| Author | Thread |
|
|
10/31/2003 09:51:35 AM · #1 |
John's post on 7 minute exposures got me thinking about the 10-D. More specifically, how the 10-D handles digital noise and hot pixels at longer exposures.
Below is a 17 minute 46 second image of some powerlines and their support. These things were sizzling and crackling like crazy, but I thought it was picturesque enough for my needs so I shot it anyway. Lighting conditions were LOW cloud cover and farms about a mile away in any direction.
Here are the particulars:
Canon EOS 10D
Shooting Date/Time: 10/31/2003 2:21:41 AM
Shooting Mode: Manual
Tv( Shutter Speed ): 1046 seconds (17min./46sec.)
Av( Aperture ): 2.8
ISO Speed: 100
Lens: 16mm Zenitar Fisheye (manual)
If you check out the 2nd thumbnail below "Coronal Loss Detail" you'll see a full res crop of the digital noise as well as 2 of the 10 hot pixels in the image. You may also notice some blue light around the insulators and on the actual power lines themselves. These are not hot pixels, these are sparks...apparently causing the crackling/buzzing sound (also called Coronal Loss). Yes the image is a bit blurry, but the Zenitar isn't what you'd call sharp glass. ;)
Of course this is only one image of some clouds with a mediocre lens so it isn't much of a test, although it sure is fun!
Message edited by author 2003-10-31 09:52:07.
|
|
|
|
10/31/2003 10:50:44 AM · #2 |
Hey Q-jet, I see you've got the 16mm Zenitar Fisheye! How do you like the results? I have the 8mm Peleng myself, and I'm fairly impressed, especially for the price...
Really cool image of the coronal discharge! A close-up of this would have been killer in the Science challenge. Could you actually see any of the coronal discharge with the naked eye? I'd think not, but not sure.
Might be interesting to try it with multiple exposures stacked to reduce noise, and maybe a dark frame subtraction to banish the hot pixels.
BTW, the first thing that comes to mind when I see your "handle" is a GM carburetor...
|
|
|
|
10/31/2003 11:06:25 AM · #3 |
Is that supposed to happen? Kind of freaky! Great picture. I wonder what the electric company would think of that?
|
|
|
|
10/31/2003 11:10:03 AM · #4 |
Originally posted by kirbic:
Hey Q-jet, I see you've got the 16mm Zenitar Fisheye! How do you like the results? I have the 8mm Peleng myself, and I'm fairly impressed, especially for the price...
Really cool image of the coronal discharge! A close-up of this would have been killer in the Science challenge. Could you actually see any of the coronal discharge with the naked eye? I'd think not, but not sure.
Might be interesting to try it with multiple exposures stacked to reduce noise, and maybe a dark frame subtraction to banish the hot pixels.
BTW, the first thing that comes to mind when I see your "handle" is a GM carburetor... |
Ya know, I still have an 8mm peleng on order...apparently they forgot to send it to me or something. I should really look into that sometime. The Zenitar works pretty well for a $150 lens. I shoot quite a bit of BMX/Freestyle stuff and that pretty much requires a really wide angle lens, wider than the 16mm + 1.6x (25mm) of the Zenitar. Do you have any examples using the 8mm?
Hmmm, a close up of the discharge...great idea! I'll try to get back out there soon, before it gets TOO cold!
About my user name, you're right! I named myself after the carb on my '76 Caprice.
|
|
|
|
10/31/2003 11:43:19 AM · #5 |
I'm curious why you chose to use an aperture of f/2.8 for the original photo? I bet if you did the same photo with a smaller aperture, your noise level would be lower. Just a guess though...
|
|
|
|
10/31/2003 11:50:51 AM · #6 |
Q-jet, I will try to post some samples from the 8mm sometime soon.
|
|
|
|
10/31/2003 11:51:19 AM · #7 |
Originally posted by jmsetzler:
I'm curious why you chose to use an aperture of f/2.8 for the original photo? I bet if you did the same photo with a smaller aperture, your noise level would be lower. Just a guess though... |
But a smaller aperture is going to mean an even slower shutter speed for an equivalent exposure - so more noise, not less. |
|
|
|
10/31/2003 12:00:05 PM · #8 |
Originally posted by jmsetzler:
I'm curious why you chose to use an aperture of f/2.8 for the original photo? I bet if you did the same photo with a smaller aperture, your noise level would be lower. Just a guess though... |
Like Gordon mentioned, I figured I could get the most bang for my exposure by going with a larger aperture. It would be interesting to find out if there's much of a difference at that interval using a smaller aperture.
