Digital camera buying tip from an engineer


Last week my digital camera (Canon A610) died, after only two years of light service. (It turns out that a large batch of cameras made by Canon late 2005 had a bad CCD connector which tends to die after a year or two.) While I was obviously frustrated, part of me was also secretly happy since it meant I could have the fun of shopping for another camera.

After a fair amount of research (but not enough, as it turns out) I ended up buying an 8 million pixel Olympus camera (SP-560UZ) to replace the 5 million dead ones on my old camera. Progress, right? Not really. I was surprised to find that the 8 MP camera, made over two years after my dead camera, produced images of lesser quality. Sure, they were higher resolution, but that was about it. How is that possible?

I looked a little more into the sensor elements used (called CCDs, for charge coupled devices) and it seems that the digital camera companies generally increase pixel counts without actually making the sensors any bigger. In fact, the CCD on the Olympus is actually about half the area of the Canon! [Note: as pointed out by Leonid, below, they don’t have to much leeway to do otherwise in the megazoom cameras.] So the pixels get much smaller, and since each pixel requires a certain amount of circuitry (that can’t shrink any further) the sensor actually becomes even less sensitive as a whole. Furthermore, the amount of noise experienced by each pixel doesn’t shrink as quickly as the pixel size (for reasons that are a bit complicated), so an 8 MP sensor experiences significantly more noise than a 5 MP sensor. They make up for this with somewhat better CCD technology and clever image processing, but there is only so much that can be done.

Another, perhaps even worse, result of smaller pixel size is that the maximum number of photoelectrons that can be stored in each pixel is lowered. A pixel on a CCD acts like a bucket for electrons. (It’s sometimes incorrectly stated that CCD pixels store photons.) A photon of light hitting the CCD pixel has a certain probability of causing an electron to be “freed” from the silicon and dropped in the bucket. While this is just a metaphor, electrons do actually follow a lot of the same rules as water filling a bucket do. Once the bucket is full, the electrons spill out, often into a neighboring pixel that isn’t yet full. Furthermore, if the bucket (pixel) is shrunk, it can’t hold as many electrons.

The increased noise plus the smaller capacity to hold electrons means that each pixel can’t handle a very large difference between light and dark in a scene. In other words, as you increase the exposure, the pixels “fill up” much quicker than if the pixels were larger. The ability to measure large variatation between light and dark in a scene is called dynamic range. A lack of dynamic range shows up as washed out highlights and lack of detail in shadows. An example of this is shown in the following zoom from a picture taken with the Olympus:

Washed out sky

It looks like a cloudy day, but this picture was actually taken at 3 pm on a nice, sunny day. The blue sky was not bright at all to the eye, but it was enough to cause the pixels which saw the sky to all max out and overflow. The fact that they maxed out is indicated by the pure white that resulted, and the overflow of photoelectrons into neighboring pixels is evidenced by the “bleeding” of the white into the tree branches. Admittedly, this is not strictly proof of anything, as I would have to provide a picture of the same scene taken with a better camera for you to be able to truly verify my claims, so you’re just going to have to trust me that this day wasn’t. Here is the full picture, to show that the rest of the picture was not overexposed:

The Olympus is a great camera in most every way, but it appears that they have pushed the pixel count so high that picture quality has suffered. All manufacturers of cameras in this class appear to do the same thing, and not one of them seems to have the guts to say “enough.” From an engineering standpoint, this pixel race makes absolutely no sense. At some point, adding pixels is counterproductive and actually lowers the effective resolution for most situations, due to the effects of increased noise and loss of detail due to dynamic range reduction. In my opinion, this point was reached on the small sensors used in compact point-and-shoot cameras at about 4-5 MP. Unfortunately, engineers don’t run companies, marketing types do. And the marketing lemmings invariably decide that putting a sticker that says “8 MP” on the side of the camera is more important that the quality of the images it produces. Most likely, they don’t even understand the trade-offs involved in doing so, and only hear the first three words when the engineering manager says “Yes, we can do that, but…”

People buy into this because they, understandably, assume that companies couldn’t possibly be so crass and cynical as to intentially fool people into paying more for an inferior product. There was probably a time when that was a fair assumption, but those days are long gone in the Persian bazaar that is the consumer electronics industry.

The counterintuitive upshot is that you can actually get a better quality image from a $200 low-end camera than from the higher-end $400 model from the same brand. If you’re buying a digital camera, consider intentionally buying a 5 or 6 MP model (if you can still find them) even if you can afford the 8 or 10 MP version. Check the specifications and buy the one with the largest CCD you can find. This means avoiding the cute pocket cameras, if you care at all about image quality. If you need to print poster-sized enlargments that require more than 5 MP, you just need to bite the bullet and splurge for a digital SLR; they use much larger sensors that operate on a fundamentally different read-out principle, and as a result they can produce 10 MP images with incredibly low noise. The idea of a small consumer level point-and-shoot camera with 8 MP is a bit crazy, if you ask me, and a terrible engineering choice.

