Karroly: You forget to say that increasing the diameter of the "tubes" is like waiting longer under the rain : this will increase and make more equal the number of raindrops caught by each tube, thus reducing the "signal-to-noise" ratio. For sensors, it means, putting aside electronic noise, bigger pixels will catch more photons and average the random nature of light you are talking about here.
This is why, given the same resolution, a big sensor with big pixels will ALWAYS be better in (extremely ?) low light than a small sensor, even though there were no electronic (or thermal) noise at all...
However, in practice, the true question is, and your article does not give an answer, under which pixel size and scene brightness, does a pixel catch too few photons so that the difference between adjacent pixels becomes noticable ?Is this "noise" greater or smaller than the electronic or thermal noise ?In other terms, do we really have to care with the random nature of photons with today sensors ?
Using a larger test tube is just like adding together all the water collected in the smaller tubes. From a shot noise point of view there's no difference.
There is a read noise difference, but that's the topic of part 2, which is why it's not covered here.
Exactly when you're limited by shot noise and when you're limited by read noise will vary, situation to situation and camera to camera, but again, that's something for part 2.
Mssimo: In a few cameras with "ISO Irreverence". Would you recommend not to use anything but base ISO up to the point where the camera starts to use hardware amplification?
I like ISO Irreverence, I might have to use that.
That's a subject for part two. Which I need to finish writing and illustrating.
Marcelobtp: So now i can understand why have a "sun" inside my studio with very high shutter speed; leaf shutter please... Burn the model and get the finnest image quality possible.
The very fast shutter will cancel-out the benefits of mounting a star in your studio.
Karroly: "As a result, when you shoot two different sized sensors with the same shutter speed, f-number and ISO, the camera with the smaller sensor has to produce the same final image brightness.. from less total light."
I am sorry, I do not agree at all.
Please, let me put this another way. If I use an F:2.8 FF lens on a APS-C body, it is still an F:2.8 lens and the picture taken (at same aperture/speed/ISO AND PIXEL SIZE) is just a crop of the FF sensor. The APS-C area of the FF sensor gets the same amount of light (either total or per area unit) than the APS-C sensor and thus the signal-to-noise ratio is the same...
For a rough example, [compare the Nikon D7000 to the D800](http://www.dpreview.com/reviews/image-comparison/fullscreen?attr18=daylight&attr13_0=nikon_d7000&attr13_1=nikon_d800&attr13_2=nikon_d7000&attr13_3=nikon_d800&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=1600&attr16_1=1600&attr16_2=1600&attr16_3=1600&normalization=full&widget=1&x=0.10529966390216211&y=0.49552852974905925)
The APS-C region of the D800's sensor is around 15MP, rather than the D7000's 16MP, but it's pretty close. In this example both are shot at f/5.6, 1/640 seconds.
At native resolution you can see that the pixel-level performance is very similar (as theory would suggest).
Now click on the 'Print' button at the top right of the widget, so that you can compare them at the same scale. The camera that's captured more signal looks better, even though the pixel-level performance was very similar.
You may not agree, but it remains true, I'm afraid.
You're right that the APS-C crop from the full-frame camera will be identical to the native APS-C image (same light per unit area, same area). It's also the case that the pixel-level quality will be the same because the light per unit area is the same.
However, as soon as you view all three images at the same size (either by blowing-up the APS-C crops or downscaling the full frame image), the full frame image will be cleaner.
The full frame image is made up from more signal, and this grows faster than noise. There'll be more of a discussion of this in part two (or possibly three) when we look at the effect of combining data from multiple pixels.
DirkL: Given the same amount of light, a larger sensor will be illuminated less than a smaller sensor.See the definition of lux.
That's true, but you'll need to state a lot more of your assumptions for that statement to be photographically meaningful.
