The Truth About HD DSLRS

It’s safe to say that since HD video entered the DSLR arena, no other feature has garnered as much attention from both photographers and videographers, or seen as much improvement in quality and control. Now, if you’re a pro photographer looking to shoot an award-winning video or merely looking to add HD video recording to your services in order to remain competitive, the question you should be asking is whether the latest pro DSLR models can deliver the best of both mediums or if there are more affordable alternatives for capturing high-quality HD video.

The answer may depend on your personal shooting style and business strategy. For example, if you’re a wedding photographer currently shooting with a 21-megapixel Canon EOS-1Ds Mark III (pre-HD DSLR era), should you invest $2,500 in the full-frame, 21-megapixel EOS 5D Mark II in order to add full HD 1080p video-recording capability? Or does it make more sense to keep the EOS-1Ds Mark III as your primary camera and dedicate a more affordable, lighter backup camera such as the 18-megapixel Canon EOS Rebel T3i ($900) to shoot 1080p HD video? If you choose the EOS T3i, you can get away with using its lower-cost, image-stabilized EF-S 18-55mm kit lens without lowering video quality. Plus, the T3i features a dedicated movie button, a swiveling three-inch LCD monitor (ideal for overhead or low-level shots), a Movie Digital Zoom function that can turn a prime lens into a 3x-10x zoom while maintaining HD video resolution and a Video Snapshot function that lets you capture a high-res still photo during video recording.

HD DSLR Cameras Compared

If you’re a Nikon shooter, you might consider the $1,200 Nikon D7000 (with 1080p video, stereo input jack and active AF in video mode) as your primary video camera instead of the more expensive $5,200 D3S (with 720p video, no stereo input and manual focus in video mode), or even the semipro D300S ($1,700, with 720p res and manual focus only in video mode). Sony shooters can only look toward the advanced DSLR models for video, as neither of Sony’s full-frame, pro-level cameras offer it. A good choice is the new 16.2-megapixel Sony A580, which produces stabilized 1080p video, has an articulating three-inch LCD and a dedicated video record button, and stores video in AVCHD or MP4 formats. For a few hundred dollars less, the 16.2-megapixel A55 enables high-speed phase-detection autofocus in video record mode (thanks to translucent mirror technology), but features a high-resolution electronic viewfinder (EVF) instead of an optical viewfinder.

Olympus shooters may be wondering why the pro-level E-5 only features up to 720p video, but it shares that limitation with a few pro-level DSLRs from Canon and Nikon. And, for the moment, it’s the only Olympus DSLR with HD video recording of any kind.

For Pentax shooters, the pro-level, 16.3-megapixel Pentax K-5 ($1,700) is a better choice than any of its advanced DSLR siblings. It records stabilized 1080p video at up to 25 fps, is weather-sealed and cold-proof (to 14º F), and has a stereo input jack and a high-res, three-inch LCD monitor.

What Do Pro DSLRs Offer On The Video Front?

With the video advantage tilting toward most advanced DSLR models, are there still reasons why a pro-level DSLR may be the best tool for your video projects? Perhaps.

Most pro-level DSLRs are more durable and weather-resistant, feature higher-capacity rechargeable batteries and include advanced image-processing circuits to reduce noise caused by heat or shooting at higher ISOs. That’s the case whether the pro camera uses an APS-C sensor, a Four Thirds sensor or a full-frame sensor. In this last category, only two full-frame pro DSLR models from Canon can output up to 1080p video—the EOS 5D Mark II and the EOS-1D Mark IV. These deliver an advantageous full field of view from an ultrawide-angle lens, plus an increased depth-of-field “separation” at large apertures compared to smaller APS-C or Four Thirds sensors. According to extensive RAW data comparisons found at, the larger pixels found on full-frame sensors contribute to lower-noise images and increased shadow details (especially at higher ISO settings), and it’s likely these benefits are passed on to the video frames.

