Many outdoor photographers have long admired nature documentaries like those from the Discovery Channel and National Geographic. Yet when they tried shooting in traditional video, the results were often disappointing. Video wasn’t known for its detail and color rendition.
Now a lot of people have big plasma and LCD screens showing off the best in sports television. Show a nature video on them, and it looks even worse—unless it was shot in high definition, or HD. That’s definitely the future of video.
We’ve been looking at HD digital video lately for three reasons:
1 We’re excited by the image quality for nature work.
2 High-quality still images are possible from HD video.
3 We’re launching a new magazine called HDVideoPro (which will be out this spring).
I’ve been working with a couple of Canon HDV camcorders and I’m impressed with what’s now possible in video. HDV is a recording format for HD video and is used by Canon, Sony and JVC in their most affordable HD video cameras. HDV is capable of pro results, though it’s a highly compressed format that won’t match the big, expensive HD cameras that Hollywood uses.
The HD standard is set for all cameras because they must be able to play back video on all HD sets and monitors. Cameras produce resolutions of 1920 x 1080 pixels or 1280 x 720 pixels, commonly shown as 1080i or 720p (i and p refer to interlaced and progressive scans, which we’ll return to shortly). This is hugely better than the standard digital video of 720 x 480.
These cameras aren’t cheap by any means, but considering what you get, the price is remarkable. A compact, three-chip, pro-level camera will run about $4,000 to $6,000. A compact, one-chip camera designed for consumer use can be had for $1,000 to $2,000. To get anything close even a few years ago would have required an investment of $20,000 or more.
The difference between one-and three-chip cameras (the "chips" refer to the number of sensors) is in the ability of the camera to capture tonal range, color and light in darker conditions. In bright light, you’d be hard-pressed to see the difference between these cameras, though the three-chip camera will typically have a slight edge in color rendition. In low light, a three-chip camera will capture scenes with less noise, better tonality and richer color.
I recently worked with a one-chip Canon HV-10 and three-chip Canon XH-A1. My friend Michael Guncheon, HelpLine columnist for PCPhoto, took a look at the HV-10 footage on a high-end studio monitor (he’s a partner in a video postproduction company) and said it looked amazing when shot in good light. That’s a testament to the capability of this class of camera. The A1 is even better and understandably is becoming a popular choice among pros who want a compact, one-piece HD camcorder.
(A side note: I can’t recommend the HV-10 for serious video shooting, however, because it has no microphone jack. Good audio is critical for video and you need a separate microphone, whether a shotgun or wireless, to achieve that. However, Canon has announced the HV-20, essentially the same camera with a different shape—and a microphone jack!)
Both cameras are 1080i camcorders, meaning they produce interlaced video (though the A1 and HV-20 have progressive options at certain speeds). You can stimulate a debate among video people by asking if interlaced or progressive is better. Video is produced by scanning a signal at very high speed across the monitor. Interlaced units do it faster by scanning alternate lines in two passes, which appear in two "fields" that combine into a frame of video (video’s default speed is 30 frames per second). To some, this makes the video look a little smoother.
Progressive units scan the whole image at once so one frame of video is complete. This is a real advantage when taking still shots from video because each frame is done as a single unit that doesn’t have to be blended. Also, progressive video can look sharper than interlaced, which is why 720p looks very similar to 1080i (1080p offers a very high quality but demands a camera with very high processing capabilities at normal speeds of 30 frames per second). Taking stills from interlaced video is trickier since action will emphasize little lines that aren’t part of the scene (they look like a comb on the edge of the action)—artifacts from the twice-scanned interlacing process.
Some HD camcorders offer a high-megapixel image capture separate from the video capture. This can be effective. However, you also can capture still images directly from the HD video with good results. I’ve made captures from video from the XH-A1 that I blew up to very good 10x18-inch prints that would look great framed on the wall.
