The Megapixel Myth

Now I’m not saying I know a damned thing about photography. I know what aperture does to your depth of field, I know why high speed film is grainy (even when it’s high speed digital “film”) and I know why a fast 1200mm lens makes an excellent christmas gift, but that’s mostly about technology, not photography. The kids in Born Into Brothels didn’t know any of that stuff and they shot more powerful photos with dollar store 35mm point and clicks than I will likely ever manage with cameras far more spiffy than I will ever buy. Photography is about taking the things that you see in the world and capturing them so that other people (including future instances of yourself) can see them the same way you did. In a sense this is, of course, impossible for a host of reasons but basically you take a photo so that later you can look at the photo and go, oh yeah, remember that?

Since the end goal is pictures, printed or on-screen, a certain contrarian attitude has emerged among the would-be elites of the digital photography circuit. It goes, to a first approximation, like this:

The human eye can only recognize 200dpi or so, so even for an 8×10 print, which is as big as I would want to get, [the argument is always phrased in a run-on sentence with comma-splices] you’re only looking at 1600x2000pixels, which is 3.2 megapixel, therefore the idiots buying 5, 7, or 10 megapixel cameras are just tools of the corporate whore neo-fascist consumerism brainwash, and their purchasing of these cameras affirms their sheephood. Burn.

The problems with this argument are myriad, but since it tends to come up a lot I figured it was worth laying out at the very least some general outlines. First off, let’s agree with some of the spirit of what’s being said, to wit: a lot of twits with more money than brains think they can buy themselves status by deciding they are somehow purer of heart than the other dilettantes, and buying professional grade SLR backs and lenses and tripods and cases and light meters and battery packs… and then leaving them carefully packed away at home because they’ve lost interest. This is a very valid observation, and all right minded people should join hands, light candles and, much as they did in the seventies, beat the living shit out of these people and take their cameras. But this observation does not, in itself, invalidate the market for cameras above 3.2 megapixels.

We all agree that actual pros are perfectly justified in owning the top of the line, because we imagine their print needs are far beyond our own, and we’re not entirely wrong, though a lot of pros don’t need 12MP on every assignment, or maybe ever. Herein lies the first problem with the myth of the 3.2 MP barrier which, as I have already anticipated in the title, I will henceforth abbreviate to “the megapixel myth” or just “duh.” The fact is that pros are also more demanding on their camera’s engineering, optics, and feature set, and are willing to pay for it, so those are the cameras that get the best work. Buying a higher MP number doesn’t automatically get you a better piece of kit, but it often does, and it’s relatively rare if it gets you a worse one (at least from the same manufacturer, and with the note that there are exceptions to this.) The sensor might not be worth it to you, but the quality definitely might be.

The quality argument goes further too. A higher pixelcount often means imperfections in the lens(es) become more evident, both in the quality of the glass, and in the construction of the barrels. Maybe you think your 1.5MP phone cam takes fine pictures insofar as you can still recognize people and all, but it’s amazing how much easier it becomes to do little things like focus, when your optics are helping instead of hurting you. Quality, in photography, often means becoming invisible — not assisting or altering so much as failing to detract. Obviously camera makers are incented to put that extra work into cameras with higher MP counts where a) people might actually notice and b) there is any pricing margin whatsoever to squeeze it in.

But the big one, the huge elephant in the room when it comes to people making this argument, has nothing to do with professionals, in fact it almost makes it sound like pros are the only ones who should be allowed to own low MP digitals. Here’s the thing: most humans who pick up a camera and point it at things can’t frame a shot to save their lives. Don’t blame them, they’re just trying to snap pics of things that are interesting, but every one of them gets home and looks at their photos and is disappointed at how small the leaning tower of Pisa is, or how the great shot of their niece at the beach is ruined by that guy peeing in the background. The reason more pixels is good is so that you can decide which 3.2 million of them to use for your print. This isn’t hard any more. With the advent of things like picasa everyone can do easy cropping jobs to get in closer, remove distractions, and make the picture match the experience it’s supposed to be capturing. Remember how that is what photography is about?

