5D noise at ISO100?

gdanmitchell said:

Here is a 100% crop from my 5D:



Details:

Canon 5D
400x400 section
24-105 f/4 IS lens at f/8 and 24mm
Histogram curve centered
Crop from near center of the frame.
Converted directly from RAW via ACR with no sharpening or other
adjustments.

Click to expand...

Looks good, I have some like this too, not all my images require noise reduction, in fact, only a few. I hope my original post did not imply that.

The particular image I posted was actually ISO200, (my mistake) and is at an edge. The image required a -0.60 exposure reduction to avoid blowing the surf. There was too much DR. I could also have done HDR and left the sky brighter. Or I could use neat image, which I did.

gdanmitchell said:

Dan Mitchell

--
---
G Dan Mitchell
SF Bay Area
http://www.gdanmitchell.com/

Click to expand...

--
http://www.pbase.com/roserus

Ben

Ben_Egbert said:

pprior said:

You're introducing too many variables. You can't take a picture,
upsize it and then complain about noise. As others note it could
be from anything.

Look at the picture under original crop and post THAT image, then
we can talk.

Click to expand...

Well it was a question, not a complaint. Running neat image is not
a problem. Also, given that my monitor displays an image that is 21
inches wide, the uprez is not totally unreasonable.

Click to expand...

Wrong choice of words on my part - what I meant to say was if you are doing the uprez any observation about noise may be from other factors than sensor noise. The upres is not unreasonable, but may be contributing to the noise that you find irritating.

Hope that sounds less confrontational. I just find it strange that you find a problem at such a low ISO with such a good camera.

Cheers.

--
Jeremiah 1:5

Ben_Egbert said:

I agree, the posted image is to show the noise, it is never that
large either on screen or in a print. I just wanted to show that it
is really there. The truth is, this image does not print well, the
trees are muddy looking.

Click to expand...

Your image may print "muddy" because your colors are off. Your trees do look muddy because they are. Your scene is mostly dark with white waves and a white lighthouse - very high DR in the shot.

If you want to find noise you can find it in anyshot if you blow things up to 800% and underexpose your image which is clearly what you have done here. My fixed up version of your image (and yes you did under expose your image probably because of the very bright parts you have in it): http://albums.photo.epson.com/j/ViewPhoto?u=4298510&a=32506611&p=75463278

It may look slightly different in a browser than it does where I adjusted it with PSE3 on a calibrated monitor.

pprior said:

Ben_Egbert said:

pprior said:

You're introducing too many variables. You can't take a picture,
upsize it and then complain about noise. As others note it could
be from anything.

Look at the picture under original crop and post THAT image, then
we can talk.

Click to expand...

Well it was a question, not a complaint. Running neat image is not
a problem. Also, given that my monitor displays an image that is 21
inches wide, the uprez is not totally unreasonable.

Click to expand...

Wrong choice of words on my part - what I meant to say was if you
are doing the uprez any observation about noise may be from other
factors than sensor noise. The upres is not unreasonable, but may
be contributing to the noise that you find irritating.

Click to expand...

Could be I am too sensitive to noise. I hated film grain too, never did like anything over ISO100 film and preferred ISO25. But I suspect my 5D blows away any slide I have. I never printed any slides this big either.

pprior said:

Hope that sounds less confrontational. I just find it strange that
you find a problem at such a low ISO with such a good camera.

Click to expand...

Well it is a new camera, I just got home with 3 DVD's worth of raw images and only found 40-50 worth processing. Most was my fault, I had CFN-1 set for focus and was pushing the wrong button to focus. I caught it some times, but not others. I have this turned off now. Anyway, I also was trying out my Cokin-P system and had lots of vignetting at wide settings, even with my 70-200. I also found that my hyperfocal system needs rejiggering.

Just learning a new system, I guess most of the photos I have taken are just practice.

I had my lapttop along, but it is nearly worthless for judging IQ. It does show vignetting, but sharpness or grain are not easy to see.

pprior said:

Cheers.

--
Jeremiah 1:5

Click to expand...

--
http://www.pbase.com/roserus

Ben

Hellashot said:

Ben_Egbert said:

I agree, the posted image is to show the noise, it is never that
large either on screen or in a print. I just wanted to show that it
is really there. The truth is, this image does not print well, the
trees are muddy looking.

Click to expand...

Your image may print "muddy" because your colors are off. Your
trees do look muddy because they are. Your scene is mostly dark
with white waves and a white lighthouse - very high DR in the shot.

