"Noise" in digital pictures

I've seen a lot of threads here dealing with "Noise" in images taken by the Olympus E-10 camera. I'm not convinced that there is anything here other than user error.

First, in the only thread that mentioned it, the commentor was storing the images in jpeg format. Right there could be the cause of the "noise". Jpeg is a lossy format. That means that what comes out, is NOT the same as went in. Some information WILL be lost. TIFF images as stored by the camera are at least stored without compression, so you have a better chance of seeing, that which is due to the camera's limitations rather than the compression engine's limitations.

Hello Kim,

Don't feel "attacked" by this reply - it's constructive!
I'm only using jpg and I don't have complaints about noise.
So, the jpg is not (as only) responsible for possible noise.
BTW, I thinks most complaints about noise were when using RAW-files.
There's a large difference between "noise" and "jpg-artefacts".

So, in my opinion, there are some E-10's around that produce noise. The reason why isn't cleared for the moment.

I also don't think that the E-10 is bought by a lot of people that don't have a "digital photographic" background (so without experience).

People who are doing digital imaging for quit a long time really can appreciate the E-10 quality (the have something to compare ))
BTW, here's an image with real noise: (no, not E-10
It's from my former Canon Pro 70:



That's what I call noise (or jpg-artefacts?)! -
The Pro 70 always has had problems with a nice, blue sky.

Jaja
http://www.belgiumdigital.com

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

You must view the image at full resolution to see any noise, you may not see any in print unless you go real big. The samples are E10 and D30 from Steves Digicams. There are many more. Look at the sky and the glass at 100% or greater and you will see... The E10 noise is typical off all small CCD cameras, I personally wouldn't worry about it.

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

more, of coures the days and times are different...









There are differances but they may not be visible with normal size prints or reduced resolution...

peczenij said:

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

Click to expand...

E10

http://www.steves-digicams.com/e10_samples.html

D30

http://www.steves-digicams.com/d30_samples.html

peczenij said:

There are differances but they may not be visible with normal size
prints or reduced resolution...

peczenij said:

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

Click to expand...

Click to expand...

Yes...look at these two photos ! The d 30 is far better than the e-10 ! Just the difference in color depth is shocking. There is also less noise in the d 30 sky. How can anyone deny the huge difference in these two photos? The color of the shy tells the tale.

peczenij said:

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

Click to expand...

Hello,

Don't shoot at me! Just a thought ...

When I was looking at D-30 photos, they all seem 'not that sharp' to me, compared to the E-10 photos. The D-30 photos are all much more softer.
(The examples showed in this thread also - I removed them here).
Could that have to do something with the (sometimes) visible noise?

Just a thought ... Someone mentioned here to use the camera with less sharp settings in the menu ...

Jaja
http://www.belgiumdigital.com

Jan Sanders said:

Yes...look at these two photos ! The d 30 is far better than the
e-10 ! Just the difference in color depth is shocking. There is
also less noise in the d 30 sky. How can anyone deny the huge
difference in these two photos? The color of the shy tells the tale.

Click to expand...

THe D30 photos are much softer and require setting the camera to sharp and/or software sharpening (sopposedly that's the preferred method to in-camera sharpening). The software method will require extra effort for processing. The E10 photos are perfectly sharp right out of the camera...

Jaja - Belgium Digital said:

Don't shoot at me! Just a thought ...
When I was looking at D-30 photos, they all seem 'not that sharp'
to me, compared to the E-10 photos. The D-30 photos are all much
more softer.
(The examples showed in this thread also - I removed them here).
Could that have to do something with the (sometimes) visible noise?
Just a thought ... Someone mentioned here to use the camera with
less sharp settings in the menu ...

Jaja
http://www.belgiumdigital.com

Jan Sanders said:

Yes...look at these two photos ! The d 30 is far better than the
e-10 ! Just the difference in color depth is shocking. There is
also less noise in the d 30 sky. How can anyone deny the huge
difference in these two photos? The color of the shy tells the tale.

