Luminous Landscape Forum
Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: BernardLanguillier on August 25, 2003, 09:57:31 pm
-
Isn't there also an effect on the noise removal alorithm used by the camera?
Different ISO inducing different noise patterns/spatial frequences?
Just guessing though.
Best regards,
Bernard
-
To answer one of Danno's questions, a few sample figures for the number of electrons that a photo-site can hold, from Kodak's sensor web site, are
- about 50,000 in the 6.8 micron photo-sites of the Kodak full-frame style CCD sensor in the Olympus E-1, and also in the 8.8 micron photo-sites of an 11MP 35mm sized interline CCD sensor.
- about 100,000 in the 8.8 micron pixels of the Kodak full-frame style CCD sensor in their 16MP digital back (the one Michael recently reviewed and bought).
50,000 already needs 16 bits to count, but the thermal noise levels are around 20 electrons so there is not much point counting much more accurately than to with 20, which is about what 12 bits can count. (Probably this is why the 16-bit A/D converter used with that Kodak 16MP sensor in the Imacon Express back does not improve image quality noticably; it justs measures the noise better.)
Most sensors have a true ISO close to the lowest ISO setting. At that ISO, a photo-site full of electrons generates the maximum level in the A/D output; at one stop higher, a half-full pixel generates the maximum level, and so on. As I have said in the neglected twin of this thread, everything above the lowest normal ISO setting is "electronic push processing" with inherent loss of dynamic range.
Beware that the noise tests at DPReview look at a region of uniform and average illumination; maybe suitable for predicting the appearance of skies, smooth skin and such, but a very different thing from assessing noise levels in shadow regions.
Has anyone assessed shadow noise levels at increasing ISO's on something like the 10D?
Conceivably, a camera's software could be clever about detecting regions of uniform illumination and smoothing out noise there, but still have highly visible noise in regions of low but more variable illumination. A bit like the way Sony used to fake out S/N tests in some of their CD players by having the D/A converter "shut off" when it detected the pure zero input used in early testing.
-
Even so, I still prefer to think of the ISO setting on a digital camera as analogous to the size of the bucket. If I have a small bucket, I can't add much water (light) before it overflows! Conversely, if I have a large bucket, the same amount of water (light) will leave some of the bucket capacity unused (e.g. an underexposed image).
Jim,
Well, let me first say I'm just as guilty as you regarding misleading analogies. Rethinking the matter, I don't believe ISO can be sensibly compared to water pressure in a hose. I seem to have inadvertently thrown in a red herring. (I knew I shouldn't have had that glass of wine :D ) The water pressure is more analagous to the light intensity of the scene being photographed.
But I'm at a loss to understand why you would think ISO has anything to do with the size of the bucket. The bucket size is a fixed property of the sensor. If we exclude additional lighting (flash or studio lights etc), there are only two parameters that affect exposure - aperture and shutter speed. ISO is really a misnomer on a digital camera and has nothing to do with sensitiviy, but rather is an instruction to the electronic processing part of the camera telling it whether or not to compensate for under exposure and by how much. In other words, it's a bit like the 'push/pull' processing instruction you give to the lab when handing them a 100 ISO film that you've exposed as though it were 400 ISO.
It's also interesting to note that this extra processing required of the camera, to compensate for underexposure, results in larger RAW file sizes. My 1GB microdrive holds 141 RAW D60 images at 100 ISO but (from memory) about 136 at ISO 400 and 130 at ISO 1000.
-
Ah, I have made a terrible assumption! I was thinking this whole time that the sensitivity of the sensor is fixed. Partly true, but the sensitivity can be adjusted by increasing or decreasing the "gain" of the ANALOG signal. This is done by adjusting the voltage provided to the chip. The gain is exponentially proportional to the voltage! So a small change in voltage can really crank up the gain. The drawback is that increase gain means increased noise - which explains the noise part. The ADC treats the data as it usually does because it still sees a range of 16-bit or 12-bit values.
Reference: http://www.roperscientific.com/library_enc...chip_gain.shtml (http://www.roperscientific.com/library_enc_on-chip_gain.shtml)
So in terms of our bucket analogy... It kind of breaks down at this point. Unless we say that we use magic to increase the amount of water in each bucket at higher ISOs! If we were to change the analogy slightly, so that each bucket was being filled with air-filled balloons, we could then say that increasing ISO meant we were making the balloons bigger by increasing the pressure in the balloons or by decreasing the air pressure in the buckets. Yeah, that's really strange, and maybe not very good, but that was the best I could do.
