Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: Noise/ISO underexposure  (Read 1944 times)
wmchauncey
Sr. Member
****
Offline Offline

Posts: 358


« on: September 16, 2013, 12:53:51 PM »
ReplyReply

My Ds3 has a limit, by my standards, of ISO 400>if I take an underexposed image, which was taken at ISO 100, and must apply a +1.00 in LR to bring it up to proper exposure,
 the amount of noise is equal to what I would see from a ISO 400 image without LR exposure adjustment.

Are newer camera bodies that can shoot at higher ISO values handle badly missed exposures more capably than the older bodies...can you boost exposure in LR by +1.00,
assuming an initial setting of ISO 100, without adding visible noise to that image?
Logged

The things you do for yourself die with you, the things you do for others live forever.
A man's worth should be judged, not when he basks in the sun, but how he faces the storm.

My stuff...http://1x.com/member/chauncey43
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1615


« Reply #1 on: September 16, 2013, 01:13:02 PM »
ReplyReply

Are you asking if newer bodies allows you to underexpose more at ISO100 without visible noise becoming bothersome?

I believe that many new-ish Nikon/Sony cameras has higher DR at base ISO, due to lower read noise (and also less visible banding). To the degree that some (e.g. Nikon D7000) are nearly "ISO-less", i.e. you gain (more or less) nothing in raw-file-quality by increasing ISO even when this results in severe underexposure. You might as well always shoot in ISO 100 as long as you shoot raw and have the possibility of raising brightness in your raw developer (and do not care about sensible in-camera preview). As an added bonus, you should have more highlight headroom.

-h
« Last Edit: September 16, 2013, 01:20:56 PM by hjulenissen » Logged
Vladimirovich
Sr. Member
****
Offline Offline

Posts: 1147


« Reply #2 on: September 16, 2013, 01:17:47 PM »
ReplyReply

My Ds3 has a limit, by my standards, of ISO 400>if I take an underexposed image, which was taken at ISO 100, and must apply a +1.00 in LR to bring it up to proper exposure,
 the amount of noise is equal to what I would see from a ISO 400 image without LR exposure adjustment.

Are newer camera bodies that can shoot at higher ISO values handle badly missed exposures more capably than the older bodies...can you boost exposure in LR by +1.00,
assuming an initial setting of ISO 100, without adding visible noise to that image?

also remeber that each camera model and each dcp (DNG) camera profile in ACR/LR might have hidden (not visible in UI) expocorrections (a sum of one for camera and one in profile, that sum might not be zero) - so may start your push/pull from already adjusted data.
Logged
PhotoEcosse
Sr. Member
****
Offline Offline

Posts: 545



« Reply #3 on: September 16, 2013, 02:35:46 PM »
ReplyReply

Not directly related to your question but a couple of observations that might be of interest:

Up to ISO 6400, my D800 is less "noisy" than my D3s - but there is another benefit that is not often commented upon. The "quality of the D800 high-ISO noise is also better than that of the D3s. More like old-style film grain than conventional high-ISO noise. Above ISO 6400, the D800 noise does become more noticeable than that of the D3s.

The Dynamic Range of the D800 sensor is also higher than that of the D3s, meaning that (apart from a number of other benefits), you can pull a bit more detail out of the extremes of the image using the Highlights and Shadows sliders in Lightroom.
Logged

************************************
"Reality is an illusion caused by lack of alcohol."
Alternatively, "Life begins at the far end of your comfort zone."
wmchauncey
Sr. Member
****
Offline Offline

Posts: 358


« Reply #4 on: September 16, 2013, 04:18:12 PM »
ReplyReply

Okay, a follow up question...can a camera's ISO performance degrade over time?
Logged

The things you do for yourself die with you, the things you do for others live forever.
A man's worth should be judged, not when he basks in the sun, but how he faces the storm.

My stuff...http://1x.com/member/chauncey43
PhotoEcosse
Sr. Member
****
Offline Offline

Posts: 545



« Reply #5 on: September 17, 2013, 04:22:05 AM »
ReplyReply

Okay, a follow up question...can a camera's ISO performance degrade over time?