The point made in the article is: when trying to take the same photo (same shooting position and framing, ie: equivalent focal lengths), then in the same scene illumination, with the same f-number and shutter speeds, the large sensor will see the same luminous flux per unit area as a smaller one and hence will be more illuminated in total (since it is measuring that flux over a greater area).
pidera: As pointed out by others, the summarizing phrase "There are three factors that affect how much light is available for your sensor to capture: your shutter speed, f-number and the size of your sensor." is a bit unambitious. Shutter speed and physical aperture do the trick, no need to bring in sensor size except for the fact that sensor size and available options for physical apertures are somewhat linked (albeit more in the past than now).
I agree but it made more sense to me to talk in terms of sensor size (which people can relate to), rather than aperture diameter or equivalent aperture, which some people push back against.
I wanted the largest possible group of people to be able to engage with the article, so I tried to avoid anything that would distract from the main points I was trying to make.
sans culotte: As in mentioned equivalence article here are some important details omitted."the full frame camera will see four times as much light as a camera with a Four Thirds sensor"The camera itself doesn't deal with light. Light is recorded by sensor pixels. So pixel density & efficiency is the key.FF sensor area is 864 mm^2 & APS-C sensor area is 329 mm^2. That means Fuji 16MP sensor has lower pixel density (16000000/329 = 48632 px/mm^2) than new Canon (or Sony, who knows?) 50MP sensor (50000000/864 = 57870 px/mm^2). So in this example with FF you get more detail with relatively same noise, it's still nice, but easily shows that bigger sensor doesn't automatically means more light.Also there is sensor efficiency which has an impact on final result.All that means that the only reasonable way is comparing specific sensors noise rate.
Because the effect of smaller pixels involves read noise, it's something we'll either look at in part 2 (which is about read noise), or part 3, if we end up extending the whole thing in response to the range of interesting examples raised in the comments.
cpt kent: Seems to be a lot of folks exclaiming brilliance without questioning. Sources? References? Research? Further reading?
This was purely intended as a beginner's guide to... so there are no sources, *per se* beyond some conversations Rishi and I have had.
In terms of further reading, though, I believe [Emil Martinec's much more thorough exploration](http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/) is the place to go.
Andy C Knight: I'm having difficulty understanding this line...
"There are three factors that affect how much light is available for your sensor to capture: your shutter speed, the size of your aperture (not f-number) and the size of your sensor."
I keep thinking... Shutter, F number & sensor size or...Shutter, physical aperture size & focal length.
Can someone explain where I am going wrong?
Andy C Knight, you're absolutely right. I've corrected my mistake.
Mikity: Good article. I think the changing exposure/changing ISO section is needlessly convoluted though - and it's not obvious you can hover over the images!
I take your point, I'm just not sure what I can do to make the 'buttons' more obvious, without breaking our style sheet.
Anastasiadis Lazaros Thessaloniki Wedding: I read the article and I personally don't agree with a lot of things. Raindrops inside tubes don't behave like light in photos, sensor size does not matter the technology and quality of the sensor does not have to do with its size. This article tries too hard to impress the amateur photographer with magic tricks but actually it does not offer anything new except maybe some confusion about how you should take photos without much noise.
The technology and quality of the sensor certainly have a role to play, but that's what part 2 of the article will be about.
However, with the exception of some quite specific examples, the dominant source of noise in most tones in most photos comes from shot noise: not the camera. And the sensor size *generally* plays more of a role than the specifics of its technology.
falconeyes: Interesting and important article.
However, it should have used fewer words. The article makes a simple matter look more complicated than it really is. And may discourage some to read it.
Everybody thinking that noise is (mostly) a camera artefact should read the article tough.
It's something I've been thinking about recently (and whether we should expect this from a camera company or from independent developers). The Auto ETTR stuff that Magic Lantern are doing support my theory it'll probably be the latter.