Other still photo advantages of full-frame sensors, such as typically higher resolution and improved color and tonal gradations, unfortunately may be lost in translation when recording video. The reason is that for full HD 1080p resolution, only a fraction of the pixels found on a typical DSLR sensor (approximately 2 megapixels) are needed for each video frame, so a full-frame, 21-megapixel sensor may not produce video that’s any sharper than a 12-megapixel APS-C or Four Thirds sensor (assuming all use a high-quality lens and are focused accurately). In addition, as video is being recorded and saved, the white balance and color bit depth get locked in, removing useful tonal data. What you wind up with is closer to 24 or 30 JPEGs per second recorded to video instead of RAW files containing additional tonal data. Someday, HD DSLRs may include a RAW video recording mode, but for now the volume of data produced by capturing 30 RAW frames per second at 1080p resolution (approximately 1440 Mb/s, or 11 GB a minute!) is too much for even the fastest memory cards and would bring most computers to a standstill in postprocessing.

Shrinking the pixel data captured by the entire sensor in a DSLR down into individual 16:9 aspect ratio frames requires a mixture of in-camera cropping and pixel-sampling techniques. On most DSLRs, pixel sampling merges the exposure and color data from multiple pixels into each pixel used in a video frame. This results in improved light sensitivity, lower image noise at high ISO settings, greater shadow and highlight details, and truer colors. Performing this magic at a constant 24 or 30 fps, however, requires a super-fast video-processing engine, which produces heat as a by-product. Unless that heat is removed effectively—which is more likely in a pro DSLR body containing a magnesium-alloy frame—the image quality of the video will begin to suffer. Some DSLRs, such as the Pentax K-5, have heat-warning indicators and will shut down automatically at a threshold temperature in order to protect sensitive circuits.

The bottom line: For the price, full-frame DSLRs and pro-level APS-C DSLRs may not provide significant HD video quality advantages compared to semipro and advanced DSLRs, but they’re still a better choice for their rugged construction and reliability.

Where Has All My Memory Gone?

How does your camera compress the video it records, and at what data rate does it store video on your memory card? These questions are important to video shooters as they greatly affect image quality and the length of your videos. To a lesser extent, both may affect the compatibility of your video with editing programs and sharing services.

While nearly every imaging device supports the codecs (compression/decompression algorithms) used in standard JPEG photo files, there isn’t a similar standard for video. As a result, a variety of compression algorithms are used to shrink the video data on different DSLRs into a manageable size for storage and playback. The Nikon D90 and many early HD DSLRs used Motion JPEG
(MJPEG) compression, which results in a data bit rate of up to 48 Mb/s for 720p format video including sound. (Note: Divide Mbits by 8 to get megabytes/sec). MJPEG-compressed movies can be opened by most video-editing programs, but create fairly large files. Pentax still uses MJPEG compression in its pro 16.3-megapixel K-5 DSLR, averaging 62 Mb/s for 1080p video with stereo sound. To keep video files smaller and reduce bit rates without lowering image quality, manufacturers such as Canon, Nikon, Panasonic and Sony are using the advanced H.264 compression codec in many models. That codec can squeeze 1080p video down to about 17 Mb/s to 44 Mb/s, depending on the camera. The Canon EOS 5D Mark II uses H.264 compression to achieve bit rates of 42 Mb/s in .MOV format. Nikon’s D7000 uses H.264 to achieve 44 Mb/s in .MOV format, and Sony’s A580 and Panasonic’s G-series cameras use H.264 compression within the AVCHD format file, resulting in a bit rate of 17 Mb/s.

Not all manufacturers list the compression format or bit rate in their camera specs. Instead, they list the maximum length of a single video clip or the maximum size in gigabytes of a single clip—which is usually 4 GB due to limitations in the FAT32 file structure used on most memory cards and hard drives. That equates to up to nine minutes of video per clip on the Pentax K-5, up to 30 minutes for the Sony A580 and up to 12 minutes for the Canon EOS 5D Mark II. If you think 4 GB per clip is a limitation, perhaps you’re better off with an HD camcorder—since it’s more likely your audience will fall asleep watching any continuous clip longer than five minutes.

