Buying guides 2015

News Reviews Articles Buying Guide Sample Images Videos Cameras LensesPhonesPrinters Forums Galleries Challenges

What is equivalence and why should I care?

Published Jul 7, 2014 | Richard Butler

1 Depth-of-field equivalence

The concept of 'equivalence*1' is still somewhat controversial and not always clearly understood. We thought it was about time we explained - and demonstrated - what equivalence means and what it doesn't.

What is equivalence?

Equivalence, at its most simple, is a way of comparing different formats (sensor sizes) on a common basis. This is already the way most lenses are talked about: it's quite common to say that a compact camera includes a '28-120mm lens' but the key and (often unspoken) word in that description is 'equivalent.' It's a simple way of describing the range of fields-of-view that the lens offers, cancelling out the effect of sensor size by using a common reference point.

A 100mm equivalent lens on a small-sensor camera will give the same framing and perspective as an actual 100mm lens does on a full-frame camera, regardless of sensor size, because they are equivalent.

Camera	Sensor Format	Sensor diagonal	Crop factor	Focal length	Diagonal Angle of View *2
Canon EOS-1D X	Full Frame (864mm²)	43.3mm	1	100mm	24.4deg
Olympus OM-D E-M5	Four Thirds (224mm²)	21.6mm	2	50mm	24.4deg

It's this logic that the idea of 'crop factors' is based on. The 'Four Thirds' sensor format has a diagonal very close to half that of a 'full frame' sized sensor. And, sure enough, if you calculate the angle-of-view of a 50mm lens on a system with a crop factor of 2, it's the same as for a full frame camera with a 100mm lens.

Equivalence helps compare different lenses and cameras by using 35mm 'Full Frame' as a reference point. This doesn't mean that Full Frame is the best, or even the optimal format, it's just a reasonably well understood point of comparison.

However, it's not just focal lengths that can be thought of in equivalent terms. In our recent reviews of fixed-lens cameras, we've tended to include a chart showing 'equivalent' aperture values and how they change as you progress through the zoom range. In the rest of this article, we'll discuss and demonstrate how apertures and sensor sizes interact.

The reason we do this is because it's become quite common, for any given part of the market, to have to choose between cameras with different lenses and sensor sizes. This wasn't such a common problem in the film era since 35mm was the de facto standard. As such, most people didn't need to directly compare quality and characteristics across those different formats. 35mm was better than 110, medium format was better again and large format was still better. In the digital era, the significance of (and difference between) sensor sizes isn't always appreciated, not helped by a rather opaque naming system.

Apertures and F-numbers

It turns out, the relationship between sensor size and aperture is very similar to that between focal length and sensor size. It's the physical size of the aperture that defines depth-of-field, not its F-number. Two lenses offering the same angle-of-view with 25mm diameter apertures will give the same depth-of-field of shot at the same shooting distance and when viewed at the same output size.

So, let's consider the effect of that 25mm aperture in the lenses we've been discussing. Both lenses give the same angle-of-view, so you're likely to shoot them with the same shooting distance - since that way you'll get the same framing in both photos.

Now think about what that means in terms of F-number (focal length/aperture diameter). The 25mm aperture in the 100mm lens would be considered to be f/4 (100mm/25mm), whereas the same-sized aperture in the Micro Four Thirds lens would be an f/2 (50mm/25mm). The table below shows the calculated depth-of-field for the two lenses, shot at the same working distance:

Camera	Focal length	Aperture diameter	F-number	Depth of field *3	Equiv. F-number
Canon EOS-1D X	100mm	25mm	f/4	Near 16.1m Far 26.3m Total 10.1m	f/4
Olympus OM-D E-M5	50mm	25mm	f/2	Near 16.1m Far 26.3m Total 10.2m	f/4

As you can see, although the lenses are quite different, the 50mm f/2 lens is giving the same framing and the same depth-of-field as a 100mm f/4 lens is on Full Frame. As such, you can say that a 50mm f/2 for Micro Four Thirds is equivalent to a 100mm f/4 Full Frame lens in terms of both field-of-view and depth-of-field.

We don't expect you to take our word for this. Conveniently, it's become fashionable for manufacturers to produce 85mm equivalent, F1.2 lenses, which makes it relatively easy to show the ways in which they are, and aren't equivalent. We'll demonstrate this on the third page of this article.

*1. Not to be confused with 'Equivalents' - an influential series of abstract photographs by Alfred Stieglitz, and an approach to photography that they inspired.
*2. Calculated using TawbaWare's Angular Field of View calculator
*3. Calculated using DOF Master, with the lenses specified, focused to a distance of 20m

Tags: article, equivalence, feature, fundamentals, learn, sensor-size, technology

Next page 2 Equivalence of total light

TABLE OF CONTENTS

Comments

Total comments: 2154

...and I thought this was going to be about Alfred Stieglitz!

My age is showing.

0 upvotes

That's why I included a footnote after the fourth word of the article.

0 upvotes

roygbiv - no, pixel size & pixel count is not the determining factor, sensor size is - or rather, sensor size + larger physical aperture presented for the same exposure. Pages and pages and article after article explain this, as will any sensor or lens designer who frequents this site, so feel free to do the research yourself. For instance, consider the D800 vs. D7000 - same pixel size, same sensor generation, different sensor size, D800 has 1.3 stops advantage at the same exposure.

Edited 17 seconds after posting

0 upvotes

for Nikon 1 V3 if we set the aperture to 5.6 not the "equivalent", then the ISO required should be roughly 3200 which is same as canon 1DX. when we look at the noise performance do we need to compare DOF as well ? For film camera, large format is superior in quality because the actual file is bigger, given same emulsion, for same print size large format is much better because it requires less enlarge ratio. To get a print size of 8x10, a 8x10 field camera just need to contact print, but you need to enlarge 35mm by about 53 times which is the same analogy as pixel size determines noise performance. you won't be able to see any grains on a 8x10 print from a 8x10 negative even they have the same emulsion and processing

0 upvotes

p4 real world total light demonstration - again, if the 4 cameras are of similar MP count then of course the smaller the sensor the worse noise performance because each pixel is smaller. let's say a full frame sensor is 20MP then a MFT sensor of 5MP should give equivalent noise performance. Of course, a 20MP MFT is much worse than the full frame.

0 upvotes

If you have the same sized pixels, you get the same noise performance at the pixel level but the larger sensor produces a less noisy image overall.

For instance, if you compare Nikon D7000 and D810, which have essentially the same sized pixels, the D810 still produces cleaner images when compared at the same viewing size.

0 upvotes

again pg 2 total light
sensor size is small so the "total" light capture is less than larger sensor because the sensor size is only 1/4 of the the larger sensor. It is like you are painting a small piece and a large piece, they look the same (quality, exposure) but you use less paint on the smaller piece i.e. you have same exposure same ISO same aperture on large or small sensor although the actual amount of light reaching the larger sensor is more because of it is physically bigger like you need more paint to paint a bigger piece. But noise performance still not because of just the sensor is bigger, it is the pixel size that determines. If we stitch 4 same pieces of small sensor together to get a bigger sensor, the noise performance is the same because pixel size has not changed

0 upvotes

It all comes down to whether you're considering things at the image level or the pixel level.

At a pixel level, then yes, pixel size and F-number matter. Pixels of the same size will usually give roughly the same performance, regardless of what size sensor they're part of.

However, as soon as you consider things from a whole image level, this changes. With the large sensor (assuming the same pixel size), you have more pixels. As soon as you view the images from the two sensors at the same viewing size, the one from the larger sensor will look better.

Yes, if you take four images with the smaller sensor and stitch them together, you can create the same image, with the same image-level noise. But that's because you've taken four exposures and captured four times as much light as with a single shot.

0 upvotes

pg.2 And, because the more light you capture, the less noisy your image, this is key to why large sensors generally give better image quality than small ones.

I think larger sensor captures better image quality because given same number of pixels, pixel size on larger sensor is bigger thus able to hold more light giving better noise to signal ratio and not simply because you capture more light. We do not need to capture more than enough light for each frame, we need the right amount of light to give correct exposure regardless of sensor size ?

Would like to know more about the equivalent magnification of lens on different sensor size though, it is a bit confusing: a 50mm on MFT gives same angle of view as 100mm on full frame, what about the "actual" magnification like we want to capture distant objects with a long lens, if magnification is the same, then it is just a matter of cropping the frame then one should really considers using full frame if can afford 2 cameras in 1

0 upvotes

Pixel size makes much less difference than intuitively you might think. There's a more in-depth explanation here.

It doesn't matter whether you compare two different sized sensors with the same pixel count or the same pixel size - the larger sensor wins and it wins by about the same amount.

0 upvotes

Take only one SAME FF body with possibility to take pictures with a chosen crop mode (Many FF bodies brands can do that now), so that you get a FF and an APSC 1,5 camera in only one body (2 in one) ON EXACTLY THE SAME SENSOR !!!
On this FF body, install one FF lense with: f2 maximum aperture
Let’s take 2 photos with the same body, same lense, same manually chosen f2 for example, same sensor !! , same exact moment, same subject lighted exactly the same way.
Picture 1 in FF mode and with chosen speed, let’s say: 1/200
Picture 2 in APSC crop mode...WHAT WOULD BE THE SPEED FOR THE PHOTO # 2 ? WOULD IT BE : 1/200 divided by 1,5 = 1/133 (Slower speed in order to let the same light intensity of the same lighted subject go to the APSC cropped same sensor, knowing that the same lense will become a f2 multiplied by 1,5 = f3 in APSC mode (Slower lense than in FF mode as said !! OR, WILL it be SAME 1/200 IN ORDER TO GET EXACTLY THE SAME LIGHT EXPOSITION ON THE SAME SENSOR ON PHOTO #2 RESULT?

0 upvotes

Someone with a signal processing background has to re-write this. FF's advantage with respect to light capture has less to do with the cone of light, and more with size of photo-cell and pixel count. The reason that the real-life total light demonstration doesn't appear to display a linearly-related drop-off in noise performance is because it has little to do with the "total light capture" that this article (and so many well-meaning people) are blaming the lost performance on the light that falls outside the sensor. Unless the lenses are made of drastically different material capable of significantly different light refraction (which they aren't) , the light density is roughly equivalent. What's lost is the sensitivity of the photocells, and their ability to discern signal from noise. The smaller they are, the more difficult it is to discern. Drawing the light cone and showing that FF's light cone is larger than the APS-C light cone is a red herring.

5 upvotes

Hi! I am a media technology engineer and photographer with background to this topic. I respect that you have this understanding of sensor sizes in relation to signal capture, due to the vast amount of inaccurate information on the topic. However, I must inform you that sensor size and sensor "resolution" is quite irrelevant when it comes to capturing light. While it is true that each photo cell receives less signal if the resolution is higher, you receive more samples and the resolution can easily be traded for improved SNR. The only variables which determinine the amount of light captured is the lens and the shutter speed. An f2 lens on an iPhone has similar light gathering capability as that of an f16 full frame lens (as well as DoF). When determining the light gathering capability, simply multiply the f-number with crop factor (for the iPhone approx. 7 I think). I hope you will try and research this further. I spent this time just to help you!

Edited 37 seconds after posting

2 upvotes

RoyGBiv: I agree 100%. I thought exactly the same thing when reading the article. It started out OK explaining focal length and crop factors then lost a lot of credibility in my mind when it made statements like "... because the more light you capture, the less noisy your image, this is key to why large sensors generally give better image quality than small ones".