The next time I go out, I'll try to make multiple exposures @ 15 min and use different apertures. Ahh if I could only find the time.
|
|
|
|
10/31/2003 12:24:50 PM · #9 |
| There isn't really a reciprical relationship failure for digital, so the aperture/ shutter speed relationship still holds - for the same exposure if you decrease the aperture by one stop, you double the shutter speed, so to get the same exposure at f4 for example, it would be a 34 minute shot. |
|
|
|
10/31/2003 12:47:42 PM · #10 |
| Seems to me like you'd want to make your aperture decisions on an individual shot basis. Wild idea? I don't have much experience with long exposures, but depth of focus and lens sharpness at different apertures would of course have an impact just as in shorter exposures. Compositionally Quadrajet's powerline shot benefitted from a wide aperture for the shorter exposure time and the reduced evidence of any dust on the sensor (there's always something) at the expense of lens sharpness from being slightly stopped down. A static subject with lots of sky is a lot different than a street shot, for example, where moving objects dictate that exposure time would be of the utmost importance, and aperture and ISO would then be adjusted accordingly. |
|
|
|
10/31/2003 12:58:11 PM · #11 |
but for anything over 5 minutes, the noise issues start being more critical than depth of field and so on. Doubling a shutter speed from 1/60s to 1/30s might make a difference in stopping motion, but there is no real noise impact. 17 minutes to 34 minutes is a huge difference in noise.
I found with my D60, that 5 minutes was the threshold for useful results, going beyond that produced far too much noise in the image.
Here's a close-up from a 345s exposure
//www.pbase.com/image/21744374/original
Message edited by author 2003-10-31 12:58:56. |
|
|
|
10/31/2003 01:05:50 PM · #12 |
Are you sure that's the case with the 10D? Though I've seen nothing on a side-by-side comparison it seems like there's significantly less noise on the 10D exposures than on your D60 exposures. No battery noise on that powerline shot at all, whereas your bridge shot had significant blooming on the battery side. Might well be less general sensor noise, too, though not sure.
Originally posted by Gordon:
but for anything over 5 minutes, the noise issues start being more critical than depth of field and so on. Doubling a shutter speed from 1/60s to 1/30s might make a difference in stopping motion, but there is no real noise impact. 17 minutes to 34 minutes is a huge difference in noise.
I found with my D60, that 5 minutes was the threshold for useful results, going beyond that produced far too much noise in the image.
Here's a close-up from a 345s exposure
//www.pbase.com/image/21744374/original |
|
|
|
|
10/31/2003 01:09:02 PM · #13 |
For anything longer than a few minutes, I would without doubt break the exposure into multiple shots, and stack them. The difference is very significant.
|
|
|
|
10/31/2003 01:22:26 PM · #14 |
Originally posted by jimmythefish:
Are you sure that's the case with the 10D? Though I've seen nothing on a side-by-side comparison it seems like there's significantly less noise on the 10D exposures than on your D60 exposures. No battery noise on that powerline shot at all, whereas your bridge shot had significant blooming on the battery side. Might well be less general sensor noise, too, though not sure.
|
Like you said - you aren't seeing side-by-sid comparisons. The 17 minute shot above was shot in Minnesota, in late Fall - I'm guessing its cold at night there. ( based on temps tonight at 2AM the temp was about freezing - that's a 30C ambient temp difference for a start!) The 6 minute D60 shots were taken in summer in Texas - it was probably still about 80F when I took those shots maybe closer to 90F - I was in shorts/ t-shirt certainly. I'd also been shooting for about an hour so there is significant heat generated from the battery - especially as I'd probably taken about 30 minutes worth of continuous long exposures by that point.
The 10D is less noisy, but I doubt very much less. It would be interesting to see a 100% sample from that 17 minute shot if that's possible, and with out any despeckling or neat image adjustments.
The temp differences would account for about 10% more noise straight away, just for kT/C noise
Message edited by author 2003-10-31 13:32:48. |
|
|
|
10/31/2003 01:29:36 PM · #15 |
Interesting...though if this is true (which I'm sure it is) then we can go back to saying that noise from battery and sensor will still indeed be less of a factor in shots in cold conditions and where the battery is not heated significantly, and again say that aperture and shutter speed might be used in the same manner as in a shorter exposure ;).
|
|
|
|
10/31/2003 01:35:18 PM · #16 |
Not really, because one of the major noise sources on CMOS sensors is the dark current shot noise - which is basically dependent on exposure time.
So in cold locations, the thermal noise is reduced but the DCS noise isn't changed.