The Trickle-Up Effect

Someday we may see a liquid-cooled HD DSLR that solves the over-heating problems in video mode, and it’s likely that feature will show up in an advanced-amateur-level DSLR first. That has been the trend lately, and examples include the Nikon D7000’s active AF capability in Movie mode and the Canon EOS T3i’s Movie Digital Zoom feature. Nikon’s active AF uses contrast detection and actually can track a slow-moving subject during video recording. Canon’s Movie Digital Zoom function magnifies the center of the video frame to the equivalent of a 3x-10x optical zoom while maintaining the 1920×1080 pixels per frame of full HD video, effectively turning a 50mm ƒ/1.4 prime lens into a 50-500mm ƒ/1.4 zoom (with a jump from 50mm to 150mm) for video. While it’s possible that optical resolution, light sensitivity and color gradations will change as the magnification progresses due to fewer pixels being sampled, it’s still an innovative feature that we bet will trickle up to pro Canon DSLRs soon.

We’re also confident that, given the growing interest in HD DSLRs, many of the current shortcomings found in them will be addressed and fixed in future models. Top on our list is even faster AF capability and motion tracking during video recording—without using a translucent mirror that lowers the camera’s light sensitivity or forces you to use an electronic viewfinder.

We also expect designers to figure out ways to quiet down the internal lens aperture and focus movements, as well as image stabilization and other camera noises. Until then, on-camera microphones will continue to pick up every camera noise, forcing you to use some kind of dual-source audio system to get the microphone closer to the subject. At minimum, that requires a stereo input jack (now fairly standard), an accessory mic attached by long cords or a wireless RF microphone system. A more elegant solution would be a DSLR featuring built-in Bluetooth compatibility (or optional Bluetooth integrated into a battery grip.) This would allow for the use of securely paired, remote Bluetooth microphones and headphones—a much cleaner solution.

Other missing features that are found in most camcorders include a blinking video recording light and built-in video light. Both serve a purpose, but given the design constraints around a DSLR, adding a video light may not be possible. Perhaps LED lights could be integrated into the camera’s external flash?

Time will tell if any of these features will become reality. However, if you aren’t willing to wait for these improvements to arrive in a DSLR body, perhaps you should consider an interchangeable-lens HD camcorder with an APS-C sensor for your video projects, such as the Sony NEX-VG10. It offers all of the creative advantages of a large-sensor DSLR, but looks more like a camcorder and lacks RAW file capture and other DSLR still photo advantages.

Michael J. McNamara has been reporting on imaging products and trends since 1989. His blog and a portfolio of movies and still photos can be found at

Video-Optimized Lenses On The Horizon?

What are the best lenses for recording HD video from a DSLR? If extreme depth-of-field control is your priority, a prime wide-angle 24mm ƒ/1.4 may do the trick—and it will provide better low-light performance for still and video shooting. For the majority of situations, however, an ultrawide zoom lens such as the Pentax DA* 16-50mm ƒ/2.8 ED AL SDM or the Canon EF 16-35mm ƒ/2.8L II USM will deliver the best combination of wide-angle views, depth-of-focus control and price.

Still-camera lens manufacturers have yet to release video-optimized (VO?) lenses with a smooth, electronic zoom similar to those found on HD camcorders, unfortunately. This type of lens would allow for automatic rack zoom and rack focus initiated by touching two points on the LCD preview monitor. Currently, rack focus is found on the Sony A55, but rack zoom is only found on pro-level HD camcorders.

In addition to electronic zoom lenses, we’d like to see more affordable hybrid lenses with longer zoom ring and focus “throws” for creating more stable, professional video. The problem now is that most modern lenses are optimized for still shooting using the camera’s AF system, requiring a relatively small twist of the zoom ring to go from wide to tele quickly, and a relatively small turning radius on the focus control to move from near to far focus. But when shooting video, slow zooms and moderate, precise focus changes produce better videos, so having a longer throw for zoom and focus rings is a plus. The ironic twist is that many older manual-focus lenses had longer zoom and focus throws than the current offerings.

Leave a Reply