The small sensors are not noisier because they are smaller – they are noisier because each light gathering photo-cell on the sensor is (nearly always) smaller than those on a larger sensor… and that assumes the same level of technology; some sensors are better than others re signal to noise ratio (eg. compare today's sensors with one from 15 years ago). If the crop sensor had the same size light gathering elements as a larger sensor, it should perform the same. The individual photo-cells on the sensor don’t “know” they must perform better just because they are part of a full frame sensor.

0 upvotes

roygbiv - no, pixel size & pixel count is not the determining factor, sensor size is - or rather, sensor size + larger physical aperture presented for the same exposure. Pages and pages and article after article explain this, as will any sensor or lens designer who frequent this site, so feel free to do the research yourself. For instance, consider the D800 vs. D7000 - same pixel size, same sensor generation, different sensor size, D800 has 1.3 stops advantage at the same exposure.

Edited 1 minute after posting

0 upvotes

I'm a mFT user, but even I can admit that a bicycle is not a plane. Sure my bicycle still gets me from point A to B but at the end of the day it is not a plane.

0 upvotes

What does a 36x24mm have a crop factor of 1? Just because Oskar Barnack used standard available film when designing a small camera?

2 upvotes

Because 35mm was a dominant format for long enough that when digital photography started using other sensor sizes, virtually all (if not all) manufacturers kept using it as a reference point. Quoting focal lengths by reference to the 35mm format is the absolute standard.

8 upvotes

Good answer.

0 upvotes

As all experienced film photographers that enlarged their own shots know, the more you enlarge an image the more noise/grain will be in your print. What they also knew, because they were handling the film/sensor image everyday, is that the same shot taken on a 35mm with a standard lens 50mm, and a 2¼ MF with a standard lens, 75mm, produced images on film/sensor of the same subject at very different absolute sizes. The 35mm film is physically smaller. Therefore, for scenes framed the same way, the absolute size of your subject on the film was much smaller. If you took a portrait of a subject with a 35mm camera, to create an 8x10 print you had to enlarge it 8 times to produce the print. Taking the same shot with the 2¼ you had to enlarge it 4.5 times to produce the print. The less you enlarge the less noise/grain in your print. It doesn't make a difference whether that "print" is on screen or not. Smaller sensors have to be enlarged much more, therefore they reveal more defects/noise.

Edited 3 times; latest 6 minutes since posting

1 upvote

Continued: I believe that is all anyone need understand about sensor size as long as you are comparing sensors of equal quality to each other. As a former Ansel Adams Zone System devotee who shot 4x5 in. film on a view camera, I admit to having been a geek of the first order. But today we have sensors that record much more detail than my film ever has. And also today we have sensors that are vastly smaller than 35mm film, aka FF. The more you enlarge an image the more noise you will see. How should you pick a camera as far a IQ is concerned? Start with sensor size. Compare images with different sensor sizes. Pick the the sensor size with the best IQ you can afford. Next, pick a brand/model that has that sensor size and the ergonomics that suit you best. But also pick a camera system that will let you grow and expand in the future.

Happy shooting!

Edited 2 times; latest 13 minutes since posting

2 upvotes

Excellent presentation, but I really don't care and most people shouldn't either. Crop factor (for FOV) was (still is) a handy idea back when everyone shot FF film and were migrating to digital. This mindless fixation with EXACT FF equivalence in every aspect isn't helpful or even useful for the vast majority of (all) shooters who will never (nor should they) even own a FF camera. I have an OMD and a 6D and I can tell you with absolute certainty that when I'm out shooting with the Olympus - I don't ponder how I should adjust anything to match exactly what I would get with the 6D. There are alot of good reasons for using a crop body camera and with a little practice should be capable of producing great results and obsessing about exact DOF, total light or noise equivalence is just silly. Cameras are getting so good today that we are arguing about less and less. Get the camera that makes sense and use it. Don't drink the Northrup kool-aid.

Edited 10 minutes after posting

4 upvotes

@Richard Butler: Excellent article!

This topic has confused many minds and it is awesome to see it explained that well with such accuracy, conciseness, and supported with actual images.

Only two remarks:
Page 3: You speculate that the more contrasty blur discs from the 1"-type sensor images come from lower LCA. Is it not much more likely, that the claimed "f/1.2" of the lens are not a true "f/1.2"? Have you tried to compare the "f/1.2" exposure to a "f/4" exposure and see whether you actually get the expected increase in intensity? The lens itself may even be "f/1.2" but extreme incident angles on the sensor could mean that the latter actually just "sees" an effective "f/1.8", or so. Camera manufacturers combat this "blindness" of sensors to very wide apertures with hidden ISO increases.

Page 4: You write "generally larger pixels...will tend to give cleaner images", but in one of your comments you demonstrate that you know that the pixel size is immaterial. I'd adjust the text.

1 upvote

I love small sensors. I love them because in theory they solve all our problems, and that is why manufacturers continue to strive to eliminate the one glitch-NOISE- from the equation.

Depth of field freaks need not worry at all because all you need do is use hyperfocal distance and put your subject at the back third of it, or just get closer.

But this misses the point of what these sensors will be one day soon, able to do.

You'll be able to use the best aperture giving the greatest depth of field to produce all in focus shots. You will be able to see a scene from front to back captured at high shutter speeds, in poor light as you can now only in blinding sunlight with a full-frame sensor.

You will have problems though unless you use Leica M, and Voightlanders, because nobody else can be bothered to inscribe large scale accurate depth of field scales on their lenses, even though their lenses are huge compared to M lenses and there are acres of space thereon.

1 upvote

As an enthusiast with little real talent, the ability to take photographs with a shallow dof is very important !

This is the easiest way I can take impressive pictures of my informal subjects that look like I'm 'good at photography' when I post them on facebook.

As a bloke, I simply love endlessly comparing performance figures when considering a purchase choice, so I can feel a bit smug that 'my camera is the best'.

Now, at the summer bbq today, If I pull out my 5D and shoot at 105 f/4 I'll get nice enough piccies but I'll startle my subjects into posing.

My tiny GM5 M43 with equally tiny 45 f/1.8 gives me similar pics while looking like I'm waving a point n shoot around.

The LX100 gives me 35 f/2.3 , nearly as shallow dof and the gaddamn thing is a zoom lens and shoots in the dark without a flash !

heh heh just the truth fellas, I'm guessing we *are* all fellas ?

Everyone knows megapixels went too far. Who's got the shallowest dof in the smallest camera ? Now you're talking ;)

2 upvotes

I'd like to see where the medium format cameras fit in. Is a Pentax 645 worth the additional cost?

Great article, BTW.

1 upvote

Well, if you're going for the shallowest depth-of-field, full-frame might be your 'sweet spot', simply b/c - for all practical purposes - full-frame is the only system that offers f/1.4 equivalent lenses.

Now, yes, there are exceptions - you could put a rare f/1.0 lens on APS-C to get f/1.4 equivalent, or use a Metabones Speedbooster on APS-C with a f/1.4 lens to get f/1.4 equivalent DOF.

But when it comes to medium format, I'm not aware of any f/1.4 lenses off the top of my head, so please let me know if I'm wrong.

Take a Pentax 55mm f/2.8 lens for the 645Z system, for example. Its crop factor is .79x, which makes it ~44mm equivalent in terms of full-frame (FF). If you do an equivalent aperture calculation, it comes out to f/2.2 (down from the original f/2.8 of the lens).

So you're essentially getting 44mm F2.2 equivalent on the 645Z. Point being, you can get to much shallower DOF equivalents (f/1.2 anyone?) on a FF system, b/c of the vast diversity of lenses for FF systems.

5 upvotes

It seems like we could use another page here on diffraction as well Mr. Butler. People are not understanding that the equivalent aperture is also a consistent way of expressing the diffraction limit of every system for a given resolution. I think people generally understand the diffraction limit of FF cameras but don't understand for matching resolutions that diffraction limit also exists on any format at the same equivalent aperture.

These portrait lenses may not offer a maximum aperture small enough but it would be helpful to show diffraction starting at different f-stops but the same equivalent f-stop as you move from format to format.

Equivalent aperture helps in comparing total light and depth of field for moderately sized sensors but it's arguably even more important to consider when dealing with small high resolution sensors with slow zooms or slow ILC lenses in front of modern high mp sensors.

Edited 2 times; latest 2 minutes since posting

1 upvote

mosc I'm sure that you know that diffraction happens at ANY aperture. Any, even f1. The smaller the aperture the bigger the effect. Then you choose an ARBITRARY limit to decide if the diffraction is relevant or not. This limit is often related to the circle of confusion that you are working with or to the pixel pitch. So the "diffraction limit" depends on the pixel pitch or on the intended usage of the image (large print, web content, etc.). Neither of these is fixed for a sensor size.

You talk about "matching resolution" but I know of no cameras in the real world with matching resolution across different format. For example a 64MP full frame sensor to match a 16MP m43 one.
So, like noise, this is an extremely approximated equivalence.

Moreover I doubt anybody is ever going to compare and choose different systems on account of the diffraction difference.

3 upvotes

Interesting opinions, all could be correct in their own aspect..

1 upvote

Please have a look at the caption of the Oly 300/4 picture:

http://bit.ly/1N4ULqh

"Although not as fast as a 600mm F4 would be for full-frame". Is this sentence correct, wrong or just confusing? What does "fast" means in this context? Can I use it to shoot indoor sports or not? Is this equiv. reminder relevant in this article context?

I can see the value of this idea when comparing compact cameras with different sensors (LX7 vs RX vs LX100). But with ILS users usually know how the system(s) they use performs.

I suspect, and see around, that many people get this from this article: smaller sensors have more noise AND lenses are slower (it's not your thesis, I know).
I agree that you have to understand FoV, diffraction and "noise" equivalence, but not at the cost of muddying the waters with the exposure basics. Beginners will get lost in this debate.

And the elephant is still there: tech differences/choices and a few years of ageing can have a huge impact on the noise comparison.

1 upvote

Nobody ever obsessed about equivalence when there were only 35mm, a few medium formats, and a few large formats. Equivalence is real and must be understood when moving between formats, but it isn't helped by marketing organizations like DPReview, or the quote you listed. Aperture equivalence is the real groaner! When speaking of lenses, e.g. 600 f/4, the two numbers refer to the focal length focused at infinity and the aperture ratio. They are 'absolutes'. F/4 in this 'context' refers to brightness and not to any effect on depth of field! Depth of field is altered through variables, and as an 'absolute' is only determined by one factor: image magnification at the film/sensor plane.

1 upvote

Hi Richard I think that this "equivalent aperture" idea is doing only harm and confusion.
Try this: take three different cameras with different sized sensors and shoot a moving subject with the same aperture number, shutter speed and ISO and equivalent focal lenght.
You'll get three "identical" pictures with the same "brightness" and the same amount of motion blur. Do you agree?

Yes, DoF will be different in a precise amount and you can calculate the "equivalent DoF aperture". And yes, the noise will be different but it vastly depends on the actual sensor you are using: try a 5D classic vs a Sony A7S. Do you agree?

So the "equivalent aperture" is NOT equivalent with respect to exposure and motion blur and it is equivalent for DoF and, with a big approximation, for noise.
What is more important? Exposure and image content or noise and out of focus areas?

In which why this "equivalent aperture" concept is better then saying that there is a two stops difference in DoF and less noise?

1 upvote

I don't agree the images will be identical, since you then spend time explaining the ways in which they're different.

Please note that the article doesn't at any point suggest you should ever use equivalence to work out exposure - it is only suggested as a way of comparing two systems. This is particularly relevant when considering buying a camera in part of the market where multiple sensor sizes are available (enthusiast compacts, for instance. APS-C and Micro Four Thirds ILCs). Understanding the effect of sensor size on FoV, DoF and total light (and hence, ~noise) is useful.