A couple examples might help, but understand that the purpose here is to demonstrate the benefit gained by focusing on a portion of the original, not to exhibit my photography prowess, to which we’ve already established I make no claim. If you draw from this series a general principle that shots get better when you get closer to your subjects, then you are right.

Don & Marg - Before
Don & Marg - After

Church - Before
Church - After

And in a particularly extreme case of cropping in to get at a detail
Plane - Before
Plane - After
Hello Mr. Guy in Seat 1A!

The takeaway here is not that you are stupid to buy a camera you can afford, and must instead buy some monstrosity. High resolution has its own problems, not the least of which is the very practical matter of being able to fit fewer images on a card. And if you buy a camera more expensive than you can tolerate placing in harm’s way, you will leave it at home and take exactly NO high resolution pictures of anything. All of that would be much worse than just shooting with a low res camera. All I’m asking is that you not deliberately go cheap in service to this notion that a 5MP or 7MP camera is somehow frivolous conspicuous consumption. I came by mine honestly enough, someone broke into my house and stole my old one – the Canon S70 is a delightful little 7MP gem, but I didn’t pick it, my insurance company did. Even if you did buy it with your own money though, I promise I wouldn’t disapprove.

[Credit for the linked photos at the beginning of this rant is gladly and gratefully given to fantasygoat, though he is likely not the photographer for any of them.]

23 thoughts on “The Megapixel Myth

  1. Interesting read, and hopefully I (an amateur) got some interesting to add:

    My father is a pro and as such owns a Hasselblad H2D (22 MP). I discussed the pixel story with him and he had added to the above that 8 MP is about the best ratio for the chip size that are in most mid range Digital SLR’s which have a 22.2×14.8mm chip. The new full sized 36 x 24mm could benifit from more MP. The 22 MP camera’s of Hasselblad are only more interesting because the chip size is also bigger (36.7 x 49.0 mm).

    The recently introduced Hasselblad with 39 MP is pretty much b-s because the chip size did not increase. Why is this bs? Apparently (and again I’m an amateur) a single light been is larger then the pixels on the chip. this means that adding more pixels to a same-sized chip does not actually increase the image resolution or anything.

    However, If this stuff is of interest to you I suggest you take a look at the multishot technology of Hasselblad. That stuff will blow your mind!

    My ideal camera would be a Canon amateur (nice and light) D-SLR with a 36 x 24mm chip (so it can handle all my old lenses) with 16 MP. thats more then I’ll ever need.

  2. You make some good points, but the megapixel myth was really along the lines that more pixels != BETTER pixels. This was about the time that low end consumer cameras starting cramming in 5-6 MP sensors, when the existing SLRs (Nikon D1, Canon D30 etc) were still at 6mp or so.

    This has leveled off to a degree, but there will always be a threshold of quality that a cheapo pont and shoot can’t traverse relative to a good SLR, and regardless of the MP count, a shooter who wants better pixels will pay for it, regardless if a cheap camera has the same pixel count.

    Much like a BMW might have the same horsepower as a Dodge, but the experience is much different.

  3. This is a similar problem to the one that computer sound card people have had for years. The Soundblaster Pro reached the human hearing limit sometime in the mid 90’s yet every year or two since, Creative have come up with a higher quality sound card promising sound like you have never heard. Well they all now offer sound you cannot hear (meanwhile your dog is going nuts!) because not many of us know we can’t hear it and it looks good in store to have a bigger better product. It might be true that we only need a 3.2MP camera, but what would they have sold us for the last few years if we all knew that.

    So you can’t see all the deatail on your photos, you can’t hear all the sound from your computer, you have to sit no more than 5 feet from an HDTV to get the benefit. This isn’t going to stop. I’m just interested in how far big companies manage to take it.

  4. I never thought about the “better equipment” angle before. The cropping argument has always been obvious to me, though sometimes I do like to print really large photos (16×20, etc.).

  5. I’m having trouble with your argument that more pixels on the same camera is always better. There are concrete cases in which this is wrong.

