D800 ISO

[Iliah Borg]

D1H - D2H - D2Hs - D3 - D3s

Don't seen an H after D3. So much for the H line.

The wonderful part of it is that D1, D1H, D1X, D2X, and D2Xs are all based on the same sensor design.

--
http://www.libraw.org/

 

[TOF guy]

which is the primary contributor to visible noise in mid-tones of high-ISO images.

That and the fact that application of gamma will make the noise more obvious in mid-tones

But it's also about dark areas of the image. First because while the "shot noise" increases with signal, the noise / signal ratio decreases with it. This impacts very dark areas of the image because of the way our vision works (Weber's law, the reason why we can see faint stars over a dark background at night time, but we can't see them in daylight - yet they're out there. In fact we see them, we just don't distinguish them).

Regardless of the electron/photon ratio (Ne/Np) that you achieve, shot-noise limited SNR cannot be better than SQRT(Np).

Exactly !

--
Thierry

 

[Marianne Oelund]

Except that the D2H was not very good at high ISO ...

D2Hs worked just fine for me. Of course, at that time 1600 was considered "high."

There is nothing the D2H is good at that the D2X could not achieve.

Speed: 8fps without cropping down. Perhaps it's a detail of no consequence for some, but it completely ruled the D2x out for my purposes. Now, if there had been an f/2 zoom available, the D2x HSC mode might have been an option (just as it could make the V1 a viable option today).

The D2X is indeed the next DX flagship after the D2H and in this regards is its successor.

Not for my applications - and I'm not alone. I still own and use both, and they are not used for the same purposes.

 

[Marianne Oelund]

But it's also about dark areas of the image.

Yes, but to discuss that, I would then have needed to get into some detail about the region where read noise starts to dominate. In terms of DN, that's dependent on ISO setting and camera model - just more detail than I have time to delve into now. Please excuse my over-simplification.

 

[bobn2]

Bob,

Thank you for admitting that you are arguing out of context.
Where is it written that a premise must be contained within a single sentence?

Mike, your behaviour is pathetic and childish. What you said was absolutely wrong and I have shown that it was with a simple example. Everything else in that post, which flows from your faulty premise is also wrong. Since it appears that you require me to knock it down point by point, i will now do just that. It will take a couple of posts...

In other words, as long as you don't use the additional pixel count of a 36MP sensor to make a larger print than you would make with a 24MP sensor having the same dimensions, you can enjoy a higher resolution in that print, as seen at a given viewing distance, without having to use a wider aperture (smaller f-Number) to prevent diffraction from inhibiting that higher resolution.

As I showed, this is wrong. You can use some of the extra resolution to print larger while still having higher resolution.

But if you try to make a larger print with the 36MP sensor than you would make with the 24MP sensor having the same dimensions and you're intent is to deliver the same resolution in the larger print as you have enjoyed in the smaller print, for viewing at a given viewing distance (i.e., you do not anticipate an increase in viewing distance that's proportional to the increase in print size), you will have to shoot at a wider aperture (smaller f-Number) to prevent diffraction from inhibiting your desired print resolution.

As above this is also untrue, you can get more resolution and a larger print without using a smaller f-number.

In short:

1) When like-sized sensors with dissimilar pixel counts are used to produce like-sized prints (identical enlargement factors) the higher-density sensor will yield a higher resolution print while shooting at the same f-Number. There's no need to open to a wider aperture to avoid diffraction when using the higher-density sensor, because the diameter of diffraction's Airy disks will be the same size in both prints, after enlargement, at any given f-Number.

Not wrong, but not exactly news, either.

2) When like-sized sensors with dissimilar pixel counts are used to produce proportionately dissimilar print sizes (dissimilar enlargement factors), the higher-density sensor will yield the same print resolution in the larger print as was had with the lower density sensor in the smaller print ONLY if the higher-density sensor uses a proportionately smaller f-Number (a wider aperture). Only by shooting with a smaller f-Number will the diameter of diffraction's Airy disks be the same size in both prints, after applying dissimilar enlargement factors.

Wrong, as stated above. It's simply a third repetition of your faulty premise.

