Full Frame Image Quality, Myth or Reality?

[mostlyboringphotog]

Larger sensor image quality can be derated to equivalence of any crop sensor size and show that the image quality is not very different per the same magnification. However, to view the large sensor image compared to the same FoV crop sensor image, the larger sensor image will not be enlarged as much as the crop sensor. To me, this is one absolute advantage of a larger sensor that a crop sensor could not over come.

 

[En Trance]

Larger sensor image quality can be derated to equivalence of any crop sensor size and show that the image quality is not very different per the same magnification. However, to view the large sensor image compared to the same FoV crop sensor image, the larger sensor image will not be enlarged as much as the crop sensor. To me, this is one absolute advantage of a larger sensor that a crop sensor could not over come.

In practice this is never done. One either steps back to frame the same view in favor of the FF, or shoot with a wider angle to eliminate the FF advantage.

 

[FingerPainter]

The subject area is constant, it does not change. The sensor area is constant, it does not change. These are true no matter where the camera is moved.

Yes, but when you move the camera, the image of the subject on the sensor does change size.

The light reflected off of the subject has no idea that it is going to pass through a lens and then onto a sensor. It is more or less intense exactly according to the inverse square law.

Double the subject distance, and you'll tell me the light intensity will be quartered. But double the distance and the area that same light covers on the sensor will also be quartered. A quarter the intensity over a quarter of the area means it is still the same intensity per unit area. That's why exposure doesn't change with subject distance.

But that larger sensor at the closer distance that has the same intensity per unit area, has more area, so it has more light total. It doesn't have more light because of the shorter subject distance. It has more light because it has a larger surface area. You seem to accept that more light results in a better SNR, so, there you have it: Larger sensors collect more light, and thus give better SNR.

Larger sensors have another advantage too. A lens doesn't have to be as sharp to yield a given level of sharpness for a given megapixel count. Or a lens on FF yields a sharper image than that same lens on APS-C for the same portion of the image circle.

Sharpness is essentially a measure of how much of the light from a singel point source falls on the same pixel. When 24 MP are spread over a FF sensor, each pixel covers more than twice the area of a pixel on 24MP "HF" sensor. So a lens only has to be 1/1.5 as sharp on a FF sensor to resole the same as a lens on an APS-C sensor. Or alternately, the same lens mounted on a 24MP FF sensor will result in a sharper image than that lens mounted on a 24MP APS-C sensor, bcaseu it will be able to direct more of th light from a point source onto a larger pixel.

You've talked abour MP count more than once. It has no significant impact on noise level across the image. The 12MP Sony A7s, the 16MP Nikon D4s, the 22 MP Canon 5DIII, the 24MP Nikon D610, and the 36MP Sony A7R all have roughly the same sensor size, and all have nearly identical SNR performance at most ISOs.

In contrast, similar techonolgy APS-C sensors from all three conmpanies have about one stop worse SNR performance at all ISOs. That's pretty much what you'd expect if sensor size was driving noise performance.

The Canon sensor has inferior DR at low ISO. That is because it's older design introduces read noise between the sensor and the ADC. This read noise obscure low level signals, thus reducing DR.

 

[En Trance]

The subject area is constant, it does not change. The sensor area is constant, it does not change. These are true no matter where the camera is moved.

Yes, but when you move the camera, the image of the subject on the sensor does change size.

The light reflected off of the subject has no idea that it is going to pass through a lens and then onto a sensor. It is more or less intense exactly according to the inverse square law.

Double the subject distance, and you'll tell me the light intensity will be quartered. But double the distance and the area that same light covers on the sensor will also be quartered. A quarter the intensity over a quarter of the area means it is still the same intensity per unit area. That's why exposure doesn't change with subject distance.

But that larger sensor at the closer distance that has the same intensity per unit area, has more area, so it has more light total. It doesn't have more light because of the shorter subject distance. It has more light because it has a larger surface area. You seem to accept that more light results in a better SNR, so, there you have it: Larger sensors collect more light, and thus give better SNR.

Larger sensors have another advantage too. A lens doesn't have to be as sharp to yield a given level of sharpness for a given megapixel count. Or a lens on FF yields a sharper image than that same lens on APS-C for the same portion of the image circle.

Sharpness is essentially a measure of how much of the light from a singel point source falls on the same pixel. When 24 MP are spread over a FF sensor, each pixel covers more than twice the area of a pixel on 24MP "HF" sensor. So a lens only has to be 1/1.5 as sharp on a FF sensor to resole the same as a lens on an APS-C sensor. Or alternately, the same lens mounted on a 24MP FF sensor will result in a sharper image than that lens mounted on a 24MP APS-C sensor, bcaseu it will be able to direct more of th light from a point source onto a larger pixel.