The net result of all my research is the realization that I had a great thing in that little Canon A610, which makes its loss even worse. It was one of the last models where Canon used a relatively large 1/1.8 inch CCD, and to my eye it struck the right balance between resolution and image quality. After all this, I’m just going to try to find a used A610.


26 responses to “Digital camera buying tip from an engineer”

  1. See, you think that no one will be interested, but this is fascinating and useful. I never would have known, and I would have thought something along the lines of “what is my freaking problem, thinking the cheap camera pictures were better?”
    But I’m curious what you did – keep the 8, or take it back. I suspect they’d hit you with a restocking fee.
    This is a very sleek comment screen, with the black and gray background and the light type. I quite enjoy it.

  2. Mere: I am going to return the 8 MP camera, and will likely replace it with a $150 Canon compact. For a while I’ll have both, and will post about the comparison between the two.

  3. > I looked into the sensor elements used
    > (called CCDs, for charge coupled devices)
    > and it seems that to save money,
    > the digital camera companies generally
    > increase pixel counts without actually
    > making the sensors any bigger.

    To save _your_ money, dude. 🙂 If the size
    of the sensor is increased, then the size of
    lenses and its focal length has to be
    increased. Can you imagine what would be
    lens size, weight, and cost of a camera
    with 27-500mm focus length equivalent with
    a sensor, say, twice bigger?

    > The counterintuitive upshot is that you can
    > actually get a better quality image from
    > a $200 low-end camera than from the
    > higher-end $400 model from the same brand.

    Nothing counterintuitive. If you are not going
    to use zoom more than 3x, stay away from
    ultra-zooms. You pay extra $$ not for image
    quality, but for ability to have a pocket
    telephoto. If you are going to use zoom
    in the range 5-18x, you have to be prepared
    to have somewhat worse image quality. How much
    worse? Can you tollerate it? How important for
    you to make shots at high zoom? This is
    the trade-off worth considering.

  4. Leonid: Your point is very well taken that the decision to have 18x zoom makes it impossible to use a large CCD. It’s not so much that I’d like them to increase the CCD size, as that I’d like them to quit with the pixel arms race. I think it’s ludicrous for a consumer level camera to have 8 million pixels, and yet even the low end cameras these days have over 7 MP. It’s a fundamentally flawed idea to compromise image quality for zoom length, I think.

  5. I agree that for usual picture (even to view on computer), you don’t need a very high mega pixel camera and that is why I have Fujifilm Finepix F20 with 6 Mega pixel and it gives excellent results.

    But at the same time for photography enthusiast who needs wide angle and more optical zoom, such expensive and super zoom cameras can solve their problem.

    The answer of your concerns about adding more pixels is that now the companies are also increasing the CCD sensor size from standard 1/2.5

    A similar kind of issue was raised some 5~6 years ago about increasing the hard disk capacity. People at that time also raise the same point that adding more sector/disk to increase the HDD capacity would result in more bad sectors since the size of HDD cylinder was kept same but the time showed that such problems were solved.

  6. Stewart:

    Sorry for the late reply. I’m afraid I’ll have to weasel a bit on this answer, since not every company uses the same CCD technology, and thus pixel size isn’t the only variable. (In fact, at the risk of this comment stream turning into a Fuji commercial–I swear I don’t have anything to do with them–I should mention that Fuji has developed a CCD which uses dual pixel types with different sensitivities, which overcomes a lot of the dynamic range limitations of small pixels. This may have filtered down into their lower-end cameras by now.)

    Having made that disclaimer, in my research on cameras prompted by my disappointement in the otherwise great Olympus, I estimate that a good guideline for the minimum ratio of CCD size to pixel count is
    $latex \frac{s_{\mathrm{CCD}}^2}{n_p} > 0.05 \,\,\mathrm{inches}^2/\mathrm{million}$,
    where $latex s_{\mathrm{CCD}}$ is the size of the CCD (as given in the technical specs) and $latex n_p$ is the number of pixels in millions. Note that it’s actually the ratio of the length squared to the pixel count (which is proportional to the area of a single pixel). For example, the Olympus I purchased has a ratio of about 0.016, which is probably why I’m not happy with it…

    By the way, thanks for providing me with a good excuse to use the LaTeX plugin!

  7. Hi — This is a really great post, and I like the comments too. I teach classes about digital photography to adults and will link to this post. So many folks have the mindset, as far as megapixels are concerned, that more is better — and the manufacturers have no reason to discourage this type of thinking.

    I have one comment about the photo you showed. It *is* a great example of pixel bleeding and a blown-out sky — but I think maybe it’s not totally fair to that particular camera. It actually is not showing the extremes of dynamic range (there are very few shadows/darks/blacks in the shot) — it’s only showing one end of the extreme, the highlights. I’m guessing that the camera was set to a matrix metering mode (ESP in Olympus cameras) and pointed at the roof. Had you used a center-weighted average, or spot metering, and pointed the camera at the sky, you would’ve gotten a very different result.(Although… if you used spot/cntr-weighted-avg metering and pointed the camera at the roof, you could’ve ended up with this shot as well.)

