D60 noise test

[Bob Reim]

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera at long exposures. This will give an indication of how good the imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this is called a dark frame). Set your ISO speed at the slowest level and take a shot in bulb mode. You can have your noise reduction on or off (if the camera has that option) so you can see the difference. Take a shot of several minutes or longer. The more noise the imager has the more 'snow' the final image will have in it. You can also do this at different ISO levels to see how the camera does. At high ISO values it should get noisey very quickly. I am interested to see the results of this test. It would also be interesting how other D60 will compaire at the same settings. Also please note that image noise doubles for every 8 degree C rise in temperature so it would be nice to know about what temp the shot was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg C below room temp. This way they can get exposures over several hours long. It would be interesting to see how the D60 does.

 

[Phil Askey]

You're wasting your time here, the D60 has built-in automatic noise reduction, you can't disable it. In any case the D60 has nothing like the noise you may have seen of other digital cameras (I think you may be thinking along the E-10 / E-20 lines).

Check out my 4 minute exposure:

http://www.dpreview.com/reviews/canoneosd60/page13.asp

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[Ed Freedenburg]

Yesterday I did the exact test, Cap on iso 100, bulb 4 mins and it looked like a technicolor Milky Way. I'll try to upload the image for you.

I was expecting like your 4 min exposure about 10 pixels out of whack, mine had one out of 200 out of whack.

Check out my 4 minute exposure:

http://www.dpreview.com/reviews/canoneosd60/page13.asp

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[Ed Freedenburg]

This is my first try for uploading images to pbase. They 0002.jpg is 4 minutes at iso 100, 00004 is 5 minutes at iso 100.

You have to downlod the target and view at 100%. First try here so help me out.

http://www.pbase.com/efreedenburg/d60_test_shots

I was expecting like your 4 min exposure about 10 pixels out of
whack, mine had one out of 200 out of whack.

Check out my 4 minute exposure:

http://www.dpreview.com/reviews/canoneosd60/page13.asp

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[no name 3]

Ed,

When I clicked on your link, I got two images that looked pretty much like pure black on my monitor. Of course, it may be that my monitor isn't completely calibrated, but I expect to still see some noise if it's as bad as you say.

I think on the D30, a four minute exposure would've given you an almost white screen that looks more like TV static.

You have to downlod the target and view at 100%. First try here so
help me out.

http://www.pbase.com/efreedenburg/d60_test_shots

I was expecting like your 4 min exposure about 10 pixels out of
whack, mine had one out of 200 out of whack.

Check out my 4 minute exposure:

http://www.dpreview.com/reviews/canoneosd60/page13.asp

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[Bob Reim]

The reason I am considering this camera for astro work is due to your 4 minute exposure. Some cameras do give the option for turning off the noise reduction, I prefer to leave it on for long exposures. My current camera does a excellent job on 60 second exposures with noise reduction (can't be turned off). I am interested to see just how good this camera is. I am very familure with imagers (CCD and CMOS) as I did research on this several years ago when I was designing a digital camera (I am an electrical engineer). I knew all the major manufactures of imagers (well about 2 years ago anyway) and I spoke with many of the design engineers of these imagers about specs. Anyway the fact that Canon uses CMOS and can do long exposures is very impressive. The quantum efficiency of CCD is better than CMOS so I would expect that it would take a longer exposure with a CMOS imager to get the same result of a CCD with the same pixel size. This noise test can tell a lot about the quality of the imager. There are several sources of noise besides dark current (which this noise test will show) and it will also tell how good the noise reduction is.

Ed: Thanks for taking the time to do this test. I thought the results looked very good. I thought you said that the results did not look very good? Is the images you posted the same one you said looked very noisy? I would like to see longer exposures too if you have the time. Unless the noise reduction is removing almost everything (meaning will not leave any signal left over for an image) this camera seems very good. Another test would be to take a 5 minute shot on a known object (like a dark room where the light is very controled) and see what happens. There should be just a little light in the room (very little or to the point where you really can't see your way around at all) . This may show if the noise reduction is removing everything (including signal) or the imager is just really good.

Thanks.

Check out my 4 minute exposure:

http://www.dpreview.com/reviews/canoneosd60/page13.asp

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[Jason Hutchinson]

Are you kidding? Your samples are way better than what you typically see from a digicam. Heck, I doubt there is a 1D that could ever do that.

Jason

I was expecting like your 4 min exposure about 10 pixels out of
whack, mine had one out of 200 out of whack.