You've talked abour MP count more than once. It has no significant impact on noise level across the image. The 12MP Sony A7s, the 16MP Nikon D4s, the 22 MP Canon 5DIII, the 24MP Nikon D610, and the 36MP Sony A7R all have roughly the same sensor size, and all have nearly identical SNR performance at most ISOs.

In contrast, similar techonolgy APS-C sensors from all three conmpanies have about one stop worse SNR performance at all ISOs. That's pretty much what you'd expect if sensor size was driving noise performance.

The Canon sensor has inferior DR at low ISO. That is because it's older design introduces read noise between the sensor and the ADC. This read noise obscure low level signals, thus reducing DR.

Stop changing my words so you can write your script. I said that the sensor area does not change. You are lying and saying that I said that the image does not change when the camera is moved. What benefit are you trying to gain from lying?

The area of the subject, say a model's body, is constant. The sensor area, say 24mm x 36mm does not change ever. Light travels between these areas no matter where the subject and sensor are and along its path, it (Light) follows the Inverse Square Relationship Always.

 

[Abe R Ration]

For dummies.

For the rest.

The story is that a very large image sensor with the Mystical, Magical size of 35mm film, has Image Quality that is Greatly Enhanced. The assumption is that pixels and voids are optimized for Image Quality.

Is this true? Or are some manufacturer's struggling to keep up with the sensor game and simply marketing their inferior, gigantic, sensor as a benefit, as they claim that Resolution (MP Rating) is not important for Digital Camera Design? (These aren't the Droids that you are looking for!)

Are crop sensors simply aimed at the lens sweet-spot and reaping the benefit of not using the more challenging periphery of the frame?

Please, no questions about the question. (What do you mean by Blue? What planet are you shooting on? When you say 35mm do you mean 35.00 or 35.00001? Because I need to know that before I give my script to the world.)

Interested in your thoughts..., Maybe!

 

[FingerPainter]

The subject area is constant, it does not change. The sensor area is constant, it does not change. These are true no matter where the camera is moved.

Yes, but when you move the camera, the image of the subject on the sensor does change size.

The light reflected off of the subject has no idea that it is going to pass through a lens and then onto a sensor. It is more or less intense exactly according to the inverse square law.

Double the subject distance, and you'll tell me the light intensity will be quartered. But double the distance and the area that same light covers on the sensor will also be quartered. A quarter the intensity over a quarter of the area means it is still the same intensity per unit area. That's why exposure doesn't change with subject distance.

But that larger sensor at the closer distance that has the same intensity per unit area, has more area, so it has more light total. It doesn't have more light because of the shorter subject distance. It has more light because it has a larger surface area. You seem to accept that more light results in a better SNR, so, there you have it: Larger sensors collect more light, and thus give better SNR.

Larger sensors have another advantage too. A lens doesn't have to be as sharp to yield a given level of sharpness for a given megapixel count. Or a lens on FF yields a sharper image than that same lens on APS-C for the same portion of the image circle.

Sharpness is essentially a measure of how much of the light from a singel point source falls on the same pixel. When 24 MP are spread over a FF sensor, each pixel covers more than twice the area of a pixel on 24MP "HF" sensor. So a lens only has to be 1/1.5 as sharp on a FF sensor to resole the same as a lens on an APS-C sensor. Or alternately, the same lens mounted on a 24MP FF sensor will result in a sharper image than that lens mounted on a 24MP APS-C sensor, bcaseu it will be able to direct more of th light from a point source onto a larger pixel.

You've talked abour MP count more than once. It has no significant impact on noise level across the image. The 12MP Sony A7s, the 16MP Nikon D4s, the 22 MP Canon 5DIII, the 24MP Nikon D610, and the 36MP Sony A7R all have roughly the same sensor size, and all have nearly identical SNR performance at most ISOs.

In contrast, similar techonolgy APS-C sensors from all three conmpanies have about one stop worse SNR performance at all ISOs. That's pretty much what you'd expect if sensor size was driving noise performance.

The Canon sensor has inferior DR at low ISO. That is because it's older design introduces read noise between the sensor and the ADC. This read noise obscure low level signals, thus reducing DR.

Stop changing my words so you can write your script. I said that the sensor area does not change. You are lying and saying that I said that the image does not change when the camera is moved. What benefit are you trying to gain from lying?

You really are something else. Where did I say that that you said the image does nor change when the camera is moved? Nowhere.

