DSLR (Digital Single Lens Reflex) cameras with video capability are all the rage these days. It seems that every announcement for a new DSLR includes the camera’s formidable video recording capabilities, as if you don’t need to do anything other than point your camera to make great video. If only it were so.

There are many reasons to use DSLRs to make video for eLearning, not the least of which are the stylish quality of the video they can shoot, the range of lenses, and, of course, the additional control of composition you can gain. Recently, I suggested that readers shouldn’t use a DSLR for video production in eLearning. Am I recanting my previous statement? In a way, yes. I’m now advocating the use of DSLRs for creating e-Learning video, but there are still lots of caveats. And at the same time, it remains a lot easier to use your handy dandy camcorder to make video because DSLR’s are not ready for easy video production.

The key word is “easy.” Each brand of DSLR has its quirks and individual video formats, and not all DSLRs can record broadcast-standard video in SD or HD (standard or high definition). If your end result is for Web consumption only, then SD and HD are not a consideration. But when capturing your video into your editing program, it can be painful if you don’t know the exact digital video format. Canon, Nikon, and Pentax all record in different sizes and formats. In fact, even for a given manufacturer, different camera models record in different formats and image sizes.

What DSLRs can do

The thing to remember is that DSLRs are first and foremost devices for creating still photographs (although that’s changing). Can they really record video and do it well? Do they have enough digital “horsepower?” The answer to both questions is “yes.”

Doing the math

Mid- to upper-end DSLRs can take four to seven or so frames of still images per second in RAW format. (Editor’s Note: explanations of a number of technical terms in this article are in the last section.) Most DSLRs run out of buffer space after 14 to 20 or so images (this varies widely from camera to camera). That’s the number of frames in about a half second or so of video, hardly the 24, 25, or 30 frames of progressive video a camcorder easily shoots.

Does this mean a DSLR doesn’t have the horsepower to shoot video? No. There is a tremendous amount of information recorded in a still image file in RAW format — video does not contain nearly as much data. Let’s compare the number of pixels in one second of full HD video to the number of pixels in the five full-frame still images the same camera could record in one second.

Full HD video (also called 1080p) at 30p (thirty progressive frames of video per second) requires a DSLR to process and write to memory 62,208,000 pixels of information per second. If that DSLR has a 14 megapixel (MP) sensor, at full frame rate (5 frames per second) the camera needs to process 70,000,000 pixels of information per second. Obviously, the DSLR can handle the raw pixel processing power.

So what’s the difference? A single 1080p frame consists of 2,073,600 pixels. If you’ve got a 14MP sensor, you’ve got a whole lot more pixels. (Almost seven times as many!) The sensor must use the same pixels each time it records a frame or you would get a kind of strange stroboscopic shift because the camera would record the frame from a different spot on the lens as it focuses on the sensor. This is enough to give me a headache, but the bottom line is that the video recorded can be beautiful, even if it is a little more complex to create.

The benefits of the DSLR for video

Now that you know the technical difficulties, what are the advantages of using your DSLR camera to shoot video? There are five main benefits.

  1. Low light sensitivity: DSLR cameras have a setting for ISO (light sensitivity) and each generation of DSLR does a better job of handling high ISO (low light) settings with less pixel noise in the backgrounds. In practical terms, this means you can shoot with less light in any given scene.

  2. Sensor size (see Figure 1): Sensor size matters because sensors and their interaction with light are kind of funny things. At the top end of the price and quality range (at least for cameras that cost less than $25,000), most camcorders have tiny sensors, even when they have three. At the other extreme of the range, you can use a point-and-shoot camera to record video. Such a camera generally has a teeny (1/2.3 inches) sensor, enough for about 10–14 megapixels these days. In the middle of the range, cameras use an APS-sized sensor that has 14MP. Or you can use a full frame camera. “Full frame” refers to the size of a 35mm film frame from the days when you still loaded film into your camera. Full frame sensors run from about 14MP to 27MP.

 

infographic on devices and their varied screensizes

Figure 1. Relative sensor sizes in digital cameras.

