| Author | Thread |
|
|
04/10/2006 12:34:38 PM · #1 |
Since I underestimate the exposure and meter a lot, I found that the noise level due to ISO settings is more predominant in even the slightly underexposed scenes - and almost non-existant in the dead on exposures; but it has to be dead on.
Is this observation just a coincidence or is this common? If it's common what can some of the techies here attribute it to?
-Rick |
|
|
|
04/10/2006 12:44:08 PM · #2 |
yes. you get more noise4 in teh shadow areas.
why are you underexposing? How are you metering?
First, use incident metering only, as reflective metering is inacurate (black vs white surfaces will give different reflective readings, but with a meter on incident you get one correct reading). So the in-camera metering is wrong all the time - it's just guessing. Your 20D does not have spot metering...spot metering reflective wise is not too bad if you can estimate 18% gray or are trying to expose for a specific spot ( use skin (face), a highlight or hot spot and things are usually pretty good)
Shoot raw - you get about 2 stops either way. I have not tested the noise issue on a RAW image that I pushed 2 stops vs a JPG, but I suspect it's tons better.
|
|
|
|
04/10/2006 01:11:41 PM · #3 |
Originally posted by CarpeNoctem:
Since I underestimate the exposure and meter a lot, I found that the noise level due to ISO settings is more predominant in even the slightly underexposed scenes - and almost non-existant in the dead on exposures; but it has to be dead on.
Is this observation just a coincidence or is this common? If it's common what can some of the techies here attribute it to?
-Rick |
No, it's true. The dark areas will show more noise, always. There's an explanation for it, having to do with photosite sensitivities and how much of the gamut is captured at low exposures, or something like that, but I'm not the one to explain it. Kirbic is. Kirbic knows all this stuff.
R.
|
|
|
|
04/10/2006 01:26:32 PM · #4 |
I've noticed that, and that overexposed areas don't have the noise. I suggest reading a few of the many threads about "expose to the right", or read //www.luminous-landscape.com/tutorials/expose-right.shtml and //www.adobe.com/digitalimag/pdfs/linear_gamma.pdf
Message edited by author 2006-04-10 13:32:09. |
|
|
|
04/10/2006 01:28:47 PM · #5 |
Very simply put: signal to noise ratio. The noise is actually very consistant across the entire sensor but is more evident where the signal amplitude is low, ie. darker areas. Think of each pixel in the sensor as a bucket. The signal consists entirely of electrons in these buckets produced by photons. But there are other sources of electrons that continually leak into the buckets that we perceive as noise, such as thermally generated random electrons. Longer exposures, elevated temperatures, and high ISO settings* will agravate the production of noise and reduce the signal to noise ratio.
*Higher ISO settings introduce other spurious noise from the processing electronics. |
|
|
|
04/10/2006 01:41:14 PM · #6 |
If anybody remembers tape recording, you have the same effect, where there's more evident noise during a soft cello passage than listening to Jimi Hendrix ...
Turn up the volume control and you'll really hear the hiss.
Message edited by author 2006-04-10 13:42:18. |
|
|
|
04/10/2006 01:43:37 PM · #7 |
Originally posted by ElGordo:
Very simply put: signal to noise ratio. The noise is actually very consistant across the entire sensor but is more evident where the signal amplitude is low, ie. darker areas. Think of each pixel in the sensor as a bucket. The signal consists entirely of electrons in these buckets produced by photons. But there are other sources of electrons that continually leak into the buckets that we perceive as noise, such as thermally generated random electrons. Longer exposures, elevated temperatures, and high ISO settings* will agravate the production of noise and reduce the signal to noise ratio.
*Higher ISO settings introduce other spurious noise from the processing electronics. |
Remember also that the eye sees logrithmically, while the sensor sees linearly. So the eye sees a greater percieved difference between the two darkest shades that the sensor measures than between the two brightest shades. (We meter 18% grey because this is what we percieve as halfway between black and white).
Noise affects all readings equally. If a sensor has values of 1 (black) to 100 (white) (not a very good sensor at all), and the noise at one pixel is "3", you probably won't notice it if the reading is 99 when it should be 96, but if the reading is 4 and it should be 1, you'll percieve a huge difference.
