I'm sure this topic has passed many times.
So this topic is redundant, but it will only be short-lived :-)

Here goes:

Most current digital backs have a dynamic range of 12 f-stops.

Current DSLRs promote a 14bit-A/D conversion, MFDB are in the 16bit range.
My understanding is that this bit-talk referes to color depth, so my question is:

How many f-stops can a DSLR handle? And why are they more limited than MFDB's?

Authors note: This topic is in no way a do-I-really-need-a-25k-digital-back-when-I-can-buy-a-5k-canikon? So please, don't go where I didn't :-)

thanks

Mike

Dynamic range and bit depth are two different spec's. Higher bit depth doesn't equate higher dynamic range.

Dynamic range is the range from highlighted to shadow. Bit depth is the number of steps. Think of dynamic range as a staircase of a fixed height and bit depth as the number of steps in that staircase.

or in other words, "bit depth" describes the number of grey shades = tonal values (steps) available/possible for each colour channel. Therefore, for a given dynamic range (exposure latitude) you can have more or less tonal values (grey shades) going from white to black, depending on the bit depth of the digital device (256 for 8 bit; 1024 for 10 bit; 4096 for 12 bit; 16384 for 14 bit; 65536 for 16 bit).

Thierry

Ok, everything so far seems to be as I expected. So long story short -> what kind of dynamic range (expressed in f-stops) can you get out of a dslr?

I'm curious about this since all (or most) MFDB brands state 12 f-stop dynamic range in their product descriptions, but dslr-brands don't.

regards

Mike

Hi, DSLRs range from 6,5 or 7 (like Nikon D2x) up to 8,5 - 9 for the better ones (Canon 1ds II and a bit less for Canon 5d), an exception are the Fuji S3 and S5 pro which have a dual pixel approach with one set of (smaller) extra pixels placed between the regular pixels only for the highlights.

I had the S3 and it works, it just comes at a price: RAW files are huge (double as large as normal) and slow down the camera while not holding that much of detail (actually little more than 6MP). But after getting my Canon 5d I was dissapointed to see again how easily highlights are brutally burned out.

I am also curious to know what the reasons are why MF backs can theoretically achieve higher DR. It cannot be pixel size alone, as the more recent backs with resoluton more than 30MP don't have larger photosites than certain DSLRs.

Has anyone actually really measured real life DR or the backs?

I think that DR is the future of DSLR progress. Pixel count will probably top out at 18-20 for 35mm sensors, but DR is far from ideal.

Bernie

An interesting and somewhat complex set of questions. Your first question "How many f-stops can a DSLR handle?" is difficult to answer because dynamic range is a moving target on any digital sensor, camera or back. It changes based on a number of factors. It can be influenced by ambient temperature, cooling capability, exposure time and package electronics–among other factors.

For instance, for every 6.3°C increase in sensor temperature of the Kodak 39mp sensor the dark current noise will double and decrease dynamic range by about one stop.

The measure for dynamic range is not the bit-depth of the A/D converter. The A/D converter can not replace dynamic range that the analog sensor (CCD or CMOS) is unable to capture. Knowing the bit-depth of the A/D converter does not tell you the dynamic range of the capture device.

Also, the specification given by manufacturers of "f-stops of dynamic range" is not standardized. It is based on the maker's standard for a given set of conditions.

The measure of dynamic range is signal-to-noise ratio at a specific temperature. As mentioned above, increases or decreases in sensor temperature have a dramatic effect on dynamic range (signal/noise ratio). So, depending on a host of variables, the dynamic range varies.

Behind this link is a lot of information about that: http://www.normankoren.com/digital_tonality.html

Best,
Johannes

I recently switched from a Canon 5D (excellent camera) to a P30+ on a Mamiya 645AFDII.

I was not sure if the DR claims were real or just hype. According to the specs the DB should have 2-3 stops more dynamic range than the 5D.

After using the back for a few months I am still amazed at the increase in dynamic range. I have found both shadow and highlights to have more accurate detail. For example, if I have shadow areas on the 5D that need to be lightened you will see noticeable noise. With the P30+ there is less noise and more detail in those shadows.

Imagine you are shooting landscapes where you would normally use a 2 stop ND grad filter or take two exposures and layer them. With 2-3 stops extra range, simply forget about it and shoot.