So to sum up, what happens when you change ISO on a CCD device is that you are amplifying the analog signal on the chip by increasing the operating voltage to the sensors. The drawbacks are increased noise, but there is no loss of dynamic range. The camera/chip makers determine through experimentation what voltage/gain levels correspond to a given ISO.
It would appear the Jonathan W. had the right answer all along and I pretty much ignored him. Sorry! :)
-
Someone asked about Kodak's sensor information: there is lots of good technical stuff at their Solid State Image Sensors site
http://www.kodak.com/global/en/digital/ccd/ (http://www.kodak.com/global/en/digital/ccd/)
(the "ccd" is a misnomer; they do CMOS as well!)
Despite being a company site, the information seems solid and reliable, probably because it is mostly intended for technically knowlegable customers like scientific users of sensors, and so there is little room for disinformation. It goes beyond product descriptions, with published papers on their technologies.
Kodak gives out more complete information than the other sensor selling giant, Sony, which is the only reason I never use Sony spec's in my discussions.
Roper Scientific is another good site of course.
-
Ray,
your observations fit fine with the (very flexible) theory; they suggest that noise in the D60 comes mostly from the sensor itself and no additional fancy noise reduction is done for higher ISO settings, so that noise gets amplified in proportion to the signal when higher ISO settings are used. Still, ISO pushing on the D60 does in one easy step what would otherwise require all the extra work that you describe (first underexpose at base ISO, then push up in raw conversion or image editing software.)
On the other hand, I recall that a while ago someone in this forum did similar experiments with a 1Ds, and reported getting lower noise by using a higher ISO setting than by underexposing at ISO 100 and then compensating in the digital domain. (That was in response to my skepticism that ISO pushing in camera could help very much in conjunction with full 12-bit output, so we have almost come full circle!)
I suppose this just tells us that there are significant differences between cameras in the details of their noise sources and noise processing!
-
I was wondering what changing the ISO setting on a digital camera does, exactly. Does is actually somehow change the sensitivity of photoelectric sensor sites, or does it just add a scaling factor in front of the ADC when it's reading out the result?
-
It changes the amplification of the ADC converters in the sensor. This changes the light level range that registers as minimum and maximum numeric values.
-
Frankly I'd like to know too! After a lot of testing in a D30 I concluded that whatever Canon does, it does it extremely cleverly. It's not just a matter of jacking up on-chip amplification. There's no way I could replicate an ISO 1600 image by starting with an ISO 100 image and post processing, regardless of linear or nonlinear conversion from RAW. An electron well fills up with x electrons. You jack up the amplification, you chop off the high end. Result: dynamic range at ISO 400 should be one quarter that at ISO 100, ie x/4. But that's not the outcome either. It's a combination of processes that still eludes me...
-
Danno,
you have to remember that ISO is a measurement of Illuminance and time
Maybe at some level this is true, but for me, it's more convenient to think of ISO as a measure of sensitivity, independent of time. I like to use the analogy of filling a bucket (photosite) with water (light) using a hose (light path, especially aperture) connected to the water supply (the scene to be photographed) with a valve (shutter). The longer the valve is open, the more water will be dumped into the bucket. Likewise, the larger the diameter of the hose, the more water will enter the bucket. Think shutter speed and aperture! Okay, how much water do we *want* in the bucket? The size of the bucket is analogous to the sensitivity of the sensor (including the photodiode and amplifier) - the ISO, if you will. A small bucket will fill quickly - a high ISO. A large bucket will fill slowly - a low ISO. When I make an image, the goal for the highlight area would be to "fill the bucket with water", but not to overfill it. If I overfill the bucket, that's the equivalent of a blown highlight. Can you see how the size of the bucket (ISO) is related to the length of time I run water through the hose (shutter speed)?
Well, if you've read this far you may want to take a look at this pretty good discussion of what goes on inside a digital camera:
Anatomy of a Digital Camera (http://www.extremetech.com/article2/0,3973,15466,00.asp)
This article is a bit dated, but covers the basics pretty well. Note that it's a multi-page article, and the good stuff is several pages in.
