For digital cameras, quality is about more than just megapixels. Digital sensors are complex devices, but camera manufacturers learned long ago that camera buyers equate more pixels with better quality. To some extent, that's true, but mostly it's not.
In the early days, cameras produced images that were 640 pixels by 480 pixels. A camera with more pixels (1024 by 768, for example) would certainly be better. But once cameras advanced beyond about 3000 pixels by 2400 pixels (7.2 megapixels) the actual pixel count became irrelevant.
Why? 3000 by 2400 is sufficient to produce a 10 by 8-inch print at 300 dots per inch. Most people share images on the screen these days and that's also sufficient for a rich, textured, detailed image on even a large screen. Even basic low-end point-and-shoot cameras today offer at least 10 megapixels and some cameras that are intended for pros are in the 50 megapixel range. These cameras provide sufficient visual information to allow for greater enlargement and cropping.
Now a sensor's dynamic range is arguably more important than the number of pixels the sensor has, but that's a topic for another time.
While we hear a lot about needing 300 dots per inch for acceptable photographs, that isn't always the case. An acceptable wall-size enlargement (36" by 24") can be made from a digital image that's 2560 pixels by 1920 pixels. If you do the math, you'll see that this is the equivalent of just 71 dots per inch on paper. The print is acceptable because a wall-size print will be viewed on a wall. Closer viewing distances require higher resolution printing. To take this to absurd extremes, consider highway billboards. These are viewed from a distance of at least 50 feet and the resolution is usually 12 dots per inch or less.
Just about any digital camera sold today will have more than enough pixels, but there are other things to be concerned with. One of those is the physical size of the sensor.
Click any of the smaller images for a full-size view. To dismiss the larger image, press ESC or tap outside the image.
Sensors come in a variety of sizes. The largest measure about 50mm by 40mm and the smallest measure about 4.5mm by 3.5mm. It's possible to buy a 20 megapixel camera with a 5mm by 4mm sensor, a 36mm by 24mm sensor, or a 50mm by 40mm sensor. Same number of pixels, but a different sensor size. So what's the difference? The obvious answer is the the pixels (they're called "photo sites") on the smaller sensor will be smaller than those on the larger sensor.
What difference does that make? Understanding this requires some knowledge of the incredibly complex technology used. A pixel actually consists of 4 photo sites: 1 for red, 1 for blue, and 2 for green. Our eyes are more sensitive to green and that's why most of today's sensors use this pattern.
Without getting too far into the weeds, let's just say that smaller sensors with the same number of pixels as a larger sensor won't produce as good an image. When photo sites are smaller, noise increases because not as much light reaches the sensor. So that suggests that larger sensors are always better than smaller sensors.
Not so fast, though. The two most common sensor sizes for digital SLR cameras are called "full frame" and "cropped sensor". Full frame sensors are the size of 35mm film, 36mm by 24mm, and cropped ("APS-C") sensors are either 24mm by 16mm (Nikon, Sony, and Pentax) or 22mm by 15mm (Canon). When choosing between full frame and APS-C, larger isn't always better. The smaller APS-C sensors will have more digital noise than the full frame sensors, but the difference will be so small that it will be visible only in laboratory tests.
Too many amateur photographers buy cropped sensor cameras because they generally cost less and look forward to the day when they can buy a full frame sensor. The results are not always what the buyer expected. First, any lenses that have been designed for a cropped sensor camera won't work properly on a full frame camera. It's just physics, not some nefarious plot by manufacturers. But more important is the fact that a full sensor has a different view.
I own a Canon digital SLR with a cropped sensor. Canon's APS-C sensor means that a 50mm lens will act like an 80mm lens on a film or full frame digital SLR. Sometimes this is referred to as a "magnification factor", but that's not what it is. The cropped sensor sees only part of the light that the lens captures. That's great if you're a photographer with a telephoto eye (that's me), but not so good if your vision tends more to wide angle views.
