Alley in Pisa, Italy

by Lars Kotthoff

submit your photo


Hall of Fame
View past winners from this year

Please participate in Meta
and help us grow.

Take the 2-minute tour ×
Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It's 100% free, no registration required.

This is going off of today's iPhone 5S announcement, when Phil Schiller said that "bigger pixels" is the key to a better picture.

apple says bigger is better other parts are larger too

  1. What does this mean?
  2. Is it true or just a marketing ploy?
share|improve this question
3  
It means Apple is following Nokia, but they will make it look like they are ahead of Nokia. Sensor area of iPhone5 is 15,5 mm2 and Sensor area of Lumia1020 is 58,1 mm2. (source) What will be this new sensor in iPhone5S? –  Esa Paulasto Sep 11 '13 at 5:47
    
@EsaPaulasto handwaving calculations based on both the "15% larger sensor" and "1.5 micron pixels" give a size of around 4.9 mm x 3.6 mm = 17.6 mm^2. –  Philip Kendall Sep 11 '13 at 8:03
1  
    

5 Answers 5

up vote 37 down vote accepted

What does this mean?

This is referring to the size of the photodetectors, which represent the pixels, on the camera's sensor.

In short, the more light reaching these photodetectors, the more accurately they can measure light levels, thus reducing noise. To simplify things, noise is usually a factor in low-light photography.

There are two ways to get bigger photodetectors on a sensor:

  1. Have fewer pixels.

  2. Make the sensor larger.

The former is a trade-off against resolution, whereas the latter is a overall gain. In the iPhone 5S, they've gone with the latter.

Small sensors such as those in cellphones and compact cameras suffer the most from having small pixels, because their pixels are tiny compared to those in DSLRs and other professional/prosumer cameras.

The megapixel race

The desire to cram more and more pixels into camera sensors has harmed the sensitivity of the sensors, particular small sensors in cellphones and compact cameras, resulting in more noise at the same light levels.

Unfortunately, it was driven by the mentality that "more pixels = better", as megapixel count is very easy to market your camera by. In reality, diminishing returns ensure that the more pixels you add, the less important it is to add more, and we're getting to the point where a lot of sensors, particularly the tiny ones inside cellphones, can't even resolve enough resolution to justify the megapixel counts anyway, making those extra megapixels more a burden than a benefit.

Advancements in technology such as microlenses and back illumination, along with more advanced noise reduction algorithms, have gone some way toward making up for the lost sensitivity due to increased pixel counts in such a small sensor, but not far enough: today's cellphone cameras are still horribly noisy due to their small light-gathering area.

It is very difficult to market your camera based on sensitivity or other aspects of image quality, when megapixel count remains the primary marketing statistic associated with digital cameras, however misleading.

So why is Apple marketing it this way?

The uninformed public still expects that if a camera improves, its megapixel count rises.

Apple, however, is releasing a "better" camera in their cellphone, but with no rise in megapixel count.

Technically, this is a good idea for image quality. However, it's difficult to explain to people why, when they've been conditioned to believe that megapixel count is the primary metric for image quality.

Marketing the iPhone 5S camera in this way is an attempt to:

  1. Shift the marketing of camera sensors away from pixel count and towards sensitivity and other aspects of quality.

  2. Distract from the fact that megapixel count, which many people still assume determines image quality, has not risen.

Unfortunately they are not the first to attempt this same revolution and won't be the last. The megapixel myth is very firmly entrenched.

Note: in response to some comments below, I define the "megapixel myth" as the myth that pixel count is the primary factor determining image quality and that more pixels is always better quality - not that pixel count doesn't have any benefit. Also inherent to the problem of the megapixel myth is that it distracts from important statistics such as sensor size, which is virtually never even talked about for cellphones and compact cameras.

Is it true or is it a marketing ploy?

It is technically true, but in this case the difference is so small as to be insignificant.

It is true that keeping the same number of pixels and making each pixel bigger results in better sensitivity, and that's what they've done here. This should result in lower noise (or, if the camera can use faster speeds as a result, possibly less motion blur in low light). Their actual claim that "bigger pixels = better picture" is a simplification of this and could potentially confuse people into thinking their outdoor daylight pictures might also be improved, which is not the case. But in low light scenarios the improved sensitivity will provide a benefit.

However, they haven't made them very much bigger. With only a 6.6% increase in sensor size, any resulting gain in sensitivity is going to be tiny: so small that it insignificant. To put it into perspective, a DSLR has a sensor that is approximately 500-900% as big as a compact camera. That would result in a significant sensitivity gain of multiple f-stops. A 6.6% increase is very insignificant. It represents about one sixth of one f-stop.

So it's kind of a marketing ploy as well. To make any significant gains they would have to increase sensor size a lot more. It's also a marketing ploy because previously their marketing did focus on megapixel count, and yet they've now switched the way they market because they have opted not to increase the megapixel count and still want to be able to boast that they have improved the camera.