Here's a formula for calculating the aperture at which diffraction will begin to impact an image at a specified combination of desired print resolution for a 10-inch viewing distance (in lp/mm), enlargement factor, and anticipated viewing distance (inches):

N = anticipated viewing distance / desired print resolution for a 10-inch viewing distance / enlargement factor / 0.0135383

Well, a fairly simple restatement of the Airy disc formula, but ignores the fact that diffraction, by itself does not determine the MTF of a lens until one gets to diffraction limited apertures (in the proper sense, not this diffraction limited by pixel size nonsense), so pretty much a formula without useful application unless you're working above f/11 or so.

If you prefer to always assume someone may view your prints as closely as 10 inches, you can use this formula:

N = 1 / desired print resolution for a 10-inch viewing distance / enlargement factor / 0.00135383

Note: The two versions of the formula have different constants (a 10x difference).
Thanks for letting us know how to multiply by 1. Will be useful for some.

I recommend you select a desired print resolution no lower than 4 lp/mm and no higher than 8 lp/mm.

Thank you for your recommendation.

So, if you know your desired print resolution and how much larger the diagonal of your final print will be than your sensor (or film) diagonal, you can use this formula to calculate the f-Number at which diffraction will begin to inhibit your desired print resolution for a 10-inch viewing distance.

But what about the cases when the resolution is not diffraction limited? Or the cases where the lens is diffraction limited but the resultant MTF is high enough for the cameras MTF to make a worthwhile contribution?

Whenever someone declares that diffraction isn't a problem for a given camera until a certain aperture is selected, do yourself a favor and IGNORE what they are saying if, in the same breath, they have not QUAILIFED their statement with a specified combination of enlargement factor, anticipated viewing distance, and desired print resolution. There are so many possible combinations of these variables that any attempt to state that diffraction will become "visible" at some constant aperture f/N is ludicrous.

What isn't ludicrous is that for most of the aperture range for most of the cameras we use, the overall system is not diffraction limited, and increases in camera MTF can deliver useful gains in overall resolution because the diffraction contribution to the MTF is high enough for it not to act as a 'limit' in the presence of other 'limiting' factors, the aberration MTF and the camera MTF.
... to be continued
--
Bob

 

[bobn2]

It's possible to have a resolution requirement that's low enough, and/or an enlargement factor that's small enough, and/or a viewing distance that's great enough to permit the use of even the smallest apertures (largest f-Numbers) available on your lens without concern for diffraction.

OK, but again, not news.

The degradation caused by diffraction is never a problem until the Airy disks are large enough in the final print, after enlargement, to inhibit your personal goal for print resolution at an anticipated viewing distance.

Here's a little news for you . The Airy disc is an abstraction, the average photographer will never recognise an image as a superposition of 'Airy discs', so really, this is not useful information.

While I'm on my soap box... Anyone who says they shoot with smaller apertures (larger f-Numbers) than those calculated using the above formula for a desired resolution of 4 lp/mm at a 10-inch viewing distance and finds the results to be satisfying, either has poorer than average vision -OR- does not understand what they are losing when stopping down too far.

Or has different standards for acceptable sharpness than you? More of an issue would be the people who use your formula thinking they can produce humongous prints at large apertures, only to find that they cannot ignore resolution loss to lens aberrations or camera.

A print can have accutance (edge sharpness) even though it lacks resolution (detail). Detail lost to diffraction at the time the light is striking the sensor can not be restored. At best, all you can do is sharpen the image afterwards. It is impossible to restore subject detail that has been lost to diffraction.

Well, that depends. It depends on which contrast level you are working with as 'cutoff'. Still, the point is that while we need to understand the effects of diffraction, your discussion is oversimplified to the point of uselessness, and ignores the relevant issues discussed above.

Enlargement factor is the enemy here and in digital photography, that equates to pixel density.

This is nonsense. Enlargement factor does not, at all, equate to 'pixel density'.

The more pixels a manufacturer squeezes onto a sensor of a given size, the greater will be the enlargement factor necessary to actually exploit all those pixels.