As soon as you start thinking through the implications of a true ISO-less camera, you realise how much of conventional camera design you would probably end up throwing away (which I think makes it more likely that these changes would come from outside the industry).
tlinn: Great article, Richard. A couple questions:
1) Does shot noise present as primarily luminance noise or color noise too?
2) Am I correct to infer that part 2 of this series will answer the question of whether or not it is beneficial to ETTR at ISOs other than the base ISO?
1) My understanding is that it appears as both chroma and luminance noise, but there may well be people able to give a more precise answer than this.
2) Part 2 will address why the best approach varies between cameras and will clarify the sections of our reviews that look at this.
Thank you. And I must admit it's ended up longer than I planned.
Rex1227: Does that mean the a shot at a specific ISO cannot get any cleaner up to a point since the amount of photon noise cannot be reduced?
For each sensor size, f-number and shutter speed, there will be a certain amount of shot noise associated with each tone in the image. That will act as a hard limit on how good a particular ISO setting can get.
jon404: ***** Five-Star Article!Thanks, Richard. Your writing is so clear and informative... VERY much appreciated.
Credit should also go to Rishi. This article stems from many, many discussions about noise, sensors, the ISO standard and how people use cameras. It wouldn't be the same article without his input.
RichRMA: "Output size" I thought was determined by pixel count?
"In turn, this is why we talk about different sensor sizes representing an image quality/size/price balance: because, so long as the sensor's electronic performance is similar, the effect of shot noise means that sensor size is the major determinant of image quality. Yes, pixel count can make some difference, but shot noise tends to play a much bigger role, if you compare images at the same output size."
If you cut a FF sensor in 4 parts, essentially making it a 4/3rds size sensor, the individual pixels will still render the same noise as when it was a full-frame sensor. Overall light collection does not effect the response since the pixel size does not change and therefore the noise will not change. You could do an experiment by inserting a physical cropping mechanism in front of the FF sensor and see this is the case. All that will happen is the angle of view of the scene will be cut in half, which has no impact on noise.
Output size is entirely up to you. View images on your monitor or print them out and you've probably rescaled. It's certainly unlikely that you'd choose the size based on the captured image's native dimensions.
Take a 1/4 crop from a full frame image and yes, the pixel-level noise will be identical to any other part of the full image (and to a 4/3rds size sensor). However, scale (up or down) the image to the same viewing size as the whole sensor image and the noise will be more apparent.
Photato: About Sensor Size.Strictly speaking Sensor Size should not be a factor in Noise, because lenses can concentrate the same light (photons) in a large or small area.So much so that you can start a fire concentrating Photons with a magnifying glass in a small area.
What happens is that normally smaller sensors have a higher Density of pixels making it less efficient in photon collection.
For instance. A Small Sensor filled with 8 Micron Pixels should be able to collect the same amount of Photons than a Bigger Sensor filled with 8 Micron pixels. The difference is that the Bigger Sensor would have higher resolution.
Photato: there are always complicating factors (for instance a pixel that captures too much light means you can run out of bit-depth to fully encode everything it's capturing in the shadows), but *broadly speaking*, shot noise plays a larger role in most situations.
Part 2 will talk more about the role of electronic noise (though it'll need a much more length piece or, perhaps, a book to detail every possible source of noise covering every possible case).
mpgxsvcd: he was suggesting 'condensing' the light that would usually cover a large sensor down onto a smaller size.
Which, as you say, changes the net focal length and hence the f number.
Eg: 35mm F1.8 on FF vs the same lens with a 0.66 condensing lens behind it. The net focal length of the 35mm lens and condenser would be 23mm. On this shorter focal length, the (unchanged) aperture would now give an f-number of f/1.2.
Consequently, a 1.5x APS-C sensor would see the same field-of-view as the full frame sensor, but it would experience greater light per-unit-area if the lens was left wide-open. So you'd get the same field of view and the same depth-of-field but, since you're essentially now using a 23mm f/1.2, you'd be more prone to over-exposure if you kept the same shutter speed.