[JinDesu]

On a per-pixel level, the D800 and the K-5 have the same noise level.

On a noise in a picture level, the D800 trounces the K-5 if DOF is allowed to change.

At the same DOF the D800 and the K-5 will have the same noise, you can see that it's within spittin' distance in the chart below.

Dr. Camera: F/it.doesnt.matter

Or if you normalize them to the same resolution - then the D800 will trounce the k-5/D7000. But per pixel, they are the same (given those reqs).

[Parallax]

The only non constant in my example is the area of the sensor being used.

[jsherman999]

But the D610/D600 beats the D800 in high ISO too.

It actually does not, except at 100%

At the same display sizes, the D800 and D610 have about the exact same noise
.

Yes.

[JinDesu]

It actually does not, except at 100%





Yes.

I said D800. Not D800E.

[jsherman999]

I said D800. Not D800E.

They are effectively the same, abut 4% difference in the DXO score. The point is, there is no low-light advantage in the image there just because the pixels are bigger - that's a misconception. There can be an advantage given some other two sensors, but it's going to be because of sensor efficiency not simply because of pixel size. Here, have a look at this.

---------- Post added 04-16-14 at 04:55 PM ----------

So if size is "THE" factor then a photo shot with a D800 in full mode will have less noise than one shot of the same scene, same lighting, same exposure, and same lens in APS-c mode, right? If size is "THE" factor, it would have to be, right?

Unless I misunderstand your question, the answer is yes - as long as you are enlarging the aps-c crop and displaying at the same size. If you just take a D800 image at say 8x10 and crop it 1.5x so it's now 5x7, that 5x7 area will still have the same noise it did before you cropped it out of the 8x10 - but it will remain a smaller image.

When you shoot the D800 in 'crop' mode, you are cropping the image 1.5x and enlarging it to the same dimensions as the uncropped 'full' mode would have had, and the crop mode has about a stop more noise as a result. In fact, aps-c crop mode on the D800 equals almost exactly the D7000/K5 native output, in resolution, noise, FOV, DOF and DR. This is why it never made much sense to have a D800 for most things and a D7000 for 'reach' - you could get exactly the same image by just cropping the D800 image.


.

[ElJamoquio]

The only non constant in my example is the area of the sensor being used.

FOV, DOF, SNR don't remain constant...?

[Na Horuk]

The only non constant in my example is the area of the sensor being used.

In the equivalence world, you need to use a smaller sensor, with the same settings to achieve the same FoV and then they can say the DoF is wider and that the noise is worse and so on. Basically, the idea is to set up a scenario where a the FF sensor will come out on top. The problem with equivalence measures is that they are based on the wrong premises and give biased results. The math.. is just math. But the conclusion is still wrong

[ElJamoquio]

In the equivalence world, you need to use a smaller sensor, with the same settings to achieve the same FoV and then they can say the DoF is wider and that the noise is worse and so on. Basically, the idea is to set up a scenario where a the FF sensor will come out on top. The problem with equivalence measures is that they are based on the wrong premises and give biased results. The math.. is just math. But the conclusion is still wrong

That's wrong, starting with the first sentence.

[northcoastgreg]

To some people, that matters - to some, not, but to say that if it doesn't matter to you so it shouldn't matter to anyone is silly

Where did I say it shouldn't matter to anyone because it doesn't matter to me? All I said is that most photographers are more concerned with FOV equivalency than with aperture equivalency. That statement is not based on my own inclinations, but on observation. I'm acquainted with scores of photographers, including professionals and prize winning advanced amateurs. I've never run across any photographer "in the wild" who knows or cares about aperture equivalency or who practices a type of photography that would benefit from such knowledge. Indeed, from discussions I've had with these photographers and from remarks and questions I've heard in Q & A sessions, I would hazard to guess that many of them aren't even aware that larger sensors are capable of yielding narrower DOF. In critique sessions at the local camera club, the critique leader, a professional portrait photography, is constantly getting on people for not getting enough of the subject in focus (he usually shoots at f16 on FF). He has never gotten on anyone for not getting a sufficiently blurred background.