This is also nonsense. The extra pixels can be used to realise additional resolution, microcontrast and freedom from aliasing at the same print size.

As enlargement factor goes up for a given combination of viewing distance and desired print resolution, the f-Number at which diffraction begins to inhibit that resolution goes down.

That is true, but as said, ignores the other factors, and in context leads one to believe that pixel density has something to do with this, which it does not.> Big, low-density sensors give us more diffraction-free stops to work with,

complete balderdash. What gives you more diffraction free stops is faster lenses. It is a fact of life that for a given print size, diffraction blurring at a given depth of field is the same, whatever size or pixel density sensor you use. Big sensors do not break that rule. The question is therefore (if we ignore the rather important point that for most of the f-number range the system is not remotely diffraction limited) which system offers the shallowest DOF, which isn't invariably the one with the largest sensor. In digital, it is generally 135 FF of the Canon persuasion.

in addtion to lower noise.

And another thing is that the amount of light collected, and therefore the noise, is the same at the same print size and DOF for any sensor size and pixel count. The noise therefore depends on sensor efficiency not size at fixed DOF. Sensor efficiency as a trend tends to drop as sensors get bigger, with the D3s being a notable exception.

The vast majority of photographers are unknowingly content with amazingly low resolution, but high acutance prints.

If they are content with that, why should they not be.

For an explanation of the difference between resolution and acutance, see Sean McHugh's tutorial on sharpness:
http://www.cambridgeincolour.com/tutorials/sharpness.htm

One of his less misleading and misinformed tutorials.

Here's an Excel spreadsheet I put together, titled: "Sensor (or Film) Size vs. Print Size vs. Diffraction"
http://home.globalcrossing.net/~zilch0/tools/Sensor1.xls

Thanks for the effort, but given the faults in the premise it is based on, not as useful, is it?

Watch out folks, he'll be attacking individual words, next...

Well, it was you who objected to me only finding the glaring fault in your opening premise. Are you happier now I've pulled the lot apart?

Frankly, Mike, pretending that resolution is solely about diffraction simplifies things past usefulness. Providing non-factual 'guidance' based on your faulty premise is hardly helpful.
--
Bob

 

[teodorian2]

Here is a classification (scroll down to the bottom of the page) which clearly validates the facts I have written:

Two lines of "Flagships" as I said.

http://en.wikipedia.org/wiki/Comparison_of_Nikon_DSLR_cameras

I want to see this classification on a Nikon site. All the rest is just the view of a person.

The reality is that Canon had a well defined sport oriented camera, the 1D. Nikon never had that line and its cameras straggled the lines. Now Canon itself is giving up on the cost of a separate APS-H sensor, on a $,4000 + camera that sells to a relatively small market, and is introducing a 1DX which is more of a jack of all trade, like Nikon.

Proof in the pudding: you still haven't answered my question: what is it that you can do with the D2H which you cannot do with a D2x. All you can do is refer to a somewhat arbitrary classification. But you can't named a photographic activity. Your lack of true answers speaks volume.
--
Thierry

You claimed that the D2X was the successor of the D2H and I have prooved that statement to be false. Nothing more. Nothing less.

 

[TOF guy]

You've "proved" nothing. D2X was the next flagship generation after the D2H. That's all.

 

[TOF guy]

Except that the D2H was not very good at high ISO ...

D2Hs worked just fine for me. Of course, at that time 1600 was considered "high."

Neither were good at high ISO. That was just not a feature where Nikon put the emphasis at the time.

There is nothing the D2H is good at that the D2X could not achieve.

Speed: 8fps without cropping down. Perhaps it's a detail of no consequence for some, but it completely ruled the D2x out for my purposes. Now, if there had been an f/2 zoom available, the D2x HSC mode might have been an option (just as it could make the V1 a viable option today).

Maybe the HSC did not work for you. But it's presence means that Nikon was again trying to make a camera straggling the lines. The D2x tried to do sports (in fact it was specifically introduced right before a sport event, and Nikon was lending the camera during that event). It was the flagship successor to the D2H, even if for you both cameras were usable and in different situatinos

The D2X is indeed the next DX flagship after the D2H and in this regards is its successor.