Combine the DR with no AA filter, better optics, higher resolution and greater bit depth and the difference can be dramatic even for 16x20 prints.

8 bit jpeg: 256 levels
12 bit raw: 4096
14 bit raw: 16384
16 bit raw: 65536

I also switched from a Canon 1DmkII at 12 bits to the 1DmkIII at 14 bits. The difference with just a 2 bit increase (4X) makes substantial improvements in image quality. Multiply that difference by 16 when comparing 12 bit to 16 bit images. More levels means cleaner grads and more realistic detail. Makes the announced Canon 1Ds MK III look very interesting and will split the advantages of high end DB image quality.

That analogy (which comes from Bruce Fraser's writings, and perhaps from other sources) is often quoted, but I don't think that it is entirely true when referring to modern digital cameras. This is discussed in this Link. Higher bit depth does not guarantee increased dynamic range, but is necessary for it. With linear encoding 12 bits can encode 12 f/stops and 14 bit encoding can give two additional stops. The extra two bits can decrease quantization noise.

Bill

I think I'll take credit for that analogy if I can. I dug up a slide of a presentation I did back in January 1999 (note how crude it is visually <g>). I have Bruce's Real World Photoshop 3 (1996) signed by him (the first of many books he signed for me), page 298 has a great paragraph on dynamic range and dMax, but nothing about stair cases. Even earlier in my old book collection is Real World Scanning and Halftone by David Blanter and Steve Roth (that goes back to 1993!). Page 107 also has a very good discussion of this topic and they point out, as they should that "not all 8-bit and 24 bit scanners can capture the same amount of information", and discuss the fudge factors in measuring dynamic range due to noise issues (where to begin?).

Possibly in the context of photography, but in the general context of electronics and analogue to digital conversion it has been around for a long time.

I use a digital back alongside a DSLR, I'm sceptical about the claims for significantly greater latitude from digital backs. Maybe a half or two-thirds of a stop, but I've never seen the evidence for much beyond that.

I may be wrong here, but I seem to recall that Phase One base their 12 stop claim on the latitude up to the blooming point (ie where light spills over into adjacent pixels), if so then it isn't latitude as I or most photographers would understand it, and it may explain why I don't see the roughly four stop advantage over DSLR's that their claim implies.

I'll take anything I can get <g>

"That analogy" was used by my neighbor when explaining "tonality" to me as one distinguishing characteristic between enlargements from 35mm & large format.

This was in the late 70's. I'm sure it wasn't the first use... Perhaps she knew Bruce back then?

Edit: Now that I think about it, the comparison was between a ladder and a slide.

More levels can not be appreciated unless the blackframe noise is below about 1.25 ADU. Any time the noise is above 1.25 ADU, the number of levels used is excessive. 1Dmk3 ISO 100 RAWs do not suffer any loss of DR by dropping the 2 extra bits. If you are getting better results from the mk3 at ISO 100, it is probably for some other reason than the two extra bits of noise, like a converter being forced to use two more bits of precision (which it could have done with 12 bit data), or different defaults used by the converter for the camera. At higher ISOs, the mk3 has up to 1/2 stop more DR, but this does not require the 2 extra bits.

This does not clarify the issue for me, Andrew. I can appreciate that bit depth could be analagous to the number of steps of fixed height, but dymanic range as it is usually expressed (in terms of number of f stops) is more like a staircase with each step being twice the height of the previous step.

If we could have a dynamic range consisting of steps of equal height we could increase DR dramatically. Those steps at the top are too steep to climb .

isn't it rather those steps at the bottom which are to narrow to step on?

Or is it both at top and bottom that is critical? Whereas with lower bit depth top and bottom would render as white and black without distinguishable details, the larger bit depth allows detail to be rendered in these regions due to the more steps of the ladder there.

I am not sure, but... does this make us perceive it as greater DR, while it is actually not? Is this what has been a key difference between MFDBs and DSLRs?

How will as example files from a 1Ds Mk 3 compare to those from an equivalent MP MFDB?

Much thanks.