If you want to find noise you can find it in anyshot if you blow
things up to 800% and underexpose your image which is clearly what
you have done here. My fixed up version of your image (and yes you
did under expose your image probably because of the very bright
parts you have in it):

Click to expand...

Well you need to understand this is 500 miles from home. I was there three days running looking for light, clouds in the sky and some surf. I had to take what I could get. It is a difficult scene. The trees need to be illuminated by sun for a decent shot. My shot was not, the sun was behind clouds. I got better images the next day, but the light in the lighthouse was truned off (I like to get the light in the image).

Hellashot said:

http://albums.photo.epson.com/j/ViewPhoto?u=4298510&a=32506611&p=75463278

It may look slightly different in a browser than it does where I
adjusted it with PSE3 on a calibrated monitor.

Click to expand...

Your version has very dark trees. I probably muddied them up trying to get them lighter. I do like your colors better. For some reason, the WB was red, I moved it down from 6000 to 5500.

Thanks for doing this, it gives me some ideas. But I don't really need this image, I only posted it as an example of noise, not an image that needs to be salvaged. I have two more that day with noise, but the first and last day are fine with respect to noise.

I took maybe 200 images of this scene with 17-40, 28-135 and 70-200 lenses. I used a CP no filter, and also ND grads. The biggest problem was I used CFN-1 and was pushing the wrong button. The red square came on and I thought, yes, it is in focus. Well no, it was not in focus, and I could not tell until I got home. The DOf was deep enough to make it look ok on the LCD and my laptop, but not when I got home and could really look.
--
http://www.pbase.com/roserus

Ben

Ben_Egbert said:

Well you need to understand this is 500 miles from home. I was
there three days running looking for light, clouds in the sky and
some surf. I had to take what I could get. It is a difficult scene.
The trees need to be illuminated by sun for a decent shot. My shot
was not, the sun was behind clouds. I got better images the next
day, but the light in the lighthouse was truned off (I like to get
the light in the image).

Click to expand...

These excuses don't change the data in the image you are using as an example. Bright white in the middle of dark green trees under sunlight is a DR nightmare.

Ben_Egbert said:

Ben_Egbert said:

http://albums.photo.epson.com/j/ViewPhoto?u=4298510&a=32506611&p=75463278

It may look slightly different in a browser than it does where I
adjusted it with PSE3 on a calibrated monitor.

Click to expand...

Your version has very dark trees. I probably muddied them up trying
to get them lighter. I do like your colors better. For some reason,
the WB was red, I moved it down from 6000 to 5500.

Click to expand...

Changing the WB temperature isn't always enough to properly balance the colors of an image. I think you also had a blue color cast in your picture besides being too red.

They are not "very" dark on my calibrated monitor (but they are darker than your version), is your monitor calibrated? I had to add a lot of contrast because as you see in comparison, your version is very low in contrast - everything looks flat. Proper contrast makes pictures come alive.

Ben_Egbert said:

Thanks for doing this, it gives me some ideas. But I don't really
need this image, I only posted it as an example of noise, not an
image that needs to be salvaged. I have two more that day with
noise, but the first and last day are fine with respect to noise.

Click to expand...

Again, this image really doesn't have noise if you have to view it at 200% or bigger especially when it was under exposed to begin with as everyone else is telling you too.

Ben_Egbert said:

David J. Littleboy said:

Try really heavy-handed noise reduction before upres: I bet you'll
still see the noise after the upres.

Click to expand...

Actually a normal run of neat image tames the noise nicely. I don't
mind running it, I was mostly just asking if others have this issue
on some images.

Click to expand...

My claim here is that the noise that you are seeing isn't camera noise.

So my idea was that if you NeatImaged the bejesus out of hte image before upressing and still saw the noise, you'd definately know that it wasn't camera noise.

Anyway, my experience is that that much noise in the skies at ISO 100 is seriously anomalous.

--
David J. Littleboy
Tokyo, Japan

Yes Ben, there is noise with the 5D at ISO 200 and even 100. It is noticeable on the computer display at 100%, especially in somewhat underexposed skies. I find it much less obvious when I process RAW using C1 instead of ACR. C1 does a bit of noise reduction but doesn't seem to sacrifice any fine detail compared with ACR.

At any rate, your 5D noise is normal. Some are just less sensitive to this issue.

--
Amin

Have a look at this......BIG IMAGE!
http://forums.dpreview.com/forums/read.asp?forum=1032&message=22700899

There are several things going on here I believe.

One of them is that the color range of a monitor is often inferior to that of a printer. The monitor often has fewer discrete colors available with which to display the image, and so what might be a smooth tone on a print becomes rather blotchy on the monitor. This is particularly true with LCD monitors which seem to struggle more with continuous tone than CRTs.