Not for my applications - and I'm not alone. I still own and use both, and they are not used for the same purposes.

Say the rumored 24 MP D400 comes out, but you still like the D2H for a specific usage where you want small images: it would still not mean that the D400 is not the current flagship DX camera, succeeding in this role to the D300(s), itself succeeding the D2X(s), the latter succeding the D2H(s).

--
Thierry

 

[Iliah Borg]

Dear Thierry,

Except that the D2H was not very good at high ISO ...

D2Hs worked just fine for me. Of course, at that time 1600 was considered "high."

Neither were good at high ISO. That was just not a feature where Nikon put the emphasis at the time.

To the best of my knowledge Nikon were trying to do their utmost with high ISO on D2H/D2Hs. Also, for me D2H was quite usable for low light, and I continued to use it even after I got D2X - because even with downsampling D2X images D2H was still better, not to mention the frame rate available with D2H which served me better for catching the right phase of the motion and helped handholding.

I do not know for a fact if Nikon themselves thought of D2X being sort of D2H flagship successor, but what was the timeline with D2Hs? Was it out after D2X?

--
http://www.libraw.org/

 

[TOF guy]

I do not know for a fact if Nikon themselves thought of D2X being sort of D2H flagship successor, but what was the timeline with D2Hs? Was it out after D2X?

Hi Iliah,

Using dpreview's database the D2Hs was announced ~5 months after the D2X.

I stand corrected.

--
Thierry

 

[TOF guy]

The D2X is indeed the next DX flagship after the D2H and in this regards is its successor.

Not for my applications - and I'm not alone. I still own and use both, and they are not used for the same purposes.

After checking the database I stand corrected: the D2Hs came after the D2X.
--
Thierry

 

[Marianne Oelund]

Model introductions overlapped:

The D2H was introduced July 2003, the D2Hs in Feb. 2005 (after D2X). The D2Hs "inherited" a number of advancements that first appeared with the D2X. Final firmware updates for both D2Hs and D2X occurred in Sep. 2006, only about 1 year before the lines ended.
The D2X was introduced Sep. 2004 (before D2Hs), the D2Xs in June 2006.

Both lines continued until the availability of the D3 in late 2007, i.e. the D2H/s had a longer run than the D2X/s did.

The D2Hs ($3500) and D2X/s ($5000/$4700) were at different price points.

There is no question that, for some users, the D2X/s was a welcome replacement for the low-resolution D2H, but for others, the D2H/s replacement didn't arrive until the D3 release. Nikon recognized that fact by continuing D2Hs availability throughout the D2Xs product life.

There is little sense in those two groups of users beating each other over the head about their preferred product lines.

 

[TOF guy]

After checking the database I stand corrected: the D2Hs came after the D2X.
--
Thierry

 

[Iliah Borg]

Dear Thierry,

Using dpreview's database the D2Hs was announced ~5 months after the D2X.

Thank you.

--
http://www.libraw.org/

 

[Mike Davis]

Bob,

You're thinking through a straw. It's as if you're only capable of digesting one or two sentences at a time and as long as you continue to do that, you'll continue to argue out of context.

Bob,

Thank you for admitting that you are arguing out of context.
Where is it written that a premise must be contained within a single sentence?

Mike, your behaviour is pathetic and childish.

Wow, were your hands shaking when you wrote that?

What you said was absolutely wrong and I have shown that it was with a simple example.

Your example was out of context.

Everything else in that post, which flows from your faulty premise is also wrong.

My "faulty premise" being the first sentence of my post, right? Well, at least I've got you reading past the first sentence, now.

Since it appears that you require me to knock it down point by point, i will now do just that. It will take a couple of posts...

Joy!

In other words, as long as you don't use the additional pixel count of a 36MP sensor to make a larger print than you would make with a 24MP sensor having the same dimensions, you can enjoy a higher resolution in that print, as seen at a given viewing distance, without having to use a wider aperture (smaller f-Number) to prevent diffraction from inhibiting that higher resolution.