Enjoy!
-- Jim
-
Okay, how much water do we *want* in the bucket? The size of the bucket is analogous to the sensitivity of the sensor (including the photodiode and amplifier) - the ISO, if you will. A small bucket will fill quickly - a high ISO. A large bucket will fill slowly - a low ISO.
Jim,
I like the bucket analogy but I think you've confused matters by equating the bucket size to the sensitivity of the sensor. This is perhaps a hangover from the film era when we could choose the sensitivity of the film.
With a digital camera, the sensors have a fixed sensitivity. Correct exposure for optimum, noise-free results is at one ISO only. Increase the ISO and you are basically sending the camera a forewarning, "I'm about to send you an underexposed image. Make the best of it." At which point, all the electronic wizadry the camera can muster whirs into action to often produce a surprisingly good result.
In fact, the result is often so good it's difficult to believe the sensitivity of the individual photosites has not been increased. However, I think part of this success is due to 'early' processing. I know from my own experience when scanning a seriously underexposed slide, if I scan at default settings, I know I'm never going to be able to compensate for that underexposure adequately in Photoshop. However, if I make the right adjustments at the scanning stage, which will include moving the middle slider on the histogram from a value of 1.00 to maybe 3.00, as well as increasing the analogue gain, I can get quite remarkable results. And after applying the noise reduction program 'Neat Image', I've got what I consider to be a remarkable transformation of a 'write-off' slide.
To get back to the bucket analogy, I would say the size of the bucket equates to the size of the photodetectors, the # of photons to the quantity of water, the diameter of the hose to the diameter of the aperture, the ISO to the inverse of the water pressure and the shutter speed to the time the water is flowing.
As far as I can see, these analogies will work across all digicam formats. For example, a really tiny digicam like the Minolta DiMage xt, will have tiny photosites or buckets (about 2.5 microns). It doesn't take much to fill them, therefore the diameter of the hose has to be small in proportion to the diameter of the lens aperture. The 3x zoom of the xt has apertures ranging from f2.8 to f3.6, but the format is so small (about 1/36th of the area of 35mm cropped to the 4/3 format), the aperture diameters are proportionally small. A fully exposed image on the Dimage xt, all buckets full (say a white wall) will be equivalent to an underexposed image on the Canon G2 with buckets only half full as a result of double the ISO. The noise levels will be similar, ie. xt at 100 ISO = G2 at 200 ISO, exposure compensated.
Of course, these figures are not in any way precise and different levels of technology in different cameras will appear to change these relationships. I'm rather amazed at how good the 3MP Dimage xt is, considering how tiny the camera and its pixels are. I'm thinking of getting one as a second camera
-
Thanks guys for the info. I think I have a better understanding. Admittedly, I knew my take on it was a bit off. The idea of filling the bucket half as much for the next higher ISO is a good analogy. The ADC must then scale up the values and eliminate noise.
But I must insist that ISO is time dependent. When I jump up from 100 to 200 ISO, I'm not decreasing the size of the sensor, I'm merely cutting the time I allow it to catch photons in half. In film, it just means that the chemicals react twice as fast requiring half the time of exposure. The result is basically the same, although some cameras do a better job of treating noise than some films treat grain.
BJL, where did you find that info on the CCDs? I looked all over and couldn't find it. Can you provide a link or two? Thanks. I love learning about this stuff! :)
-
Danno,
But I must insist that ISO is time dependent. When I jump up from 100 to 200 ISO, I'm not decreasing the size of the sensor, I'm merely cutting the time I allow it to catch photons in half.
It seems to me that shutter speed, not ISO, is how we characterize the time we allow the sensor to catch photons. The ISO is a way of communicating the sensitivity of the system (not just the photosite).
What am I missing?
Enjoy!
-- Jim
-
How about these for analogies between bucket size and "ISO" setting
a) the contents of each bucket [photosite] are measured by pouring into a graduated cylinder [A/D converter] and reading off the level from the graduations on the side. Higher ISO corresponds to using a narrower cylinder of the same height, so all readings are higher but sufficiently full pixels lead to overflow.