I'm not lying. I'm agreeing with you. Perhaps you are reading more into my comments than what is there.

You didn't make a claim on whether the image changes size when you move the camera. I started with the two true statements you made, and then added another one you didn't make (nor did you contradict it).

The area of the subject, say a model's body, is constant.

Yes.

The sensor area, say 24mm x 36mm does not change ever.

Yes.

Light travels between these areas no matter where the subject and sensor are and along its path, it (Light) follows the Inverse Square Relationship Always.

OK. I'm agreeing with you. Let's see what the implications of your correct observations are.

Take a shot of your model from two metres away from her/him. (You say you usually shoot with six inches above and below the model. If you used a 36x24mm sensor in portrait orientation, and your model was an unusually short 5'0", then the model's image on the sensor would be 30mm tall. Now step back 1m further from the model and take another shot. Don't change the focal length (i'm sorry that this will leave more than 6" above and below the model), the lighting, or the exposure settings on the camera. How tall will the model's image be on the sensor now? 20mm. Because we didn't change the lighting or exposure settings, the model's image in the second photo will be just as bright as in the first photo. That means the same number of photons per unit area made the model's image on the sensor for each photo. However, the model's image (not the whole sensor image, only the portion covered by the image of the model) on the second shot is only two-thirds the height (and width) so only 4/9 as many photons made the second image of the model. Since the amount of noise in the light varies with the square root of the amount of light, the SNR across the model in the second shot is only (4/9)/(2/3) = 2/3 what it was in the first shot. IOW, the second image of the model is noisier.

 

[edispics]

http://en.wikipedia.org/wiki/Image_noise#Effects_of_sensor_size

or:

"The size of the image sensor, or effective light collection area per pixel sensor, is the largest determinant of signal levels that determine signal-to-noise ratio and hence apparent noise levels, assuming the aperture area is proportional to sensor area, or that the f-number or focal-plane illuminance is held constant. That is, for a constant f-number, the sensitivity of an imager scales roughly with the sensor area, so larger sensors typically create lower noise images than smaller sensors. In the case of images bright enough to be in the shot noise limited regime, when the image is scaled to the same size on screen, or printed at the same size, the pixel count makes little difference to perceptible noise levels – the noise depends primarily on sensor area, not how this area is divided into pixels. For images at lower signal levels (higher ISO settings), where read noise (noise floor) is significant, more pixels within a given sensor area will make the image noisier if the per pixel read noise is the same.

For instance, the noise level produced by a Four Thirds sensor at ISO 800 is roughly equivalent to that produced by a full frame sensor (with roughly four times the area) at ISO 3200, and that produced by a 1/2.5" compact camera sensor (with roughly 1/16 the area) at ISO 100. This ability to produce acceptable images at higher sensitivities is a major factor driving the adoption of DSLR cameras, which tend to use larger sensors than compacts. An example shows a DSLR sensor at ISO 400 creating less noise than a point-and-shoot sensor at ISO 100."

 

[PK24X36NOW]

There is no Science associated with designing a digital sensor to be the same size as a filmstrip. That is like designing an accelerator pedal to be the exact same length as a whip.

It isn't science... it was a business decision.

You could make a digital sensor any size you want to. In the early days, they were made very small, because the fabrication and material cost was so high. When the cost came down, some manufacturers (Canon, Nikon and Sony) wanted to use their EXISTING AF lens designed for film cameras on their new digital cameras.

They had been using these lenses for their smaller APSC sensored cameras, for the very same reason, but since the lenses were designed for a larger negative (or sensor) the perspective was changed. Once they started making "full frame" DSLRs they had lenses that matched the bigger sensor perfectly.

So the "full frame" 36mm x 24mm digital sensor was born.

This was a practical decision because it enabled the use of literally a thousand lenses that were designed for that filmstrip you speak of. Of course, those lenses could be used with sensors smaller than 36mm x 24mm, and they often were, but by matching the sensor size to the negative size the lenses would precisely the same perspective as they had with film.

The manufacturers got an instant lens catalog for their high end digital SLRs by using the lenses they had already designed for 35mm film SLRs. Think about it. Zero cost for R&D.

 

[PK24X36NOW]

The subject area is constant, it does not change. The sensor area is constant, it does not change. These are true no matter where the camera is moved.

Yes, but when you move the camera, the image of the subject on the sensor does change size.

The light reflected off of the subject has no idea that it is going to pass through a lens and then onto a sensor. It is more or less intense exactly according to the inverse square law.