Click to expand...

Click to expand...

peczenij said:

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

Click to expand...

OK guys (and gals), I think I will jump in here and offer my two cents worth. First, I have been monitoring this forum(s) for the past month or two; trying to soak up all of the info. I have had a couple of SLRs in the past but have not been active for five or so years. Digital has made it enticeing to resume photography especially since I don't need a darkroom - my laptop will do quite nicely.

The E-10 seems to be a good compromise at the moment especially at the $2K price point. I am sure the big boys will produce a few models in the near future that will blow its doors off but, for now, perhaps OLY has a winner.

As for the two above mentioned pictures, and I am assuming the first is the E-10 and the second is the D30, it is apparent that the E-10 is noisier in the shadows. Not only are the windows not as clean as the D30 image, but other shadow areas are also a bit "grainy". Even the asphalt of the road is not quite as clean. The colors seem to be about the same saturation except for the sky, which leads me to my second point.

The E-10 seems to capture a slight overcast - notice the suggestion of wispy clouds whereas the D30 simply blows away that detail, over saturation perhaps. From the position of the shadows cast on the windows, I am assuming the pictures were taken within minutes of each other. Is it possible that the E-10 picks up a bit more detail than the D30, i.e., a bit more dynamic range? If so, is there a bit more detail in the shadows despite a noisier image? I am not PhotoShop literate, but might a bit of tweaking in the blue channel smooth out some of the noise?

I enjoy the banter on this forum but, since I am a "newbie", don't chew me up too badly.

Happy Holidays!

D1



E10



big CCD and small CCD
$2000 vs $5000+lens

peczenij said:

http://www.steves-digicams.com/e10_samples.html

D30

http://www.steves-digicams.com/d30_samples.html

peczenij said:

There are differances but they may not be visible with normal size
prints or reduced resolution...

peczenij said:

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

Click to expand...

Click to expand...

Click to expand...

I'm not sure, but it seems to me that the two original photos were taken at different times. Looking at the parking lot, one shot has a white car and a bike, the other has neither. Now maybe it is only a few minutes, maybe it's hours. I know that at least here in NY, the sky can chance color over time...

Also, am I just picking the wrong topics, or do allot of contributors here like to bicker back and forth like high school kids..?

Mike

Jan Sanders said:

peczenij said:

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

Click to expand...

Click to expand...

Albert said:

The E-10 seems to capture a slight overcast - notice the suggestion
of wispy clouds whereas the D30 simply blows away that detail, over
saturation perhaps. From the position of the shadows cast on the
windows, I am assuming the pictures were taken within minutes of
each other.

Click to expand...

They were taken on different days. If you look on the right side of the building you'll see that a huge stove vent in the D30 picture is gone!

I have been looking into that problem today after finding out that a daytime portrait taken in diffused shadow light on a sunny winter day in California was unusable - I left 360 ISO setting by mistake.

It sould not happen, I thought, even if you use ISO 360.

So I compared "far field" test images from http://www.imaging-resource.com

1. "Soup dish"-type cameras are bad. Nikon Coolpix990 was the worst I saw.

2. Of all better cameras (D1, D30, S1, and E10), E10 was the noisiest. D30 was the smoothest. However you cannot have your cake and eat it: D30 gave a nasty color artifact to white windows on the house on the test shot.
It comes to processing algorithms, I decided.

So E10 is in a really good company, and being the cheapest camera of all big ones, it justifies its performance.

My suggestions are two:

1. Blur the noisiest channel, while keeping the other two sharp to reduce noise. E.g. in incandescent light, your blue channel will be the noisiest, red and green will remain the same.

1a. Use gaussian blur at a small radius setting (I usually use

value in Gimp; experiment in Photoshop). It gets rid of jpeg compression artifacts and noise without affecting the printability of the picture.