Message edited by author 2006-04-10 13:45:01. |
|
|
|
04/10/2006 01:44:42 PM · #8 |
Originally posted by ElGordo:
Very simply put: signal to noise ratio. The noise is actually very consistant across the entire sensor but is more evident where the signal amplitude is low, ie. darker areas. Think of each pixel in the sensor as a bucket. The signal consists entirely of electrons in these buckets produced by photons. But there are other sources of electrons that continually leak into the buckets that we perceive as noise, such as thermally generated random electrons. Longer exposures, elevated temperatures, and high ISO settings* will agravate the production of noise and reduce the signal to noise ratio.
*Higher ISO settings introduce other spurious noise from the processing electronics. |
I have a fair understanding of the 'signal to noise ratio' concept - but as far as a digital camera is concerned I was under the impression that the noise portion was added to the signal (a junky analogy).
The ISO settings above the first (100) is basically an amplifier. So, at high ISO 1600-3200 (H) we have high amplification which this concept is really 'noisy'. When a photo site starts filling - it fills on top of the noise?
I know this really doesn't matter, but I find it interesting. Interesting that if the exposure is dead on, the image is virtually noiseless - even at 3200.
Originally posted by Prof_Fate:
why are you underexposing? How are you metering?
|
I love the histogram, too much. I have found that it's easy to shoot by histogram. But 90% of the time shooting by histogram results in an technically improper exposure.
Consider trying to photograph a low key portrait. You want your black velvet background to show completely black with no detail and you want you subject's face to expose in Zone 6. Taking a shot with the histogram for a proper exposure 'probably' will show a lot of blown shadows and a spike at Zone 6.
Shooting by histogram may dictate you adjust exposure to efficiently use the dynamic range of the camera and you may open up a 1/2-1 stops to 'open' the shadows. When in fact, this is a technical over exposed image.
I do this ALL the time. This waterfall I LOVE to photograph I am always underexposing it, because there are specular highlights where the water crests the top of the falls,and more importantly the sky through the trees become blown due to the nature of the surroundings.
Or, indoor arena's have windows and lights in the background that blow out, and I compensate by underexposing.
In all of these, the blow outs are unimportant because they are small lights and it's the way it is, or the blown sky is to be cropped out anyway.
So I have a tendency to under expose when I shoot by histogram.
-Rick |
|
|
|
04/10/2006 01:50:43 PM · #9 |
Originally posted by CarpeNoctem:
I love the histogram, too much. I have found that it's easy to shoot by histogram. But 90% of the time shooting by histogram results in an technically improper exposure.
Consider trying to photograph a low key portrait. You want your black velvet background to show completely black with no detail and you want you subject's face to expose in Zone 6. Taking a shot with the histogram for a proper exposure 'probably' will show a lot of blown shadows and a spike at Zone 6.
So I have a tendency to under expose when I shoot by histogram.
-Rick |
How about if you use a gray card (or even better, a black, grey and white card) to set your exposure for the shot? |
|
|
|
04/10/2006 01:57:27 PM · #10 |
The cameras electronic processing cannot distinguish between electrons produced by noise sources and those produced by photons. Yes, they are additive. That is precisely why the noise is more evident in darker areas, showing up as lighter pixels. The ratio of the electrons produced by the photographed scene and those produced by other sources is a critical factor in producing a useable image.
Message edited by author 2006-04-10 13:58:20. |
|
|
|
04/10/2006 01:59:22 PM · #11 |
Originally posted by hankk:
How about if you use a gray card (or even better, a black, grey and white card) to set your exposure for the shot? |
Got some, but it's not always prudent. Especially with a 'histogam' *wink wink*.
It's not that I don't know how to expose, its that I am trusting the histogram too much. Also, I'm not putting the same amount of thought into each shot as I did film.
What I need to do is observe the scene and note that I may have blown shadows or highlights and determine if that's what I want or not.
-Rick |
|
Home -
Challenges -
Community -
League -
Photos -
Cameras -
Lenses -
Learn -
Help -
Terms of Use -
Privacy -
Top ^
DPChallenge, and website content and design, Copyright © 2001-2025 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 12/31/2025 03:36:09 PM EST.
