Finally, reason is about to prevail over marketing machismo. Specifically, Canon and Sony are coming up with more advanced cameras featuring less pixels. Why? In these new cameras, less pixels translates into better pictures in low light. (You might want to refer back to two Monday Notes on digital photography: Pixels Size vs. Number and More on Sensors Digital Photography.)
So far, the selling argument has been more pixels equals better pictures. A higher number conveys an image, so to speak, of higher quality. This is not entirely untrue: it’s nice to have lots of pixels when you need to “crop”, to throw out a large fraction of the original image in order to concentrate on a key detail, a face for example. If you have enough pixels left in the “crop”, it will print or display with good detail. But there is an important downside: for a given sensor size, more pixels means smaller pixels. In turn, this means each pixels will receive less light energy, less photons to be converted into electrons. The smaller number of electrons will have to “fight” against the background electrical noise in the sensor. The lower signal-to-noise ratio means lower quality pictures. This is particularly true in low-light situations where, to begin with, the number of incoming photons is smaller.
Go to the excellent Digital Photography Review site and try one of the Camera Database pages such as the one for Nikon cameras. There, for each camera, you’ll see a very useful number: pixel density. It varies from 42 MP/cm2 (megapixels per square centimer) for a low end camera, to 1.4 MP/cm2 for the top-of-the-line D3, both for the same pixel count: 12 million. (You’ll find similar results for Canon cameras.) The ratio is more than 30 to 1 between the low-end (compact) machine and the big DSLR (reflex) camera. As a result of its fatter pixels, the higher-end camera doesn’t need a flash to get remarkably good pictures at night, on the street or in “social” context, the latter meaning people at dinner parties.
But for compact cameras, ‘social photography’ equals flash. Go to Flickr or Facebook and see how many pictures involve a social event, a party, a dinner, and how many of those are taken at night, using the camera’s flash. With less than flattering results. The faces closest to the camera are often flattened, overexposed, the rest of the scene is in the dark because the camera’s flash is too weak to illuminate the breadth and the depth of the recorded scene. If you don’t use the flash, the small, high-density sensor produces noisy, muddy images. If you use the flash, you’re likely to get partially illuminated and unflattering pictures. The “obvious” solutions involve using higher-end cameras with or without their sophisticated flash systems. This isn’t practical for most of us: these cameras are big and expensive. (Paradoxically, the latest big anvils are extremely easy to use, often easier than “consumer” compacts. You get very good pictures in “Idiot Mode”, just point and shoot, trust this klutz.)
Enter the Sony WX1 and Canon S90 “less is more” cameras. These are two new compact cameras, good for social photography, with less pixels for better low-light performance. Using the Canon camera as an example, previous models sported 12 megapixels sensors, the new one offers “only” 10 megapixels. Using a slightly larger sensor than before, pixel density is now 23 MP/cm2 vs. 43 MP/cm2 in models such as the SX200 IS. The result is, we’re told, usable images at 3200 ISO. Coupled with a more “open” lens, one that takes in more light (for geeks: max. aperture is F2.0), we now have a camera that will take really good dinner party pictures without flash.
Sony’s WX1 uses what they call a new “Exmor R” sensor to produce similar results in an equally compact camera, with a wider (24mm min. focal lens) but less open (max F2.4) lens and a clever panorama capture mode.
These cameras are said to be available in October for less than $400.
If we know anything about the digital camera industry, it’s that it’s a fiercely competitive one. We can be sure to see Panasonic/Leica, Casio/Pentax, Samsung and others follow suit with similar “less is more” compact cameras. This might take a little while: Sony and Canon make their own sensors, I don’t believe the other manufacturers do or that they have the advanced capabilities the two leaders enjoy. (Interestingly, Sony supplies the sensor for the Nikon D300/D300S models, see the dpreview.com camera database. When, for a given camera, the database says “unknown” for the sensor manufacturer, you can be sure it’s not the camera brand.) — JLG
Bonus feature for geeks:
Automatic twilight flash photography
The problem: You take pictures of your friends, in a street, at dusk. Your camera’s flash “correctly” illuminates the group but the stores in the background, a few meters away, don’t get enough flash light. As a result, the background looks unnaturally dark, the picture doesn’t reproduce what your eyes (or, rather, your brains) saw.