    If two cameras have an identically-sized imaging chip and different pixel counts, the area on chip dedicated to a single pixel will be smaller for the higher-megapixel camera. This means less photons available per pixel. Since you want similar light sensitivity in both cameras, you’ll end up amplifying the signal from the higher-megapixel camera more than you do for the lower-megapixel camera. This amplification is the major source of camera noise.

    Just look around for Panasonic camera owners. Pana has 3 cameras with very similar features and the same chip size, but one is 4MP, one 5MP and one 8MP (FZ15, FZ20, FZ30 if you’re curious). The higher the pixel-count, the worse the images come out in low-light conditions. If you’re doing indoor shots, you’re for all reasons better off with the 4MP model than with its bigger brothers.

  6. You have completely missed the point of most arguments that say a consumer shouldn’t bother with a camera over 5MP or so (And by the way, your “ceiling” of 3MP is incredibly outdated. There are very few people nowadays who would say that 5MP is not better than 3MP).

    Anyway, back to what I was saying. The main argument for not going above 5MP on a consumer level camera is that you tend to see a dramatic rise in image noise because consumer-level cameras have small image sensors. The sensors they use nowadays are the same size as when the average consumer camera was 2MP. As manufacturers keep cramming in more pixels in the same sensor size, the pixel density keeps increasing and noise increases, as Adolf said. Basically, you reach a point of diminishing returns after 5-6MP.

    I have seen very few people make the argument that anyone buying more than 3MP is a “corporate whore”. This is an extreme argument. More often, you hear it said that “3MP is enough to print an 8×10 photo, so the typical consumer shouldn’t need more than this”, but this argument is aimed more at setting a guideline for Joe Average to spend his money wisely.

    However, the digital camera manufacturers keep stuffing in more pixels only because it is very easy to market a camera with higher pixel counts and because they need to be seen to be always “advancing”. This trend only ends up hurting consumers who end up paying more money for a camera that really doesn’t take better pictures with all those extra pixels, and this is what most people are saying nowadays.

  7. The internet has such a wonderful way of making everyone sound like a prick, doesn’t it? This is me giving you the benefit of the doubt.

    I am, of course, quite aware of the arguments about the relative shortcomings of manufacturer focus on increasing MP count through higher sensor density. Let’s talk about the fact that by keeping sensors small and not investing in quality high speed glass, manufacturers are often giving you pixels beyond the resolving power of their lenses, so that your 12MP are all blurry megapixels. Let’s talk about the fact that mechanical aspects of the camera like flash placement and usability issues like menu design are often bad enough that you can’t take the pictures you want or enjoy the process even when the tech doesn’t get in the way. Perhaps the root of your criticism here is that I didn’t use this post to digress on to a tangent about the entire mechanic and technique of point and shoot digital photography?

    If I have given the impression that I think you can buy yourself perfection in a $120 5MP (interpolated) Panasony with a recycled-pop-bottle lens, then I am sorry. My goal was to tackle a specific consumer myth which, unlike you, I have encountered more than occasionally. And at the root of it is, I feel, an interesting insight about the perception of photography and how it is changing in the digital age, i.e. that cropping is no longer the secret of the professional, but a democratized, trivial way to make better pictures, and something to consider in camera purchases. I do of course appreciate your taking the time to make sure I was aware of these issues, tone notwithstanding.

  8. Good article 🙂 I don’t often read all the comments on blogs, but your final reply made it worthwhile 😉

    Keep up the good work!

  9. well its not just mega pixels, its also the size and quality of the sensor. more light per sensor site the more dynamic range or whatnot. less noise to actual detail. many cheap cams can only really get their full pixel count with massive amounts of light, indoors or dim settings and you get a load of noise.

    and not to mention the problem of digital cameras sorta fibbing on the pixel counts to begin with. because of the bayer filter and the way that works. your actual resolution is lower than just the pixel count. there was a 3 layer sensor being developed but that is mired in allsortsof trouble so its a non issue, to get real 7mp of actual detail you need more than 7mp.

    and yes u need a good lense.

  10. 3 or 4 ‘pixels’ (R or G or B) == 1 full colour ‘spot’. I think so anyways.