Aperture equivalence is something that seems to have been emphasized only recently. Most of the articles that I've found on it through google were written in the last few years; the oldest mention I could find was Ken Rockwell from 2009. Articles before 2009 that discuss equivalency seem to mostly concentrate on FOV equivalency. I suspect that concern over aperture equivalency arose following the introduction of the Nikon D3, which helped stir up interest in and advocacy of FF to a fever pitch.

Regardless of the technical or scientific merits of equivalency, some have used it as an instrument of advocacy to bash smaller formats, and this probably explains most of the hostility that has arisen against it. Most photographers, however, are probably merely indifferent to aperture equivalency, as they are to most of the other technical issues that are quibbled about on internet photography forums.

[nomadkng]

... I've never run across any photographer "in the wild" who knows or cares about aperture equivalency or who practices a type of photography that would benefit from such knowledge. Indeed, from discussions I've had with these photographers and from remarks and questions I've heard in Q & A sessions, I would hazard to guess that many of them aren't even aware that larger sensors are capable of yielding narrower DOF. In critique sessions at the local camera club, the critique leader, a professional portrait photography, is constantly getting on people for not getting enough of the subject in focus (he usually shoots at f16 on FF). He has never gotten on anyone for not getting a sufficiently blurred background.

Aperture equivalence is something that seems to have been emphasized only recently. Most of the articles that I've found on it through google were written in the last few years; the oldest mention I could find was Ken Rockwell from 2009. Articles before 2009 that discuss equivalency seem to mostly concentrate on FOV equivalency. I suspect that concern over aperture equivalency arose following the introduction of the Nikon D3, which helped stir up interest in and advocacy of FF to a fever pitch.

Regardless of the technical or scientific merits of equivalency, some have used it as an instrument of advocacy to bash smaller formats, and this probably explains most of the hostility that has arisen against it. Most photographers, however, are probably merely indifferent to aperture equivalency, as they are to most of the other technical issues that are quibbled about on internet photography forums.

which summarizes my whole questioning of this "narrow depth of field is the holy grail" logic. If the model's eye is in focus but not the nose, this is a good image? If one orange but not the banana is in focus, this is a good image? I guess maybe it's the world of photography I grew up in, but since when is shooting with 2" of focus plane the norm? When I do pet portraits I shoot f22 if I can get away with it, (even cats if you want to be snarky about dog muzzles being long). Macro photography needs f11 at min, f16 more comfortably. The whole premise of landscape shooting used to be f64. I think the only time I get anywhere around f5.6 is when I'm shooting wildlife in low light because I'm fighting desperately to get fast enough shutter speeds and keep the ISO under 1600.

On a further note, I did a test shoot on my last hike, I actually shot some landscape scenes at f7.1 and f8 just to see how they would turn out. I also captured the same scene at my usual f13-f22 range. To be honest with you, I preferred the f13-f22 images even at first glance. The loss of resolution to diffraction was not nearly as abrasive as the loss of my hyperfocal range. The blurred OOF branches/rocks etc were a distraction compared to the entire image being in focus but a slight loss of resolution because of diffraction.

So in that sense, APS-C actually helps me, and probably a lot of other photographers. Deeper DoF also gives us a greater margin for error.

[jsherman999]

Where did I say it shouldn't matter to anyone because it doesn't matter to me?

You didn't, really, I was using your post as a jumping off point there to address a common theme. A lot of folks do have that attitude, and wander in here (FF forum) to express it in various angry ways.

.

Aperture equivalence is something that seems to have been emphasized only recently. Most of the articles that I've found on it through google were written in the last few years; the oldest mention I could find was Ken Rockwell from 2009. Articles before 2009 that discuss equivalency seem to mostly concentrate on FOV equivalency. I suspect that concern over aperture equivalency arose following the introduction of the Nikon D3, which helped stir up interest in and advocacy of FF to a fever pitch.