 

Is there a difference? In two words: “You bet!” The first reason is the size of the pixel site. A full size or APS-C sensor is between 9 and 40 times the size of a typical small camera sensor. If the number of pixels is the same, what’s the difference? It’s the size of the pixel site. If a given teeny sensor has 14MP and a big sensor has the same number of effective pixels, then simple math shows the pixel site is far larger. Why does this matter? It all has to do with photons, particles of light. A photon striking a pixel site on a tiny sensor will bounce around a bit before being absorbed by the baffles around the sensor. The light can “bleed” over into another sensor because the site is so small to begin with. This creates noise in the image because an individual sensor will probably be struck more than once. A larger sensor gets less photon “bounce,” so in theory it has less noise. It also works out in practice, especially in the case of pixel sites that are 12 (or more) times larger on the bigger sensor. Less noise means a clearer picture and greater low light sensitivity. So large sensors create images with much less noise than small sensors.

There is one additional reason larger sensors work better for videography. It has to do with the combination of sensor and lens. A small sensor can’t show the depth of field a lens can record. There are many reasons for this. We generally use a 35mm frame size as a reference point. A small sensor, say 1/45th the size of a 35mm frame, crops a picture a lot with a given lens focal length.

Think of it this way: the field of view with a 55mm lens (considered a “normal” lens in 35mm photography) is X-degrees wide for any given lens or aperture. Aperture is the amount of light the lens lets through. The higher the aperture number, the lower the amount of light and the greater the depth of field. Long (telephoto) lenses have shallower depth of field than wide angle lenses. When you couple a long lens to an APS or full frame sensor, you can get a very shallow depth of field, which allows you to make the area of focus the important part of the picture. Small sensors, since they use only a smaller part of the sensor, have HUGE depth of field compared to the image from a large sensor. An additional difference is that a small sensor on a camcorder can’t use different exposure values, so it can only shoot practically with a lens aperture of about f5.6 or so. And since that small sensor only uses a small angle of view, sometimes called the “crop factor,” it makes a wide angle lens, which usually would have a wide field of view, into a telephoto lens, in a manner of speaking. This means subjects both close and far are in focus. In other words, the smaller the sensor, the smaller the angle of view. When you shoot with a small sensor, almost everything is either in focus or out of focus.

  1. Lenses: If you have a DSLR (or an old film SLR for that matter) and an investment in lenses, you understand this benefit. Lenses are an incredible asset for creating video. A good DSLR (or film SLR) lens is generally far better quality than a camcorder lens. If they’re in your equipment bag, you can use true telephoto, macro, or wide angle lenses to create images that you really can’t create with a camcorder. And you can do it with the same style you’d get when shooting still images. The idea of using a telephoto lens to shoot a person from a long way off is intriguing. The person would be in focus and the background out of focus (at fairly wide aperture settings, say up to f8 or so). There is a somewhat overused word for this kind of focus on your subject. It’s called “bokeh,” which refers to parts of the image being blurry, which brings us directly to the next benefit.

  2. Style: The above benefits together provide an immense gain in image control. If you want the whole image to be sharp and in focus, then use a wide angle lens with a focal length of 24mm to 35mm. If you want selective focus, anything above an 80mm lens and higher with an APS or full frame sensor will give you the kind of soft backgrounds or foregrounds you want. “Telephoto” usually refers to a lens that’s over 120mm focal length. But a 200mm or 300mm lens is even better and has less depth of field. This is control. And if you want to create great images, you need control. The range of lenses available for DSLRs is staggering. In addition to control over focus, this range greatly increases the number of situations in which you can shoot video. There are lenses that excel in low light situations. There are other lenses that are ideal for outdoor settings where there is so much light that overexposure is a concern.

  3. Manual control: One of the features of shooting in video mode on a DSLR is that the camera switches to manual mode. The mirror goes up and you can only use the viewfinder on the back of the camera or an external monitor. This is a downside for some people, but it adds to the amount of control you have.

For example, you control white balance when in manual mode. White balance is a camera setting that adjusts for lighting so that white objects appear white in your video. Most modern camcorders have “automatic” white balance, which means it can change anytime you’re taking a shot.

You’ll also do your own focus because autofocus is often shut off when your DSLR is in video mode. (Note that the new models from Canon and Nikon, however, can use the autofocus when shooting video.) Personally, I like to focus manually, as it gives me more control. You’ll need to or should set the aperture as well. All this manual control goes back to style; you control the entire exposure process.

The drawbacks to DSLR

As you’ve seen, there are lots of plusses when shooting video with a DSLR. What are the minuses? There are four that you should pay attention to:

  1. Additional equipment: Yes, you’ll need some additional equipment. I’ve included a small list of essentials in the next section, but you can go pretty crazy with additional “accessories” for your DSLR. Some people spend $5,000 (or more) on additional equipment.