The solution: Modern cameras use one number, one pre-flash exposure and on-board software.
Let’s say the distance between the camera and your friends is two meters. The background lies four meters further away, for a total distance of six meters. Physics say illumination falls with the square of the distance. So, the background being 3 times (6 meters) further away the your friends (2 meters), if your group is correctly exposed, the background will get 9 times (3 times squared) less light. That is why, with old-style flash exposure the background looks bad. Or why using a compact camera’s flash on a distant scene, such a sport stadium or a concert, is totally useless.
Now, on higher-end cameras, the on-board computer knows how far your friends are, it has access the lens’ focus distance. Then, when you press the shutter release, the camera fires a pre-flash before the actual flash. It most cases, the whole sequence is so quick your won’t see it. The computer “looks” at the scene and the software detects a bright foreground, your friends, and a dark background. This uneven illumination wasn’t present when the software was looking at the scene’s light values before the pre-flash: your friends and the background, without flash, showed approximately equal light values, low, perhaps, but comparable. Having compared the light values in the scene with and without the pre-flash, the next step is for the software to try and balance the background and the foreground exposure.
The source of light for the background can’t be changed, it’s ambient light. The source for the foreground can’t be modified: you need this much light at this distance for correctly illuminating your friends.
The camera’s computer opens the shutter for a long enough time to let enough non-flash (ambient) light produce a correctly exposed background. Sometime during that exposure, the computer fires the flash. The result combines a pleasantly exposed background and a correctly illuminated group of friends. (This computerized balancing of background and foreground exposure also works in the reverse situation, a “contre-jour”, a dark subject against a brightly illuminated background, such as a face against a bright sky.)
There are limits, of course. In practical terms, this works better when the pre-flash doesn’t produce too large a difference between foreground and background illumination. You can’t keep the shutter open forever just so you capture enough background light, the picture would be blurred. That’s why I refer to twilight or dusk. But the results can be surprisingly pleasant, especially because this works in Idiot Mode. Unless the user screws up by trying to set up extra modes. Trust this klutz. Fortunately, factory spies have caught on and many cameras now have a “Restore Factory Settings” function.
You can get such features on DSLR (reflex) cameras starting at $400 (with a basic lens). Starting at $500-$600, you’ll have many choices.
Why not on compact cameras? The flash, the battery and the on-board software too limited, so far. —
- The End Of Megapixel Wars – Part II – The Canon S90 TweetLast August, I wrote about picture quality finally winning against macho marketing. In other words, it seemed Canon, Nikon and Sony were giving up the simplistic escalation: my camera has more pixels than yours, therefore it is better. In the P&S (Point & Shoot) category especially, the facts were that more pixels ended up producing [...]...
- Pixels: Size vs. Number TweetOMG, says the blogger, the next iPhone’s camera will have 3.2 million pixels instead of today’s measly 2 million! The blog entry gave me the final push for an occasional, meaning at irregular intervals, series of columns on digital photography. The idea is to find insights into what’s really going on in this very dynamic [...]...
- App Cameras TweetIn an August, 2010 Monday Note titled Smartcameras In Our Future?, I wished for smartphone-like apps running on a nice compact camera such as Canon’s S90 (now replaced by the S100). At the time, in-camera photo processing was limited and wireless connectivity required accessories like Eye-Fi, a clever but not so easy-to-use SD card with [...]...
- Crazy Patent Wars TweetEvery week we see more companies filing more lawsuits over patents. Microsoft and Oracle sue Google, Apple sues a long list of companies…and they all countersue in a new kind of circular firing squad, a real MAD (Mutually Assured Destruction) conflagration. Soon, high-tech companies will have more lawyers than engineers. First, a word on the [...]...
- Smartcameras in our future? TweetI have two cameras in front of me: My smartphone and a Canon’s S90. And I wonder: Why isn’t there an app store for this neat compact camera? I can download any number of third-party, post-processing photo applications to my smartphone. I can crop, filter, stitch, frame… And there will be more applications tomorrow. With [...]...