Can't think why it should.

As an amateur, I have never (so far) made more than about 25,000 exposures on any digital camera - and never noticed any degradation of high-ISO performance over that somewhat limited range.

What might happen is that one's expectations change over time. I was looking at some Raws taken with my D80 six or seven years ago. At the time, I thought that the images were "noiseless" at ISO 800. When I look at those same images today, I can definitely see a little noise.
Logged

************************************
"Reality is an illusion caused by lack of alcohol."
Alternatively, "Life begins at the far end of your comfort zone."
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1615


« Reply #6 on: September 17, 2013, 05:08:29 AM »
ReplyReply

Okay, a follow up question...can a camera's ISO performance degrade over time?
I doubt it. Any electronics components will have a finite expected life span, but I would expect problems to manifest themselves as "either-or", rather than gradual degradation. Perhaps colour filters and AF mechanics are exceptions?

-h
Logged
Greg D
Full Member
***
Offline Offline

Posts: 169


« Reply #7 on: September 17, 2013, 07:38:11 AM »
ReplyReply

I believe that many new-ish Nikon/Sony cameras has higher DR at base ISO, due to lower read noise (and also less visible banding). To the degree that some (e.g. Nikon D7000) are nearly "ISO-less", i.e. you gain (more or less) nothing in raw-file-quality by increasing ISO even when this results in severe underexposure. You might as well always shoot in ISO 100 as long as you shoot raw and have the possibility of raising brightness in your raw developer (and do not care about sensible in-camera preview).
-h

True, you can underexpose then brighten without excessive noise. But you'll still have the drawbacks of any underexposed image, i.e. low contrast, dull colors, etc.  Still no substitute for "ETTR" if possible.
Logged
wmchauncey
Sr. Member
****
Offline Offline

Posts: 358


« Reply #8 on: September 17, 2013, 08:39:45 AM »
ReplyReply

I can appreciate the fact that there my be a bit of heightened expectations as my skills have increased, but as the joy stick is broken I'm going to send it in to have it fixed and have them check it out.
Thanks for your feed back guys!      Wink
Logged

The things you do for yourself die with you, the things you do for others live forever.
A man's worth should be judged, not when he basks in the sun, but how he faces the storm.

My stuff...http://1x.com/member/chauncey43
Vladimirovich
Sr. Member
****
Offline Offline

Posts: 1147


« Reply #9 on: September 17, 2013, 08:50:21 AM »
ReplyReply

True, you can underexpose then brighten without excessive noise. But you'll still have the drawbacks of any underexposed image, i.e. low contrast, dull colors, etc.  Still no substitute for "ETTR" if possible.
shooting @ base ISO does not mean underexposure... ISO is not a part of exposure... ISO (or rather gain) is a "postprocessing" parameter that you select before the exposure which for a raw shooter controls the quality (and amount) of applied (if any - it might be just a tag in raw files) analog/digital gain... so you select the optimal gain and then you just expose properly (ETTR if you like)...

PS: that is assuming that your raw converter does a better job with push during raw conversion than ADC/firmware in camera w/ gain/multiplication... does not matter for tag only approach.
« Last Edit: September 17, 2013, 08:53:06 AM by Vladimirovich » Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1615


« Reply #10 on: September 17, 2013, 01:21:00 PM »
ReplyReply

True, you can underexpose then brighten without excessive noise. But you'll still have the drawbacks of any underexposed image, i.e. low contrast, dull colors, etc.  Still no substitute for "ETTR" if possible.
I believe that there is no physical reason to expect those files to be inherently low-contrast, dull colors etc.

Depending on your raw converter, there may be practical reasons why this happens.

Multiplying a number digitally by some number >1 before or after storing it as a raw file should basically be the same thing as long as quantization noise is not an issue (which it seems not to be in this case).