Regards
Anders

John, do you have data on the 1DMIII or is this merely conjecture? From Roger Clark's analysis of the noise characteristics of the 1DMII, there is reason to believe that a 14 bit ADC would improve the DR of that camera at low ISO. The black frame noise at ISO 100 (this is for very short exposures and is not to be confused with thermal noise present with long exposures of many seconds or minutes) is strongly influenced by the ADC noise. The SNR of an ideal ADC is related to bit depth, as explained here. The SNR of an ideal ADC is 6.02N+1.76 decibels (db), where N is the bit depth of the device. For a 14 bit ADC the ideal SNR is 86 db and it is 74 db for a 12 bit ADC.

Look at Roger's figure 8a. At high ISO the bit depth of the ADC has little effect on noise, since read noise of the sensor predominates. However, at ISO 100 ADC noise predominates and a high quality ADC with more bit depth would reduce this noise. Of course other components in the imaging chain affect DR and merely sticking a 14 bit ADC in the chain may not improve performance.

Bill

For me DR is the most important thing in the whole MFBD discussion. The abrupt way a DSLR (in my case 5D) goes from a defined area to a blown out area is the most disturbing and unaesthetic point in digital photography for my work. As long as a subject is in controlled light, there are little problems, but a reflex or a bit backlight might destroy the whole atmosphere of a photo. My three or four year old back behaves much better (film-like) here and is definitely worth the effort. If Canon and Nikon came along with sensors capturing the same DR I would probably drop the MF thing for a faster and more comfortable working.

But given that 14 bits doesn't change the actual DR, wouldn't those extra 2 bits go a long way to smooth the transitions at the point between blown and some detail?

The point is, the stair case has a fixed height. That the steps (bits) are ½ inch apart or a foot apart doesn't change the height of the stair case.

I`m a photographer. I prefer to talk about aesthetics, not about bits. My older Leaf back is quite good. My 5D isn`t that good for my personal expectations. If others have different expectations, I don`t care. I wouldn`t like to take a picture of a celebrity with a 5D and have a look at that image in five years.

Bit depth and dynamic range really are the same thing, provided that
1) System noise is low enough that bits are not wasted digitizing noise
2) The ADC is linear, so each bit represents twice as much signal as the previous bit.

The "steps on a ladder" analogy is flawed, because the ADCs we are talking about are linear. With a linear ADC, extra bits necessarily extend the length of the ladder, as well as providing smaller steps.

Here is a thought experiment to illustrate the connection between bit depth and dynamic range. Suppose we photograph a perfect gray scale, with discrete steps of exactly one f/stop in brightness, using linear ADCs with 8 bits or 12 bits, in a camera with no noise.

The 8-bit ADC will record 8 steps of the gray scale, with output values from maximum white (8-bits = 255) to minimum not-quite-black (1-bit = 1), for a total dynamic range of 8 f/stops. The 12-bit ADC will record 12 steps of the gray scale, with output values from maximum white (12-bits = 4095) to minimum not-quite-black (1-bit = 1), for a total dynamic range of 12 f/stops. Extra bits increase the dynamic range which the ADC can record.

Depending on the relative exposure, the extra steps of the gray scale recorded by the 12-bit ADC can be darker or brighter than the steps recorded by the 8-bit ADC. Let's assume we expose optimally in each case, so that the brightest step of the gray scale which we are interested in recording barely saturates the sensor, turning on all bits in the ADC output. Then the 12-bit ADC will record four steps of the gray scale which are too dark to be recorded by the 8-bit ADC.

With this exposure, each section of the gray scale will be represented with much finer gradation by the 12-bit ADC than by the 8-bit ADC. This is easy to see mathematically -- just divide the output of the 12-bit ADC by 16. Now every discrete step of the gray scale will be numerically the same for both ADCs, but the step size of the 12-bit ADC will be 16 times finer than the step size of the 8-bit ADC.

Another way to think about this is to replace the discrete gray scale steps by a continuous brightness ramp. The 12-bit ADC will record any small section of the continuous scale with 16 times as many discrete output values as the 8-bit ADC. Extra bits provide more resolution across the entire dynamic range.