However, if you did, for whatever reason, choose to shoot two cameras at equivalent focal lengths, apertures and ISOs (and therefore, the same shutter speed), you' d get the same brightness and motion blur and the same depth-of-field. Not necessarily the same noise (as this can depend on sensors, as state in the article).

4 upvotes

also, in case of zoom lens having fix maximum aperture, example: the canon 24-70mm f/2.8L, used on a full frame, when keeping the maximum aperture the same (at 2.8) while changing the focal length, will it change the aperture diameter as well and affect the amount of light going through the aperture? I do the maths, and at 24mm f/2.8, the aperture is opened at 8.57mm diameter, while at 70mm f/2.8, the aperture is opened at 25mm. will zoom in and out while keeping the aperture the same will change how much light my sensor receives? I have this question because I shoot contemporary dance concerts and more than often choreographers love moody pieces so the lights are extremely low. I can't compensate shutter speed because i need to capture dancers' movements. Also i hate noisy photo so i try not to boost ISO too much, the one that is left is aperture so I want to know if a zoom lens is appropriate in this situation where the aperture diameter change while I'm zooming in and out.

1 upvote

As the article says, equivalent aperture is only relevant when you're comparing across sensor sizes.

If you're mounting the f/2.8 lens on a single camera then the effect of f/2.8 is the same, regardless of focal length. The f-number ends up dictating light per-unit-area (illuminance), and, since you're not changing the amount of area that you're capturing that light on (you're keeping the sensor size the same), the effect is the same.

It's important to bear in mind that the ISO setting doesn't make that much difference to the noise. It's usually the shutter speed and aperture you use.

Edited 15 seconds after posting

2 upvotes

So I read the articles and have a few question:
_ so let say I'm using a 50mm at f/2 on a APS-C camera, which means that the aperture is opened at 25mm diameter. then I switch to a full frame camera, using a wider lens 35mm, if I want to achieve the aperture diameter of 25mm, my lens has to be opened at 1.4 (35/25=1.4).
My questions are:
1. is the depth of field maintain the same between the two settings above?
2. Also, because the full frame will capture more light, does it mean that having the aperture open at 25mm will make brighter image on a full frame compared to the cropped sensor?
3. if yes to question 2, let say i want to capture the same amount of light for the setting i set on the APS-C on a full frame, should i go with a smaller aperture than 25mm? and yes then how can i calculate the correct aperture diameter?

0 upvotes

Should I assume you've got your 35mm and 50mm back-to-front? (Did you intend to have the same field-of-view with both cameras?). I'm going to make the assumption that the difference between a 35mm F1.4 lens and a mathematically ideal 33mm F1.3 lens is negligible.

1) An APS-C camera with 35mm F1.4 and a full frame camera with 50mm F2 shot from the same position and focused on the same thing will have essentially the same depth-of-field (just as the comparable examples in the article have).

2) In this instance, if you had the same shutter speed, the full frame camera would receive less light per unit area but the same total light (because you're measuring more area)

3) I don't think I understand your question here.

1 upvote

I know it's an old article but i hope someone can explain some of my questions.
Here is how i understand it:
I have a 5d and a 450d and a 50mm 1.8. Both cameras have identical flange distance, so the lens is producing the exact same image at the sensor plane.
450d captures a smaller part of this image. I think everyone agree with me so far.
The part i don't understand is why do some insist that the picture that 450d sees is darker than what 5d sees? it is a smaller amount of the total light that is entering the lens but it is also used to illuminate a smaller area. I assume that the amount of light that each pixel (photo cell etc.) receives is the same between the cameras.

1 upvote

You're right: the two cameras will receive the same amount of light on every square mm of their sensors.

However, the larger sensor has more square mm capturing that same brightness, so captures more light in total. And, as this article tries to explain, the more light you can capture, the less noisy your images will appear.

1 upvote

5D has larger pixel size so each pixel is brighter. (Imaging a bigger tube collecting rain) To make up the difference, ISO100 on 450D actually amplifies more than ISO100 on 5D.

1 upvote

Thanks Richard.
Imagine that the two sensors that we are comparing have the exact same pixel size and arrangement, (only different pixel count), don't you agree that each individual pixel receives the same amount of light? so why is it that one pixel should be noisier than the other?
remember that i don't care about diffident FOV. two cameras are at the exact same place and I'm not touching the lens either,

Edited 6 minutes after posting

1 upvote

I'm stuggling here as well. Provided the small and large sensor have identical pixel areas they collect the same number of photons +/- √n. This means that the random 'noise' is identical in the top figure for both sensors. Is not the reason that FF sensors tend to be less noisy than part frame because the pixel size is smaller?

In the top figure the APS-C sensor captures the same light flux per unit area as the FF sensor, but is not capturing the same image, maybe this is where the difference comes?

1 upvote

@sina_hml - If the pixels are the same size, then an APS-C crop from the full frame sensor will be identical to an APS-C image, taken with the same settings.

However as soon as you use all of the full frame sensor, its extra light capturing area will put it ahead.

2 upvotes

esasjl - pixel size makes very little difference in most cases. It's the total area (and hence total light) going into making up the image that accounts for the differences between sensor sizes.

I don't know what you mean by 'the top figure,' so I'm not sure how to respond.

0 upvotes

@richard "If the pixels are the same size, then an APS-C crop from the full frame sensor will be identical to an APS-C image, taken with the same settings."

Hence the name "crop sensor". The light equivalence should have been described in this way in those articles. The "total light" discussion caused even more confusion and the belief among many that what you stated above is not true.

0 upvotes

@richard "Panasonic FZ200 has a 4.5-108mm F2.8 lens

In full-frame terms that would be equivalent to a 25-600mm F15.6 lens.

This means 108mm, F2.8, ISO 100 on a small sensor would be the same as 600mm, F15.6, ISO 3086. Which, as you can imagine, gives you a lot of depth-of-field but is usable in daylight, but almost unusable in low light."

Would it not have been clearer to state that 108mm f2.8 on a small sensor is equivalent to 108mm f2.8 on a FF sensor cropped to a FOV of 600mm.

There are many examples of usable FZ200 images in low light. Aren't the IQ differences between a FF crop and the FZ200 images largely due to pixel size and sensor characteristics/age?

0 upvotes

@nigelht - there will be some differences in performance because of sensor differences but they're likely to be fairly small.

Pixel size has very little role to play (have a look at this, where the full frame and APS-C camera have the same size pixels but the full frame sensor still performs better, thanks to simply having more sensor.

I'll address the reasons for this in part 2 of the noise articles we're publishing (should be later this week).

0 upvotes

@richard "full frame sensor still performs better, thanks to simply having more sensor."

Having more sensor AND having more lens. It isn't "simply" due to larger sensor size.

The presumption that a longer lens exists for FF is not always correct. For example, while you can use the Canon 85mm f1.2 on a M43 there isn't a 170mm f1.2 lens for FF.

Or when using a 400mm f2.8 on the Nikon V3. There isn't a 1080mm f2.8 lens for your D810 to utilize the entirety of the extra sensor area. The best you can do is a 800mm f5.6.

The clearest way to describe light equivalence of crop sensors is to describe it as a crop of a FF sensor because as you state:

"If the pixels are the same size, then an APS-C crop from the full frame sensor will be identical to an APS-C image, taken with the same settings."

Describing equivalence as "600mm f15.6 ISO 3086" is convoluted and leads to the wrong conclusion...such as the camera is unusable in low light.

Edited 43 seconds after posting

0 upvotes

You're right, I didn't express all my assumptions there. You do need a comparable lens to exist. In that case, where a full frame lens with an equivalent field of view exists, the difference is due to the total light.

Note that the article makes clear that equivalent ISOs are not absolute, and serve only as a guide to performance.

However, why would 600mm F15.6 ISO 3086 (if you were confident that equivalent ISOs were going to give you a close-enough answer) make you conclude that the camera is unusable in low light?

Thinking in terms of crops may be easy for you, but being able to plot equivalent aperture graphs makes it easy to express the capabilities of otherwise comparable cameras with different sensor sizes.

0 upvotes

"However, why would 600mm F15.6 ISO 3086 (if you were confident that equivalent ISOs were going to give you a close-enough answer) make you conclude that the camera is unusable in low light?"

Um...I don't. It was a comment in this thread you made 4 months ago regarding the FZ200.

"Which, as you can imagine, gives you a lot of depth-of-field but is usable in daylight, but almost unusable in low light."

Granted you say "almost" but there are many images in the Panny forum of low light shots taken with the FZ200. It can't be that hard to use and for a bridge camera appears to do reasonably well with low light.

Did I misunderstand the context of your statement and you me ant something else?

0 upvotes

Reasonable for a bridge camera is probably fairer than 'almost unusable' - I take it back.

Modern small sensors tend to perform better (in proportion to their sensor size) than large sensors, as a very rough rule-of-thumb, which is part of why I shouldn't use 'equivalent ISO' as anything other than a very rough guide.

At F2.8 and ISO 100 the FZ200 is creating its image from the same amount of light as a full frame camera is at F15.6 and ISO 3086 however, the performance won't necessarily be identical.

1 upvote

Some people here seem to think that they can produce the same level of clean image (low noise) with a GH4 as with an A7R when shooting at iso100 with the FF camera. Not possible.

Edited 2 minutes after posting

4 upvotes

True,

But take into account that you can't use ISO100 compared to 100 when you need the same depth of field and shutterspeed with the A7r.

in that scenario the a7r would be at ISO 400 and in turn have the same noise(in theory since there are more factors like sensor technology)

1 upvote

BarnET,

The case you describe is correct, but it omits the fact that FF users have an _option_ to shoot at larger (physical) apertures and lower ISOs where as smaller format users such as m43 users are limited to "fast zooms" at 1/5.6f on FF terms and "ISOs" yielding higher level of noise (comparable to 4x ISO @ FF).

FF user can choose to use the same parameters (field-of-view, depth of field, shutter time, image level noise (ISO)) than m43 user, but m43 user cannot choose to take photos at all the parameters FF user can choose from.

My first camera was 43s (same sensor size as m43) and I sold it away since its image quality was nowhere near what I wanted. Yes, I carry bulk, but that is the trade-off I _consciously_ made. One can take GREAT photos with m43 - no doubt about it. And m43 systems are far easier to carry around. But let's be clear that those systems are NOT equivalent in pure image quality terms.

DPreview did fantastic job in writing the Equivalence article.

Edited 1 minute after posting

1 upvote

Agree with all your points Jylppy.
But then it won't cost as much and it's a LOT smaller.
Even when compared to the fullframe mirrorless models from Sony.
At least when you compare the lenses.

It's a trade-off. And DPreview did a good job in that articl to explain that very well indeed. Beside the difference in bulk with the old 43 system was not as obvious as it is now.

So back then it was a much easier decission to make.

Edited 1 minute after posting

1 upvote

Hi BarnET,

Actually, while the MLC FF bodies (Sony A7 etc) are far smaller than FF DLSRs (flange-distance, no need for pentaprism), the lenses are not that different in size - at equal aperture and focal length. I have read that from lens design perspective, the current DSLR flange distance is quite close to optimal for FF sensor size. Once one makes "fast" lenses for FF MLC the lenses are roughly at the same size to FF DSLR lenses. Maybe this is the reason Sony does not make 1/2.8f zooms for their FF MLC (the lenses are for SLT and need a adapter for their mirrorless FF moiunt).

There is speculation that maybe one day Canikon drops mirror from DSRL - but keeps the Flange distance the same AND maintains a smaller sensor MLC ( Nikon 1, APS-C) MLC line that is more compact in size. So it seems, Fuji made the right calls in their MLC approach.

But you are right, MLC FF bodies are smaller vs. FF DSLR.

1 upvote

I was talking about m43 vs FE-mount.
But i agree with you.
again :-D.