The Sigma 150mm-600mm zoom lens I use at zoos on an APS-C camera gives me the equivalent of 240mm-960mm. That's a plus for the times when I would want to use a longer lens. On the other hand, I have a Sigma 10-20mm lens that would be extremely wide on a film camera or a full-frame digital camera. On my cropped-sensor camera, it's the equivalent of 16mm-32mm. My point here is that the sensor needs to be matched to your eye.
Maybe you're wondering what I mean by "normal". Well, it's subjective at best. A "normal" lens is one that reproduces a field of view that appears "natural" to a human observer -- neither wide angle nor telephoto. Wikipedia has a long article on the subject.
| Sensor Type | Width (mm) | Height (mm) | Area (mm²) | Crop Factor | ||
| 1/3.2" (Apple) | 4.54 | 3.42 | 15.5 | 7.61 | The entry for 35mm film (digital full-frame) is shown in red because it is the primary reference point for lenses that are used on digital cameras. Manufacturers often describe lenses as "35mm equivalent". That's because 35mm equivalents are familiar, not because 35mm (or digital full frame) is better. The other lines in bold are the most common sizes for digital SLR and mirrorless cameras. | |
| 1/3.0" (Apple) | 4.80 | 3.60 | 17.30 | 7.21 | ||
| 1/2.6" (Samsung) | 5.50 | 4.10 | 22.55 | 6.31 | ||
| 1/2.5" | 5.76 | 4.29 | 24.7 | 6.02 | ||
| 1/2.3" Type (Canon, Olympus) | 6.17 | 4.56 | 28.07 | 5.64 | ||
| 1/1.7" (Canon) | 7.44 | 5.58 | 41.51 | 4.65 | ||
| 1/1.7" (Pentax) | 7.60 | 5.70 | 43.30 | 4.55 | ||
| 2/3" (Nokia, Fujifilm) | 8.80 | 6.60 | 58.10 | 3.93 | ||
| 1" Type (Sony, Nikon CX, Panasonic) | 13.20 | 8.80 | 116 | 2.73 | ||
| Micro Four Thirds | 17.30 | 13.00 | 225 | 2.00 | ||
| APS-C (Canon) | 22.20 | 14.80 | 329 | 1.62 | ||
| APS-C (Nikon, Sony, Pentax) | 24 | 16 | 284 | 1.50 | ||
| 35mm (digital full-frame) | 36 | 24 | 864 | 1.00 | ||
| Kodak Medium Format | 49 | 37 | 1803 | 0.70 | ||
| Hasselblad Digital Format | 54 | 40.30 | 2159 | 0.64 | ||
| Medium format (film) | 60 | 60 | 327,184 | 0.05 | ||
| Sheet Film (4"x5") | 102 | 127 | 1,290,320 | 0.03 |
Once (long ago) press photographers used 4" by 5" sheet film cameras. A "normal" lens on a camera like this would be about 150mm. Then smaller cameras came along: Roll film cameras were available in several sizes (7cm by 6cm, 6cm by 6cm, and 6cm by 4.5cm) and the "normal" lens was around 80mm. For 35mm full frame film cameras the "normal" lens was 50mm. For today's APS-C digital SLR cameras, "normal" is about 36mm.
You don't have much choice about sensor sizes in phones, but if you're in the market for a digital camera, you have 5 primary choices: 1", Micro Four Thirds, APS-C (two slightly different options), and full frame.
If you've examined the chart, you may be wondering about the sensor sizes shown as 1/(something). These goofy measurements date back to old TV tube sizes that refer to vacuum tubes that were used in early television cameras. The more important numbers are the height, width, and area values that show the actual size of the sensor.
It's clear that a camera with larger sensor area will produce a better image because it gathers more light, but it will require larger, heavier lenses. Today, even smart phones and tablets can provide sufficient detail. An Ipad Pro comes with 2 cameras: The back camera is a 12 megapixel device and the front camera has 7 megapixels, but instead of a sensor with an 860 square millimeter area, it just 16 square millimeters.
So although theoretically a larger sensor is a better sensor, that doesn't always hold up in practice. For those who need to take photographs in dark locations that require extremely high ISO ratings, the larger sensor will provide images with less digital noise.