Clark's "sensor DR" figure of 14 tops for the 1D Mark II is dubious when you notice that the nice low read noise of 4 electrons is only at ISO 3200, where high DR work is usually not done. At lower ISO speeds read noise gets higher, reaching 16.61e at ISO 100. I think this means that most read noise is introduced after the pre-amplification done at each photosite on Canon CMOS sensors.

Clark computes a more trustworthy "camera DR": at each ISO, based on the maximum signal and read noise at that ISO setting. The result for the 1D MkII is a S/N ratio of 3190, or 11.5 stops, within the gamut of a 12-bit A/D convertor. (Ref: Figure 5 of http://www.clarkvision.com/imagedetail/dig...nce.summary/and Table 1b of http://www.clarkvision.com/imagedetail/eva...1d2/index.html)

To obtain Clark's higher "sensor dynamic range" of about 14 stops would require increasing the "headroom" in the pre-amplifier and A/D conversion. Getting that 14 stops would require using ISO 3200 (i.e. massively amplifying at each photosite, which reduces read noise in electrons) even though the image contains bright highlights, at a level more suited to low ISO speed, and having pre-amplifiers that could apply this high amplification to those large electron counts. Normally, this high pre-amplification is only applied to the dimmer on-sensor images (lower electron counts) given by high exposure indices like ISO 3200.

There is a serious question as to whether this high amplification could be applied to the strong signals from nearly full electron wells without amplifier clipping or other problems, especially since the amplifiers are tiny ones at each photosite.


In other words, the S/N ratio of the tiny pre-amplifiers at each photosite is one DR limit not assessed by Clark.

As EricV explained in a previous post, the height of the stair case is not fixed. With a linear ramp and integer encodeing, the height (dynamic range) depends on the number of bits used for encoding. Under these conditions, 12 bits linear can encode a maximum of 12 f/stops and the darkest f/stop would contain only one level, which would be unsuitable for photographic continuous tone reproduction. If you require 8 levels in the darkest f/stop, the DR is reduced to 9 stops as Norman Koren demonstrates in a table on his web site.

Moreover, with linear encoding (rather than gamma), the steps are of unequal size, as shown in the graphic on Greg Ward's web site. The steps in the shadows are much larger than those at the top of the range. Look at the table below the stair step graphic, which shows the dynamic range with various encodings. The scRGB being proposed by Microsoft is linear with a bit depth of 16. According to Ward's criteria (5% error), it can encode only 3.6 orders of magnitude (engineers use log base 10, while photographers prefer f/stops, or log base 2). This corresponds to 12 f/stops.

If you use log or floating point encoding, then some of these limitations are removed, and for a given bit depth one can record a much higher dynamic range.

Bill

The 5d is only 12 not 14. My point is that the new technology might render images more to you standards. (might....)

Once my head clears from a taxing image editing project, I'll certainly check this all out.

It sounds like this is different for digital linear capture than gamma corrected scanners of the past.

I think we should remove quantization errors in non linear encodings from this discussion before we go crazy. Let's assume for the benefit of the discussion that sensor readouts are counting photons plus or minus noise.

Edmund

If it does, I´m open. It would be nice. If not, then maybe the next generation.

Why is it so difficult to produce a system with a DR of 12 stops ? What makes the difference? The sensor? The AD conversion? Can it ever be achieved by a 35mm DSLR ?

A DR of 14 stops is quite dubious, since a bit depth of 12 can only encode 12 stops with linear integer encoding as we have been discussing. As I read Roger's essay, the higher dark frame noise at ISO 100 is due mainly to ADC noise and could be reduced by an ADC with a better SN, i.e. higher bit depth if other factors remain the same. The gain with the 1D MII is 13 electrons/12 bit data number, and it would be only a fourth that with a 14 bit ADC. An error in the least significant bit would have less effect.



It will be interesting to see what he reports for the 1D MIII. Already, he has noted that the 14 bit 40D falls short because of in inferior ADC.

Bill

DR is about the range of brightness covered and bit depth about smoothness.

Let's paint a picture. For simplicity sake a black and white picture.

Let's say on a 8 fstop System you are allowed to use 0% white as black and 100% white as eh... white.
If you have only 4 fstops you would be restricted to use something like 0% to 6% or 22% to 28%.

The fstop limits the difference between min and max brightnesses you can use to paint.

The bit depth defines how smooth you can make a transition between minimum and maximum brightness.