    Explanation: You usually have a pattern of color dots on the sensor like R-G-B or R-G-B-G (yes two greens), repeated over and over. Each triplet or quadruplet forms a full colour ‘spot’, but each individual single-colour ‘pixel’ is counted as one of the megapixels. Each and every single-colour ‘pixel’ does contribute a B/W (luminance) value (very crisp for pure B/W), but there is 1/3 or 1/4 as much colour information. So a 10mp sensor really only has 2.5mp’ish of colour resolution.

    If you remember Foveon (where did they go?), they actually stacked the colour sensors so that EACH PIXEL had R+B+G. Was very good at the time. Other companies have tried alternative grids&distributions like hexagonals with different total RGB counts.

    At least there is a mitigating factor: colour data usually varies at a lower frequency than luminance data in the real world (and always in our eyes). So we see the ‘crispness’ in edges based on the birightness, and fake the colour detail. And I shouldn’t forget the ‘Cleartype’ effect, which certainly applies to sensors.

    So more MP’s can be a help with better colour detail, as long as the higher resolution is properly balanced with size, quality, etc..


  11. Doesn’t matter how many megapixels you have in a camera, if you’re a crap photographer, ALL your photos are gonna come out looking like shiat! 😉

    Soo many people think they are top-notch photographers these days (care of the ability to delete crap digital photo’s on-the-fly) but if you count the number of crap photo’s that were removed from the camera compared to the number of good photo’s, the ratio will define that persons ability (or lack thereof). If you took 4000 photo’s, you can be safe to guess that there should be at least 2 good shots in there (somewhere!)

  12. The cropping secret is an incredibly handy tool/skill. Many pro snappers who sell there artworks in galleries are known for cropping an image and then blowing it up… then selling it for big bucks.

    Like in amateur radio, the noise floor or sensitivity of electronics is what really governs the quality judgement criteria. Megapixel myth was foisted on to the consumer just like Intel gave MHz speed only marketing when years later it had to reverse this trend due to technology barriers.

    When consumer camera’s mature in years to come, the count of megapixels won’t be made much of but, the quality of them. Just like AMD can sell superior CPU’s than Intel with lower clock ratings…

  13. Amen to this!
    Check out my website. I have a standard 3.2 megapixel cam and it works like a charm.
    I’ll be getting a Canon eos 30d very soon but for now i’m MORE than happy with my Canon Powershot A70

  14. Pingback: deer stag
  15. I’m very happy with my Fuji FinePix A202, a 2MP camera. It was a good purchase in terms of bang for the buck. I always wondered why other people’s higher MP cameras were turning out soft/blurry shots, even with the camera held still, but my Fuji turns out sharp, crisp images with good colour and I haven’t seen any purple flange or whatever it’s called as I’ve seen with the other brands. A lot of my friends use low end, but high MP! laugh, brands like Sony “Cybershot” and Kodak. I can’t stand to look at those pictures as the colour is way off. Kodak accentuates the reds (hue is was off) while Sony is just too soft. My friends brag about the MP and think that’s everything there is to a quality camera. Some laugh at the fact I have a 2MP camera…a lowly camera to them, but let me share some of my images with you sometime. See for yourself before you buy into the “high megapixels = better” myth. Mostly high MP only means bigger size prints, NOT better quality.

  16. The human eyes see between 72 and 35o dots per square inch, this can be shown as pixels… so, when the eyes see 200 pixels per inch, we see a precise image (taken, for example by 8Megapixel camera).
    The difference you will see will not be in the number of pixels (against, say than that from a 3Megapixel camera) your eyes DO see a sharper portayal of the image with higher range SLR’s.
    This is where a higher resolution camera is better, but also bear in mind that the eyes will also see more physical artifacts from things like dust in the air etc… some particles in the shooting range are usually too small for the human eye to see. This is where the lens comes into use as a magnifier.

    a good photographer will usually add an unsharp mask to a digital image to enhance artifacts more; resulting in more work in the photo software. Unless of course, the studio is a vaccum space with no living thing in the air.

    I reckon technology and the progression of our evolution depends on us being able to see more.


  17. an example of an image cropped after resizing in 5% increments on a canon 350d

    i’d love to see this done on the bigH 😀

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s