Double the subject distance, and you'll tell me the light intensity will be quartered. But double the distance and the area that same light covers on the sensor will also be quartered. A quarter the intensity over a quarter of the area means it is still the same intensity per unit area. That's why exposure doesn't change with subject distance.

But that larger sensor at the closer distance that has the same intensity per unit area, has more area, so it has more light total. It doesn't have more light because of the shorter subject distance. It has more light because it has a larger surface area. You seem to accept that more light results in a better SNR, so, there you have it: Larger sensors collect more light, and thus give better SNR.

Larger sensors have another advantage too. A lens doesn't have to be as sharp to yield a given level of sharpness for a given megapixel count. Or a lens on FF yields a sharper image than that same lens on APS-C for the same portion of the image circle.

Up to now, you were doing good...

Sharpness is essentially a measure of how much of the light from a singel point source falls on the same pixel. When 24 MP are spread over a FF sensor, each pixel covers more than twice the area of a pixel on 24MP "HF" sensor. So a lens only has to be 1/1.5 as sharp on a FF sensor to resole the same as a lens on an APS-C sensor. Or alternately, the same lens mounted on a 24MP FF sensor will result in a sharper image than that lens mounted on a 24MP APS-C sensor, bcaseu it will be able to direct more of th light from a point source onto a larger pixel.

D'Oh! Now you're off the rails. The additional sharpness you get has nothing to do with "bigger pixels" whatsoever. It has to do with the fact that the LENS resolution in lp/mm is fixed, and is applied to more "mm" on the FF sensor which is more than twice as big. (150% as big in linear terms).

You've talked abour MP count more than once. It has no significant impact on noise level across the image. The 12MP Sony A7s, the 16MP Nikon D4s, the 22 MP Canon 5DIII, the 24MP Nikon D610, and the 36MP Sony A7R all have roughly the same sensor size, and all have nearly identical SNR performance at most ISOs.

In contrast, similar techonolgy APS-C sensors from all three conmpanies have about one stop worse SNR performance at all ISOs. That's pretty much what you'd expect if sensor size was driving noise performance.

The Canon sensor has inferior DR at low ISO. That is because it's older design introduces read noise between the sensor and the ADC. This read noise obscure low level signals, thus reducing DR.

 

[PK24X36NOW]

The story is that a very large image sensor with the Mystical, Magical size of 35mm film, has Image Quality that is Greatly Enhanced. The assumption is that pixels and voids are optimized for Image Quality.

Is this true? Or are some manufacturer's struggling to keep up with the sensor game and simply marketing their inferior, gigantic, sensor as a benefit, as they claim that Resolution (MP Rating) is not important for Digital Camera Design? (These aren't the Droids that you are looking for!)

Are crop sensors simply aimed at the lens sweet-spot and reaping the benefit of not using the more challenging periphery of the frame?

Please, no questions about the question. (What do you mean by Blue? What planet are you shooting on? When you say 35mm do you mean 35.00 or 35.00001? Because I need to know that before I give my script to the world.)

Interested in your thoughts..., Maybe!

Oh goodie...another "Since I chose a smaller format, the benefits of a larger format must be purely mythical" rationalization thread.

Image quality of FF is better relative to smaller formats because the lenses would have to be much better for the smaller formats in order to keep pace. Specifically:

An APS-C lens needs 150% of the linear resolution of a FF lens, just to keep pace.

A FT/MFT sensor needs 200% of the linear resolution of a FF lens, just to keep pace.

And so on.

Basically, the smaller the sensor, the better the lenses have to be to keep up with the quality achievable with a larger sensor. The smaller format lenses aren't good enough on a relative basis to make up for size difference between FF and the smaller formats, and are not ever going to be, since the cost of such lens quality would make the price unacceptably high.

The "cropped" APS-C format owes its existence to the fact that FF 35mm sensors couldn't be made at affordable prices when digital imaging was a newly developing and immature product, nothing more. If they could have produced 35mm sensors for the same price as APS-C sensors when the first "affordable" DSLRs were introduced, you would have never seen an APS-C format DSLR. There was never any "advantage" to them aside from the small-enough-to-make-the-product-sellable price tag.

Technology will not "solve" the quality differences, because the same "tech" can be (and is) used in 35mm sensors as in APS-C and MFT sensors.

Now before anyone gets wound up, just to be clear...

I am NOT saying you can't take a good image with a smaller format.

I am NOT saying you can't take an excellent image with a smaller format.

I am simply saying that you will get an even better result with the larger format. The format you choose and the related compromises you make as respects image quality level that you accept as "good enough" is up to you. If you're happy with what you chose, be that APS-C, MFT, 1" or a f___ing thumbnail sensor, good for you. Just stop trying to rationalize your decision by suggesting that FF's quality advantage over those smaller formats is "mythical," when it is anything but.

