Megapixels in photography, what influence?

Today, I decided to talk about a more technical subject, something I do less on this blog but should interest you. I’m talking about megapixels in photography, on a camera body. What are they really? How do they influence our pictures? Is it just a marketing number? Why are so many people interested in these famous numbers? In short, I’ll try to answer all these questions in this article. If you are interested in learning more, I invite you to read the articles on the basics of photography.

A question that many beginners, and even experienced photographers, keep asking themselves when buying a new camera: should I get a sensor with more pixels or will it affect performances? Because, for many years now, the definition of sensors has been a characteristic often promoted by manufacturers, and many true and false ideas circulate about the influence of this number of pixels, especially when it comes to cameras with a number of pixels higher than average.

First of all, what does this number of MegaPixels (MP) represent? Without going into too many technical details, there are on a sensor what are called photosites which capture the light, digitally these photosites become pixels, which means that a sensor with 20 million photosites will provide 20MP images. We can imagine a mosaic, on which the sensor will “paint” each fragment (or each pixel) so that, once all the fragments painted, the image is formed. Once your image is saved, you will be able to display it on a screen or print it.

On forums, blogs, YouTube and other platforms, you may regularly come across wild discussions or articles/videos explaining all that a large number of MPs provides… But above all, all that it will make you lose! Although it’s a bit more nuanced than that, what comes back generally is that having more pixels will allow you to print bigger and crop more your images. And that’s about it, while the list of drawbacks will grow longer as you are surfing the Internet, with randomly: lower ISO sensitivity, obligation to use higher-end lenses, heavy files to transfer, store and process, a more limited burst rate, smaller photosites capturing less light, increased risk of motion blur etc… You’ve got to know that in this list, 4 assertions that come up very often are already false!

One thing you need to know right now and that you will soon realize is that most cameras, DSLR or mirrorless cameras and even compact or bridge, which are intended for beginners or even experts, have very similar definitions, even with different sensor sizes. Thus, a Sony RX100 will rely on 20MP, and usually the majority of cameras with 1″ sensor like the Canon G7X III, the Panasonic FZ1000 or the Sony RX10 II will have the same definition, simply because Sony provides this sensor to other brands. Most of the recent mFT mirrorless cameras will offer you a 16 or 20MP sensor like the Panasonic GX9 or the Olympus E-M5 II. Most of the recent APS-C will provide 24MP images like the Nikon D5600, the Canon 2000D or M50, the Sony A6000 or the Fuji X-T20. And even with full-frame sensors, most cameras will have sensors with resolutions between 16 and 30MP. However, all the cameras mentioned here, except some low pixelized but very expensive FFs like the Nikon D5 or the Sony A9 II, are rather intended for a public of amateurs who can’t afford or don’t need to buy a more high-end camera.

Cameras that will really have a quantity of pixels significantly higher than the others will therefore be in Full Frame format (the D800/Z7 range at Nikon, 5DS at Canon, A7R at Sony and S1R at Panasonic) or in “Medium Format”, with sensors ranging from 50 to 100MP. That’s quite a lot of references, but I let you go and see their prices out of curiosity and prove that these are the cameras that will interest most of you and if you would be ready to invest as much! If you’re not curious, most of the full frame cameras mentioned start at 2500-3000€ and the cheapest Medium Format camera costs 6000€ – so for a huge majority of you, these are cameras you won’t even look at, especially if it’s your first one. Good news: all these considerations on the number of pixels will have almost no impact on the choice of camera for most of you!

But then why would I talk to you about it? Simply because, if you end up being interested in it (why else would you be reading about it, eh!), it is important that you know how to sort the truth out from the lies. Besides, many of those who will tell you “this camera has a sensor full of pixels, it’s a marketing argument, it’s useless, get another one! “shows that marketing actually works well since they try to convince you not to take a camera, with the only argument of the number of pixels and without mentioning any other feature of the camera in question. It doesn’t matter if it has the best Autofocus, the best ergonomics, the best features… You shouldn’t take it because it has too many pixels! Curious, isn’t it? A bit like if someone told you “there’s an extra spare tire in this car, it’s just marketing, buy another one”.

The first, and perhaps the only thing to remember is that the more you enlarge an image, the more you accentuate its flaws. Expressed this way, we can think that if the sensor provides more pixelated images, these images will be larger, and we will see the flaws more and more. But you should also remember that you are going to look at these images on a support (screen or paper) which will often not display all the pixels of your image at their “real” size. For those who don’t know, a FullHD screen will only display on its surface 2MP, 8MP for a 4K screen which is also about what an A4 print will require to display a very high quality. Small precision: for a print, we talk about resolution because it is not a total number of pixels but a number of dots per inch, the famous dpi. But with a mathematical calculation, we can easily define the definition an image must have in order to be printed with a very high-quality resolution (generally between 250 and 300dpi).