As I showed, this is wrong. You can use some of the extra resolution to print larger while still having higher resolution.

Your inability to honor context is astounding. You've quoted a paragraph that describes a scenario where like-sized prints are made from like-sized sensors and yet you rebut by suggesting the extra pixels of the higher-density sensor can be used to make larger prints at higher resolutions. Dude - please try to stay inside the ring - I can't hit you if you keep jumping over the ropes!

But if you try to make a larger print with the 36MP sensor than you would make with the 24MP sensor having the same dimensions and you're intent is to deliver the same resolution in the larger print as you have enjoyed in the smaller print, for viewing at a given viewing distance (i.e., you do not anticipate an increase in viewing distance that's proportional to the increase in print size), you will have to shoot at a wider aperture (smaller f-Number) to prevent diffraction from inhibiting your desired print resolution.

As above this is also untrue, you can get more resolution and a larger print without using a smaller f-number.

(Hallelujah! We're on the same page - at least for a brief moment!)

Not at the same aperture! Just look at the formula. The f-Number at which diffraction begins to inhibit a desired print resolution for an anticipated enlargement factor can be calculated as follows:

f-Number = 1 / anticipated enlargement factor / desired print resolution in lp/mm / 0.00135383

Note that if the enlargement factor increases (as it does in the scenario you are quoting), the f-Number must also decrease if we are to maintain the same print resolution in the larger print as was had in the smaller print. The greater resolution you claim is possible in the larger print could only be had by using an even smaller f-Number!

In short:

1) When like-sized sensors with dissimilar pixel counts are used to produce like-sized prints (identical enlargement factors) the higher-density sensor will yield a higher resolution print while shooting at the same f-Number. There's no need to open to a wider aperture to avoid diffraction when using the higher-density sensor, because the diameter of diffraction's Airy disks will be the same size in both prints, after enlargement, at any given f-Number.

Not wrong, but not exactly news, either.

Yay!

2) When like-sized sensors with dissimilar pixel counts are used to produce proportionately dissimilar print sizes (dissimilar enlargement factors), the higher-density sensor will yield the same print resolution in the larger print as was had with the lower density sensor in the smaller print ONLY if the higher-density sensor uses a proportionately smaller f-Number (a wider aperture). Only by shooting with a smaller f-Number will the diameter of diffraction's Airy disks be the same size in both prints, after applying dissimilar enlargement factors.

Wrong, as stated above. It's simply a third repetition of your faulty premise.

Oops! There you go, jumping the ropes, again.

(Continued in my next post, below...)

 

[Mike Davis]

(... continued from above.)

See the formula. When one or both of the divisors increase, the quotient must decrease.

Here's a formula for calculating the aperture at which diffraction will begin to impact an image at a specified combination of desired print resolution for a 10-inch viewing distance (in lp/mm), enlargement factor, and anticipated viewing distance (inches):

N = anticipated viewing distance / desired print resolution for a 10-inch viewing distance / enlargement factor / 0.0135383

Well, a fairly simple restatement of the Airy disc formula, but ignores the fact that diffraction, by itself does not determine the MTF of a lens until one gets to diffraction limited apertures (in the proper sense, not this diffraction limited by pixel size nonsense), so pretty much a formula without useful application unless you're working above f/11 or so.

Amazing analysis, Bob! Do you see anything in that formula addressing pixel size? Answer: No. By using the word "nonsense" as you have, you've once again proven that you do not understand the context of my argument. You're thinking through a straw.

This formula only tells us the f-Number at which diffraction will begin to inhibit a desired print resolution at an anticipated enlargement factor. If diffraction alone is sufficient to inhibit the desired print resolution at a particular f-Number, no other factors that might improve print resolution in the absence of diffraction need be considered - including lens MTF!

Let me beat the context over your head again:

2) When like-sized sensors with dissimilar pixel counts are used to produce proportionately dissimilar print sizes (dissimilar enlargement factors), the higher-density sensor will yield the same print resolution in the larger print as was had with the lower density sensor in the smaller print ONLY if the higher-density sensor uses a proportionately smaller f-Number (a wider aperture). Only by shooting with a smaller f-Number will the diameter of diffraction's Airy disks be the same size in both prints, after applying dissimilar enlargement factors.