By the way. having a shorter flange does benefit Sony to make smaller lenses. But only at the really wide-end. It seems that from a design perspective making lenses that have a wider focal length then the flange distance is harder. and most of the time requires a more design.

for instance a 24mm prime would be easier to make on FE-mount.
But a 50mm F1.4 not so much.

Now since fullframe camera's need longer focal length lenses the flange distance is less of a problem. Since the flange is ideal for 50mm standard lenses.

Which is a short telephoto on Aps Dslr's already. So you are again right. For fullframe the 46-47mm flange compared to 19mm isn't as big of a deal as it is with smaller sensors.

Edited 1 minute after posting

1 upvote

Yes, that is my understanding too. Shorter flange-distance of FE-mount helps in the wide-end. The issue you mentioned might be the reason that e.g. Canon's wide-angle zooms are relatively sizable lenses.

1 upvote

You can open dpreview studio comparison tool & try to find any noise difference at iso100. You won't, you will only find sharper details on sony due to higher resolution.

0 upvotes

If you have the latitude to adjust the exposure (ie. a faster lens) or can use a slower shutter speed say with image stabilisation, then the principals of equivalence make your statement untrue Elaka, of course the you can do the same tricks on FF as well which rather defeats my argument, but a slow Kit or an inexpensive telle Zoom on an A7R can be compensated performance wise by using a fast fixed focus prime on an Oly allied with Oly's effective image stabilisation to use a slower sutter speed as well, may be why so many OM owners are so happy with their gear

0 upvotes

Err.. Is that an argument at all? In many cases you are shutter speed limited (shooting hand-held regardless of IS, moving objects). And FF user can do exactly the same, juts close down aperture and crank ISO up (FF@ ISO 400 has same image noise as m43@ISO100).

And about Sony A7R vs GH4 IQ comparison at ISO100. The difference is massive for Sony's benefit:

http://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=panasonic_dmcgh4&attr13_1=sony_a7r&attr15_0=raw&attr15_1=raw&attr16_0=100&attr16_1=100&normalization=print&widget=129&x=0.7672763615448794&y=0.6571904024827748

But m43 is more compact and best you can get at that size.

0 upvotes

"Err.. Is that an argument at all? "
yes is so far that you said
"Some people here seem to think that they can produce the same level of clean image (low noise) with a GH4 as with an A7R when shooting at iso100 with the FF camera. Not possible"
my point is that equivalence says it is possible, if you have the latitude to crank up the exposure on your MFT rig and you don't on your FF set-up, I do concede that practically its not going to happen much due to the (non) availability of fractional F number lenses and low ISO MFT high res sensors, and compliant subjects.

0 upvotes

Attomole.

There is still a way to make multiple shots of the same scene at the lowest ISO's.

Then combine all these images and let software normalize the noise. The same low noise can be achieved albeit at a much longer time. And ofcourse any movement will cause complications.

Stitching images when zoomed in further is a way to get higher resolution images. I've personaly used these techniques for a Accountant bureau wall print of 2.5m by 1.5m.

Had a 60mp tiff-file of 2 gb while editing. That was pretty much more detailed then a D800 shot. It was done with a GX7 on a tripod.

0 upvotes

One more article explaining why this "equivalence" concept is a complete rubbish:
https://photographylife.com/sensor-crop-factors-and-equivalence

3 upvotes

Totally agree. And it's the one that sealed the deal: I've sold all my cameras and lenses since my iPhone has a fast f/2,2 lens. I'm set. What are you waiting for?

6 upvotes

That article has been given the treatment it deserves here: http://www.dpreview.com/forums/post/54845745

1 upvote

No it is not. Just accept the facts and move on. If you MFT user (did I guess it right?) cannot digest the fact that your "f/2.8 MFT" lens gives only "f/5.6" level of DOF/Bokeh, try to improve your mood by thinking how all-the-FF-users get sleepless nights of their envy to PhaseOne users whose f/2.8 is f/1.8 in FF...

1 upvote

f-stop is not about measuring death of field, it's a measure of relative lens opening. For example i can shoot with my OM-D (yes i'm shooting m43), check wether the exposure is correct & use same parameters for my film camera (so i'm also shooting FF).
My main criticism is that physical explantation in the article is complete rubbish. And examples proove that - x-trans sensor has less noise than all other sensors & Canon EOS-1D X has very small sensor for FF & less noise than most other FF cameras. Compare the sensor woth more common cameras like Canon 60D & Canon 5D markIII & you would get completly other results.
You can check the results of comparing here - http://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=oly_em5&attr13_1=nikon_d800&attr13_2=canon_eos5dmkiii&attr13_3=nikon_d3200&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=1600&attr16_1=6400&attr16_2=6400&attr16_3=3200&normalization=full&widget=1&x=-0.06869428202812776&y=-1.0670590553314674

Edited 11 minutes after posting

1 upvote

@sans culotte - f-number is a ratio of entrance pupil to focal length, equivalent aperture is the ratio of the entrance pupil to the equivalent focal length.

Because this takes sensor size out of the equation, it allows you to compare exposure in terms of total light used to make up the image (which is the primary determinant of noise).

In the conventional light-per-unit-area way of thinking about exposure (which we don't, for a moment advocate abandoning with current cameras), ISO is essentially a fudge-factor to make it work across different sensor sizes. (ISO 100 on one camera does not mean the same thing as ISO 100 on another - it's simply whatever it takes to give the same image brightness, despite the difference in light capture).

3 upvotes

In this article we used, to as great an extent as possible, sensors of comparable performance at each sensor size (Sony-made Bayer sensors, wherever possible).

I'm not sure what you mean by the 1D X being a small sensor for full frame Canon specs list it as approx 36 x 24mm, which is standard size.

Canon's sensors have tended to have higher noise floors than recent Sonys so, as you say, they do not perform (in terms of low ISO DR) as well as Sony sensors, but they can be better at high ISO.

Similarly, either intentionally or as a side-effect of the way X-Trans is demosaiced, it gives oddly low levels of noise.

These differences between sensor technology are why we say equivalence should only be used to give an idea, rather than an absolute figure for the performance difference between cameras.

0 upvotes

It would save us a lot of time if we didn't have to test any of these differences and could assume sensors were always solely shot-noise limited and performed exactly in proportion to size.

0 upvotes

@sans culotte,

You have so many facts wrong in your claims that calling the article "complete rubbish" is an unfounded claim.

As Richard pointed out Canon 1D X has _standard FF-sized_ sensor. Canon 1D X has lower noise, but it also has lower MPx count vs 5DIII and especially vs D800. Therefore pixel-level noise is lower, but I assume others will be quite close once you scale to same image size. Btw. O-MD 5 has the same noise at ISO 3200 than others have at ISO 6400 - exactly as the article demonstrated.

X-Trans sensor does not use Bayer-array so the "RAW comparisons" do not make justice. Fuji's RAW conversion seems to reduce noise by default and that you can see from softness of the RAW photos.

1 upvote

If I understand the article correctly, total light gives an idea of the quality of the images so a camera with a smaller sensor might do as well or perhaps better than one with a larger sensor IF it has a faster lens (AND using a slower shutter speed is not an option, forcing the camera with the larger sensor to use a higher ISO)? Pixel density is not important? This allows one to understand the relative importance of lens speed vs. sensor size (in terms of total light and IQ).

This changes the common wisdom that a small sensor camera performs pretty well in daylight but not so well in low light? But the daylight situation has not changed, and the low light situation has improved only where the camera has a faster lens and a slower shutter speed can not be used (allowing it to use a lower ISO?)

On cameras with smaller sensors does the adjusted aperture refer to the actual size of the aperture? In the cropped example the adjusted aperture does refer to the used part of the aperture.

1 upvote

All cameras use all of the aperture, they just can't always catch all the light that's projected through the lens.

Pixel density makes remarkably little difference - higher resolution images appear noisier at the pixel level because you're looking at the noise in more detail. Broadly speaking, you can scale a high-res image down to the size of a lower res image and get similar noise levels.

A compact camera is just the same as a 'cropped sensor' camera (indeed you can calculate all 'full frame' lenses in terms of Medium Format lenses/sensors if you prefer: it just happens that the 35mm film format is pretty well understood as a reference point).

A compact camera will perform well when there's lots of light because there's lots of light however, in low light it'll perform exactly as well as you'd expect with a tiny sensor and tiny maximum aperture.

1 upvote

Eg:

Panasonic FZ200 has a 4.5-108mm F2.8 lens

In full-frame terms that would be equivalent to a 25-600mm F15.6 lens.

Even though the FZ200 has a relatively bright lens, it's not nearly bright enough to allow its small sensor to operate at the same level as large-sensor cameras.

Edited 2 minutes after posting

1 upvote

"...Panasonic FZ200 has a 4.5-108mm F2.8 lens...In full-frame terms that would be equivalent to a 25-600mm F15.6 lens..."

No, that would be equivalent to a 25-600mm F2.8 lens.

Equivalency of Sensor-Area, does not apply to Light-Intensity which is a physical constant of number of photons per given time, per given spot.

With a given light intensity, say, "1 billion photons per second", it would remain the same light intensity, on a small or large sensor, or through a small or large diameter lens.

1 upvote

Your spot is too small.

1 upvote

Yuvalm - I'm standing by my original statement:

Panasonic FZ200 has a 4.5-108mm F2.8 lens

In full-frame terms that would be equivalent to a 25-600mm F15.6 lens.

A 4.5mm F2.8 lens on a 1/2.3" type sensor will allow exactly the same amount of light through as a 25mm F15.6 lens would on full frame, if given the same shutter speed.

You're absolutely right, the light intensity would be lower on the larger sensor, but that would be exactly in proportion to how much bigger the full frame sensor is. So, for instance, if both sensors had the same pixel count, each pixel would receive the same amount of light.

This article includes examples that demonstrate exactly this point.

Edited 1 minute after posting

4 upvotes

The light intensity allowed by, say, an f/2.8 aperture, will remain the same light-intensity, for a smaller or a bigger snsor, s sensor which is merely a designated area on that sensor-plane.

It's obviouse that the area on that plane cannot influence the light intensity radiating fro the lens - there simply isn't any influential mechanism between the two completely separated factors.

Sensor-Area as divided by camera's Pixel-Count, resulting in Pixel-Size;

Now Pixel-Size is indeed critical for the picture's quality, but only in terms of SNR, meaning the CPU receives a clearer, mored difined sampling signal with lower noise-level, allowing for the level of ISO possible to use, without attached noise.

To the sensor however, the whole issue of light intensity is a given, because in simple physics, the lens casts a given light-circle on a plane, and any designated area on that plane, cannot influence any of The cast Light-Circle's optical qualities, including it's Light-Intensity.

Edited 3 minutes after posting

0 upvotes

Nobody is claiming that the sensor size has any impact on the light intensity.

However, light intensity is a measure of the amount of light cast on a particular area and sensor size gives an understanding of the area that you are sampling.

I repeat: an 25mm f/15.6 lens would clearly allow lower light intensity to the sensor than a 4.5mm f/2.8 but the difference between the two is proportionate to the different areas over which the light is being captured.

So: If both sensors have the same number of pixels, then each pixel receives the same number of photons during the same exposure time. The larger sensor experiences less light per-unit-area but has larger pixels.

Pixel size plays a surprisingly small role in image quality (if you resample images to a common size). Sensor size plays a much larger role because it determines how much light you can potentially capture (because it defines how much area you have measuring any given intensity).

1 upvote

Why Pixel-Size matters in terms of Picture Quality?

CPUs don't sample sensors, but instead, they sample long arrays of individual pixels we may call "A sensor" referring to the whole array.