  2. Know your camera: You really need to know your camera and what it can and can’t do. When I got my first DSLR, I immediately decided I wanted to use manual control as much as possible when shooting stills. I normally shoot SD video and use a Canon XL-2 for that purpose. In manual mode, I have complete control over the camera. And I almost always shoot in manual mode, so I’m used to shooting that way. By the way, there are excellent light meters in DSLRs, so you can see if the image is under- or over-exposed at the ISO setting you use. But manual is a downside if you just want to point and shoot.

  3. Codecs: This goes back to knowing your camera, but you really must know the codec you’re shooting with and whether you have any control over it. You usually won’t have any control, at least not yet. Nikon has been using a form of QuickTime with the H.264 compressor, Canon has been using mpeg4, and Pentax has been using .avi (compressed) and motion jpeg. All very confusing. But once you get used to one camera — it’ll change again!

  4. Manual control (see number 5 in the list of benefits): Some people really don’t want to shoot manually. Sometimes I’m one of those people. Sometimes the situation doesn’t allow me time to set up shots, figure out if I want a large aperture with shallow depth of field or a small aperture, etc. If I just need to shoot and get the shot, I leave the whole gizmo (DSLR or camcorder) on as much automatic as I can. It’s really up to you.

The list of things you must have

Here is a short list of items that are absolutely essential for making good DSLR video.

  1. A camera brace: If you want to do anything handheld, you need this. This takes the place of a tripod. Holding a big, heavy DSLR and keeping it steady, while looking at the screen on the back of the camera, is almost impossible without a brace. Frequently braces go over the shoulder.

  2. Focus and f stop puller: This is a little gizmo that slips over the lens so that you can change focus during your shot without holding onto the lens. It’s just a different part of the lens for changing f stops. There are simple ones and elaborate ones with gears and all kinds of bells and whistles. I don’t think any of them are turbocharged, though.

  3. Monitor, or at least a viewer of some sort: This can be as simple as the camera’s back panel viewer (perhaps with a piece of cardboard to shield the viewer from sunlight). Or it can be a video monitor that hooks up to the HDMI port on most DSLRs. That’s the connector you usually hook up to your HD TV, but what’s a monitor, if not a TV?

  4. Microphones and some sort of mobile interface: The built in microphone on any DSLR is abysmal. You need to use the same mics you ordinarily use for video production and you need a way to hook them up to your camera. There are many interface boxes for this that usually mount where the tripod or camera brace would mount.

  5. Lights: You need lights. Camera mount lights make flat light, and if that’s what you want, go for it. There are lots of different varieties of lights out there and my favorite are the least expensive … good old photo flood lights. Flood lights are way cooler than halogen, they create less color hassle than LEDs, and they make beautiful light. Replacement bulbs are cheap too, unlike those lovely, but hot, halogen bulbs.

  6. Memory cards: Last, but far from least, you need lots of memory. A typical DSLR SD card or Compact Disc (usually used with the more expensive cameras) is about 2 GB. That gets used up in a matter of minutes. On some cameras, 2 GB is only good for about 4 minutes of video. There are fairly inexpensive and good 16 and 32 GB SD cards— these will give you enough time to record the video you need.

There’s lots more you can get, but this is a good start for video production with a DSLR.

Seven rules for shooting with a DSLR

  1. Know your camera — it’s important that you know how to really operate your camera. By that, I mean an understanding of how the controls and programming work on your camera, where the controls are located, and which settings are on which menu.

  2. Know your lenses.

  3. Get the right additional equipment. (See the list above.)

  4. Understand lighting.

  5. Think like a cinematographer. What do you want to focus on? How do you want your subject to look?

  6. Understand your subject and know when shallow depth of field works and when it doesn’t.

  7. Understand yourself and your capabilities and desires to have this kind of stylized control over your video.

A parting thought

One last note: For video post-processing, use as powerful a PC or Mac with as much RAM as you can afford. HD video is a resource hog. It’s not so much that the trouble comes with the editing, but rather in the rendering. And if you’re shooting HD, you might as well render in HD, then reduce the size to what your project calls for.

Some definitions

RAW: RAW images consist of the “data dump” from the sensor. RAW files contain all the information the sensor recorded at the snap of the shutter. Video files are never recorded in RAW format. All the different brands of DSLRs record to video standards, but no video frame contains as much information as an individual picture shot in the RAW format.