[simpler, less accurate] each doubling the ISO setting causes the "usable bucket size" to he halfed, as if an overflow valve is opened part way up the side: the readout of the A/D converter is the fraction of the remaining available bucket space that is filled.
-
Perfect sense! This has been very helpful.
At the risk of sounding like a broken record, once again, I must put forward the idea that time is important, even in our analogy of buckets. The bucket size doesn't change, but how long you let the bucket fill up determines the amount of water (luminance value) in that bucket.
Assuming that our shutter speed and f-stop are fixed, an exposure will give us a quantifiable amount of light on the sensors - in other words, a fixed amount of water in the buckets. Increasing the ISO one step of the digital camera doesn't change the sensitivity of the sensor (the size of the bucket). It tells the metering system and the ADC to expect half of the light (half the water) as it usually does. The meter appropriately increases the necessary shutter speeds and/or f-stop, and the ADC "resses up" the data after its been received.
Which makes me think that we may be wrong about thinking that we are losing dynamic range when increasing an ISO; that is the difference between the lights and darks will still have the same range of luminance values AFTER the ADC converts them back up to the full range. What is lost is the smooth transition of data from one luminance value to the next. Instead of 16-bit luminance values, we only have 8 bit-values to work with. The ADC resses up the values then uses an anti-aliasing algorithm to smooth out the transitions in the values the same way that video cards smooth jaggies or the same way that audio converters smooth out audio that has been digitized. But then I wonder if this is correct because if we half the number of luminance values every time we increase an ISO, then by the time we get to ISO 1600 we only have one bit per sensor! Hmmm, maybe it doesn't effectively halve the number of bits, perhaps it effectively drops off only one bit per ISO? Anyone know?
Our analogy is working, but it still doesn't completely explain how the system provides effective ISO levels when the sensor has a fixed sensitivity that is time dependent. Maybe this is more detail than some wish to know, but I for one enjoy this!
-
Vindicated at last...
-
Correction: Increasing the ISO setting is bound to reduce signal-to-noise, but reduction of dynamic range might not be apparent if the scene that's been photographed is a low dynamic range scene well within the capabilities of the camera. S/N however, will be apparent in all areas of the image.
-
An excellent link, BJL. We can all now study these technical papers assiduously and have an informed debate. :D
After a quick perusal of the material presented, I was unable to find anything that directly addresses the subject of this thread. There is a paper on ISO which contains some broad definitions which relate to our topic. For example, on the sensitivity of sensors . . . . "The sensitivity of an imager is determined by a combination of the pixel quantum efficiency and the charge-to-voltage conversion factor of the output amplifier and is specified as the imager responsivity."
and again . . . . "There are two basic types of ISO: saturation-based and noise-based. The saturation-based ISO is also
referred to as the ‘base ISO.’ The type of ISO of interest depends on the application at hand. In applications where the lighting conditions are controlled, like studio photography, exposure index settings are selected to give the
best possible image, with the image highlights falling just below the saturation level. This exposure situation is
described by the saturation-based ISO.
When lower than ideal lighting conditions are expected, a noise-based ISO is more useful. In this calculation the signal
to noise ratio (SNR) that gives a REASONABLE IMAGE for the application of interest is used to calculate the ISO. This ISO
corresponds to the maximum EI (exposure index) setting to achieve that SNR.
It is common in photography to designate SNR=40 as an ‘excellent image’ and SNR=10 as an 'acceptable image'
and calculate a corresponding noise-based ISO for each."
The above extract is really telling us no more than what we already know. If you want the best image quality, use the base ISO (usually 100). Every other setting is a compromise.
It's still not clear to me where the ISO adjustments take place; prior to A/D conversion; during A/D conversion or after A/D conversion?
-
BJL,
Well, so much for the theory. I was out shooting off a few frames with my D60 today, in good light, and managed to fit in a bit of experimentation. I took 4 identical shots at 400 & 800 ISO including 2 at 100 ISO at same f stops and shutter speed so the 100 ISO shots were underexposed by 2 and 3 stops. Converting from RAW, Breezebrowser allows exposure compensation up to +2 stops. This is not enough to get the levels up to that of the 800 ISO shot, so I made further levels adjustments in PS then compared the 400 and 800 ISO images with the corresponding underexposed 100 ISO images.