On that array, each individual pixel has given area to be sampled. and now all of a suuden, size count: the more photons collected per given time, the clearer the sampling signal, the more accurate and reliable reading and storing of that signal

The bigger the pixel-area on that sensor, the more photons are colllected for sampling of that pixel - the clearer the sampling signal for further process.

Big pixel-area leads to Lower Noise-Floor (meaning higher ISO level can be available without noise), and more accurate colour sampling (meaning smoother tone-gradation), all resulting from the collection of more photons per given time.

The bigger the Pixel-Area, the clearer and more accurate the sampling-signal is for further processing

Edited 4 minutes after posting

0 upvotes

This is only true if you compare images at the pixel level.

Even with pixels of the same size, as in this example, the larger sensor still looks better, when compared at a common output size.

I'm not sure you can generalise that the noise floor (the point at which electronic read noise contributes in addition to the shot noise) is lower for large pixels.

0 upvotes

Hi Richard

Yes, great article, but still it appears incredible how difficult it is to get the message across - judging from some of the comments :-/

However, perhaps arguing “backwards” may do the trick?

I think all people could agree that to get identical looking images from different formants you would have to
1) take the pictures from the same distance
2) use lenses with the same angle of view
3) use lenses with the same physical aperture diameter
Hence, i.e. a 100 mm and 50 mm lens would produce the same angle of view on a 24x36 mm imager and a MFT imager, respectively.
If both lenses have a physical aperture diameter of 25 mm, then they would be f/4.0 and f/2.0 respectively.

Case closed.

Edited 3 times; latest 4 minutes since posting

1 upvote

Great article. Thanks.

1 upvote

There's a neglected aspect to equivalence hiding here -background blur equivalence. the same lens on APS-C due to magnification of the final image produces more background blur than on full frame - akin to greater magnification of final prints highlighting focus inacuracies.

1 upvote

Simply, No!
It produces narrower angle of view, which one can perceive as shallower dept of field, but that is not the case. If you get the same frame with the FF(say, walk 1.5 times closer) than the dept of field will be exactly the same(of course with different backgrounds)

1 upvote

struth what a range of comments, interpretations, assertions and debate.

I actually did a BSc in Photographic Science back in the days of analogue, I struggle to follow the mish mash of statements, what is accurate and what is not, plus the way they are expressed.

I prefer that good old fashioned thing - a text book. The simplest I could find is The manual of photography 7th ed pub 1978, chapter 5 page 89. Or for the current 10th ed 2011 chapter 6 page 103 onward is a good start, by the aptly named Sidney Ray who was my lecturer on the subject.

This is basic science so frankly there is no discussion on what is actually right or wrong it is more about people not understanding or going from sources on the internet which are not written by subject matter experts.

Plus you do need to be careful as often the definitions and diagrams may use a simple lens - not a compound lens which does slightly alter some of the definitions, plus there usefulness in this type of discussion,

Edited 35 seconds after posting

4 upvotes

http://newcameranews.com/2014/06/02/trolling-for-dummies/

"Equivalence. One of the classics. Watch people that barely made it through high school math battle it out with physicists and real life rocket scientists."

: )

8 upvotes

tresise,
Thanks for the reference to "The manual of photography."

0 upvotes

Equivalence is getting the f-stop wrong . With thanks and apologies to 'Educating Rita'

0 upvotes

Equivalence is about the focal length that gives the same framing and the f-ratio that gives the same aperture diameter. For example, 50mm on mFT gives the same framing as 100mm on FF (for a given perspective) and f/2 on mFT results in the same aperture diameter as f/4 for equivalent focal length (50mm / 2 = 100mm / 4 = 25mm).

The same aperture diameter results not only in the same DOF and diffraction, but the same total amount of light on the sensor for a given exposure time, and thus the same noise for equally efficient sensors.

So, if anything is wrong, it's not understanding the difference and significance between the relative aperture (f-ratio) and the virtual aperture (entrance pupil).

1 upvote

I like many took a while to accept the principles of equivalence...and yeah I've shot in some different formats to the M4/3 one I mainly shoot in now...but as I've said before to prove a point. Strap a 35mm lens on an APS-C or a M43 and a Speed Booster adapter, and it Gains around an Fstop (as well as increasing FOV), proving simply that the entrance pupil of the lens does matter.

Edited 2 times; latest 2 minutes since posting

0 upvotes

You can think of Equivalence as a TC (teleconverter), FR (focal reducer), or cropping.

For example, adding a 2x TC to a lens has the same effect as cropping out the middle 25% of the photo or mounting the lens on a sensor half the size. However, cropping reduces resolution because only 1/4 the pixels will be on the subject, and a TC reduces resolution because of optical aberrations in the TC.

So, unless the larger sensor has the same size pixels as the smaller sensor and/or the TC is aberration free, using the lens on the smaller sensor will have greater resolution (how much greater will depend on quite a few factors).

Likewise, adding a 0.5x FR to a lens designed for an image circle twice the size of the sensor has the same effect as using the lens on a sensor twice the size, with the same caveats with regards to resolution.

0 upvotes

It took me a long time to get my mind around the equivalence debate but I think I have a pretty good grasp of it. This is a good article for those new to the debate and will help give them an understanding of this concept.

I have to comment about those that are always pointing out that f4 on FF is the same as f2 on f 4/3, they always forget to mention that as far as exposure f2 is f2. If you are shooting at f4 what I'm shooting at f2 then you obviously have to raise your ISO, partially negating the noise advantage.

Of course the detractors will point out that a FF sensor at f4 has the same amount of light falling on the sensor as a 4/3 at f2 which is true. You need to remember the amount of light per mm of sensor real-estate is the same, which is why f2 is still f2 unless we are talking DOF and noise.

The above said, I'm not a fan of m43 because I hate the tiny little cameras, they don't fit my large hands. Maybe the EM1 with a grip will be large enough for my needs in the future.

Edited 3 minutes after posting

4 upvotes

"they always forget to mention that as far as exposure f2 is f2."

It's not so much that they forget to mention it, it's that they think it goes without saying, and they think it goes without saying because they don't realize that everyone is misconstruing equivalence as a shooting tool designed to replace the exposure triangle. That's something that only becomes clear once you start trying to make sense of some of the responses.

I'm not sure I completely understand why that happens so often to be honest, but I have nonetheless learned to make it clear from the outset when engaging in such discussions.

I've also found it useful to decouple equivalence from FF crop factors and speak in 4/3"-type equivalent terms (or even 1"-type equivalent terms for that matter) when there are no larger formats in the equation. This can help get everyone past the perception that equivalence is little more than a declaration of FF superiority, which is the other major objection.

Edited 1 minute after posting

1 upvote

--> cainn24 It's not so much as they (ie people like me) "don't realize that everyone is misconstruing equivalence as a shooting tool designed to replace the exposure triangle." Equivalence tells me that the small sensor is NOT noisier than a bigger sensor (assuming the same tech), its just not getting enough light. Once you start looking for equivalent noise in different sized sensor, then the standard 'exposure triangle' goes out the window. If all you're concerned about is the same light intensity (same f-stop and shutter speed) then equivalence is useless. But the resulting images will not be the same, so what is the point?

0 upvotes

"Equivalence tells me that the small sensor is NOT noisier than a bigger sensor, its just not getting enough light."

At a certain point there's nothing you can do about that. If you try to expose a small sensor to more and more light you'll just end up with overexposed images. Large sensors have a noise advantage because they can use more light to create images, and more light = less noise.

"Once you start looking for equivalent noise in different sized sensor, then the standard exposure triangle goes out the window."

It is only when people mistake equivalence for a shooting tool and the exposure triangle for a cross format performance comparison tool that absurdity ensues.

"If all you're concerned about is the same light intensity (same f-stop and shutter speed) then equivalence is useless."

Again, equivalence is not really a tool that you use when you're just out having fun taking photos. It's a tool that you use when you want to compare performance across different formats.

0 upvotes

"At a certain point there's nothing you can do about that."

Agreed. It's called base ISO and one stop above (for a 1/4 sized sensor).

"If you try to expose a small sensor to more and more light you'll just end up with overexposed images."

At base ISO and one stop above (for a sensor 1/4 the size). At any other place, the can get identical noise characteristics.

"Large sensors have a noise advantage because they can use more light to create images, and more light = less noise."

... or you can increase the intensity on the smaller sensor and therefore get more light ...

"Again, equivalence is not really a tool that you use when you're just out having fun taking photos."

Agreed.

"It's a tool that you use when you want to compare performance across different formats."

Agreed.

0 upvotes

Me: Large sensors have a noise advantage because they can use more light to create images, and more light = less noise."

You: ... or you can increase the intensity on the smaller sensor and therefore get more light ...

This is incorrect. The point at which it is possible to realize a noise advantage by using a larger format is the same point at which you can longer increase the exposure on the smaller format without overexposing your images.

Further, the difference between sensor A, and sensor B (which is a quarter of the size) is closer to 2 stops in favour of sensor A.

0 upvotes

"This is incorrect. The point at which it is possible to realize a noise advantage by using a larger format is the same point at which you can longer increase the exposure on the smaller format without overexposing your images."

Want to bet?

FF = 400 ISO
4/3rds = 100 ISO but with longer shutter speed or larger aperture.

The only place this doesn't work is when FF is inside of two stops of the 4/3rds base ISO. Then you would be right. But, of course, base ISO is not usually the place where noise is a concern.

0 upvotes

I don't think you're really paying attention. Instead you're playing games by denying the FF system the same shutter speed and fnumber, which would give it a noise advantage that the MFT system couldn't match (albeit at the expense of DOF).

0 upvotes

I'm not denying it. What I'm saying is "So what?"

If you're going to artificially restrict the smaller sensor to less light its going to be noisier. But why would you? What in the photographic world is stopping you from giving the smaller sensor and the larger sensor the same light?

If you give the smaller sensor the same light, it will be just as clean as a larger sensor - all other things being equal. Hence, the smaller sensor is not noisier ... unless you are crazy enough to put artificial restrictions on it that automatically deprive it of light.

The way we ensure the sensors between systems are being compared with the same light is called ... drum roll please ... equivalence.

0 upvotes

Look, you started this by saying: "Equivalence tells me that the small sensor is NOT noisier than a bigger sensor, its just not getting enough light."

That's like trying to argue that your small bucket is not more empty than a larger bucket, it's just not getting enough water in it. But if your small bucket is already full to the brim, then in what sense is your argument meaningful?

The fact remains that as you try to push the performance envelope there comes a time when you simply can not expose the smaller sensor to more light without blowing out your images. That's a limitation, and it's a limitation that is inextricably connected to size.

If you want to argue that small sensors, which produce noisier output for a given shutter speed and fnumber than large sensors, aren't actually technically "noisier" at all, then you are driving at some sort of pedantry that seems counterproductive in the context of getting a good handle on issues of cross-format performance comparisons.

Edited 2 times; latest 3 minutes since posting

0 upvotes

Okay, last post.

- Light is not water.
- Sensors are not 'buckets'
- If you insist on using the same f number and SS, the smaller sensor will be noisier. Equivalence explains why and what to do to make the smaller sensor create a file with the same noise levels as a larger sensor. If you don't care about this or DoF, then equivalence is not for you.
- the same f number will create a smaller effective aperture on a system with a smaller sensor. Use the same sized effective aperture and you will have the same noise levels.
- if you can't grasp this, ask Great Bustard because I'm getting really bored here.

Edited 56 seconds after posting

0 upvotes

Apparently you've wandered in here and created a problem that wasn't here before you arrived.

I understand equivalence. I advocate for it. And nothing I've said here is inconsistent with it. And I'm not quite sure why you think otherwise. In fact from my perspective it seemed to be you who was taking issues with such principles.