SD video: Standard Definition video. This is sometimes called NTSC (National Television System Committee) video, which was the old analog TV signal in the US. (In Europe, the standard was called PAL.) When digitized, SD video has dimensions of 720 pixels across and 480 pixels high. Now, even though the digital signal is more complex in a way, it’s easier to get a TV image on a computer, whether it’s a PC or a Mac. But the number of pixels can sometimes get confusing. There are several different flavors of SD out there. In fact, 720 X 480 aren’t the only dimensions specified for SD. There’s also progressive SD, which has pixel dimensions of 720 pixels high by 1280 wide. (I wasn’t kidding about the confusing part). When I refer to SD, I refer to 720 X 480. What’s most important, though, is that your computer can show the video and you can edit it pretty automatically. And when the video is shown on a computer screen, the rest is irrelevant … the computer doesn’t care what the dimensions are.

HD video—High Definition video: This is the new standard of video … sort of. It comes in many flavors, all of which are identified by the height of the video. True HD video is called 1080p (the “p” is explained below), which means it has dimensions of 1920 pixels across by 1080 pixels high. Films are mostly shot in 4K, which refers to the number of pixels (4096) across. This is not for an average user, and takes a lot of horsepower to edit and render. 4K has 12,582,912 pixels in each image. Compare this to the 691,200 pixels in a 720 X 480 NTSC image and you begin to see the computational power you need to make a video. And on a computer screen, it’s lost. There’s even a format called 4320p. We don’t want to go there — it’s not necessary or important for our work.

p (or i, for that matter): When you see video labeled 1080p or 720p, the “p” refers to progressive scan. Regular old fashioned video is called i (sometimes 720i) where the “i” means interlaced. Interlacing means that to make a whole frame of video, the scan sequence is the 1st line, followed by the 3rd line, followed by the 5th line until the 719th line is reached. Then the scanner goes back and records the 2nd line, the 4th line, etc. until it gets to the 720th line of video. In the old analog world, video was recorded at 29.97 frames per second, so each half frame was recorded in 1/60th of a second (rounded up). The two interlaced frames made the whole frame. Progressive scan video is recorded one full frame at a time, 1st line, followed by the 2nd line, etc. Sometimes it’s also called 30p … or 25p (European PAL video) or cinema-like 24p, which also happens to be the frames per second rate of real movie film that comes in long strips of plastic with sprockets holes in it.

ISO: A measure of film speed. At one time in the US, it was called ASA. A long time ago, the sensitivity of film was hard to control during the manufacturing process. Since sensitivity was hard to control, there had to be a measurement of how sensitive to light a film was. Some films are more sensitive and some films less sensitive. A film that’s more sensitive to light can create a picture with less light or less exposure. The higher the number, the more sensitive the film (now sensor) is. Since nobody (OK, almost nobody) uses film anymore, ISO is now a measure of the relative sensitivity to light to which the sensor is set. It’s a programmatic thing. The problem with sensors and light is that the more sensitive the sensor is programmed to be, the more noise gets into the image the sensor records.

APS: Originally, APS stood for Advanced Photo System and referred to a film size that Kodak first manufactured in the mid-1990s. The film was smaller than 35mm film, and it came in cassettes that you could drop into a camera, shoot, and take to the processor (usually a drug store) to have processed. The cassettes were made by many film manufacturers, and cameras were made by Nikon, Canon, and several others. As photo sensors were developed, a sensor matching the dimensions of an APS frame was roughly 1/3 or so the size of a “full” 35mm image. This means an APS-size sensor is far less expensive to manufacture than a full size 35mm sensor, and it is way larger than a sensor in a point and shoot camera. This size has been embraced by DSLR manufacturers because it’s familiar to film (or former film) users and lenses already exist for it, but the main reason is cost. A 35mm sized sensor is the gold standard for DSLRs, but DSLRs with “full frame” sensors start north of $2,000. If you’re a really dedicated photographer, this makes sense, but for most of us, a DSLR with an APS-size sensor is perfect.

f: This refers to “f-stop”: how much light gets to the image sensor or film in your camera through the iris of your lens. Originally, f-stop was called “focal stop” or “f-number.” There’s a very complex formula that tells the physical size of the hole for a specific lens, but it boils down to this: the amount of light that reaches the sensor in your camera is determined by several different things: the ISO (sensitivity) you set, the shutter speed (and yes, even though your video is shot at 24, 25 or 30 frames per second, there’s a lot of shutter speed variation that you can set), and the f-stop. The actual “f” in f-stop doesn’t have any dimensions, per se, until the formula is applied to the lens you’re using, but think of it as the iris in your eye, the little hole that gets wider or smaller. Your eye compensates for the amount of light by opening and closing the iris in front of your lens.