PASSION is a good thing! You win the prize. But Reading is Fundamental. The Myth is that something is superior about 24x36 when in actuality it is merely a coincidence. There is nothing Magical about it.

No, there is most definitely something superior about 24x36 relative to smaller formats, it is no myth.

There is nothing "magical" about APS-C either, and its emergence as a digital format is even more of an "accident" or "coincidence" than the emergence of 35mm as a film format.

The thread digs a little deeper into the mindset of a manufacturer who can not do 24MP and then convinces an industry that Resolution is not an important aspect of Digital Camera Design!

No, this thread is rationalization and trolling by you.

If I am buying MP and someone can put 24 of them in a Half Container and their Competitor can put 22 of them in a Container that is twice as big, at best, and has to charge $twice as much money for that, then I may be led to believe that one manufacturer has his sensor game on point.

Hey - news flash - if MP is all you're buying than somebody already put 41MP in a sensor a fraction of the size of an APS-C camera (in a cell phone cam). So what's better about your APS-C camera sensor - sensor size?! LMAO I guess the argument is only valid for YOUR chosen format.

The half sensor ignores the additional light

And thus is at a distinct image quality disadvantage (noise). Applies only when taking pictures with different DOF though.

that you are claiming for the full sensor and that dirty little half sensor has to be forced to take it down the throat by using a wide angle lens.

Still like its $dirty, $little, $price.

Which has nothing to do with image quality.

A digital camera, lens, sensor combination that is optimized by design has much more merit than a sensor size that is designed to maintain use of some old film photographers lens collection.

ROTFLMAO.

APS-C DSLRs are nothing more than FF cameras with undersized sensors in them. They were designed around existing 35mm lens mounts with existing 35mm register distances. There is absolutely NOTHING "optimzed" about APS-C DSLRs whatsoever.

Think Chevy Camaro with a 4-cylinder engine sitting in the engine bay designed for an 8-cylinder engine.

You would be the troll who is making an unwanted appearance on my thread. Guess I will give you a pass since you appear to be hyper excitable. Maybe a shot of scotch will help you. Its Friday, have one on me.

I'm no troll (that would be you from post numero uno in this thread), just someone who presents you with information that challenges your rationalizations, thereby causing you some apparent distress (that you begged for with this thread right from the start).

Sorry, but I have heard enough of you for this life. Gave you a shot, but just look what you have done with it. I have heard the last from you in bold lettering. Guess I don't need to hear more. Check this out, CLICK!

Guess you can't even muster anything resembling a comeback at this point, which is hardly surprising given the vacuous argument you started (and persisted) with.

I'm quite sure the majority of the people participating in this thread feel precisely that way about you. Did you really expect a flood of people telling you how brilliant you were when you started this tripe thread?!

 

[Sonyshine]

Bish! Bash! Bosh!

 

[PK24X36NOW]

Larger sensor image quality can be derated to equivalence of any crop sensor size and show that the image quality is not very different per the same magnification. However, to view the large sensor image compared to the same FoV crop sensor image, the larger sensor image will not be enlarged as much as the crop sensor. To me, this is one absolute advantage of a larger sensor that a crop sensor could not over come.

I think you're missing something here - you can "derate" noise performance by making images equivalent, but you still retain the advantage of demanding much less lens resolution compared with the smaller sensor. This advantage never goes away.

 

[PK24X36NOW]

Bish! Bash! Bosh!

--
http://www.ipernity.com/doc/i.kirk

Just like a Batman fight, huh?

 

[Sonyshine]

you even got a crop shooter siding with the FFs

"crop-shooter' Is that some kind of put down?

--
http://www.ipernity.com/doc/i.kirk

 

[Sonyshine]

Nicely put Marty!

 

[PK24X36NOW]

Larger sensor image quality can be derated to equivalence of any crop sensor size and show that the image quality is not very different per the same magnification. However, to view the large sensor image compared to the same FoV crop sensor image, the larger sensor image will not be enlarged as much as the crop sensor. To me, this is one absolute advantage of a larger sensor that a crop sensor could not over come.

In practice this is never done. One either steps back to frame the same view in favor of the FF, or shoot with a wider angle to eliminate the FF advantage.

Neither "eliminates" the FF advantage.

 

[Sonyshine]

nt

 

[Luke Kaven]

you even got a crop shooter siding with the FFs

"crop-shooter' Is that some kind of put down?

Life is a crop-shoot.

But it's better than being a share-cropper.