    I’m not in any way sticking up for this camera, and I really appreciate your engineer’s explanation of why cameras with small sensors/high megapixel counts have the issues they do. And my experience with the small-sensor Olympus cameras is certainly consistent with yours. I just think there are some photographic choices that were made with this shot that perhaps have exacerbated its dynamic range problems.

  8. Lisa:

    Thanks for writing, and sorry for the delay in this reply. The image I show is just a very small crop from a full resolution image. I agree that it doesn’t prove anything, since perhaps I overexposed the whole image. On your suggestion I have added the full frame shot from that day (reduced for size). Having said that, the question of whether or not one can tell dynamic range from a single image from a camera is actually a complicated one.

    In any scene will sufficient tonal range, the camera will basically try to map the darkest scene tone to black, and the lightest to some color with a nearly full saturation of at least one of the additive primaries. I’m being approximate here, but my point is that in a complex scene the camera will generally give you an image that pretty much uses the full range of the digital image format it’s using. The question of dynamic range is really asking what does the full range of the digital image correspond to in actual optical brightness in the original scene? Or put another way, how many bits of resolvable information is each pixel capable of measuring? (This is a function of both well depth and noise.) This can’t really be known without having a “master” picture of the scene that day taken with some calibrated imager of sufficient dynamic range. For example, you don’t really know without having been there that the sky was not an incredibly bright thin midday overcast, and not the late afternoon clear that it really was.

    So, I guess I’m not really proving anything scientifically with my single picture, even when I show the full frame. The only value, if any, is really qualitative and vague: I can tell you it was a late afternoon, and we all know from experience that a late afternoon shot on a cloudless day with a brick building should result in a sky that is not overexposed with a digital camera of any reasonable dynamic range.

    So, your comment is very appreciated, and brings up an interesting point. Does anybody know if there is a review site that actually scientifically measures the dynamic range of the digital cameras they test using calibrated equipment? This is such an important figure of merit for a camera, and yet I can’t find anybody who measures it.

  9. I also returned my 8MP camera since i think i can get just as good pictures with a 5MP camera and save some money. Thanks for a good article

  10. My wife is thinking about buying the newer Olympus SP-570UZ model to replace an older Olympus camera. When I began to look into that camera to compare it with the Nikon P80, I came across your Amazon review of the 560UZ, which led me here. Although I am familiar with the concept of CCD dynamic range and the limits on CCD full wells, I do not understand why the Olympus gave you such severe saturation problems in your tests. There are many other cameras on the market today with similar CCDs that should also be affected in the same way, and yet I’m not aware that the problem is widespread. For example, I own an 8MP Canon Powershot A720 IS, which has a 1/2.5″ type CCD. This detector is nearly identical in physical size and pixel dimensions to the one that’s in the Olympus 560 and 570 models, but I seldom run into the saturation problem you experienced. I cannot say I NEVER encounter the same problem because there are times when I find the sky is over exposed if I’ve intentionally tried to bring up the shadows. Generally speaking, however, in very brightly lit scenes in places like Tahiti or Hawaii the auto exposure of the Canon does an excellent job of keeping the exposure level under control and what little saturation I encounter almost always can be handled in post-processing. Given the similarities of the CCDs involved, perhaps something else (or in addition) is at work?

  11. That’s really interesting. I don’t know what Canon does to do so well with their detectors. Maybe their fill factor (the percentage of the pixel taken up by the actual photosensitive part is higher. They could also do tricks with the processing to make their pixels have a higher capacitance per unit area. They could also have circuitry that “leaks off” current so that that the detector has a nonlinear response that resists saturation. I don’t know how much variability in architecture there is between makers. I assumed they all got their CCDs from the same group of companies, but maybe I’m wrong. I’d find it hard to believe that the improved performance of Canon cameras comes from better software, alone.

  12. This is very informative, thanks.

    If you are still looking for a camera, consider the A570IS. The image stabilization is extremely useful in low-light circumstances. It lets you use a lower ISO rating more often, thus avoiding the worst cause of noise in small-sensor digicams.

  13. I just came across this while searching for something completely different. LOL…. you explain very well something that I’ve been trying to for a while. It seems like the megapixel race is like the faster CPU race was a while back. Just because we can get more does not necessarily mean we should.

    I’m with you… what good does increased MP do if when you view the image at 100% it’s all noisy and the focus looks soft because of that? In reality, most people don’t print larger than 4×6 on average. Thats 1200×1800 pixels when printed at 300 PPI (what is considered “print quality, though you can get great prints at much lower PPI settings). That’s 2.16 megapixels folks.

    Most of what makes today’s cameras “better” is the image processing and much improved lag times.. not more megapixels.

Leave a Reply to photo_chick Cancel reply

Your email address will not be published.