Hopefully this is simply a communication issue; a semantic disagreement rather than an actual one.

As for the analogy, they are never perfect. But when taken in good faith the intended point is usually well-received..

Edited 1 minute after posting

0 upvotes

One additional point though:

"If you insist on using the same f number and SS, the smaller sensor will be noisier. Equivalence explains why and what to do to make the smaller sensor create a file with the same noise levels as a larger sensor. If you don't care about this or DoF, then equivalence is not for you."

For as long as you say things like this a reply will be necessary. Why? Because your statement is incomplete and therefore potentially misleading. It's omits the fact that it is not always possible to achieve the same noise levels with a smaller format.

And this:

"the same f number will create a smaller effective aperture on a system with a smaller sensor. Use the same sized effective aperture and you will have the same noise levels."

omits the fact that you can't always do that either. In other words, you're framing everything as if it is possible to routinely match FF performance whenever you want.

You may think it all goes without saying, but apparently it doesn't.

2 upvotes

So, after reading this article, I feel like I still am not clear on the noise and DOF aspects.

I currently shoot APS-C (Sony A57) with popular lenses (like Tamron 17-50/2.8 and Sony 35/1.8). Much of my shooting is lower-light with available light, and I find myself disappointed with the noise levels of my shots, even at max aperture (and lowest ISO that keeps the shutter speed realistic). I shoot RAW and use LR5 for processing.

With the uncertainty of Sony's A-mount future (another topic, not for here), I have been weighing other systems to upgrade to. Given that shallow depth of field and less noise in low light are priorities for me, I am trying to determine if:
1. I could achieve that a M43 system with $$$ glass, or if I need a larger sensor like APS-C or better yet, FF?
2. It's a factor of the quality of the lens and not the aperture (since already shooting 1.8 or 2.8) and I would get less noise with more expensive glass?
3. I just need better noise software like Topaz or NIK?

Edited 38 seconds after posting

0 upvotes

Further ...
.
"That's the way we've always done it" is not enough to fool everyone into thinking large equipment is needed for quality pictures. It is FUD - there are many articles like the one below, from PROs who have switched and never looked back.
.
http://www.smallcamerabigpicture.com/how-to-overcome-fears-using-micro-43rds-cameras-in-a-professional-environment/

The "big" advantage mostly isn't. _ Both my EM-1 and my EM-5 trounce my 6x7 camera in every sharpness/grain/noise measure I can muster. The MARGIN by which these small Olympus EM cameras defeat this great (but huge) camera (Mamiya RB-67 Professional S), astonished me.
.
Large is NO longer needed except for SOME extreme technical photography.
.
Most of the larger format adherents use maths to support their quasi-religious belief that a faster lens is "needed" to compensate for photographers who have no skill or inclination to watch the background.
.
see .. http://photohounds.smugmug.com/6x7-early-80s/

Edited 2 times; latest 2 minutes since posting

3 upvotes

In other words:

"I've decided that I don't need FF, therefore no one else should need it either. Furthermore, those who do feel that they need FF, only need it because they lack my photography skills."

You are correct in that many people don't need FF, but how are they supposed to know without investing in a new format? This article gives people an easy way to know which formats are suitable for their application.

8 upvotes

It's like saying most people don't need large trucks, or big cars, because I bought a small scooter and it works better than all other vehicles I had before.
This article talks about the physical limitations of the cameras and lenses, you can't have both smaller and more light, because there can't be more light collected by the sensor than the light collected by the lens aperture. Everyone can choose what he needs, whether he wants less noise and bigger lens or a smaller lens and more noise. Arguing what other people should and shouldn't want is simply stupid.

5 upvotes

@photohounds

Re: "Most of the larger format adherents use maths to support their quasi-religious belief that a faster lens is 'needed' to compensate for photographers who have no skill or inclination to watch the background."

Are they anything like the people who say, for example, that the 40-150 / 2.8 on mFT is "equivalent to" a 80-300 / 2.8 on FF and "justify" the superiority of mFT on that basis?

7 upvotes

Pseudo logic.
.
Most people "need a truck, bus or an aircraft" at times. The intelligent ones do NOT usually buy one. NOR do they tell others their sedans are "inferior".
.
The Equivalence Police Terr0rists and their religion indeed DO imply inferiority because some wiser people choose NOT to OWN trucks (FMF) and prefer more compact gear in their baggage that produces "equivalent" results.
.
Legacy adherents ... The mere "need" of them to spruik this misuse of maths displays the religious aspect of their pedantry. I SOLD my Mamiya RB67 - the old EM-5 left it in the DUST. http://photohounds.smugmug.com/6x7-early-80s/
.
Four Thirds produces results
I shot against half a dozen FMF users at this and the other shoots in the gallery - a level or two up.
.
There were - many record shots (I have to please everyone). MY numerous keepers SOLD more than the rest PUT TOGETHER.
.
http://photohounds.smugmug.com/Performing-arts/Eurobeat-by-Supa/
.
RESULTS, not pining for legacy formats.

2 upvotes

Re: "The Equivalence Police Terr0rists and their religion indeed DO imply inferiority because some wiser people choose NOT to OWN trucks (FMF) and prefer more compact gear in their baggage that produces "equivalent" results."

What an ignorant statement. Exactly opposite how those that understand Equivalence put it:

http://www.dpreview.com/forums/post/54471622

3 upvotes

This discussion obsesses on the fact that you get more (but not excessive) depth of field. Yes it IS an obsession that neatly avoids some facts.
.
Using the 50mm/f2 vs 100mm/f4 example above ..
You ALSO get TWO STOPS BETTER shutter speed (4 times as fast).
(better sharpness)
You get QUADRUPLE the flash range (faster recycle or More range)
I once shot a Canon FF PRO user with the 50mm f2. She remarked: "I've never SEEN so much detail".
.
Further - you also get (mostly) MUCH smaller, lighter lenses and faster zooms.
.
You also get typically 1/2 to 1/3 the VIGNETTING that full Marketing Frame (FMF) lenses can manage. This is excellent if you like to shoot images with the main subject OUTSIDE the centre third of the image but still want it sharp. Generally (with few exceptions) APS-C lenses fall halfway between for good vignetting resistance.
.
Examples .. http://photohounds.smugmug.com/Performing-arts/Eurobeat-by-Supa/

Edited 3 minutes after posting

2 upvotes

Re: "Using the 50mm/f2 vs 100mm/f4 example above ..
You ALSO get TWO STOPS BETTER shutter speed (4 times as fast)."

Um, no. You can shoot any shutter speed you like at any f-ratio you choose. For example, if an mFT photographer shot a scene at 50mm f/2 1/200, there is no reason the FF photographer could not shoot the scene at 100mm f/4 1/100.

Re: "(better sharpness)"

Are you trying to imply that mFT resolves as well at f/2 as FF at f/4? I suppose if you put the best lenses on mFT and the worst lenses on FF that might happen, but...

Re: "You also get typically 1/2 to 1/3 the VIGNETTING..."

Depends on the lens. Usually not so at the same DOF. For example, consider the 35-100 / 2.8 IS on mFT:

http://www.photozone.de/m43/852_panasonic35100f28?start=1

vs the 70-200 / 4L IS on FF:

http://www.photozone.de/canon_eos_ff/431-canon_70200_4is_5d?start=1

at the same DOF (e.g. f/2.8 vs f/5.6).

8 upvotes

Re: For example, if an mFT photographer shot a scene at 50mm f/2 1/200, there is no reason the FF photographer could not shoot the scene at 100mm f/4 1/100.

I assume you meant

if an mFT photographer shot a scene at 50mm f/2 1/200, there is no reason the FF photographer could not shoot the scene at 100mm f/4 1/200.

and I would add, for clarity, and end up with a very similar looking photograph, particularly with respect to noise and image quality.

3 upvotes

@bobn2

Yes, quite right. Thanks for pointing that out!

0 upvotes

True Bobn2 - the MFT user will also have 2 STOPS of shutter speed up his/her sleeve to console him/her :)
.
Up to a point .. MOST of the time people can use MORE DOF - it is an age old photographers lament and WHY RAIL CAMERAS have MOVEMENTS to compensate.

Those fast and often tiny MFT lenses give great OOF bokeh, by the way. I have several such lenses. I had a Mamiya RB67 when you needed large cameras to make a decent picture. Sold it and all the other legacy formats. No problem.

0 upvotes

Re: "the MFT user will also have 2 STOPS of shutter speed up his/her sleeve to console him/her :)"

You can shoot whatever shutter speed you want at whatever f-ratio and/or ISO you want on any format.

Re: "Up to a point .. MOST of the time people can use MORE DOF - it is an age old photographers lament and WHY RAIL CAMERAS have MOVEMENTS to compensate."

Is that why 1" is so much better than mFT? Is that why more than half of these photos:

http://www.dpreview.com/forums/post/39537128

are substandard?

Re: "Those fast and often tiny MFT lenses give great OOF bokeh, by the way."

They do indeed. So do many fast and tiny FF lenses.

Re: "I had a Mamiya RB67 when you needed large cameras to make a decent picture. Sold it and all the other legacy formats. No problem."

Indeed -- digital has many significant advantages over film.

0 upvotes

Ah, a strawman ... Yes the restricted DOF effect can be fun. Over use of the "look I've got a fast lens" can get old pretty fast.
.
The most hilarious shot I ever saw was taken with that soft C.85/1,2 - NO EARS! In portraits a little environmental information is more often than not desirable.
.
What I object to is the constant insinuation that you cannot have shallow enough DOF unless you are using FMF. Bollox.
.
Here, http://photohounds.smugmug.com/Performing-arts/Eurobeat-by-Supa/ The portability of my gear meant that I was able to get just the right DOF - AND a good range of angles. Shot during ONE performance and I only had one body at that stage: 45, 75 and 12mm. The local photo society was also present as I shot handheld in the murky light. Their large, cumbersome gear, tripod bound, my keepers outnumbered their total sum tenfold.
.
Less BG? Easily be done in a studio or "gasp" choosing one's shooting position carefully. What is it they say about a bad workman?

1 upvote

Re: "Over use of the 'look I've got a fast lens' can get old pretty fast."

More than half of these photos were at f/2 or wider on FF (f/1 or wider mFT equivalent):

http://www.dpreview.com/forums/post/39537128

and were taken over a period of a couple of years, at least. So, at least that gives a context for your statement.

0 upvotes

Summary: You get more with bigger stuff; but it cost disproportionately to much more for what more you get, we just wanted you to know. Think about APS-C for all around value and abilities.

PS: The problem with both DSLR's and/or mirror-less CSC systems is they still need technical development (pun intended) for different things, respectively.

Summary #2: You will not get as much 3D background with the smaller cams/lenses when you are at an equivalent angle of view, AKA the same composition and at the same F/x.x setting the lenses can muster. You'll want "faster" and "need" it. So get the primes only.

PS2: You want high IQ, fast lenses and a balanced, diffused, fill-flash system(even if it's a piece of paper). Not just one or two. Two or ALL is best.

Edited 4 times; latest 9 minutes since posting

0 upvotes

Plain logic, nothing difficult here. I think why so many fights appear about equivalence is that some people simply protect format that they chose to use, often focusing only on one point.

But nice examples, could help some people adjust better if they move to different format.

0 upvotes

What I have observed is something similar to the following:

1) a MFT user hates equivalence when talking to a FF user
2) a MFT user just loves equivalence when talking to a CX format user and/or users of various smaller format fixed-lens compacts with "fast" lenses

In summary, everyone is essentially equivalent to a FF user with respect whatever format is "beneath" them. But if everyone can just manage to get over all these petty superiority/inferiority complexes it would leave them free to embrace equivalence as a tool that can be used to meaningfully compare the performance of format X to format Y, whatever those formats are.