I think it is interesting how it evolved on DPR and other forums:

Circa 2002-2004 or so ; Most of the first DSLRs were aps-c (1.5x or 1.6x (Canon),) and people had to learn how to convert the FOV to this new format in their heads, so that when they used their existing 35mm SLR lenses they would know what to expect in terms of this 'new' FOV.

All was well and good, but then a few folks started noticing that something didn't look right, DOF seemed a tad too deep for the same FOV and aperture that they were used to, and there were a few people who actually thought the digital camera was misrepresenting their F-stop Some folks like Lee Jay, Josepsh Wisniewski, Joseph James and others tried to show them why they weren't imagining things and described the physical reasons why they were seeing what they were seeing. A few articles came from that, with I think this one being the best and most continually updated.

It seems like it really started to get talked about a lot starting in 2006 when the Canon 5D hit, though, because that body sold in huge numbers and was owned by people who also shot aps-c.

I believe that you mostly still see photography instructors talking about FOV only, and in fact a lot of them probably don't even understand equivalence themselves. It's too bad, but I guess if they're teaching mostly small-aperture shooting like Landscapes it just never comes up much in their day-day. Also, they may never learned to teach in an environment where different formats could share lenses, or perhaps they didn't want to 'confuse' their students, especially if they themselves couldn't fully quantify the effect.

Joseph Wisniewski talks about how he was trying to describe equivalence (FOV and DOF, not just FOV, how the F-stops meant different things to the resulting images too between formats, not just FL) before there was a handy label for it, back in the 80's

The main reason I think it's too bad more people don't understand it is because the whole reason FF is even better in shot-noise performance than smaller formats is described by equivalence - because the linear aperture of the lens for that same FOV and F-stop on the larger format is larger. ie:

50mm f/2.8 FF => 50 / 2.8 => about 18mm linear aperture (entrance pupil)
33mm f/2.8 aps-c => 33 / 2.8 => about 12mm linear aperture

Taken from the same distance, the above allows the same FOV, and shutter speed, but the larger linear aperture brings a larger total amount of light to the sensor, which will result in less noise, and the larger linear aperture will also result in less DOF. The reason it's important to understand is that if you plan to not accept the more shallow DOF - if you know you won't like it, and will be stopping down a lot - then you lose that 'noise' advantage.

"I guess I don't like that DOF, I'm going to stop down."

50mm f/4.3 FF => 50 / 4.3 => about 12mm linear aperture. (DOF now the same - but noise advantage gone! Doh!)

Just an example of how knowing about these relationships could help you realize something like that before you make a big purchase!

.

---------- Post added 04-17-14 at 01:01 PM ----------

which summarizes my whole questioning of this "narrow depth of field is the holy grail" logic. If the model's eye is in focus but not the nose, this is a good image? If one orange but not the banana is in focus, this is a good image?

Is more DOF control at the high-end of the F-stop range a good thing?

This 'one model's eye in focus' is a constant, tired example - doesn't anyone have any imagination, can't anyone think of any shooting scenario that doesn't involved close-in portraiture?

I sometimes feel like the forum world is made up of 1) studio portrait photographers who constantly worry about both eyes in focus and 2) landscape photographers who didn't realize their lenses even opened up wider than f/8

.

[Rondec]

You didn't, really, I was using your post as a jumping off point there to address a common theme. A lot of folks do have that attitude, and wander in here (FF forum) to express it in various angry ways.

.


I think it is interesting how it evolved on DPR and other forums:

Circa 2002-2004 or so ; Most of the first DSLRs were aps-c (1.5x or 1.6x (Canon),) and people had to learn how to convert the FOV to this new format in their heads, so that when they used their existing 35mm SLR lenses they would know what to expect in terms of this 'new' FOV.

All was well and good, but then a few folks started noticing that something didn't look right, DOF seemed a tad too deep for the same FOV and aperture that they were used to, and there were a few people who actually thought the digital camera was misrepresenting their F-stop Some folks like Lee Jay, Josepsh Wisniewski, Joseph James and others tried to show them why they weren't imagining things and described the physical reasons why they were seeing what they were seeing. A few articles came from that, with I think this one being the best and most continually updated.