If you have 5 Bits in the first case every step in your transition would be 3% white in the second case it would be 0,2% white.

The second would be much smoother because the distinguishable values are much closer together.

To have a comparable "smoothness" the 8 fstop image needs 9 bits. So for comparable smoothness higher DR needs more bit depth.

DR and bit depth work just exactly like that. The DR defines the spread of the min and max values. And the bit depth defines how smooth the transition between min and max is.
Of course larger spread of min max needs more bit depth to get the same smoothness.

And of course bit depth is only "usable" bit depth. If 4 of the 12 bits are just noise, they do not count.

Depending on the subject, a shot might look better with high DR and lower bit depth, or lower DR with higher bit depth.

Regardless of subject insanely high DR combined with horrendous high bit depth (can depth be high ?) will give the best images.

Regards
SH

It sounds like you are trying to separate dynamic range and bit depth. This is possible conceptually, but as I discussed in a previous post, it is not possible in practice, at least with current linear sensors. Increased bit depth gives both increased smoothness and increased dynamic range.

Taking your own example, covering the range 0-100% with 5 linear bits gives a step size of roughly 4%, while covering the same range with 8 linear bits gives a step size of roughly 1/2%. The three extra bits provide steps which are eight times finer. But the dynamic range is also 8 times greater. DR is not calculated as 0-100%, which would be infinite when you divide by zero. Rather, DR is calculated as the maximum brightness divided by minimum non-zero brightness. Since the 8-bit system can faithfully record brightness as low as 1/2%, it has 8 times the DR of the 5-bit system, which cannot faithfully record anything darker than 4%. The system with higher bit depth necessarily has higher dynamic range.

The quantization errors we have been talking about occur in the linear part of the work flow before gamma is applied. You can't really ignore them, since at low ISO they contribute significantly to total apparent read noise and thus dynamic range. According to Roger's analysis, the main source of read noise at low ISO is ADC induced noise.

Bill

You may find this article of interest in your discussions. CCD Dynamic Range Specifications Link

It sheds some light on why dynamic range is a moving target for any device.

As I understand it in my technically limited way, the brightests f/stop (or brightest 2 or 3 stops) of dynamic range use an unnecessarilly large proportion of the available bits to describe the signal. The steps are too small, the gradations unnecessarily fine; too fine for the eye to discern.

Such a system is wasteful. There needs to be a redistribution of bits. If we simply increase the number of bits available to describe the signal, by moving from 12 bit depth to 14 or 16 bit depth, most of the additional bits will be used to quantise the brighter stops of dynamic range, serving absolutely no purpose whatsoever because we already have more than enough bits for those brighter stops.

At the lower end where the signal tends to be obscured by noise, it's not clear to me how additional bits can improve matters much because the additional bits would surely also be used to describe the noise with even greater clarity. Both signal and noise are equally enhanced. Unless there is a way of separating the noise from the signal, any improvement in the quality of the signal cannot be realised simply by providing greater bit depth.

The ideal system for great dynamic range would be one where the individual photoreceptors were able to reduce their quantum efficiency in proportion to the amount of light they received, in real time. As the wells fill, the quantum efficiency gradually decreases, in a non-linear way that more closely matches human vision, so that each brighter stop of DR requires increasingly more than double the amount of light.

With such a system, we would be able to substantiallyincrease exposure without blowing highlights and at the same time provide more light for better exposure of the darker parts of the image. This would be a more sophisticated version of Fuji's SR system which employs 2 photoreceptors of different sensitivity under the one microlens.

Ray, your points are all good. See a few comments below.


The extra bits provide finer resolution throughout the range. You are correct that the extra resolution is wasted at the high end, but it is potentially useful at the low end (depending on noise). If the system has as many bits as needed at the low end (bit resolution comparable to noise) and therefore too many bits at the high end, then a compression scheme like gamma encoding can be used to redistribute the bits, lowering the total bit count but preserving the useful resolution. This is how Photoshop manages to work with 8-bit images derived from 12-bit sensors without much loss.


If the system is limited by noise rather than bit resolution, then by definition the bit depth is already sufficient to capture all the available information. In this case you are right, more bits (also redistributed bits) are not needed and do not help.