Edited 1 minute after posting

3 upvotes

As a mft user myself i certainly see lots of MFT vs FF/APS-C talk, didn't really see MFT vs CX though. Hate comes from the fact, that it's almost always focusing on one thing and that is DoF, second most popular is IQ for pixel peepers.

0 upvotes

I see plenty of debate on the topic of fixed-lens compacts with "fast" lenses vs MFT. I threw CX in there because just like some of those cameras it's a smaller 1"-type format.

All debate aside I have certainly evaluated a whole bunch of smaller formats in MFT equivalent terms myself and I've found it useful to do so.

MFT is basically my "cornerstone" instead of FF since it's the largest format I shoot with.

Edited 1 minute after posting

0 upvotes

I like panos .. Vignetting makes this harder to stitch.
This Full Marketing Frame (FMF) lens would make it a NIGHTMARE less than 3 stops from wide open.

http://www.dpreview.com/reviews/lens-compare?lensId=sony_35_1p4_g&cameraId=sony_dslra900&version=0&fl=35&av=1.4&view=falloff
It would LOOK like a patchwork quilt.
Many FMF lenses have ~ 2 stops falloff from centre to wedge. NO THANKS.

1 upvote

This is like using a chart to decide whether to take the motorcycle, car, truck or airplane on my next overseas vacation. Of course in this scenario the airplane is my FF Canon with 24-70 L attached.

Edited 32 seconds after posting

2 upvotes

What a completely useless exercise.

1 upvote

no, to use a FF canon with a 24 - 70 L attached might be useless exercise for many people

0 upvotes

Here's a thought experiment:

Supposing I have a m43 camera and I take a shot with a 25mm lens.

Now I take 4 more shots of the same scene but this time I intentionally underexpose by 2 stops and use a 50mm lens so that I can neatly stitch all 4 frames edge to edge to come up with the same AoV as the first shot.

If I print the resulting two images at any size which one will have better SNR?

0 upvotes

Here's an answer to the same question you posed in the PST Forum:

http://www.dpreview.com/forums/post/54388088

1 upvote

“but you need to remember that ISO100 on a small sensor won't be the same quality as ISO 100 on a larger sensor.”

It almost sounds like you are saying “If I take a full frame sensor and cut it down to micro four thirds size the noise for ISO 100 will magically change”. If you crop a sensor noise does not change.

3 upvotes

The noise does not change, but rather how you view it changes

Edited 2 minutes after posting

0 upvotes

"The noise does not change, but rather how you view it changes"

Wait, are you saying that using a smaller sensor changes my eye? Now I am really not understanding you.

he image that the cropped sensor takes will be identical to the full sensor in every aspect except for resolution if you keep all other aspects the same and use an “equivalent” lens instead of the full sensor lens.

2 upvotes

If I take a photograph of a grey wall with a FF camera and crop the center out of the to m4/3 and print the original FF on a 8-10 print and then take the M4/3 crop and print it to the same 8-10 print. I have just magnified how I will view this noise in the image( see more noise)

0 upvotes

@mpgxsvcd

Are you saying that cropping out the middle 25% of a photo has no effect on the quality of the photo?

0 upvotes

"Are you saying that cropping out the middle 25% of a photo has no effect on the quality of the photo?"

No I am saying it only affects resolution. The pixels are the exact same in the cropped sensor as they are in the full sensor. There is just simply less of them. The image view will be compressed into a smaller area but the noise will be identical to what was there before.

2 upvotes

Let's say I take a pic of a scene from the same position with the same camera at 200mm f/5.6 1/400 and 100mm f/5.6 1/400. I crop the 200mm photo to the same framing as the 100mm photo and display both photos at the same size.

For what scenario will the cropped photo have "higher quality"?

1 upvote

Even more simply:

Take an image (it doesn't matter what lens, camera, format).

Crop the middle 25% out of it. Blow it back up to the same size. You've lost resolution, relative to the original one but you've also made each pixel's noise a greater proportion of the overall image (and thus increased the image's noise).

Or, looked at another way, a full frame sensor's image is likely to be noisier than that of a Pentax 645Z (44 x 33mm), using the same settings and equivalent lenses. However, use a wider lens and crop a 36x24mm chunk out of the Pentax image, blow it up to the same size you were previously comparing and you're likely to find the noise is the same as having natively shot it in the 135 format.

1 upvote

@Richard, but why would you compare a blown up image from a lower resolution but equal pixel density sensor? You have to compare these images at 1:1 pixels. Blowing it up just demonstrates the difference in resolution not the difference in noise.

Cropping the sensor loses resolution it doesn’t affect noise. The same sensor that was there before is still there. You are just simply throwing away some of the data from it. You also throw away some noise as well. Whether that noise that you throw away is in proportion to the signal you lost is debatable.

1 upvote

He does not understand that blowing up an image is called upsampling and it reduces SNR which has absolutely got nothing to do with light. He is equating the effects of resampling as if it's related to the captured light. It's physically impossible. This article is very wrong no matter how you look at it.

1 upvote

Consider this thought experiment:

Supposing that you have a D800 sensor under any lens. Keep the same lens but now cover one half of the sensor with a completely opaque material therefore exposing only half of it. Do not cover the lens; just the sensor. If the myth is true, then it follows that this sensor half (a crop) will now gather less light and therefore produce more noise. Now uncover this half and cover the other half. Again, if the myth is true then this second half would also gather less light and, just like the other half, will produce more noise. It follows that if you fully uncover the full frame sensor, it’s really just a combination of two more noisy halves! The myth is telling us that two noisy half-sensors will produce one clean full frame sensor?! This is absurd!

2 upvotes

This thought experiment is not just a thought experiment. It actually happens every time you click your shutter. At fast shutter speeds, the shutter curtain does not actually fully open and close. Instead, the shutter curtain behaves like a very small slit that glides over the sensor. Therefore, the sensor is not exposed as a whole at the same time but in chunks defined by the curtain slit size. This slit is way smaller than half a sensor. Since this slit exposes only a part of the sensor at any given time, your full frame is really acting like a combination of multiple smaller sensors! If the myth is true then a full frame sensor is really just a combination of multiple very small noisy sensors!!!

2 upvotes

@dtmateojr

+1 I agree with what you are saying. I didn't see this article until lately and I really was perplexed by some of the conclusions.

@Richard

Any chance we could revisit this topic again? There still are a lot of people with questions about your conclusions here.

1 upvote

@mpgxsvcd

I recommend starting a thread about your concerns in the PST Forum, especially as you find yourself agreeing with dtmateojr.

Here's the link to the forum:

http://www.dpreview.com/forums/1061

There's a reason people like dtmateojr are shy to post there.

0 upvotes

I'm not shy of posting there. I just don't like how clueless people like Bustard can easily get away with their ignorance such as failing to understand f-stop and resolution and resampling. Here it is more public. The world has no room for more stupidity. If you really think that you are correct then let the world know about it. You prefer forums because you can hide in there.

0 upvotes

Re: " If you really think that you are correct then let the world know about it. You prefer forums because you can hide in there."

For the world to see:

http://www.dpreview.com/forums/post/54377271

0 upvotes

You Bustard attempted to quote me incompletely and out of context just so you could propagate your lies. Very unprofessional. I'm done with stupid liars like you.
Why don't you go out and learn to shoot instead.

0 upvotes

Re: "You Bustard attempted to quote me incompletely and out of context just so you could propagate your lies. Very unprofessional."

Yeah -- like the whole post wasn't just a couple of lines above. FFS, man.

Re: "I'm done..."

Were that it was so.

Re: "...with stupid liars like you."

Funny you should mention it:

http://www.dpreview.com/forums/post/54109159

Re: "Why don't you go out and learn to shoot instead."

Oh dear -- will the self-harming posts never stop? I'll show you mine:

http://www.dpreview.com/forums/post/39537128

Now show us yours.

0 upvotes

@dtmateojr
I will reply to your "2 noisy sensors don't make one clean one" argument. However, first let me say, you should really focus on trying to understand better. Maybe, you don't have a clear understanding of what noise is. E.g., did you take into account that noise depends on the (spatial) frequency, like it does in audio? Or that noise is a genuine property of light, not pixels? Here now goes my reply to your argument:

If you add 4 sensors to form a bigger sensor of four times the surface, you end up with four times the #pixels each with the exact same pixel noise as before. I am sure you agree.

However, in order to get the same image as before, you have to combine 4 pixels into 1, in order to have the same #pixels in both images. Because you cannot compare noise in a say 8 MP and 32 MP image in the naive way you assume you can. And adding 4 pixels into 1 makes the signal 4x as big, but the noise only twice as big (taught in math). Leaving you with half the noise per signal.

Edited 5 minutes after posting

1 upvote

Pretty lame, dtmateojr. Again, my "horrendous portfolio":

http://www.dpreview.com/forums/post/39537128

Now show us yours.

0 upvotes

@falconeyes

Re: "Explained in full here: http://dtmateojr.wordpress.com/2014/05/19/megapixel-hallucinations/"

A very nice critique of that blog, and others by the same author, can be found here:

http://www.dpreview.com/forums/post/54099350

Might save you some time reading.

0 upvotes

@dtmateojr,

I won't have the time to follow your nice conversation beyond this point. But maybe, this is at the heart of the problem: Resampling vs. light gathering. You think it is two separate things while it isn't.

Maybe, start to think that every pixel holds a count of #photons which had hit that pixel (not true yet, but close enough for any thought experiment). Now, resampling 4 pixels into 1 becomes identical to 1 pixel receiving 4x #photons, 4x the light. And for your own thought experiment, you need to resample in order to compare images with equal #pixels which is a prerequisite when talking about pixel noise comparisons.

I stop here. Probably, you have committed too deeply to your way of thinking to be able to draw back now. However, if you still can and apologize, it would show how great a person you really are. We all make errors from time to time. I wish you a good time with your photography, bye now.

Edited 6 minutes after posting

2 upvotes

I got really confused reading this article. Guess I'm stupid.

2 upvotes

Send me a personal message if there's anything I can explain more clearly (that was I can try to amend the text to make it more accessible).

0 upvotes

It's good that you are confused. Your common sense is telling you that it is wrong. This equivalence-fu does not make any sense in photography at all.

Edited 15 seconds after posting

0 upvotes

You still trying to sell your QUACK FU dtmateojr

0 upvotes

I'm not selling. I'm refuting the quack-fu that dpreview is selling. It just so happened that idiots like you are exposed of your ignorance and stupidity as well. Collateral damage is unavoidable.

0 upvotes

Re: "I'm not selling."

I see. So you're using, then. To each their own.

P.S.: "Flowers for Algernon" wasn't real. Those pills you were sold -- they didn't work. Well, not as intended, anyway.

1 upvote

You still have got nothing to refute my arguments. Your stupidity is overwhelming.

0 upvotes

Re: "

Funny you mention it. 'Cause your arguments were torn to ++++ here:

http://www.dpreview.com/forums/post/54099350

Re: "Your stupidity is overwhelming."

Funny you mention that, too, 'cause yours is being dealt with here:

http://www.dpreview.com/forums/post/54377271

0 upvotes

Great post!

0 upvotes

The only people who this is really relevant to are those that used 35mm film when the only interchangeable lens cameras were 35mm and we all knew what a 50mm or other focal length lens was.

Most of the new SLR and other smaller sensor camera users have no idea of the relationship between 35mm lenses and their sensors, and why should they care. A wide angle is a wide andgle and a tele is just that. Most people use zooms anyway.

Yes it is good to refer to a common standard for an apples to apples comparison and for those that want to dig deeper all this info is useful in making better photos. The huge majority however simply want to take snap shots for viewing in electronic media and it should be a fun activity without having to bother about focal lengths etc.