It seems like it really started to get talked about a lot starting in 2006 when the Canon 5D hit, though, because that body sold in huge numbers and was owned by people who also shot aps-c.

I believe that you mostly still see photography instructors talking about FOV only, and in fact a lot of them probably don't even understand equivalence themselves. It's too bad, but I guess if they're teaching mostly small-aperture shooting like Landscapes it just never comes up much in their day-day. Also, they may never learned to teach in an environment where different formats could share lenses, or perhaps they didn't want to 'confuse' their students, especially if they themselves couldn't fully quantify the effect.

Joseph Wisniewski talks about how he was trying to describe equivalence (FOV and DOF, not just FOV, how the F-stops meant different things to the resulting images too between formats, not just FL) before there was a handy label for it, back in the 80's

The main reason I think it's too bad more people don't understand it is because the whole reason FF is even better in shot-noise performance than smaller formats is described by equivalence - because the linear aperture of the lens for that same FOV and F-stop on the larger format is larger. ie:

50mm f/2.8 FF => 50 / 2.8 => about 18mm linear aperture (entrance pupil)
33mm f/2.8 aps-c => 33 / 2.8 => about 12mm linear aperture

Taken from the same distance, the above allows the same FOV, and shutter speed, but the larger linear aperture brings a larger total amount of light to the sensor, which will result in less noise, and the larger linear aperture will also result in less DOF. The reason it's important to understand is that if you plan to not accept the more shallow DOF - if you know you won't like it, and will be stopping down a lot - then you lose that 'noise' advantage.

"I guess I don't like that DOF, I'm going to stop down."

50mm f/4.3 FF => 50 / 4.3 => about 12mm linear aperture. (DOF now the same - but noise advantage gone! Doh!)

Just an example of how knowing about these relationships could help you realize something like that before you make a big purchase!

.

---------- Post added 04-17-14 at 01:01 PM ----------



Is more DOF control at the high-end of the F-stop range a good thing?

This 'one model's eye in focus' is a constant, tired example - doesn't anyone have any imagination, can't anyone think of any shooting scenario that doesn't involved close-in portraiture?

I sometimes feel like the forum world is made up of 1) studio portrait photographers who constantly worry about both eyes in focus and 2) landscape photographers who didn't realize their lenses even opened up wider than f/8

.

It isn't like that Jay, but I find myself a lot more in the f4 to f8 range of the aperture range than anywhere else (on APS-C, of course). Even with this, it is awfully easy to find yourself with too little depth of field. I was at the zoo shooting with 200mm at f4 and it wasn't enough depth of field. Now, if I had had a 300mm f2.8 on full frame, I would have a stop less depth of field available to me, but I just don't know that it would have helped me out a lot if it already was narrow where I was at.



This photo I think should have been shot at f5.6.

But certainly, I think you would have to admit that all of the equivalency arguments tend to fall apart a little at the telephoto end of things, where APS-C at the least holds its own.

[Parallax]

From the dazed, blank stare I suspect that was taken right after he (she?) read this thread.

[ElJamoquio]

It isn't like that Jay, but I find myself a lot more in the f4 to f8 range of the aperture range than anywhere else (on APS-C, of course). Even with this, it is awfully easy to find yourself with too little depth of field. I was at the zoo shooting with 200mm at f4 and it wasn't enough depth of field. Now, if I had had a 300mm f2.8 on full frame, I would have a stop less depth of field available to me, but I just don't know that it would have helped me out a lot if it already was narrow where I was at.



This photo I think should have been shot at f5.6.

But certainly, I think you would have to admit that all of the equivalency arguments tend to fall apart a little at the telephoto end of things, where APS-C at the least holds its own.

But of course you could've just cropped your hypothetical FF, or ran your 300mm f/4 (why do you need an expensive F/2.8?) at F/8, right?

Think of the DA*300 on FF as a phenomenal zoom lens - a 200-300 F/2.6-4 zoom, with auto zoom.

[dtmateojr]

Do people ever wonder why the size of a can does not matter for a rain gauge? This should give you a very good hint as to why aperture size or sensor size do not affect SNR. Think about it.