2. Use gimp (under unix/linux) to despeckle the image, works pretty well. Photoshop despeckling is inferior.

3. NEW IDEA:

you do not need to produce FINAL pictures right from the camera. Treat them as your negatives.

So the idea: is it possible to overexpose your pictures to reduce noise (it's the shadows that are the most difficult)?

Of course, it means shooting manual and using the histogram or several spot metering readings to find out the lighting range of the scene - I mean when overexposing, you do not want to lose important detail.

(Point-and-shoot photographers and those who never heard of zones can disregard this posting and do not have a right to own a good SLR anyway)

Going to try it tomorrow.

Michael Bond
Los Angeles

Scott Larson said:

Albert said:

The E-10 seems to capture a slight overcast - notice the suggestion
of wispy clouds whereas the D30 simply blows away that detail, over
saturation perhaps. From the position of the shadows cast on the
windows, I am assuming the pictures were taken within minutes of
each other.

Click to expand...

They were taken on different days. If you look on the right side of
the building you'll see that a huge stove vent in the D30 picture
is gone!

Click to expand...

Scott, I believe that the stove vent is hidden behind the sign because the angle of the picture is different. It looks like the photographer is has moved to his left about 15 to 20 degrees. Note that the bike and the car have also disappeared but the sun angle is about the same. Same time on different days? Could be.... the sun is slightly higher in the E-10 shot. If so, the question of detail still stands - tough to make comparisons under different light conditions.

Jaja - Belgium Digital said:

Hello Kim,

Don't feel "attacked" by this reply - it's constructive!
I'm only using jpg and I don't have complaints about noise.
So, the jpg is not (as only) responsible for possible noise.
BTW, I thinks most complaints about noise were when using RAW-files.
There's a large difference between "noise" and "jpg-artefacts".
So, in my opinion, there are some E-10's around that produce noise.

Click to expand...

Okay (btw, I don't feel attacked.) I have done a little more comparisons, with some panoramic shots from a mountain in WV. I had done previous shots with a Fuji 2900 (1800x1200 images, stored in jpeg format). With the E-10, I stored in tiff format.

In full automatic mode, the Fuji images are better to the untrained eye. They have better color, more detail and less "noise".

Overall, I like the ergonomics of my E-10 better...I just happen to like the results I get from the Fuji better, when doing distance shots. Of course, I haven't fully learned how to use the Olympus yet, and once I do my preferences may change.

What about for studio portraits will the noise be in peoples faces or on dark backdrops?

Jaja - Belgium Digital said:

Don't feel "attacked" by this reply - it's constructive!
I'm only using jpg and I don't have complaints about noise.
So, the jpg is not (as only) responsible for possible noise.
BTW, I thinks most complaints about noise were when using RAW-files.
There's a large difference between "noise" and "jpg-artefacts".
So, in my opinion, there are some E-10's around that produce noise.
The reason why isn't cleared for the moment.
I also don't think that the E-10 is bought by a lot of people that
don't have a "digital photographic" background (so without
experience).
People who are doing digital imaging for quit a long time really
can appreciate the E-10 quality (the have something to compare ))
BTW, here's an image with real noise: (no, not E-10
It's from my former Canon Pro 70:



That's what I call noise (or jpg-artefacts?)! -
The Pro 70 always has had problems with a nice, blue sky.

Jaja
http://www.belgiumdigital.com

Kim Brennan said:

I've seen a lot of threads here dealing with "Noise" in images
taken by the Olympus E-10 camera. I'm not convinced that there is
anything here other than user error.

First, in the only thread that mentioned it, the commentor was
storing the images in jpeg format. Right there could be the cause
of the "noise". Jpeg is a lossy format. That means that what comes
out, is NOT the same as went in. Some information WILL be lost.
TIFF images as stored by the camera are at least stored without
compression, so you have a better chance of seeing, that which is
due to the camera's limitations rather than the compression
engine's limitations.

Click to expand...

Click to expand...