Sensors like this have been designed and built, but they have not found widespread use in photography. The most common architecture is a CMOS sensor with logarithmic voltage conversion. See for example http://www.photonfocus.com/html/eng/cmos/linlog.php.

EricV,
That seems reasonable but isn't gamma encoding done at the output stage, either in ACR conversion of RAW or in-camera conversion for viewable jpeg purposes? Does redistribution of bits after the initial A/D conversion serve much purpose regarding fundamental image quality?

From what I've read, the 14 bit A/D conversion does seem to provide some benefits in the 40D, but one can't be sure to what extent such improvements are due to the greater bit depth.

John Sheehy claims there is about 1/2 a stop improvement in the shadows, for example. Bob Atkins claims the noise reduction feature of the 40D (which can be on or off) does not reduce resolution in any discernible way and results in a slightly cleaner image than the 20D can produce in the same circumstances. However, with noise reduction off, the 40D image is actually slightly noisier than the 20D.

It would be fun to buy a 40D (or hire one if possible) to check out such factors to see what is really going on and what practical significance there might be in any pixel-peeping improvement one might discover.

Well, I mean in the shadow end of the DR at ISO 100; if the camera exposes the RAW data more conservatively, however, there could be less improvement in the shadows and more in the highlights. Metering is somewhat arbitrary.

I wonder if its time to do some real world tests.

We can argue 'till the cows come home about wether it's the bit depth that makes the difference (my opinion is that it isn't).

Back at school when we did film tests we would shoot a grey card at different exposures and plot the desities on a graph. Can't we do the same with our digital cameras and compare our results?

I'll test mine and post my results ... will you post yours?

If that higher read-noise at lower ISO speeds is indeed largely due to the noise floor of the A/D converter itself, a better A/D converter could help. However, if Clark is already accusing a 14-bit A/D converter of limiting DR to well under 14-bits (16000:1), I am a bit skeptical that he might be wrongly singling out the ADUs as the main DR bottleneck.

If instead a good part of the noise arises earlier in the trip from electron well to A/D converter, a better ADU would not help as much. And as far as I can tell, Clark's measurements cannot distinguish, since all he has is the output of the A/D converter, not its analog input signals.

Since it is unlikely that the photoreceptors or photodiodes on the 40D's sensor have greater capacity than the photoreceptors on the 20D, it would seem unlikely that the 40D could have greater dynamic range in the highlights, no matter what the increase in bit depth. Surely any improvement would be revealed in the shadows, assuming identical exposure adjusted for any variance in ISO accuracy between the two cameras.

Whenever I've tried to compare the DR of 2 cameras (such as the D60 and 20D), I've taken a series of shots with both cameras at 1/3 stop and 1/2 stop intervals, then tried to find as close a match as possible between any 2 images that appear to be equally and fully exposed to the right and which appear to contain equal amounts of highlight detail and/or equal degrees of blown specral highlights.

I've then examined the shadows for noise and detail.

Hi
I simply prefer to see real side by side pictures of the same high contrast scene that shows the difference between a P30+ on a Mamiya and the 1DmkIII at 14 bits. a true comparison that will show DR in shadows and in highlights, colors and gradation. it can save a lot of discussions about this topic and will show real life shooting instead of camera manufacturer claims or theories . CAN ANY BODY HERE CAN DO THIS SIMPLE COMPARISON AND PRESENT IT HERE???

Menachem reiss www.reiss.co.il

Bullcrap. There's also noise to consider, the LEAST significant of which is quantization error. With the vast majority of cameras out there (MFDB, DSLR, and digicams) the shot noise, read noise, dark current noise, and other forms of noise that end up in the RAW come into play before you start having quantization issues. Going from 12 bits to 14 bits is meaningless if the median RAW value is 128 (on a 14-bit scale) when shooting a dark frame, you have only 7 stops between the clipping point and the noise floor.

Dynamic range is simply the number of stops between clipping and the noise floor. Defining the noise floor can be a bit subjective, because some types of noise are more visually acceptable than others, and a simple standard deviation calculation or similar statistical analysis does not account for this. If the most significant factor in the noise floor is quantization error, then increasing bit depth will be beneficial. But if other noise sources are more significant (which in almost always the case) increasing bit depth will NOT increase usable dynamic range. If the black level in the RAW data is below 16, then you should worry about quantization error. Otherwise, you need to deal with other noise sources before increasing bit depth will provide any meaningful benefit.