share|improve this answer
3  
All else being equal, it will, because physics. More area will gather more photons. The only case in which this would not result in better sensitivity would be if the CMOS technology used went significantly backwards. We're not talking about 1960s photodetectors. See also my paragraph starting with "Advancements in technology ...". –  thomasrutter Sep 11 '13 at 7:04
4  
it is not constructive to start comparing to 1960 sensors... when discussing one parameter, "all else equal" is ALWAYS a GIVEN. –  Michael Nielsen Sep 11 '13 at 7:42
2  
The megapixel myth is itself a myth, larger pixels means better sensitivity and higher quality per pixel but more smaller pixels means better sensitivity and higher quality per image (all else being equal). I print pictures, not individual pixels, so I'd much rather have more megapixels. –  Matt Grum Sep 11 '13 at 8:41
3  
@MattGrum - In the case of the iPhone, however, the only difference is pixel size, so it should result in improvements in the IQ of the phone. This is why FF sensors generally outperform APS-C sensors with the same number of pixels when in the same generation. However to back your point, as Nikon managed to prove with the D800, current technology will permit a massive increase in the number of pixels without sacrifice to the image quality. People are still getting their heads around that concept. –  John Cavan Sep 11 '13 at 10:34
2  
Also arguments about pixel density in FF and APS-C sensors don't scale well down to cellphone size anyway. In a cellphone the sensor is so much smaller that the pixel densities required even for a modest 8 megapixels huts up against physical limitations: the minimum circle of confusion of the optical system, the relative size of the photodetector to its surrounding circuitry, and so on. 8 megapixels on a cellphone sized sensor is really pushing it while 24 megapixels on FF, or even APS-C, is really quite sparse –  thomasrutter Sep 11 '13 at 13:31

In theory having more smaller pixels is better than having fewer large pixels.

A small pixel will capture fewer photons and thus it's output is noisier, but by taking more samples the noise averages out, by simple resampling you can simulate the result of a sensor with fewer larger pixels.

But you can do better than simply matching the result of larger pixels. More small pixels together capture more information. Big pixels let you know how many photons fell in say a 5µm x 5µm area. Small pixels also tell how many photons fell in that same area, but can also reveal how many fell the upper left 25µm x 25µm region, the lower right 25µm x 25µm region etc. this extra information can be used for smarter resampling to yield less noisy more detailed images at the same resolution as a camera with larger more sensitive pixels.

In practice things are never quite equal. When read noise is taken into account reading the values from 4 smaller pixels means taking the read noise penalty four times, leading to greater noise in the shadows and reduced DR. This shouldn't be a problem with the latest Sony Exmor sensors with ultra low read noise but may be the reason Canon have as of 2013 not produced a body with more than 22MP).

Pixels aren't flat entities. Usually the photosensitive part is at the bottom with the surrounding electronics on layers above. This makes the pixels a bit like a well. When you make pixels very very small, the proportion of photosensitive area goes down and the depth of the well as a proportion of it's width gets much larger, reducing the number of angles it can receive light from and thus the sensitivity.

This is where they're at with phone camera sensors, they can't increase the number of pixels without seriously sacrificing fill fraction and angular efficiency, hence to try and appeal to consumers they're reducing the number of pixels. The best solution is to make the sensor bigger, but that would require a bigger lens and wouldn't allow for a super slim body that sells lots of phones.

There are also disadvantages to having fewer larger pixels. The reduced sampling frequency leads to increased moire, requiring stronger anti-alias filters exacerbating the problem of reduced resolution. A lower sampling rate also leads to more demosaicing errors.

Sadly the "small pixels bad" mantra is applied across format sizes and is causing camera manufacturers to become reluctant in pushing quality forward by increasing pixel counts as the "conventional wisdom" is that it's a bad idea and makes images noisier, more prone to camera shake, and requires better lenses (every one of these points untrue).

share|improve this answer
    
Don't know that I'd say Nikon has followed the "conventional wisdom" entirely given the D800... or you could argue that the conventional wisdom is starting to change given the trend of removing the AA filter being another sign of that too. –  John Cavan Sep 11 '13 at 10:27
1  
This makes me wonder if Apple reduced the pixel count in their new iPhone 5s. But I thought they did not. Didn't they make a bigger sensor instead? –  Esa Paulasto Sep 11 '13 at 19:41
    
They kept the same number of pixels and increased the sensor size (by a TINY amount, though) from 1.4 µm pixel size to 1.5 µm pixel size (1/3.2" sensor size to 1/3" sensor size). –  thomasrutter Sep 13 '13 at 5:29

It's all about catching the light that comes in as accurately as possible.

It's fairly simple, with a bigger surface (per pixel) you will catch more light and be able to make a more accurate representation of that light in the end-result.

It is true that the photo's will be better, but just 15% per pixel isn't spectacular. So it's marketing as well. Nokia has done a similar thing with their PureView. Downsampling 41 Megapixels to 5-10 megapixels will also dramatically decrease noise-levels. Nokia's PureView will be superioir to this if you ask me.

share|improve this answer

The main idea here is that bigger sensor yields better picture (if the sensors use the same technology). What Apple have done is to increase the sensor size, and by keep the same number of pixels each pixel will have a higher signal-to-noise-ratio, resulting in a better picture.

share|improve this answer

There are two main advantages from making bigger pixels:

  1. bigger pixels collect more light, so the sensor are more sensitive. This means that in low light condition you can shot with low ISO and so with low noise.

  2. with large apertures (e.g. f/2.2) and little sensor occur that many light rays interact randomly with different pixels blurring your image. And this lower the resolution.

However even if we are debunking the well know myth of mega pixel we are generalizing "the bigger the better" truth about sensors. Of course how stated above is true but the performance of a camera depends from many other factors. To project an image in a bigger sensor maintaining a short distance between the frontal lens and the sensor (as an iPhone have to do) you need more lenses, and if the quality of one of these lens are not good, all the performance of the system will be worst. So "Bigger Pixels = Better Picture" is more a slogan than a law.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.