Check out my 4 minute exposure:

http://www.dpreview.com/reviews/canoneosd60/page13.asp

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[Bryan Biggers]

Those are VERY VERY good. Best I've seen from any consumer digital camera, and the super low light performance is the only reason that I'm considering the D60 really.

I'd love to see more samples. Id REALLY like to see some star trails from a dark site. Bryan

 

[murphy]

Ed,

When I clicked on your link, I got two images that looked pretty
much like pure black on my monitor. Of course, it may be that my
monitor isn't completely calibrated, but I expect to still see
some noise if it's as bad as you say.

I think on the D30, a four minute exposure would've given you an
almost white screen that looks more like TV static.

When you go to the page and click on a thumbnail, go to the text at the bottom of the picture and click on "Original"

A very LARGE image will download. Scroll around the image and you can see the bad pixels. 99.9999% (most of them) are solid black. (as they should be) But if you look carefully, you'll see small specks.. Some larger, some smaller and different colors.

This is called "Noise" because they should all be black with the lens cap on no matter how long you take the picture. (Theoretically)

Murph

 

[Bryan Biggers]

Very interesting banding pattern on the 300 sec exposure, esp. along the bottom and left edge. Of course those things would subtract right out when you subtract your dark frame.

Question: Does the D60 take a second exposure that is as long as the first for the noise reduction? In other words did it take 10 minutes to take the 5 minute exposure?
Bryan

Those are VERY VERY good. Best I've seen from any consumer digital
camera, and the super low light performance is the only reason that
I'm considering the D60 really.

I'd love to see more samples. Id REALLY like to see some star
trails from a dark site. Bryan

 

[rcliff]

It's just like watching Star Trek Next Gen on TNN with the bad film bright right top and bottom corners. Plus you get the stars too!!! To boldly go where no one has gone before.

You have to downlod the target and view at 100%. First try here so
help me out.

http://www.pbase.com/efreedenburg/d60_test_shots

I was expecting like your 4 min exposure about 10 pixels out of
whack, mine had one out of 200 out of whack.

Check out my 4 minute exposure:

http://www.dpreview.com/reviews/canoneosd60/page13.asp

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[MarkK]

Bob,

There are 81 pixels with a luminance above 75 for your 4 min. exposure. There are 152 in the 5 min. exposure.

Not bad compared to similar numbers for a 10 second exposure on an E-10.

Regards,

Mark K.

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[Bob Reim]

I looked at the orginals and the results are good! I noticed that the 4 minute shot and the 5 minute shot had many of the same hot pixels. This can be easily removed with a dark frame. So I guess the D60 does not do a dark frame otherwise these would not be here. If I recall from Phills review he said that the camera does not do a dark frame. This is very interesting.

There are 81 pixels with a luminance above 75 for your 4 min.
exposure. There are 152 in the 5 min. exposure.

Not bad compared to similar numbers for a 10 second exposure on an
E-10.

Regards,

Mark K.

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[MarkK]

Bob,

I agree. The most pronounced (brightest) pixels are duplicated, which is good news. Dark frame subtraction should work wonderfully.

How much time did you wait between the 4 min and the 5 min exposure?

I wonder how constant the 'hot pixels' are? If they are reasonably constant, you could just use an archived dark frame of the approximate exposure time for all but the most critical work.

Regards,

Mark K.

There are 81 pixels with a luminance above 75 for your 4 min.
exposure. There are 152 in the 5 min. exposure.

Not bad compared to similar numbers for a 10 second exposure on an
E-10.

Regards,

Mark K.

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[Hartmut59261]

If you want to see a really bad "lowlight performance", try the same test with the 1D!
Regards
Hartmut

 

[Phil Askey]

You want me to be brutal?

That's a FOUR MINUTE EXPOSURE... Show me another digital camera under $5000 that can do that!

This is my first try for uploading images to pbase. They 0002.jpg
is 4 minutes at iso 100, 00004 is 5 minutes at iso 100.

You have to downlod the target and view at 100%. First try here so
help me out.

http://www.pbase.com/efreedenburg/d60_test_shots

 

[Hartmut59261]

You want me to be brutal?

That's a FOUR MINUTE EXPOSURE... Show me another digital camera
under $5000 that can do that!

Have you ever tried the same with the 1D?
Hartmut

 

[Phil Askey]

Erm, I think that's what I said.

You want me to be brutal?

That's a FOUR MINUTE EXPOSURE... Show me another digital camera
under $5000 that can do that!

