 stop fine tune exposure for better results for my shooting. Using this, I find all the well exposed areas of the shot are free of noise, areas that have a much darker exposure due to shadows but that are not yet black will generally exhibit grain/luminance noise, and whenever something hits Black, It stays that way. Any high ISO NR settings, even Low, and any truly black areas in my tests I conducted to get a feel for high ISO and NR settings have been 99% noise free.

I basically feel I can use the camera without any hesitation, worry, or thoughts whatsoever until ISO 400-500, I could shoot NR off easily. ISO 640-1000 no worries so long as I am paying a bit more attention to exposure and composition, here NR on Low is a good idea... and 1250-1600 is problem free if I work for good exposure- NR on low or at worst normal is nice. Often i'll work for the middle ground in NC with NEFs. I could shoot without NR but I like having it applied for me and then tweaking afterwards.

Could it be a bit better? you bet...I hope a FW update and the new Capture offer even better NR than we already have...I'd like a bit better resuls on edge detail, but I have no complaints...I was taking care of some ISO 1000 shots this evening and turned DOWN the NR from its on-low defaults.
--
Loving the 85mm f/2.8D PC Micro
My gallery! http://nikonmadness.smugmug.com

I have type II banding -and- I feel the noise is considerably worse than I'm seeing posted by other people. I'm having major noise at ISO 250 and 320.

I'm hoping the banding fix will also be a noise fix. At 800 and above, most of my pics won't even clean up with NR tools without pretty much destroying the shots.

--
AAK - http://www.aakatz.com

Probable other people noticed it also and for others it may sound weird, but my D200 performs better (by means of noise) at ISO 800 than at 400, this in relation to photographs I made with my D70, which I unfortunately had to sell.

Love my D200

Michel

--
Curiosity is the key to creativity - Morita Akio
http://www.fotopropaganda.com

...and is what I have experienced with that set of cameras.

I became quite convinced after months of using a D50 (along with the D2X, D70 and D200 too) that it isn't simply the result of noise reduction. There is more detail in a D50 ISO 1600 image than any of the others.

Photobug said:

Using the same or very similar sensors between D50 and D70(s), yet
they are at the two extreme of your comparison group?

Could it be the effects of NR (that smears details at the same time)?

Could it be the effects of WB settings? For this I've found a
dramatic difference with D70 while shooting under tungsten lights
(higher noise with AWB or settings other than Incandescent).

Still wonder if anyone has ever tested a few D200's to check
copy-to-copy variation???

Steve Bingham said:

Under identical conditions and lens I tested the D50, D70, D200,
and D2x. Before the D200, the D50 ruled. Now the D200 lies
somewhere between the D50 and D2x and better than the D70 at ISO
1600. (Kelvin 3400 interior flood, Nikon 17-35, f8, tripod)

Simply one test by one person. In summary, D50, D200, D2x, D70. All
small increases. All shot raw. Still surprised how well the D50 did.
--
Steve Bingham
http://www.dustylens.com

Click to expand...

Click to expand...

The point here wasn't to perform an academic research, the point was to figure out if your unit is wrongly adjusted or not.

Dark current for this purpose is more than enought if you have one D200 that generates twice as much noise in that mode than another one in similar conditions.

Well, it doesn't matter anyway. If people do believe for real that there is no way to meausure noise - heck with it, there is no way.

Bill Janes said:

Paul_U said:

I do not get it.

Just close the lens cap, set to M mode, take a series of shots on
1/60 shutter speed, open NEF in ACR with all parameters set to 0,
then run BLur-Average and you get uniform noise level of your
camera.

After that we will see for sure what noise it is and no more
guessing will be needed.

Click to expand...

Dark current is only one component of the total noise and along
with read noise is only a mihor component of noise with most normal
photographic scenes. In normal practice, photon noise is the major
component and it is related to pixel size, fill factor, and quantum
efficiency of the sensor. For more details see this article by
Roger Clark, a space scientist and avid photographer:

http://www.clarkvision.com/imagedetail/evaluation-1d2/index.html
--
Bill Janes

Click to expand...

--
Best regards from UPVStudio Photography

Where in D200's spec sheet does it mention photosite size?

In D200's manual they list 23.6x15.8mm as CCD size and 10.92 million as total pixels.

However, even if it's listed it doesn't mean much. What matters is size of photo-sensitive area in a photosite. Photosite size is merely a upper bound of such area. Some sensors have such area occupying most of photosite (high fill factor); some have such area being only a small percentage of a photosite (low fill factor).

Nowhere did I see they mentioning fill factor or photo-sensitive area in a pixel. This is what really counts. A sensor (some ancient ones, especially CMOS sensors) can have huge pixel (like 10umx10um) but tiny photo-sensitive area (like 3umx3um) and have lousy noise performance.

I haven't seen Canon or Nikon revealing such numbers (fill factor or active area size). If anyone knows of such numbers, I'll be very interested to see.

It's good that you found a good D200. But experience of many people here does point to a possibility of big variation from copy to copy.

BTW my test was properly conducted. The sharpening was always set to the same or having both sharpness and noise considered together.

Photobug

Hans Giersberg said:

I'm not oversimplifying. The published photosite size for the D200
is smaller. The size is in the spec sheet.

As to variation, I have an early vintage D200 (if there is such a
thing) and I'm not seeing the variation that you report. So I guess
my answer to your question is that there is at least one D200 in
the world that doesn't exhibit that.