Before explaining how to get high-quality image capture from HD video, I have to qualify a few things about video to stills:
1 Video is normally shot at slow shutter speeds (1/30 to 1/60 sec.). This allows a blending of action from frame to frame so we see motion better. However, it causes a problem for capturing still images; cameras must be tripod-mounted and you must be careful of moving subjects.
2 Video doesn’t have the same color space as digital cameras. It’s a more restricted space, though it can give excellent results, just not the same as that from a digital camera. You’ll need to make some adjustments in Photoshop or another image-processing program.
3 Many cameras offer high shutter speeds that can stop action, but as you shorten the shutter speed, the video can look choppy or strobed. There’s a compromise that must be made, depending on the key use of the video, or else you need to shoot two segments at different speeds of the same thing.
Whether you’re shooting video or capturing still frames, you need to download your video to your hard drive. Video takes up a lot of hard drive space, a minimum of 1 GB for every four minutes of video. You need a fast, big hard drive to work with video.
Most of the latest video-editing programs will handle and capture HDV footage, and they will allow you to capture a single still frame. I’ve found Adobe Premiere Elements, Apple Final Cut Express and Pinnacle Studio to be affordable, easy-to-use programs for downloading, editing and outputting HD video.
Following is a simple way to capture a video still (check your software’s manual for the specific controls needed). I’ve used this to get image files that make great prints.
1 Using your video-editing software, find the spot in the video that you want to capture.
2 Move the scene back and forth, a frame at a time, to find the exact point that shows the sharpest, best image with the least artifacts from interlacing. This is important, and you may be surprised by the range of image quality that appears.
3 Export the shot as a still frame and save it to your hard drive.
4 Open the shot in Photoshop or another image-processing software. If shooting HDV, you’ll find the image isn’t full width and looks squashed left to right. Use Image > Image Size, check Resample, uncheck Constrain Proportions, then set the Pixel Width to 1920 and click OK.
5 If your program has a de-interlace control, try it, though don’t be surprised if it doesn’t have a big effect. To deal with interlacing and compression artifacts, try using the Blur tool just on them and be careful not to oversharpen them. You may have to sharpen to a layer and remove the sharpening done to interlacing artifacts.
6 Process to get the best image as you’d normally work a photo.
I discovered another way of getting an even better image from video from David Leeson II, a video editor at The Dallas Morning News. His father, David Leeson, is known as an innovator in bringing video to photojournalism, and also is at The Dallas Morning News (look for an article about him in the first issue of HDVideoPro). Here’s Leeson’s approach; it will seem a little odd, but it works. You must have a large, high-resolution monitor (at least 1600 x 1200):
1Find the spot in the video that you want to capture as before, moving the scene back and forth, a frame at a time, to find the best spot to capture from.
2 Increase the window size of the scene in your software as large as possible, filling up the monitor. (Leeson finds Final Cut Pro with a 99% view gives the best results in his workflow.)
3 Do a screen capture of what you see with your computer (in Windows use the Print Scrn key; in Mac, use Shift Command 3 or go to the Grab program in Utilities).
4 Open your image-processing software (such as Photoshop) and paste the screen capture onto a blank document (when you open a new, blank document, it should be at the size of your screen capture or you can make one big enough to handle it using a resolution larger than your screen resolution). Flatten the file and crop to the image itself.
5 Make a huge, outrageous enlargement of the image to about 30 inches wide at 600 ppi (use Bicubic Smoother in Photoshop or try a plug-in like Alien Skin Blow-Up or onOne Genuine Fractals).
6 Resize the image back down to a normal size at 203 ppi (that’s what they use at The Dallas Morning News and it works for printing on an inkjet printer). This fills in some of the damage to the scene that comes from interlacing.
7 Process to get the best image as you’d normally work a photo.
Some believe that HD video is the wave of the future, that photographers at some point may just carry an HD camcorder for both video and still images. I’m not sure I buy that (there are issues, such as different shutter speed needs), but HD video definitely has interesting possibilities for nature photographers.