Edited 1 minute after posting

0 upvotes

For some reason, people seem to think that Equivalence is all about comparing to FF. This is not true -- it is about comparing different formats, FF or otherwise.

For example, if we are comparing mFT to FF, then the equivalence ratio (crop factor) is 2. If we are comparing mFT to 1.5x, then the equivalence ratio is 1.33. If we are comparing 1" to 1.6x, then the equivalence ratio is 1.69.

The point of Equivalence is to compare systems in terms of perspective, framing, DOF, total amount of light on the sensor, exposure time, and the display size of the photo. If all these are the same, the photos are equivalent. If one system can use parameters that another system cannot, and it results in a better photo, then, of course, one would do so.

However, the fact that FF is commonly used as a reference point is neither here nor there with regards to Equivalence. And, of course, if you're not comparing different formats, there's no need for Equivalence.

12 upvotes

Not quite ..
.
There were half frame and even smaller 110 interchangeable lens cameras at one end.
.
At the other, larger formats 6x4.5 6x6 6x7 6x9 and that's before we get to CUT FILM.
.
All had interchangeable.lenses - 35mm was the CONSUMER camera and good for (relatively) compact sports use

35mm was UNACCEPTABLE until the sensor (film) was good enough, I used a Mamiya BB67 for wedding formals into the nineties
.
... It completely outclassed 35mm, which I used for the rest of the shoot.
That was then.
Fact is: The OMD out-resolves the RB67 (with about 20 TIMES the negative area) by a VERY LARGE margin.

Edited 2 minutes after posting

2 upvotes

What a great article! It explained everything clearly, an excellent addendum to Tony Northrup's notorious video. Nevertheless, I am fully convinced that only 1% of photographers understand/comprehend the contents. Photographers are not (necessarily) intellectuals; in practice the huge majority is quite weak in maths, whatever great photographers they be anyway. This article and the entire subject is purely for nerds/geeks only. After reading nearly all of the follow-up comments, I fully understand the reaction and opinion of each of them: all and each is completely meaningful/reasonable from a certain aspect/perspective, and the majoroty of the perspective is that for 99% of the photographers (shallow) DoF equivalence is totally meaningless from purely practical point of view. Thank you again for the brilliant explanation and samples. Take care, Miki

0 upvotes

Re: "Nevertheless, I am fully convinced that only 1% of photographers understand/comprehend the contents."

That high, eh? ;-)

Re: "This article and the entire subject is purely for nerds/geeks only."

If, by "geeks", you mean people who wish to understand what the differences between formats are and why they are what they are, then, yes.

Re: "fter reading nearly all of the follow-up comments..."

!!!

Re: "...the majoroty of the perspective is that for 99% of the photographers (shallow) DoF equivalence is totally meaningless from purely practical point of view."

Inasmuch as DOF itself is "totally meaningless from purely practical point of view", that is true. However, it's not merely DOF, it's also diffraction and the total amount of light projected on the sensor, which means noise (so long as the sensors are close in terms of efficiency, which is usually a good bet for sensors of the same generation).

3 upvotes

Oh, I am delighted, Great! I love your explanations and comments tremendously. You really know what you are talking about. I am so much happy that eventually, I understood these concepts and the maths. Now the next task is to apply the concepts in everyday photography practice. Miki

0 upvotes

Remember, however, that Equivalence is only relevant when comparing formats, not when out taking photos. In other words, it isn't meant to be practical in terms of everyday photography.

For example, let's say you used mFT, and you find yourself thinking, "Man, I'd love a 35-100 / 2, but they only have a 35-100 / 2.8, and the Zuiko 35-100 / 2 is a beast". Well, you might consider FF with a 70-200 / 4.

Alternatively, let's say you used FF and were always stopping down to f/5.6 or so to get the DOF and/or sharpness you needed. You might want to consider mFT.

That's where the practical application of Equivalence comes in.

Edited 2 minutes after posting

6 upvotes

Perfectly and very well said, Sir! I have a Nikon V1, a Fujifilm X-A1, and I used to have a Sony A7 (now owned by my daughter), and I am in the middle of buying a m43 camera to get practical experience with that system, too. Exactly, this was my conclusion after reading this brilliant article: I have a nice legacy (FD, AI, MD, OM, PEN F) lens collection, and I'd like to use them on my m43, Sony, Fuji mirrorless systems (MF, of course). Take care and thank you so much, Miki

0 upvotes

Here is how anybody can experiment with the equivalence by himself. You can do it with any camera.

Select a few lenses with different FLs, or a zoom (preferably with const aperture) and shoot the same subject at the same distance at different crop factors keeping the same FOV adjusting the lens settings to the equivalent values. Keep the shutter speed constant, let the ISO float. At the end you should see almost identical pictures. That's the equivalence in action.

Some caveats: 1) not all sensors are ISO-less, so there might be variations due to different ISOs selected for equivalent settings; 2) lenses (and zooms) might differ optically at various FLs; 3) all lenses have some light loss, so equivalent F-stops might not correspond to the equal total light; 4) don't crop too much or pixelation becomes visible... In other words, all those imperfections that exist in practice, which we tend to overlook in the discussions, effect the equivalence in the same way.

Happy experimenting!

1 upvote

As an example of empirical data for the theory of equivalence, consider this simple graph:

http://www.dpreview.com/galleries/1681015916/photos/3002418/

It shows the noise performance of the Nikon D800 and the Olympus E-M1 for equivalent exposures. The values in the graph all come from DxOMark's tests of the cameras. As is clear, for two sensors of roughly the same generation, it all comes down to the total light!

Unfortunately, this seems to have offended some "full frame"-aficionados who think of smaller cameras as mere toys. Especially, since the E-M1 manages a fair bit better dynamic range at the equivalent exposures. I've tried to appease them by pointing out how the E-M1 is simply incapable of matching the high light performance of the D800 (without stacking), or the lack of v.large aperture lenses, but the idea that a "FF" isn't much! better than a "crop" at low exposures is simply too bitter a pill to swallow. :/

But with equivalence you have to take the sweet with the sour.

Edited 2 times; latest 3 minutes since posting

4 upvotes

What you say is true. What's important to realize is that there are two factors to the quantity of noise in a photo:

1) The total amount of light falling on the sensor
2) The efficiency of the sensor

The photon noise (noise from the light itself) is determined solely by the total amount of light falling on the sensor and the QE (Quantum Efficiency -- the proportion of light falling on the sensor that is recorded). For sensors of the same generation, the QE is all but identical, so that's a non-issue.

The read noise is the amount of additional noise added by the sensor and supporting hardware, and that can vary quite a bit depending on the ISO setting and the pixel count.

Sensors of a given generation with more pixels tend to have more read noise per area at higher ISOs, and this is what results in the lower DR for equivalent photos on the D800 vs EM1 at higher ISOs.

2 upvotes

I would like to see comparisons (and images) to some of the medium format sensors, especially WRT total light.

0 upvotes

Sure. However, what we would see is that any differences in the quantity of noise would be due to differences in sensor efficiency. That is, there is nothing special about medium format with respect to Equivalence.

0 upvotes

Yes, there is something special about MF sensors, just as there is/was something special about the old Kodak SLR/n sensor. It isn't the total light that determines noise levels. It is the sensor technology.

0 upvotes

The only thing "special" about MF sensors is their size and pixel count, which means they can gather more light for a given exposure and record photos with more detail.

And, yes, it is the total amount of light falling on the sensor, combined with sensor efficiency, that determines the noise in the photo. There are no sensors of a given generation where one sensor is twice as efficient as the other, but there are sensors that gather much, much, much more light than others.

In any case, the sensor in the EM5 and Canon 6D have radically different designs, and yet:

http://www.josephjamesphotography.com/dpr/pics/6DEM5/

What property might those two photos have in common? Hint: the same total amount of light fell on each sensor.

3 upvotes

Sorry, but I don't see any reason to believe that sensor technology and pixel size are not the determining factors. So, we'll just have to agree to disagree.

0 upvotes

You are, of course, free to dismiss all the evidence to the contrary, such as the example in the link above, but I should wonder why you would choose to do so.

3 upvotes

But why would anyone want to ruin a pet theory by examining the available evidence? That's a recipe for disaster!

0 upvotes

@ cainn24

"Pet theory", eh? Well, there was the example I gave above demonstrating noise equivalence to a "T" when comparing equally efficient sensors:

http://www.josephjamesphotography.com/dpr/pics/6DEM5/

And there are tons of examples here:

http://www.dpreview.com/forums/post/51782704

demonstrating the same principle with a variety of sensors, not all of the same generation, and of various efficiencies.

So, as I said, you are, of course, free to dismiss all the evidence to the contrary, such as the examples in the links above, but I should wonder why you would choose to do so.

1 upvote

I was referring to Kerry's pet theory about total light not being the primary factor in the noise equation (for a given sensor generation).

All the evidence is, to an overwhelming degree, on your side. In the colloquial sense equivalence isn't a theory, it's a fact.

0 upvotes

@ cainn24

My apologies for the misinterpretation.

0 upvotes

It's all good. I should have been more clear about who I was addressing.

0 upvotes

@pdelux,

Since you asked, I have been trying to point out how ridiculous it is for DPR to point out that FZ200 has equivalent aperture of f/15.5.
BTW, I don't own FZ200 so it's not personal. Also, I'm not disputing the math.

Edited 7 minutes after posting

0 upvotes

It's no more or less ridiculous to point out that the FZ200 has a FF equivalent f-ratio of f/16 wide open than it is to point out the FZ200 has a FF equivalent focal length of 600mm at max zoom.

As has been pointed out to you multiple times, at max zoom and aperture, the lens on the FZ200 is 108mm f/2.8 which is equivalent to 600mm f/16 on FF.

Simple. Clear. Correct.

If you say it 17 more times, and I give you the same response 17 more times, the comments can hit 2K, which, I suppose, is one way of looking at it. ;-)

Edited 3 minutes after posting

8 upvotes

Total comments: 2154

Editorial content

Product database

Community

About us Work for us Advertise with us FAQ Feedback Interest-Based Ads Privacy Legal

All content, design, and layout are Copyright © 1998 - 2016 Digital Photography Review All Rights Reserved. Reproduction in whole or part in any form or medium without specific written permission is prohibited.

Mobile site


FlatCam prototype eliminates lens using pinhole-like sensor mask Feb 16, 2016	Sony develops 22MP smartphone sensor with on-chip AF processing and video IS Feb 16, 2016	How tough can it be? LensRentals looks at why it's so hard to make lenses Feb 11, 2016	Panasonic and Fujifilm reveal joint sensor development, offering global shutter Feb 4, 2016

72	What is your favorite fast M4/3 Prime AF Lens?
69	Busted by client for using Panasonic cameras
55	The essence of Photography
53	Argument between friends over these 2 photos (Canon 6D VS Nikon D810)
50	For over 50 only!
45	Oly 300mm f/4 Bring on the pics!
43	Ricoh to enter the mirrorless interchangeable lens camera market?
41	GX8, Pen-F, or EM1: Best C-AF?
40	* Wed C&C "No Theme" Thread #412 2016 03 15 *
40	LX100 Colorful Tasty
40	Will DSLRs compete with mirror-less?
39	My First Wedding Shoot
37	Sigma Used My Moon Shot
36	Be Careful When Ordering Cameras From Outside The USA (or Canada)
35	Starting Afresh With Fuji in Mind
34	Baby Pen is coming soon (EPL8)
33	Pentax Greens are BAD
33	How do you avoid / remove noise on you m4/3?
31	DSLR sensors, last longer?
30	Next 645?