There is also the issue of whether any perceived benefits are due directly to the actual contents of the extra bits, or just the fact that they may force converters to use more precision.

I'm pretty sure he's wrong about that for Canons. Most other brands of DSLR clearly have a single initial read amplification, and then just use a low-gain amp to drive the ADC for different ISOs. Nikon D2x for example, and many of the older Nikons, have 15x the blackframe noise in ADUs at ISO 1600, compared to ISO 100. That small difference in blackframe noise in electrons is answered simply by the ADC noise. For Canons, however, all the evidence points to the initial read at the photosite just being noisier (in electrons) for the lowest ISOs. The 1D* cameras and the 5D all have groups of ISOs - x, 1.25x, and 1.6x that all have the same blackframe noise in electrons. And every other group of three is different from every other group of three. There are no gaps or spikes in their histograms unique to the 1.25x and 1.6x ISOs, so the scaling is done *before* the ADC, so if the initial photsite reads were all the same, then blackframe noise in electrons would be more close for 160 and 200, 320 and 400, etc.

Also, if you assume that the total blackframe noise is the square root of the sum of photosite read noise squared plus ADC noise squared, no values will satisfy real world blackframe noise at ISOs 1600, and 100, while also satisfying ISO 400, on the cameras I've tested.



Yes, I think that this is the real problem, IMO. The 400D and the 40D get read noise down to 1.65 and 1.38 ADU (12-bit) respectively, compared to around 2.0 and higher for previous cameras, so it may be a better ADC in use (the improvements are greatest at the lower ISOs) over previous cameras in the series. The "pro" cameras have had blackframe noises down around 1.26 for a while now.

That sync's up exactly with my understanding of all this, certainly with scanners and, unless someone can explain why a linear encoding would be much different, cameras too.

The part above about the useable data above noise is also the clincher. Who decides at what point we start measuring? In the old days, a company like Agfa and Microtek would have different dynamic range spce's for what was in fact the identical hardware. Let the marketing boys into the mix, the science gets a bit dicey.

Maybe Michael can do a piece on this and put to rest the controversy.

The one guy who I deeply respect to provide the skinny is Mike Collette who basically invented the scanning back (Betterlight). But I haven't run into him or heard from him in a few years, not sure if he's still around. But this guy knows his stuff and can explain it so I can understand it.

If you go back and read the previous post I was referring to, you will see that it started:

"Bit depth and dynamic range really are the same thing, provided that
1) System noise is low enough that bits are not wasted digitizing noise
2) The ADC is linear, so each bit represents twice as much signal as the previous bit."

I knew that without the noise disclaimer someone would make a post much like yours (though perhaps without the first sentence), but I guess the point has to be repeated every time. So, bullcrap to your bullcrap.

Even if both conditions are met, bit depth is merely one factor that limits dynamic range, not a measurement of dynamic range itself, just like the displacement of a car engine is a factor that limits its maximum horsepower, but is not useful as an actual horsepower measurement; there's too many other factors that come into play. You still have to specify a minimum number of levels to provide an acceptably low level of shadow posterization (most people would say betweeen 8 and 32, depending on who you you ask) and factor in the fact that white balance scaling is never 1:1:1 between the color channels (there can be over 2 stops of difference between the levels of the red channel and blue channel when shooting with tungsten light) so dynamic range is going to be 3-6 stops less than bit depth, depending on WB of your lighting and how fussy you are about shadow posterization. And then there's the fact that the highest voltage value from the sensor is usually calibrated to produce a slightly-below-maximum outputput value from the ADC, so your maximum numeric output from the ADC is usually 10% or so below the greatest possible mathematical value. For example, Canon DSLRs usually clip about value 3600 out of a possible 4095 maximum. The theoretical best DR you're going to get out of a 16-bit system is about 12 stops, even if sensor and ADC noise is not a factor at all (which is never the case). So you're wrong in theory, and even more wrong in real-world practice.

So bullcrap to your bullcrap on my bullcrap.