Hint: When like-sized sensors with dissimilar pixel counts are used to produce proportionately dissimilar print sizes, you will have DISSIMILAR ENLARGEMENT FACTORS, and thus, the higher-density sensor will be producing larger Airy disks in the larger print even though the Airy disks start out the same size at the sensor, prior to enlargement - for any given f-Number.

If you prefer to always assume someone may view your prints as closely as 10 inches, you can use this formula:

N = 1 / desired print resolution for a 10-inch viewing distance / enlargement factor / 0.00135383

Note: The two versions of the formula have different constants (a 10x difference).

Thanks for letting us know how to multiply by 1. Will be useful for some.

Joy!

I recommend you select a desired print resolution no lower than 4 lp/mm and no higher than 8 lp/mm.

Thank you for your recommendation.

You're welcome!

So, if you know your desired print resolution and how much larger the diagonal of your final print will be than your sensor (or film) diagonal, you can use this formula to calculate the f-Number at which diffraction will begin to inhibit your desired print resolution for a 10-inch viewing distance.

But what about the cases when the resolution is not diffraction limited? Or the cases where the lens is diffraction limited but the resultant MTF is high enough for the cameras MTF to make a worthwhile contribution?

And what about the case where a meteor falls from the sky and smashes me and my camera to smithereens before I can make the exposure?

I'm not going to apologize for failing to provide a universal formula for aperture selection - please try to stay inside the ring, Bob - I'm talking about the relationship between diffraction, enlargement factor, desired print resolution, and f-Number, in the context of dissimilar print sizes from like-sized sensors of dissimilar pixel density.

Whenever someone declares that diffraction isn't a problem for a given camera until a certain aperture is selected, do yourself a favor and IGNORE what they are saying if, in the same breath, they have not QUAILIFED their statement with a specified combination of enlargement factor, anticipated viewing distance, and desired print resolution. There are so many possible combinations of these variables that any attempt to state that diffraction will become "visible" at some constant aperture f/N is ludicrous.

What isn't ludicrous is that for most of the aperture range for most of the cameras we use, the overall system is not diffraction limited, and increases in camera MTF can deliver useful gains in overall resolution because the diffraction contribution to the MTF is high enough for it not to act as a 'limit' in the presence of other 'limiting' factors, the aberration MTF and the camera MTF.

Agreed! And nothing you've said contradicts my contention.

... to be continued

Roger that!

Mike

--
Mike Davis
http://www.AccessZ.com

 

[bobn2]

You're thinking through a straw. It's as if you're only capable of digesting one or two sentences at a time and as long as you continue to do that, you'll continue to argue out of context.

Your pathetic and childish behaviour is treating a simple issue of engineering fact as a measure of self esteem, to the extent that when it is called, you just can't let it go. You objected to me finding fault with only your first sentence, even though I have shown beyond doubt that what you said is not right. I went through the lot, and you are not happy with that either. The fact is, as I said, that your initial sentence, which is the premise for the rest of a bloated, repetitive and unfocussed post, is untrue and the rest really wasn't worth bothering about. You've been trading a line of 'advice' which is based on incorrect assumptions and facts. Stop doing it, learn and get over the fact that you've been in error, we all are at some time or other.

--
Bob

 

[Mike Davis]

I am encouraged by the knowledge that some people are capable of critical thinking. You, on the other hand, continue to go astray with your childish and pathetic (your choice of words, not mine) insistence that the first sentence of my post, unqualified by the sentences that follow it, can be addressed as a standalone premise, ignoring everything that follows.

It's obviously your hope that sensible, intelligent readers will follow you right off that same cliff, but thankfully, that's not likely to happen.

Happy New Year, Bob.

Mike

--
Mike Davis
http://www.AccessZ.com

 

[d3xmeister]

I read so many posts and still don't understand.

Bob, do you think a D800 with a 36mp sensor and D7000 technology could equal the D3s in high iso/dr performance ?