There's surprisingly very little difference. The 400 ISO image shows very slightly better detail in the shadows, but so slight it is a bit like comparing bicubic interpolation with Genuine Fractals.I got similar results comparing the 800 ISO image with the 3 stops underexposed 100 ISO image, except there was a slight colour shift in the sky which needed correcting with the hue/sat controls. Over all, the noise was about equal in both cases.
I repeated the experiment indoors at shutter speeds of 1/4 and 1/2 sec. Again the results showed little advantage in increasing the ISO setting.
So to get back to the original question in this thread, "What does changing the ISO do, exactly?" Short answer is, "Nothing much." (with the D60)
-
I suppose this just tells us that there are significant differences between cameras in the details of their noise sources and noise processing!
Well, perhaps a bit more if we're comparing cameras from the same stable. Going back to dpreview's noise curves for the D30, D60, 1Ds and 10D, the first two should show little improvement up to ISO 800. Beyond ISO 800, the D60's noise curve rises more steeply to its limit of ISO 1000, so no improvement there. The D30's noise curve rises less steeply to ISO 1600, so there's presumably some 'post A/D conversion' processing going on there. Likewise with the 1Ds. Below about ISO 700 (the crossover point) the 1Ds is increasingly noisier than the D60. Above ISO 700 increasingly less noisy.
But, as regards in-camera noise reduction, the 10D is the star. Up to ISO 200 noise is the same as the D60. Beyond that, 10D noise is dramatically less than that of the D60, strongly implying in-camera processing in the digital domain in addition to the pre-ADC amplification.
-
Good find, BJL. I think I have a fairly good grasp for ISO in the digital realm. But I wonder if we have reached the limit of what we can know without delving into trade secrets of "exactly how". I'm satisified with the knowledge I have gained so far because it allows me to understand how it affects taking pictures and using other ideas like noise vs. grain and dynamic range. Anything beyond this probably won't help me take better pictures! While it is nice to know how stuff works, I can't remember anyone ever having a discussion about how to make film at different ISO's. That was left to film makers - now it seems everyone wants to know how it all works and get their 2 cents in. Times are changing, I guess.
-
Sorry about the duplicate post. If someone has the ability to delete this copy of the thread, please do so. Thanks.
-
At first I was inclined to agree with you, samirkharusi. But you have to remember that ISO is a measurement of Illuminance and time. CMOS sensors are capable of detecting as little as one photon of light, the same as your eye. Over an average exposure of 1/30 sec., how many photons will hit any given photon sensor on the chip? The answer is probably alot more than you need to get some kind of reading. (Is there a mathematical formula to determine this?) But, if you want an accurate reading you have to take many samples. It's a matter of statisictics. The ADC converts to 8/12/16-bit luminance values, but I would guess that there are probably alot more photons hitting a given sensor than there are bit values in a typical exposure. The more values the ADC has to work with, the more accurate its representation can be, similar to taking a statistical sample in a public opinion poll. The more samples you take, the more accurate your estimate will be. And my biggest point here is that the photon capture process is an estimate of what really happened, storing an exact count of every photon that struck a sensor would consume too much memory.
But don't confuse the signal amplifier with the ADC process. The amp just gets those signals up to a voltage level the converter can use to make the transistors work (probably 3.3v).
With that in mind, I don't see why you couldn't simulate ISO settings by taking smaller samples (shorter time) of photons. The problem then becomes that with smaller samples the more likely you are to have "noise" or errors in your samples. That's where anti-noise software comes in, to smooth out the bumps and quirks in your sample. In a sense, it cheats by making "noisy" pixels look like their neighbors which gives the appearance of a clean photo when the data was probably not there to begin with.
Does this decrease dynamic range? I would say no if the average sensor sample is greater than the maximum bit value and the ADC is converting down to the smaller values. If the average sample was less than the maximum bit value, then yes you would have decreased dynamic range because you're not capturing all of the potential information. In addition, noise could be "louder" than the original signal distorting the image completely.
So I guess the question to ask is: Are the luminance values produced by the ADC
a) lower than the average # of photons striking a sensor area
one-to-one correspondence of # of photons striking a sensor area
c) greater than the average # of photons striking a sensor area
Is it possible that in the same exposore that two or even all three of these conditions occur?
P.S. I could be wrong about all of this, what do I know? The ADC converts to 8 bit values? Or is 12? or 16? The principle I think is the same. I've been doing my darndest to find info on this online and this is my educated guess from what I have learned so far. I'm entirely open to correction, since I'm trying to determine what actually happens. Will camera makers ever tell us, or is this a dark secret?
-
Ray,
Correct exposure for optimum, noise-free results is at one ISO only. Increase the ISO and you are basically sending the camera a forewarning, "I'm about to send you an underexposed image. Make the best of it." At which point, all the electronic wizadry the camera can muster whirs into action to often produce a surprisingly good result.
You make some good points! Even so, I still prefer to think of the ISO setting on a digital camera as analogous to the size of the bucket. If I have a small bucket, I can't add much water (light) before it overflows! Conversely, if I have a large bucket, the same amount of water (light) will leave some of the bucket capacity unused (e.g. an underexposed image).
Notice that I'm not claiming that ISO is independent of noise! I agree that one ISO setting probably represents some "native" sensitivity of the sensor/amplifier combination. (Although I don't think "noise-free" is a fair characterization. ) It seems that we both agree that increasing the ISO from this "optimum" is done by some analog and/or digital magic within the camera, with the introduction of noise. Others have suggested this might also reduce dynamic range.
get back to the bucket analogy, I would say the size of the bucket equates to the size of the photodetectors, the # of photons to the quantity of water, the diameter of the hose to the diameter of the aperture, the ISO to the inverse of the water pressure and the shutter speed to the time the water is flowing.
I prefer to cleanly separate the environment that brings light (water) to the sensor (bucket) from the sensitivity of the sensor. The water flow into the bucket is analagous to the light striking the sensor/film. The water pressure is, I suppose, part of that system - it determines how much water enters the bucket, and how fast. As far as the analogy goes, I guess water pressure would be equivalent to the brightness of the scene. The sensitivity of the sensor/film is something apart from the water pressure. To use your concept, perhaps the bucket size is a function of both native sensor sensitivity, and the amplification used to increase the ISO beyond the native sensitivity - which also increases the noise.
Does this make sense?
Enjoy!
-- Jim
-
This is a great discussion - thanks for the interaction!
Ray,
Was it red or white wine? ::
I do believe you're right - if you equate "bucket" with just the photosite. But remember, my analogy was intended to describe the sensor "system", not just the photosite:
The size of the bucket is analogous to the sensitivity of the sensor (including the photodiode and amplifier) - the ISO, if you will.
(Actually, I should also have included the A/D conversion step in my description.) When I say "bucket size", I don't mean to imply the capacity of the photosite.
Bruce,
I think you've brought clarity to the ISO issue! Your analogies address both the fixed capacity of the photosite, as well as the variable ISO by the time A/D conversion is done. I like it!
Danno,
Is this making sense - without introducing a "time" parameter?
Enjoy!
-- Jim
-
Danno,
thanks for the Roper Scientific link. I do not think it is quite true that increased ISO is achieved without loss of dynamic range though; reading carefully, they describe options of "high sensitivity" [given by the on-chip gain method] OR "wide dynamic range", not both at once.
To understand the options, you have to take account of four main stages where noise (or error in numerical values) are introduced.
a) "Shot noise": random fluctuations in the number of photons detected by the sensor, due to fundamental physics; there is nothing to be done about that except using bigger photo-sites, so this will some day be the ultimate limit on low light sensitivity at a given photo-site size.
"Dark current": thermally caused random changes in electron counts; this seems to be one stage where progress is making small photo-sites work better, perhaps by operating at ever lower voltages and hence lower temperatures.
c) "Read noise": noise introduced in the output amplifier between the sensor chip and the A/D convertor (probably another place that progress is happening.)
d) "Discretization error": rounding off to the nearest integer value in the digital output. But with 12 bit or better A/D conversion, this is small relative to earlier sources in current sensors, so I will mostly ignore it from now on.
The first two categories of "in-sensor noise" fundamentally limit the dynamic range, which cannot exceed the ratio between this noise level and the maximum countable number of electrons: any gain or amplification also increases this part of the noise. If you push the ISO, the maximum countable number of electrons is decreased (in my analogy, it take less electrons to fill the more slender graduated cylinder), so some dynamic range is lost.
At very low light levels, dark current dominates over shot noise (shot noise is lower at photosites receiving less light, while dark current is not, and instead is proportional to exposure time: sensors with good long exposure noise probably have low dark current). So dark current is the main villain in deep shadow noise produced in the sensor itself.
On the other hand, the technique of on-chip gain described at Roper Scientific happens before the output amplifier, so avoids the amplification of read noise (and discretization noise.)
So the degree of loss of dynamic range during "pushing" depends mostly on the relative size of dark current noise vs read noise.
P. S. I am fairly sure that something similar to on-chip gain is an important part of Canon's being able to produce CMOS sensors that reverse the high noise reputation of earlier CMOS sensors.
-
There are quite a few sites that offer a basic primer on CCD and CMOS processes, but I can't find anything relating specifically to the method by which changing the ISO setting on the camera increases the gain or amplification of the charge that has accumulated in each photosite as a result of photons knocking off electrons.
I would still maintain that the 'sensitivity' of the 'electron collecting' site is not changed by changing the ISO setting on the camera, but rather there is extra amplification applied to each photodetector's charge at the read-out stage. Some of this might even be done before A/D conversion.
As BJL has pointed out, it's difficult to amplify a signal without also amplifying at least some of the noise, just as for example, when you receive a weak signal on your radio, turn up the volume and get a lot of background crackle. This is called poor signal-to-noise. The options are, either increase the strength of the signal (install a better antenna) or devise some clever trick of reducing the noise.
Increasing the ISO setting on the camera effectively reduces the signal strength through the process of letting less light in. There are consequently fewer photons striking each photosite, fewer electrons that are knocked off the silicon crystals and a smaller charge than would otherwise be the case if the ISO setting were optimal (100 ISO).
The trouble with the bucket analogy is that this term is often used to describe the 'electron collecting' capacity of each photosite. Terms like 'full well' are used to describe a photodetector that can carry no further charge. Getting the wells full with respect to the highlight areas of the scene being photographed, is all part of the recommended process of exposing to the right of the histogram. This will ensure maximum dynamic range and maximum S/N and this can only be achieved if at least some of the buckets are full at the collection stage (not at some later transferral and processing stage).
Increasing the ISO setting is therefore bound to reduce both S/N and D/R. I try to avoid it
-
Ray,
thanks for digging out those guidelines of S/N>40 for exellent image quality, >10 for aceptable, they at least set upper limits on dynamic range at various ISO's.
For example, to get above S?N of 40, you have to count 40 squared = 1600 electrons just to get above shot noise. Other noise sources are comfortably less than 40 electrons under normal circumstances I believe, so you are probably OK at that level. Digging one more number out of Kodak's site, of about 50,000 electrons as the maximum that several of their DSLR sensors can count, and you have a typical dynamic range of 1600 to 50,000, a ratio of 30 or five stops; (close to transparency film?).
Probably in shadows that are more than five stops from the top (which are even hard to print without compressing contrast), the more liberal "acceptable level" is often all you can make out, and that lets you go down to 100 electrons, which suggests a dynamic range of 9 stops. This at least fits with figures in the range 8 to 9 stops I have seen is some technical articles that, however, are vague about their definitions of dynamic range.
Every stop higher ISO loses one stop at the top of the range, so loses one stop of dynamic range for high contrast subjects.
As to when ISO pushing is applied, it seems from the Roper Scientific reference that the traditional place is gain in the amplifier feeding the A/D converter, with some sensors also doing something with "on-chip" gain; in either case, before A/D conversion. I have not heard any mention of any camera doing it in the digital domain (despite my earlier speculations).
-
I see relevance for owners of the Kodak 14n in these D60 results. Since Breezbrowser does such a good job in exposure compensation (and Capture One possibly even better) it would be of little concern to me if the D60 had only the one ISO setting of 100.
Now I don't recall the D60 being criticised for being unacceptably noisy at ISO settings above the base of 100. And as I remember, at ISO 80 the 14n is as noise-free as the 1Ds. Therefore I see no reason why software processing should not be able to produce acceptable results at ISO 400 and even 800. Not sure about the long exposure limitation, though. :)