-h
Logged
ErikKaffehr
Sr. Member
****
Online Online

Posts: 6899


WWW
« Reply #11 on: September 17, 2013, 02:25:48 PM »
ReplyReply

Hi,

Much of the noise you see in an image is statistical variation in light. Very little electronics can do about it. This is called shot noise and is only dependent on exposure, that is the number of photons collected. A larger sensor can collect more photons, so with a larger sensor you see less shot noise.

There is also noise from sensor electronics (readout noise) and something called thermal noise. All CMOS sensors use a technique called correlated double sampling that can eliminate a great part of the readout noise, that technique is not possible on CCD. Further, readout noise can be reduced by tightly integrating readout electronics and sensor. Sony and others put the readout electronics on the sensor itself, and use a large number of ADCs (Analog Digital Converters), several thousands of them. So they can do a slow conversion because they do it in parallell. Older designs like Canon and Nikon D3s and D4 use a few external ADCs. On external ADCs, readout noise is much higher, but can be reduced by pre amplification that is increased with ISO. So older designs give less noise at high ISO settings while new designs are less affected by ISO setting.

Somewhat simplified...

Best regards
Erik


My Ds3 has a limit, by my standards, of ISO 400>if I take an underexposed image, which was taken at ISO 100, and must apply a +1.00 in LR to bring it up to proper exposure,
 the amount of noise is equal to what I would see from a ISO 400 image without LR exposure adjustment.

Are newer camera bodies that can shoot at higher ISO values handle badly missed exposures more capably than the older bodies...can you boost exposure in LR by +1.00,
assuming an initial setting of ISO 100, without adding visible noise to that image?
Logged

Misirlou
Sr. Member
****
Offline Offline

Posts: 543


WWW
« Reply #12 on: September 17, 2013, 04:18:11 PM »
ReplyReply

I swear that my older (40D) Canon gets noticeably noisier as it gets hot. I've noticed it when the ambient temperature is very high (which it often is here in the desert), and I think it gets worse the longer I use the camera in a particular shoot, esp. if I'm using Live View a lot.

People tell me I'm wrong about this, but the only way to know for sure would be some sort of controlled test, and I don't have the time or inclination for that at the moment.

Logged
Tony Jay
Sr. Member
****
Online Online

Posts: 2029


« Reply #13 on: September 18, 2013, 04:42:08 AM »
ReplyReply

I swear that my older (40D) Canon gets noticeably noisier as it gets hot. I've noticed it when the ambient temperature is very high (which it often is here in the desert), and I think it gets worse the longer I use the camera in a particular shoot, esp. if I'm using Live View a lot.

People tell me I'm wrong about this, but the only way to know for sure would be some sort of controlled test, and I don't have the time or inclination for that at the moment.

You are not wrong - if the sensor is hot then noise will increase.
Some high-end MF backs actively cool their sensors.

Tony Jay
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« Reply #14 on: September 18, 2013, 06:31:43 AM »
ReplyReply


There is also noise from sensor electronics (readout noise) and something called thermal noise. All CMOS sensors use a technique called correlated double sampling that can eliminate a great part of the readout noise, that technique is not possible on CCD.


Eric,

Thanks for a very good and concise summary of the sources of noise in digital sensors. One inaccuracy is the statement that correlated double sampling is not possible with CCDs. In fact this technique was first implemented on CCDs. To learn more about this technique, interested readers can reference this link dealing with scientific CCDs.

Regards,

Bill
Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3007


« Reply #15 on: September 18, 2013, 07:17:37 AM »
ReplyReply

Eric,

Thanks for a very good and concise summary of the sources of noise in digital sensors. One inaccuracy is the statement that correlated double sampling is not possible with CCDs. In fact this technique was first implemented on CCDs. To learn more about this technique, interested readers can reference this link dealing with scientific CCDs.

Hi Bill,

There is another difference between CMOS and CCD when we look at multiple read-outs, which is probably what Erik was actually referring to. The Read-out of a CMOS sensel signal is virtually non-destructive, so one could perform multiple read-outs and average the results to reduce the read-noise component further with each additional read-out. This will of course slow down the creation of each single image.

A CCD on the other hand, transports its charge out in bucket brigade fashion and collects noise on each shift to the next sensel. After the read-out, the original sensel signal is gone. So only a comparison with the Reset signal is possible.