You understand that displaying your 24MP image in full screen on a FullHD won’t make you see a 24MP image but a 2MP image, which is equivalent to reducing the real size of the image. Now, you should know that, in relation to the first sentence I highlighted in the previous paragraph, the reverse is also true. That is to say that, whether you have a full-frame sensor of 24 or 50MP, the image you will display will be equivalent to a full-frame sensor of 2MP. Of course, if you zoom into the image, you will enlarge it and the image quality will be impacted, especially the sensation of detail and digital noise. But the quality will be affected as much with one as with the other at equal enlargement, at least if they are not separated by several generations of difference. This means that at equal enlargement, the ISO increase and the level of detail in the image will be at least equal with both sensors, and you will also find that motion blur is not seen more on one than on the other…Just as you will find that at equal exposure time and aperture, in order to have the same exposure, you will have to use the same ISO sensitivity – which confirms that the 50MP photosites, although smaller, did not capture less light but as much as the 24MP ones.

To make it simple, as long as you look at the images on the same media with the same zoom, the quality between 2 sensors of the same size will be almost the same, no matter if one is 10MP and the other one 100MP. In the same way, if a lens would suit you on a less pixelated sensor, you won’t need a higher end lens either because you use a more pixelated sensor – as long as it doesn’t “require” more than 10MP, which is the case with FullHD or 4K displays and A4 or smaller prints. On the other hand, if you look at your image on a support displaying more pixels than one of the 2 sensors, for example an 8K (32MP) screen, there you will see a difference, but to the disadvantage of the 10MP sensor because in order to visualize in full screen, it will have to “invent” pixels to display those the image doesn’t have. Paper prints always require the same number of pixels: it was recommended 8MP for a very high quality A4 10 years ago, it is still the case today and it will still be in 10 years. On the other hand, even if the displays are the same physical size, their definitions are evolving which means that a 12MP image will be comfortable on a FullHD or 4K display but will start to “suffer” on a 5K and even more on an 8K. Choosing a camera with a more pixelated sensor will ensure that your images are relatively more durable.

The example of video

One of the many good reasons to opt for a sensor with fewer pixels, apart from those mentioned above in relation to the weight of the files, would be a video use, but which would be relatively intensive. Indeed, cameras having a sensor with many pixels can still make good quality videos which, for a “classic” use, will be more than enough.

However, to film in 4K for example, sensors do not need more than 10MP (8MP but, with the 16:9 ratio in video, we lose a part of the sensor). In fact, the more pixels the sensors have in relation to the definition of the video, the more processing will be needed to scale, or even crop the image, which will therefore reduce the quality in addition to losing the field of view. Moreover, this increases the rolling shutter phenomena (image distortion during fast dollying, for example), makes it more complicated to obtain fast frame rates such as 120fps, and complex and heavy processing needed to scale the image to the video will more quickly cause risks of overheating.

At this stage, you should be asking yourself “why take a sensor with fewer pixels if it doesn’t alter the image quality? ». Although some of the ideas mentioned in the 1st paragraph finally turned out to be false, others remain true and are mostly related to the weight of the images made with these sensors full of MP. Indeed, more pixels implies heavier files, so they will take longer to be recorded, transferred and processed.

This means that the camera will generally be more limited in bursts, that a larger memory card will be needed, transfer time between the card and the computer will be longer, and processing and conversions done with software will take longer to display and record. You will be able to reduce the display and processing time with a more powerful computer, but it will be more expensive as a result. And as said above, pixel-packed cameras are rarely the most affordable, so between that, the lenses you think you’ll have to stick to them, and the computer you think you’ll have to buy to have “acceptable” treatment times, selling a kidney won’t often be too much (think about renting people from your family, if necessary, haha!).

More pixels won’t degrade the image quality, but it will nevertheless impose some constraints, especially financial ones. And if you tend to do what those who spread misconceptions about quality do, like zooming in at 100% without taking into account the difference in enlargement, it will be worse because you will convince yourself that your lenses are not or no longer good. How many times can we read this sentence: “a lot of pixels requires high-end lenses”. In reality, the principle is simple: we recommend high-end lenses from the moment you buy a high-end camera, in APS-C for example, you will be told to take better lenses on a Nikon D500 than on a D3400 when it has a less pixelated sensor.

Highly pixelated sensors simply allow larger enlargements, the more you enlarge an image the more you accentuate its defects and vice versa. Their main flaws will therefore be the size of the files and obviously the price. Remember however that if more pixels can bring a greater durability to your images, it will not currently be visible, and when screens increase their definition, 20-24MP images will remain quite acceptable. In fact, 10 years ago, it was estimated that 12MP was more than enough and 24MP was useless. Today, it is estimated that 24MP is sufficient and 50MP is unnecessary. What will we estimate in 10 years?

And finally, don’t forget that cameras with very pixelated sensors are anyway not accessible to the majority of you, and whether you can afford to buy them or not, you have to first of all be interested in all the other features such as sensor size, AF performance, construction quality, ergonomics, grip, functions…which will very often be much more important.

I hope you liked this article, a bit more technical I have to admit, but I hope you still enjoyed it. Once again, I thank Alex for his clarifications and corrections. To go further in the learning process, why not read the interest of RAW and Jpeg in photography?

Sylvain