Have you ever tried the same with the 1D?
Hartmut

 

[Duane DeSieno]

When you use photoshop levels to exadurate the noise, you see what might be light leakage on the right hand side of the frame.

I get something similar on a 30 sec exposure on my Minolta d7 along the top of the frame.

Still, it's very impressive performance.

Duane

Since I did not get any replies to this last post I will try it again.

I would like to see some one do some noise testing of this camera
at long exposures. This will give an indication of how good the
imager is (for low noise). The test is simple

Leave your lens cap on so you are taking a shot of darkness (this
is called a dark frame). Set your ISO speed at the slowest level
and take a shot in bulb mode. You can have your noise reduction on
or off (if the camera has that option) so you can see the
difference. Take a shot of several minutes or longer. The more
noise the imager has the more 'snow' the final image will have in
it. You can also do this at different ISO levels to see how the
camera does. At high ISO values it should get noisey very quickly.
I am interested to see the results of this test. It would also be
interesting how other D60 will compaire at the same settings. Also
please note that image noise doubles for every 8 degree C rise in
temperature so it would be nice to know about what temp the shot
was taken.

Image noise is why cameras for Astro imaging are cooled to -30 deg
C below room temp. This way they can get exposures over several
hours long. It would be interesting to see how the D60 does.

 

[Samir Kharusi]

Bob,

I agree. The most pronounced (brightest) pixels are duplicated,
which is good news. Dark frame subtraction should work wonderfully.

How much time did you wait between the 4 min and the 5 min exposure?
I wonder how constant the 'hot pixels' are? If they are reasonably
constant, you could just use an archived dark frame of the
approximate exposure time for all but the most critical work.

Hi all! I am detecting a bit of unjustified over-enthusiasm in this thread. I am pretty well versed with astrophotography using film, an SBIG 237 astro CCD (yes, chilled) and the D30. To gauge my experience level feel free to visit my website:
http://www.geocities.com/samirkharusi/index.html

Sorry in advance if bandwidth rationing shuts you off. Try again later if you are a fellow astronerd. OK, here are my comments for the cognoscenti. I tried the usual noise tests with and without the D30's built-in dark frame substractor. It works. Perhaps the D60 just uses a built-in library for its subtracting, whatever. I have not managed to find a way to process the D30 RAW file in a fashion similar to processing an SBIG native format file. Ie you cannot shoot a D30 at ISO 100, fiddle with "Background" and "Range" like for normal SBIG CCD image files and end up with the same result as when you shoot at ISO 1600. Astro CCDs do not have the silly film-carryover-concept of ISO in their baggage! For high sensitivity you actually have to use high ISO in a D30. Using ISO 1600 and a few seconds exposure at 25 deg C you always have noise visible in the images. This noise can easily be processed out by simply shifting the toe for your histogram, basically adjusting the "Background". For low light astro images (ie not lunar/planetary) you then need to stretch contrast, ie bring down the top end of your histogram, ie restricting the "Range". This obviously will immediately make any residual noise prominent. Keeping these constraints in mind I found that the longest usable exposure at ISO 1600 in my D30 at 25 deg C to be 8 seconds. This is probably roughly equivalent to 25 to 30 seconds on ISO 800 film (taking film's reciprocity failure into account). This is enough to get a reasonable image of the central part of M42 (Orion Nebula) but NOT the outer reaches. By the way, on Daylight White Balance the central part of M42 comes out bluish, not red as I expected... Four seconds exposure at ISO 1600 is more pleasant to use. In an 8" SCT at f10 you get down to Magnitude 11 stars. Honest. Check out my photos of Vesta on my website. For a reason that I cannot as yet comprehend, adding two 4 second exposures in a D30 results in a better signal-to-noise ratio than a single 8-second exposure. Yes, I am quite aware of taking multi-exposures in astro work, see my image of M31 using the standard Canon EF 50/1.8 with the SBIG 237 using, if I recall correctly, 60 exposures.

In brief, noise in D30 images has to be talked about in the context of temperature, ISO setting and exposure length. Noise that is hardly visible in a normal night streetscape becomes HUGE when in the context of the contrast stretching that is required in bringing out a faint nebula. For these later uses I find the limit of a D30 is 8 seconds at ISO 1600 at 25 deg C. It is NOT suitable as a nebula camera. Pity. It is excellent as a lunar/planetary camera, my site has D30 samples of the Moon, Jupiter, Saturn, Mars and comparing them to film and/or the SBIG 237. Have fun, but unfortuantely not with the faint fuzzies ;-)