There's also the fact that you're sitting much closer to a monitor than you would to a print at this size.

I suspect that if you were to print a section of the sky on a good photo printer, and then view it from a reasonable viewing distance for so large a print, that you would not be so conscious of the artifacts.

Kevin

David J. Littleboy said:

Ben_Egbert said:

David J. Littleboy said:

Try really heavy-handed noise reduction before upres: I bet you'll
still see the noise after the upres.

Click to expand...

Actually a normal run of neat image tames the noise nicely. I don't
mind running it, I was mostly just asking if others have this issue
on some images.

Click to expand...

My claim here is that the noise that you are seeing isn't camera
noise.

So my idea was that if you NeatImaged the bejesus out of hte image
before upressing and still saw the noise, you'd definately know
that it wasn't camera noise.

Anyway, my experience is that that much noise in the skies at ISO
100 is seriously anomalous.

Click to expand...

Must have missed the reply where I said ISO100 was a mistake, it was actually ISO200, and of course it was blown up pretty big.

David J. Littleboy said:

--
David J. Littleboy
Tokyo, Japan

Click to expand...

--
http://www.pbase.com/roserus

Ben

Ben_Egbert said:

Let me know if you see the noise. It looks really bad on my monitor

This was shot with a 70-200 f2.8is, at 70mm

Here it is uprezed to print size (bicubic normal) then an 800x800
crop.

Click to expand...

Why do you guys UPREZ then complain?

that makes absolutely no sense whatsoever, unless you're maybe using a 1995 VGA-resolution digital camera.

Hellashot said:

Ben_Egbert said:

Well you need to understand this is 500 miles from home. I was
there three days running looking for light, clouds in the sky and
some surf. I had to take what I could get. It is a difficult scene.
The trees need to be illuminated by sun for a decent shot. My shot
was not, the sun was behind clouds. I got better images the next
day, but the light in the lighthouse was truned off (I like to get
the light in the image).

Click to expand...

These excuses don't change the data in the image you are using as
an example. Bright white in the middle of dark green trees under
sunlight is a DR nightmare.

Click to expand...

Not excuses, just statements of fact. I have been photographing this scene every year for 20 years on my annual vacation to Oregon. The key to a good photo is being there at the right time and knowing how to use your equipment. I am just learning how to use the 5D, especially DOF, hyperfocal, focal length for the lens, etc. But that aside, there was no great picture during the days I was there, period.

Hellashot said:

Hellashot said:

Hellashot said:

http://albums.photo.epson.com/j/ViewPhoto?u=4298510&a=32506611&p=75463278

It may look slightly different in a browser than it does where I
adjusted it with PSE3 on a calibrated monitor.

Click to expand...

Your version has very dark trees. I probably muddied them up trying
to get them lighter. I do like your colors better. For some reason,
the WB was red, I moved it down from 6000 to 5500.

Click to expand...

Changing the WB temperature isn't always enough to properly balance
the colors of an image. I think you also had a blue color cast in
your picture besides being too red.

Click to expand...

Maybe, the following day I got better colors. I may post one after i work on it, still not sure how to post process 5D files yet.

Hellashot said:

They are not "very" dark on my calibrated monitor (but they are
darker than your version), is your monitor calibrated? I had to
add a lot of contrast because as you see in comparison, your
version is very low in contrast - everything looks flat. Proper
contrast makes pictures come alive.

Click to expand...

I guess high contrast is a matter of taste, I don't like crunchy looking images, and I really hate the gritty look of grain or noise. I used to like a more saturated look, but my wife has helped tone me down some.

Hellashot said:

Hellashot said:

Thanks for doing this, it gives me some ideas. But I don't really
need this image, I only posted it as an example of noise, not an
image that needs to be salvaged. I have two more that day with
noise, but the first and last day are fine with respect to noise.

Click to expand...

Again, this image really doesn't have noise if you have to view it
at 200% or bigger especially when it was under exposed to begin
with as everyone else is telling you too.

Click to expand...

Well it was actually over exposed, I had to knock it down 0.60 during raw to avoid blown whites in the surf. so it is wrong to say it was underexposed, that is not true.

--
http://www.pbase.com/roserus

Ben

KCampbell said:

There are several things going on here I believe.

One of them is that the color range of a monitor is often inferior
to that of a printer. The monitor often has fewer discrete colors
available with which to display the image, and so what might be a
smooth tone on a print becomes rather blotchy on the monitor. This
is particularly true with LCD monitors which seem to struggle more
with continuous tone than CRTs.