There are no one-size-fits-all solutions in photography. You won't have a choice of sensor size in tablets or smart phones, but advanced point-and-shoot camera, mirrorless cameras, and digital SLR cameras provide opportunities to choose among the various trade-offs: If you want a small camera, most will have small sensors. Some advanced point-and-shoot cameras have APS-C sensors and a tiny few even have full-frame sensors. If you want the least possible noise, a full-frame sensor might be the right choice, but at a much higher price and the need to buy more expensive lenses. If you prefer telephoto photography, you'll get a bonus from cropped sensor cameras.
Quite a few recent posts have dealt with digital photography. It's one of my favorite subjects and I've heard from some listeners and readers that they enjoy these reports because what used to be a highly specialized pursuit is now accessible to anyone with even a modest camera and a computer. But is it too much? Opinions, suggestions, and recommendations are always welcome and that's why the website has a contact page.
You've probably already seen recommendations to reboot your router and to update it to the most recent firmware to eliminate a security issue. More than a dozen routers are known to be vulnerable, but there are questions about whether the list includes all routers with vulnerabilities. As a result, the safe course would be to reboot your router whether its on the following list or not:
Linksys E1200; Linksys E2500; Linksys WRVS4400N; Mikrotik RouterOS for Cloud Core Routers: Versions 1016, 1036, and 1072; Netgear DGN2200; Netgear R6400; Netgear R7000; Netgear R8000; Netgear WNR1000; Netgear WNR2000; QNAP TS251; QNAP TS439 Pro; Other QNAP NAS devices running QTS software; and TP-Link R600VPN.
Updating the firmware is also a good idea, but might not always be feasible. I have a Netgear R7000 router, which is vulnerable, but I cannot upgrade the router's firmware because doing so renders the network attached storage drive inoperable. The malware appears to have been developed in Russia.
See the FBI's full statement at the top of this week's Spare Parts section in the right column.
Rebooting the router disables some of the malware, but will still leave one component in place. This component could be used to re-install the parts that are eliminated by a reboot. A better solution, if you're comfortable with the process of setting up the router, involves performing a full system reset. With most routers, this involves using a sharp object such as a paper clip to press a reset switch. If you do this, you'll need to perform a full setup, so write down all the settings first.
At the very least, the router should not use the default password. For many routers, the default user is "admin" and the default password is "password", but the router doesn't force to user to change it. Also, if you've enabled remote administration, disable it. Most routers turn this off by default.
Security firm Talos has a detailed report on its website.
Leafing through news releases this week, I spotted one that described an agreement between a couple of companies to "establish a project and implementation plan to provide an autonomous vehicle platform for [one of the companies], focused on [a delivery system] with driverless cars." Has anyone thought this through?
In the past 4 months, one of the companies raised more than $23 million an initial coin offering (ICO), a recently emerged concept of crowd-funding projects in the cryptocurrency and Blockchain industries.
The companies plan to operate "a fleet of blockchain-operated, driverless vehicles controlled by artificial intelligence" starting in Las Vegas with expansion planned for Miami.
The news release says that the driverless cars will proceed to a delivery location and then notify users via text message. The recipient's identification "will be verified [and] the car will unlock" so the package can be retrieved.
This scheme raises several questions:
Have the developers pondered these questions and others?
The US Supreme Court ruled recently that a federal ban on state-sanctioned sports betting is unconstitutional. The 6-3 ruling eliminated federal ban on sports betting so now any state can permit it.
The Professional and Amateur Sports Protection Act (PASPA) had made it illegal for a state to allow sports betting. The immediate change will be the establishment of betting sites in New Jersey, but other states will follow.
Individual states will determine whether to allow betting on sports. The New Jersey Division of Gaming Enforcement expects to have regulations approved by fall and that gambling operations will be live before the end of the year. Prior to the ruling, sports betting was legal only in Nevada.
DraftKings, which runs the a fantasy sports mobile platform used by nearly 10 million on-line customers, plans to expand into on-line betting.
The company has been preparing since 2017 to launch a sports betting platform for mobile devices and is currently working with state regulatory officials to apply for operating licenses in those states that have already passed legislation authorizing sports betting.