What you may be seeing is variation in the color of the subjects
you are shooting, and that can translate into more or less noise at
the same ISO. The blue channel tends to have more noise, so if some
of your shots are of more blue subjects, I wouldn't be surprised if
the noise is more. Just a thought. Also, if you are using AWB, you
may see some variation in the colors, and thus in noise, as you
change the composition slightly and shift white balance. Try some
tests with AWB off so that you get consistent white balance from
shot to shot. Again, just a thought. And one final thought. If
you've got autosharpening set, that could cause different levels in
the noise in your images.

I guess I would explore all these avenues first before concluding
that your D200 has an issue.

--
D200 - The Best $1600 I Ever Spent On Photo Gear

Web Site - http://www.hgiersberg.com/

Click to expand...

...has anyone compared few copies of D200??? Results?

In order to take out differences in personal judgements and shooting conditions, this is the only way we can know for sure.

One interesting thing is that what Phil found was pretty much in line with my own finding (i.e., resolution similar to 20D with noise being high in shadow areas of ISO 1600 and 3200). Actually I saw more shadow noise at ISO 1600 than D70's, even when viewing both at full-screen size. Just like Phil, I also prefer to shoot D200 without NR as Neat Image did a far superior job in removing noise while preserving details.

Photobug

You think this is news?? Every camera has high noise at ISO 1600 and 3200. 20D may have lower noise than D200, but its noise still sucks. It just sucks less than the D200 noise does at those ISOs. A friend of mine has a 20D and I've looked at his high ISO shots. They suck. I've looked at my D2X and D200 high ISO shots. They suck too. I'm surprised you expected anything different.

D200 resolution was listed as better than 20D (not the same) and less than 5D. Interestingly, D2X beat 5D on resolution, with same pixel count.

I'm still looking for the pixel size. You may have me there (tho I doubt it, since I have seen photosite size in print somewhere). But I'm still not sure why you're arguing about this. Why would you expect D200 to have noise as good as the D70.

As to noise variability, you answered one of my suggestions - sharpening. You didn't answer my other suggestions tho - AWB and subject colors. I'm not saying you're wrong, but your "findings" aren't based on facts, only on opinion. You haven't even begun to identify, much less control, all the variables at play here. Until you can identify and control the variables, you can't say didly about whether the D200 noise variability is good, bad, or indifferent.

As to your suggestion that lots of others have reported the same - where? Your sample size is miniscule. Any statistician would laugh at any attempts to draw conclusions from such a small sample size.

I think you have a theory based on questionable observations, with no control over the variables and a very small sample size of "opinions". As I said above, this doesn't make you wrong, but the burden of proof is on you, and so far you haven't proven anything.

--
D200 - The Best $1600 I Ever Spent On Photo Gear

Web Site - http://www.hgiersberg.com/

Please stop misrepresenting my words. I am not surprised at cameras having high noise at ISO 1600 and 3200. But I'm curious to find out whether D200 have big variation on high ISO noises between copies. How many times do I have to say this???

Regarding controlled tests. I don't know where you got the ideas that I didn't controll WB and subject colors. I suggest you check posts I have in this thread before making baseless conclusions. It's an A/B comparion so of course subject colors are the same. And WB settings, including any other parameters that could affect results on both cameras was controlled during the tests. Even the difference in tone curves has been explored. What more do you want?

Just in this thread there're few posters having "related" experience. I'm not making conclusions here (like you did). But I do think it's worthwhile collecting experience and thoughts here to know better.

I'm asking a question and at the same time sharing my thoughts and experience. Why should I prove anything? Are you having a hard time accepting others' opinions not in line with yours; or can't stand questions about D200's QC?

Photobug

Hans Giersberg said:

You think this is news?? Every camera has high noise at ISO 1600
and 3200. 20D may have lower noise than D200, but its noise still
sucks. It just sucks less than the D200 noise does at those ISOs. A
friend of mine has a 20D and I've looked at his high ISO shots.
They suck. I've looked at my D2X and D200 high ISO shots. They suck
too. I'm surprised you expected anything different.

D200 resolution was listed as better than 20D (not the same) and
less than 5D. Interestingly, D2X beat 5D on resolution, with same
pixel count.

I'm still looking for the pixel size. You may have me there (tho I
doubt it, since I have seen photosite size in print somewhere). But
I'm still not sure why you're arguing about this. Why would you
expect D200 to have noise as good as the D70.

As to noise variability, you answered one of my suggestions -
sharpening. You didn't answer my other suggestions tho - AWB and
subject colors. I'm not saying you're wrong, but your "findings"
aren't based on facts, only on opinion. You haven't even begun to
identify, much less control, all the variables at play here. Until
you can identify and control the variables, you can't say didly
about whether the D200 noise variability is good, bad, or
indifferent.

As to your suggestion that lots of others have reported the same -
where? Your sample size is miniscule. Any statistician would laugh
at any attempts to draw conclusions from such a small sample size.

I think you have a theory based on questionable observations, with
no control over the variables and a very small sample size of
"opinions". As I said above, this doesn't make you wrong, but the
burden of proof is on you, and so far you haven't proven anything.

--
D200 - The Best $1600 I Ever Spent On Photo Gear

Web Site - http://www.hgiersberg.com/

Click to expand...