Though like for all of these sorts of discussions, nothing beats just trying it and seeing how it works out
Message edited by author 2003-10-31 13:42:22. |
|
|
|
10/31/2003 01:45:35 PM · #17 |
| For sure...I for one would like to make a comparison with and without the battery grip, too, to see if that makes a difference. Perhaps I'll get around to that soon. |
|
|
|
10/31/2003 01:52:03 PM · #18 |
Originally posted by jimmythefish:
For sure...I for one would like to make a comparison with and without the battery grip, too, to see if that makes a difference. Perhaps I'll get around to that soon. |
Yup - I've got that grip recently too and was wondering if that would help out with the edge blooming. |
|
|
|
10/31/2003 03:05:14 PM · #19 |
Originally posted by Gordon:
Yup - I've got that grip recently too and was wondering if that would help out with the edge blooming. |
Nope, it doesn't. I don't use my camera without the grip. Below is a 15 minute lens cap shot. The camera had been in use for an hour or more, but it was a relatively cool summer evening (early August in Northern Ontario). You can seem some clearly defined circular blooms (contact points?).
 |
|
|
|
10/31/2003 03:17:26 PM · #20 |
I did wonder if it was really the battery heat causing the problem or other components - the battery grip just moves the battery but doesn't obviously change where the power goes in the camera. Thanks for that picture - I guess it clears up any doubt that the grip might help much.
I notice my CF cards get warm after extended use of the camera too - it seems to be circuitry in the hand grip area that warms up certainly - as shown by that lens cap shot too. |
|
|
|
10/31/2003 03:27:24 PM · #21 |
Maybe I'm confused, but I started thinking about the "battery heating" theory, and I wanted to throw this thought out there...
If it's battery heat, it should be coming from the camera's right side. Now consider the image projected by the lens on the sensor. It is inverted and reverted (proved this to myself by projecting thru a lens onto a sheet of paper). This means that the "right" side of the camera actually corresponds to the left side of the image as viewed on the camera LCD or computer. Yet we always see the "bloom" on the right.
Unless I'm going crazy (which is certainly possible) this puts the battery heat theory out the window.
edit: get my left and right straight!
Message edited by author 2003-10-31 15:28:47.
|
|
|
|
10/31/2003 03:36:33 PM · #22 |
Originally posted by kirbic:
Maybe I'm confused, but I started thinking about the "battery heating" theory, and I wanted to throw this thought out there...
If it's battery heat, it should be coming from the camera's right side. Now consider the image projected by the lens on the sensor. It is inverted and reverted (proved this to myself by projecting thru a lens onto a sheet of paper). This means that the "right" side of the camera actually corresponds to the left side of the image as viewed on the camera LCD or computer. Yet we always see the "bloom" on the right.
Unless I'm going crazy (which is certainly possible) this puts the battery heat theory out the window.
edit: get my left and right straight! |
It could be where the power supply and external interface for the sensor are located... The sample above with the battery grip on shows it isn't the battery as such, but the thermal heat could be coming from elements within the sensor array itself or components near-by on the board.
Message edited by author 2003-10-31 15:37:28. |
|
|
|
10/31/2003 04:25:49 PM · #23 |
Okay, this post has nothing to do wiht photography or batteries, but does have something ot do with electricity.
This past summer I visited a friend in Indiana who had the HUGE power tower things next to her yard. I was trying to get a sunset picture without one of these beasts in it, so I walked out into the field underneath the line. I could here the "singing" but didn't think that much of it until I could feel my hair standing up. Then, when I rubbed my finger across my camera, it had that static feeling like a TV does. Very strange.
My friend said she has been lightly shocked occasionally while walking under the lines.
Like I said. Nothing related to photography, just felt like talking.
Continue on.
|
|
|
|
10/31/2003 04:36:07 PM · #24 |
| What the hell do you do for 17 minutes???? |
|
|
|
10/31/2003 04:39:46 PM · #25 |
That's kinda where my thoughts were going, especially after looking carefully at dwoolridge's image. There are clearly several points from where the noise "radiates". There appear to be 5, and they are almost perfectly arranged along the sensor edge, symmetricaly about the center.
Certainly there are heat sources such as the processor, which should put out more heat than the battery ever will. I'm not sure of the location of the major heat sources, but I'm surprised that the effect is so uniform along the sensor edge. It's possible that heat is being transferred up to the sensor along mounting lugs that are soldered into the board, though again I have no idea if there are even mounting points in those locations, much less soldered ones.
|
|
Home -
Challenges -
Community -
League -
Photos -
Cameras -
Lenses -
Learn -
Prints! -
Help -
Terms of Use -
Privacy -
Top ^
DPChallenge, and website content and design, Copyright © 2001-2024 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 04/19/2024 10:05:46 AM EDT.