There's also the fact that you're sitting much closer to a monitor
than you would to a print at this size.

I suspect that if you were to print a section of the sky on a good
photo printer, and then view it from a reasonable viewing distance
for so large a print, that you would not be so conscious of the
artifacts.

Click to expand...

All of this is true, in fact, I printed it and don't see noise in the 16x24 print, but I do on the monitor. In fact, what I see on the monitor is why I posted this.

Remember, this is a $1500 high end LCD, calibrated and all. It is almost as large as the print. It is my most used form of display, so I want my images to dislpay well on it.

KCampbell said:

Kevin

Click to expand...

--
http://www.pbase.com/roserus

Ben

chinch said:

Ben_Egbert said:

Let me know if you see the noise. It looks really bad on my monitor

This was shot with a 70-200 f2.8is, at 70mm

Here it is uprezed to print size (bicubic normal) then an 800x800
crop.

Click to expand...

Why do you guys UPREZ then complain?

Click to expand...

Why does a simple question get branded as complaint? Do you think I attacked your precious camera?

I just got a new camera, I love it, but I am trying to understand it by asking other 5D owners about it. I am having trpouble figuring out how to post process, I need to ask about this stuff.

I uprezed it to the same size required for printing, whats strange about that?

chinch said:

that makes absolutely no sense whatsoever, unless you're maybe
using a 1995 VGA-resolution digital camera.

Click to expand...

--
http://www.pbase.com/roserus

Ben

Juli Valley said:

I think you are pixel peeping. Print one photo to see what you
get. I don't think it will show in the print.

Click to expand...

It does not show in prints, it shows on my monitor when I use it for wallpaper. Remember, this is a really big monitor. Viewing images on my monitor is as important to me as prints. I want them to look good here.

Juli Valley said:

--
Juli
http://www.pbase.com/julivalley/galleries
Canon FiveDee, Canon 2oD, Canon Gee3, and Canon S7o, Fuji Eff30.

Click to expand...

--
http://www.pbase.com/roserus

Ben

Amin Sabet said:

Yes Ben, there is noise with the 5D at ISO 200 and even 100. It is
noticeable on the computer display at 100%, especially in somewhat
underexposed skies. I find it much less obvious when I process RAW
using C1 instead of ACR. C1 does a bit of noise reduction but
doesn't seem to sacrifice any fine detail compared with ACR.

At any rate, your 5D noise is normal. Some are just less sensitive
to this issue.

Click to expand...

Thanks for a rational reply. I guess it is always bad to ask a question like this on a forum where people hold their equipment as being infallable and perfect in every possible way.

Amin Sabet said:

--
Amin

Click to expand...

--
http://www.pbase.com/roserus

Ben

Ben_Egbert said:

Why does a simple question get branded as complaint? Do you think I
attacked your precious camera?

I just got a new camera, I love it, but I am trying to understand
it by asking other 5D owners about it. I am having trpouble
figuring out how to post process, I need to ask about this stuff.

I uprezed it to the same size required for printing, whats strange
about that?

Click to expand...

5D noise at ISO100?

^^^^ real silly topic title as it implies perhaps every digital camera made has noise at ISO100 and will be construed as whining.

i dont own the 5D.

you're uprezing but not comparing a PRINT you're comparing on a monitor with much more limitations than the print

uprezzing is BAD if you're going to pixel peep.

PhillipA said:

Have a look at this......BIG IMAGE!
http://forums.dpreview.com/forums/read.asp?forum=1032&message=22700899

Click to expand...

Yep, and I noticed another post comparing the 5D to a Mamia, and the samples at 100% crop were so noisy it was unusable, and the Mamia was far worse than the 5D.

Some people must like noise. In fact I have seen people post techniques where they use noise to enhance sharpness. Sorry, I view noise and grains as technical artifacts that need to be reduced until they are totally below our perception level. The 5D is a fine camera and has excellent noise preformance, but it stil lhas it, and one needs to take care to avoid it.

Grainy images are a fad of some sort, just like extreme shallow DOF.

--
http://www.pbase.com/roserus

Ben

Ben_Egbert said:

Ben_Egbert said:

Again, this image really doesn't have noise if you have to view it
at 200% or bigger especially when it was under exposed to begin
with as everyone else is telling you too.

Click to expand...

Well it was actually over exposed, I had to knock it down 0.60
during raw to avoid blown whites in the surf. so it is wrong to say
it was underexposed, that is not true.

Click to expand...

Just because you had over exposed whites doesn't mean that your darker blue sky and dark green trees weren't under exposed. As I said before bright white mixed in with largely dark green trees is a DR nightmare.