Cheers,
Bart
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1615


« Reply #16 on: September 18, 2013, 08:37:44 AM »
ReplyReply

There is another difference between CMOS and CCD when we look at multiple read-outs, which is probably what Erik was actually referring to. The Read-out of a CMOS sensel signal is virtually non-destructive, so one could perform multiple read-outs and average the results to reduce the read-noise component further with each additional read-out. This will of course slow down the creation of each single image.
Interesting. Does this mean that Canon _could_ have had the same 14-ish stops of DR at base ISO as Nikon/Sony seems to have, even without fundamental changes in sensor design, simply by taking the time to read sensel values multiple times (surely a faster and more predictable process than taking multiple exposures)? Or is banding noise something fundamentally limiting the benefit (averaging does not seem to be a fix for banding).



-h
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2714



« Reply #17 on: September 18, 2013, 09:45:02 AM »
ReplyReply

Hi Bill,

There is another difference between CMOS and CCD when we look at multiple read-outs, which is probably what Erik was actually referring to. The Read-out of a CMOS sensel signal is virtually non-destructive, so one could perform multiple read-outs and average the results to reduce the read-noise component further with each additional read-out. This will of course slow down the creation of each single image.

A CCD on the other hand, transports its charge out in bucket brigade fashion and collects noise on each shift to the next sensel. After the read-out, the original sensel signal is gone. So only a comparison with the Reset signal is possible.

Cheers,
Bart

Bart,

Thanks for the additional information. Correlated double sampling is one way to reduce read noise, but there are alternatives as discussed in this post which I bookmarked some time ago. Canon was the first major vendor to implement CMOS in their 35 mm sensors and their methodology was discussed in a white paper where they explained that no loss of image quality occurred, but it was unclear to me if they were using correlated double sampling or some other methodology. The Sony Exmoor sensors have very low read noise, but I have never read a clear explanation of the methodology (it may be a trade secret).

Thermal noise is often mentioned along with the suggestion that one should cool the sensor, but for most terrestrial exposures of 1 second or less thermal noise is negligible. Nikon does do some type of dark field subtraction for long exposures, but they also use hot pixel suppression, which does not please the astronomical community, because it also erases faint stars. Parallel processing allows slower readout and is one method used by the Exmoor. Cooling is often used in scientific and astronomical imaging, but these applications use slow readout to reduce read noise, but this allows additional time for dark noise to accumulate and makes cooling more necessary.

Hardware pixel binning can also reduce noise, but, as far as I know, the only company to achieve this is Phase One with their sensor plus technology. Apparently this is not feasible with CMOS. One can pixel bin in post (e.g.) 2:1 downsizing, but this involves 4 read noises rather than the one read noise that occurs with hardware binning where the superpixel is read out with only one read noise.

This is interesting to geeks such as myself and I would be interested in further discussion, but the general forum population likely has little interest in this arcane topic.

Regards,

Bill
Logged
Vladimirovich
Sr. Member
****
Offline Offline

Posts: 1147


« Reply #18 on: September 18, 2013, 11:48:34 AM »
ReplyReply

Thermal noise is often mentioned along with the suggestion that one should cool the sensor, but for most terrestrial exposures of 1 second or less thermal noise is negligible.

does that account for live view though ?
Logged
Vladimirovich
Sr. Member
****
Offline Offline

Posts: 1147


« Reply #19 on: September 18, 2013, 11:51:52 AM »
ReplyReply

Interesting. Does this mean that Canon _could_ have had the same 14-ish stops of DR at base ISO as Nikon/Sony seems to have, even without fundamental changes in sensor design, simply by taking the time to read sensel values multiple times (surely a faster and more predictable process than taking multiple exposures)? Or is banding noise something fundamentally limiting the benefit (averaging does not seem to be a fix for banding).



-h

you probably need some form of global shutter then... you have to store the goods somewhere on die, because delivery of them (for modern resolutions) off die takes fractions of second still, is it not ?
Logged
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad