Is noise the same across the board?
What I meant to ask is, would the noise levels resulting from using a lower ISO with a longer exposure be the same as that same exposure using a higher ISO and correspondingly shorter exposure?

"Test for yourself" would be a reasonable response to this question, I know, and I intend to. However, I would like to know the theory behind it, if anyone here happens to know.

Intuitively I would suspect that the noise would be different, however logically I might assume otherwise.

Thanks.

This is an interesting question. A comparison of ISO noise and thermal noise. I have a feeling that for any situation you will have complicated answers. The answer will vary a lot based on whether you use dark frame subtraction or not, ambient temperature, temperature of the sensor, etc. If I find myself bored some night I will do some testing and share some detailed results.

Phillip

Under most conditions, the total noise is related to the exposure, which determines the number of photons collected. A long exposure at a low ISO will collect more photons and will have greater signal to noise, even though the absolute noise is higher. Sensors with large pixels also collect more photons and will have less noise if other factors are the same.

A digital camera changes ISO by amplifying the signal. Fewer photons are collected and noise is higher. Some noise is contributed by the amplification and reading of the charge of each pixel. This is called read noise, which can be reduced at higher ISO. This type of noise involves mainly the darker portions of the image. If light is limited, there is an optimum exposure.

Thermal or dark current noise varies with temperature and becomes a factor only when exposure is many seconds or longer. It is not a significant factor in most situations.

For a good analysis, see Roger Clark's web site:

http://www.clarkvision.com/imagedetail/dig...rmance.summary/

Thank you!
That is precisely the explanation I was looking for.

Regards,
Phil

In Stephen Johnson's book, he talked about sensor noise from long exposures. But he was not specific what he meant by "long".

He did mention some cameras have "dark current measurement/subtraction functions" to deal with noise from long exposures. The idea seems to involve creating a "dark frame, an image of the sensor with no light striking it" (at long exposure?) and then subtracting it from the image to remove the noise.

Is there a common name this feature is referred to by the manufacturers? Is it commonly found on most models?

It depends on how long... I've shot a lot of night scenes at ISO100 (10D & 5D) and anywhere from a few seconds to 30 seconds and the scenes are amazingly low noise--nearly "grainless." However, if I crank the ISO, even a few stops, the apparent noise increases greatly.

I haven't shot any startrails but I'm sure as exposure times increases noise will increase. I'm just not sure at which point a high ISO is an advantage noise-wise over a longer exposure. I do know it's somewhere beyond 30 seconds for the 10D and 5D.

This image was shot at ISO100, F11 and 15 seconds (10D). It is virtually noise-free even as 11 x 14 print. I also shot shorter exposures of the same scene at ISO 400 and 800. They were terribly gritty.

I don't see my reply in this thread - I thought I posted one; well anyway:



The important things to realize are:

1) Dark current noise (related to exposure time and temperature) exists relative to signal in the analog sensor wells, and the ratio has absolutely nothing at all to do with the ISO setting.

2) ISO-related noise exists relative to RAW exposure (the dynamic range of the RAW data at that ISO).

3) Depending on the camera, there may be more noise at higher ISOs relative to absolute signal, the same, or less noise.

#3 is very important to know for your camera, but I don't know if this data is actually listed anywhere for specific cameras. I know that for recent Canon models, the lowest ISOs have the highest noise, relative to absolute signal. This means that you can consider absolute exposure first, and use ISO accordingly, to make sure that the ISO chosen lets your signal give a good exposure for that ISO. If your camera does not improve in noise efficiency at higher ISOs, then it may not be worth using the higher ISO, since it will not give you any less noise.

As far as the trade-off between time and signal-to-dark_current_noise is concerned, in most of my experience, dark current noise does not quite double in double the time, so the longer exposure is generally better; IOW, with the same f-stop, you are better off with 8x the exposure time at ISO 200 than 1x at ISO 1600. (ISO 100 is not always a good ISO for highlights, depending on the camera).

A more objective experiment, however, would be in order. I would expect to see any advantage for the shorter exposure (all other things being equal) to only happen at very long exposures (multiple minutes), because there is *always* more noise with a higher ISO or under-exposed/pushed lower ISO that has nothing to do with dark current. In my recent Canon 20D test, ISO 100 at 30 seconds was better than ISO 1600 at 2 seconds, but clearly part of the benefit was the lower non-dark_current noise at ISO 100. The hottest pixels were slightly hotter in the 1600 image.

Perhaps some useful links for high iso speed photos:
http://highiso.net

But there's just test samples I think.

Here are some universal principles of exposure and ISO with digital cameras.

Shoot at the lowest possible ISO that allows an acceptably short shutter speed. Shooting a long exposure at low ISO gives a cleaner result than a shorter exposure at correspondingly higher ISO. If you are shooting static subjects from a tripod, use the lowest native ISO you can.

Shooting a high ISO and exposing properly is better than shooting at a lower ISO, underexposing, and push processing.

Avoid ISO settings that are not native to the camera sensor, i.e. that are achieved by processing tricks in-camera and not by the sensor itself. An artificially low ISO setting, like 50/L on the 1Ds, sacrifices dynamic range. You're no better off than shooting at ISO 100 with regard to capturing highlights. An artificially high ISO, like 3200/H on the 1D-MkII is the same as underexposing and using a positive exposure compensation setting in the RAW converter.

Jonathan, How do you know which ISO settings are native to the camera?

Julie

As a rule of thumb, it it is an extended ISO setting that has to be enabled in the camera menu (ISO expansion, etc) before it can be used, then it's probably fake.

Well, you won't see this information in the manual, usually. The RAW data has to be analyzed to tell how it was derived. In some cases, there is little or no negative effect from using a mathematically-derived ISO setting, but in others, you can do better by using the more real ones.

From what I understand from John's previous posts on this subject, ISO settings have traditionally moved in double numbers, ISO 100, 200, 400, 800 etc.

In the latest Canon models, we have fractions of this. ISO 250, 320, 640 etc.

These are the non-native ISOs that serve no real, purpose.

What is the definition of "native ISO"? What makes ISO settings "artificial"?

From my readings, a camera's lowest ISO setting means a sensor and its surrounding circuits are operating at a sensitivity (or response to light levels) that results in an optimal signal to noise ratio. Perhaps that is a good definition for "native ISO". When ISO is raised, the sensor's signal is amplified in hardware to boost the sensor's sensitivity. This comes at the expense of moving away from the sweet spot of the optimal signal to noise ratio, or at a cost of increase in noise. If the higher ISO settings are results of hardware adjustments, I would not call them "artificial". Perhaps it is not a given that all ISO settings are done by hardware, making things even more complicated.

I am by no means an expert on this, but just trying to digest in layman's terms from my readings. I think of operating at the lowest ISO is like normal film development, and raising the ISO is like pushing/pulling film development.

They serve no real purpose for the RAW shooter, but they can be useful for the JPEG shooter with some cameras. The 30D's ISOs 160/320/640/1250 are actually better than the real ISOs that they're working off of, since they are basically exposing to the right, without any danger of clipping any more than the main ISO's JPEGs would clip.

Absolutely NOT. What I'm referring to as the "native range" are ISO settings obtained by adjusting the voltage gain of the amplifiers between the sensor and the A/D converter before A/D conversion. "Non-native" settings are those derived by in-camera mathematical manipulations of the RAW data after A/D conversion, and typically must be enabled by some type of "ISO extension" or similar menu function before they can be selected. These "extended" settings generally offer the RAW shooter no advantage whatsoever; they can be duplicated or bettered by shooting at an ISO setting within the "native range" and push or pull processing. ISO 50/L on the 1Ds and ISO 3200/H on the 1D-MkII are examples.

The "native" ISO of the sensor itself is that which involves no voltage amplification (unity gain) between the sensor and A/D converter(s), and normally is the camera's lowest ISO setting, usually between 50 and 100.

Amplifying the voltage prior to A/D conversion is of benefit because it is scaling the output of the sensor to more-or-less match the input range of the A/D convertor, which gives you as many bits of image data as possible to work with. Manipulating the bits in-camera after A/D is of very limited use; instead of working with real (though potentially noisy) image data across the range of RAW data values, one is confining the image to a smaller range of RAW values and interpolating the image from there.

The in-between ISOs on the 30D are made the same way, even if they are within what you call the "native range". ISO 160 is 200 pulled one stop, then the digitized data divided by 1.26; ISO 250 is 200 pushed one stop and multiplied by 1.26. The ISO 250 RAW histogram has a gap about every 5th value, ISO 160 has a doubly-populated value about every 4th value. ISO 250 is missing 1/3 stop of highlights that it had from before its digitization was multiplied.



The lowest ISO requires amplification, too. Just not as much as the higher ISOs. There is some constant, K, where ISO 100 has 1K gain, ISO 1600 16K gain, etc.

Even when signals are not amplified, they are still buffered, which is basically an amplification that just happens to be unity.



The native base ISO of the sensor may be somewhere between 0.7k to 1.5K, but not used by the camera as such.

I wasn't aware of that being the case with the 30D's "between stops" ISO settings. The 1-series bodies seem to use varying voltage amplification for all ISO settings in the "native range", as opposed to firmware tricks on the RAW data after-the-fact. What tool are you using to check/verify this?

I shoot at 3200 on my 30D quite a bit, are you saying that it isn't real 3200, but mere trickery?

You might want to look at this analysis by Roger Clark

In general there is not much point using an ISO much higher than that at unity gain of the sensor, which maxes out at 1600 for the Canon 5D and is probably around 1000-1200 for your camera. At unity gain one electron corresponds to one 12 bit data number in the raw file, and it does not make sense to digitize anything less. In other words, ISOs above 1600 achieve little benefit.

Bill

So...would it be better to shoot at 1600 and push a stop?

The histograms of the RAW data show clear manipulation. Gaps and spikes spaced just right to show that the data was integer-scaled.

IRIS is the easiest tool for examining RAW data. It loads all RAW data literally, with no manipulation.

Even thought the 1-series and 5D use analog amplification for the in-between ISOs, they can also be quite inefficient noise-wise, because the cameras are optimized for the higher ISOs, so a 160 that is 100 amplified by 1.6 has more shadow noise than 200.

If you shoot RAW and want the extra highlight headroom; many ISO 3200 scenes are ones with lights in them, at night, so that can come in handy.

I often shoot low light in Tv-mode, so I set the camera to 1600 and as high an EC as I can safely use; that covers a wide range. If light is ample, I can get a very good ISO 800 or even 400 exposure, and if the light is weak, I just have whatever I get.

Meh. When I'm shooting 3200 its usually because I'm shooting college sports for the university paper, meaning I'm shooting JPG anyway. I guess I'll just remember this for other more vital projects, thanks for the info.

These are much better answers. Thanks.

In summary, "native" and "native range" ISO settings are controlled by hw, while "artificial" ISO settings are altered by sw. But unfortunately, the manufacturers won't tell you which is which.

We've been through this before - there is nothing special about unity, and nothing in Roger's work proves that there is. He is using a mathematically-derived 3200, which tells us nothing about what a gain-based one might be.

The cameras aren't even close to counting electrons. The noise makes the count very uncertain. The RAW ADU unit is pretty meaningless, too. Would lowering the bit depth to 11 bits make one more stop of gain (3200) suddenly feasible? Would 13 bits make it suddenly only feasible to do 800?

The only trade-off concern of any value is the noise, and it is very clear that in the recent Canon DSLRs, graphing various noise components at various ISOs shows improvement possible past ISO 1600. Graphing 3200 does not show what 3200 would be if it were analog; it is only 1600 gain repeated, as measured in electrons.

There is no real unity in any real physical sense, even if it did mean something. There is simply a trade-off curve; additional amplification means less noise as measured in electrons, though the benefit gets smaller and smaller with each increase. So-called "unity" is only a featureless point in the curve.

Yes, we have gone over this before and I'm not certain that there is much to add. In Table 1a in the link, Roger does give data for ISO 3200 but it is not clear to me if it they are measured experimentally or derived mathematically.

Clark Table 1a


Unity gain would change with bit depth, but whatever the bit depth, it does not make sense to quantize the signal any finer than to the nearest electron, especially when the read noise at ISO 1600 for the 1D M2 is 3.9 electrons and the photon noise down 8 stops from maximum is around 5 electrons (both expressed as SD). According to Roger's table, read noise actually goes up at ISO 3200 and dynamic range is lower.

If you have other data, I would be interested in seeing them.

Bill

There is no ISO 3200 on any Canon. All Canon ISO 3200s are an illusion. They are ISO 1600, under-exposed by a stop, and then the RAW numbers doubled (or in the case of the 10D, the numbers are left alone as ISO 1600 numbers, which are 800 doubled, and it is assumed that a converter will recognize the need for a stop more software EC).

Doubling causes a stop of DR to be clipped away.



This is read noise on my 20D, total, and broken down further into horizontal and vertical line noise (banding):



The yellow line is divided by 10, so that all the data fits into the same area. The 3200 is artificial, and has the same noise in electrons, or the normalized ISO 100 ADUs that I use here (which are directly proportional to electrons).

The trend shows room for small improvement for total and vertical line noise, and room for tremendous improvement for horizontal line noise, which is by far the most detrimental noise in Canon low-light shadows. 1-dimensional line noise is extremely problematic, as it does not disappear with small prints, or downsampling as much as 2D noise does. I have severely binned down shots taken at ISO 3.2 million, and the only noise left is banding noise, in bold black and white bands for a greyscale image.

John,

The data are interesting, but I have difficulty interpreting them because I am not familiar with your terminology. What is the equation for normalized ISO 100 ADU and what are the units? Also, what is the equation for line noise, what are the units, and how is it measured?

Bill

The vertical axis is standard deviation at ISO 100 in ADUs. It is only literal for the ISO 100 line noises. The ISO 100 total noise (yellow) is divided by ten, so it fits in the same chart, because the other two would just become unreadable flat lines if they were all at the same scale.

For ISO 200, all the values are divided by 2; for ISO 400, they're all divided by 4, etc. That's to make the values in the chart meaningful relative to absolute signal. This is the same thing as measuring in electrons (which noone really does; people just convert to electrons with an assumed knowledge of how ADUs translate to electrons), the unit is just arbitrary. The unit is irrelevant to the point of the chart; the point of the chart is that noise at the higher ISOs doesn't look like it's going to level off immediately above 1600, with analog gain; especially horizontal line noise, THE most visible and annoying noise of Canon low-light shadows (even though it is only about 1/10 as strong as the total noise). Roger's analysis looks at noise monolithically, as a statistic, with no concern for the fact that all noise is not equally visible.

The methodology is very simple; I take the standard deviation of the RAW blackframe bitmaps, and the bitmaps with all the pixels in each line averaged, both horizontal and vertical.

John,
Those are convincing graphs. Do we have an issue about in-camera processing of ETTR ISO 3200 shots versus 1 stop underexposed ISO 1600 shots processed in commonly used converters such as ACR, to which one might apply a +1 stop EC?

In other words, is the boost or doubling (in ISO 3200 mode) to the RAW ISO 1600 data, equal to, better than or worse than, the job that some, or all RAW converters might do with a 1 stop underexposed ISO 1600 shot?

The unit of noise is important if one wishes to compare your data with other published data. Also, the definition of noise and how it is measured is important. Roger subtracts two frames taken under identical conditions and the difference represents random variations in the signal, the usual definition of noise. If one takes the standard deviation of an area of a homogeneous object in the image as a measure of noise, these deviations could represent random noise or systematic variations arising from nonuniform illumination or non-homogeneity of the sensor response or processing. Of course, for a black-frame, there is no illumination. Furthermore, once one begins to weight noise according to its visibility, complications imposed by the weighting method are introduced.



Is the standard deviation derived from variations within the lines or between the lines? Between line variations could be non-random and equalized by a smart algorithm. Finally, you concentrate on read noise, but most people are concerned with total noise. According to Roger's analysis, photon noise predominates except in the deepest shadows.

If you're shooting RAW, ACR, C1, etc. are probably a bit more sophisticated than the one in-camera, but the difference you see is minor. I'd say "equal to" with perhaps a very slight edge to external processing. But if you shoot JPEG, it's better to let the camera push the exposure a stop in 12-bit mode internally than to push it in 8-bit mode externally.

The main reasons that *I* don't use 3200 on my 20D are that 1600 has a stop more DR, and that most of the time I'm shooting in low light, the camera is set to Tv-pri AE, so with a positive EC, I stand a chance of getting an exposure index less than 1600 if the light improves, but I am not sacrificing much of anything by using 1600 instead of 3200 if I don't get a lot of light.

However, a couple of people have reported that ACR's line noise (banding reduction) works better with 3200. I haven't tested that, but it seems logical that this is the case, since ACR most likely works in a fixed bit depth (possibly just 12 bits), so the 3200 RAW data is expanded one more bit. In my more casual comparisons, I really haven't noticed much of a difference between the two except for the greater highlight headroom of the 1600. There may not have been much visible banding in those comparisons.