16
Integrating with Larger Camcorders

Using an HF G10 together with an XA10 on a two-camera setup will match perfectly as long as they are both running with the same settings, including program mode, exposure and white balance. There is similar compatibility between the G10/XA10 and Canon’s XF professional line of camcorders, the XF100, XF105, XF300, and XF305, which range in price from $3,000–$7,500. In fact, the G10/XA10 has so many features in common with Canon’s XF line, it could be used as a directly matched companion. G10/XA10 will also hold its own as part of a production strategy with other professional camcorders and with HDSLR (high definition single lens reflex) still cameras used for video.

XF Camcorders

Although they occupy a more compact body, the Canon Vixia G10 and the XA10 provide almost all the functionality of Canon’s XF series of professional camcorders. They share the same f/1.8 zoom lens, the same eight-blade shutter, the same -inch native 1,920×1,080 CMOS image sensor, and the same image processor (with its features like Face Detection and tracking) as the XF100 and XF105.

The G10 and XA10 are much more compact because the functions of the numerous physical switch controls on the XF100 and XF105 have been reallocated to the touchscreen. The XF100 and XF105, however, can record in slow and fast motion—in other words, at a different frame rate from the playback rate. The most significant difference is that the XF100 and XF105 can record 4:2:2 color space onto Compact Flash cards at 50Mbps, whereas the G10 and XA10 record 4:2:0 onto SDHC cards at 24Mbps. On the practical level, most viewers will not be able to see the difference between 4:2:0 and 4:2:2 except in circumstances like chroma keyed special effects, which could show fringing and edge artifacts in 4:2:0.

4:2:2 color sampling has double the color resolution of 4:2:0, providing finer color transitions and microscopically cleaner edges, which can be important in processes like chroma keying and advanced compositing effects or in printing video onto 35mm film. Otherwise, unless closely compared side by side, 4:2:0 images from the G10/XA10 can be hard to distinguish from 4:2:2 images from the XF series camcorders.

The XF100 and XF105 files are in Material Exchange Format (MXF), a professional industry standard. MXF files are similar to the MTS files of AVCHD but differ in that metadata is embedded in the video files themselves and the bit rate can go as high as 50Mbps (double the information with half the compression of 24Mbps, the highest quality mode on the G10/XA10). Editing programs that handle MTS files can usually handle MXF files and mix them on the same timeline or transcode them to a common editing codec.

Because the XF100 and XF105 use the same lens and sensor as the G10 and XA10, the image will match in optical quality and depth of field. Images from the XF100, XF105, XF300, and XF305 are more customizable in the camera than those from the G10 and XA10. The four XF camcorders can manually set a variety of image adjustments and store them as nine different presets. Up to 20 presets may be saved to an SD card. These include adjustments for gamma (the contrast response curve), black level, master black, low key saturation, knee, sharpness, overall and selective noise reduction, skin detail, color matrix, white balance, color correction, and setup level.

The highest end of the series, the XF300/305, uses three -inch chips for finer processing of each primary color. Like the single-chip XF100/105, the XF300/305 records a native 1,920×1,080 image in 4:2:2 color using MXF files on Compact Flash cards with bit rates up to 50Mbps. The XF300/305 has an 18× zoom lens, 4.1–73.8mm (29–527mm equivalent) that opens to f/1.6–2.8. The camcorder’s tools include both a waveform monitor and a vectorscope. It weighs 5.8 pounds, which is more than three times the weight of the G10/XA10.

Here, too, the G10 and XA10 will have the same depth of field and very similar if not identical lens quality to the XF300/305. In comparing the G10 and XA10 to the XF100 and XF105, viewers would be hard pressed to tell footage shot on one camcorder from another unless they were able to view the images side by side.

Integrating with Other Camcorders

At times, you may need to mix camcorders of different brands and different qualities on a multi-camera shoot or use the G10 or XA10 as a backup for a higher-end professional camcorder of another brand. (See Figure 16.1.) Here, a more aggressive attempt must be made to make the images of dissimilar camcorders visually contiguous.

Figure 16.1 Sony’s three-chip NX5 and Canon’s single-chip XA10.

If possible, shoot comparative tests to see what matches and what stands out. One of the techniques for balancing two or more camcorders to each other is to focus each on the same subject and feed their image into a common field monitor set for split screen. White balance and exposure need to be equalized before anything else is considered.

If each camcorder has an internal or external waveform monitor, they can be set so that black and white levels are more precisely matched. Other instruments like a color RGB parade (a triple waveform monitor that shows red, green, and blue waveforms in a side-by-side comparison) and a vectorscope (which shows color-saturation levels and predominant hues) are useful for measuring and comparing color balance, hue, and saturation. But even without these instruments, a split screen allows a direct comparison in which differences and similarities are immediately apparent and many adjustments can be made by eye.

For making adjustments within the G10 and XA10, start with white balance and exposure. Then, you might want to consider image effects, which enable you to customize color saturation, brightness, contrast, and sharpness. Choose FUNC > Image Effects > ON and then adjust the Color Depth, Brightness, Contrast, and Sharpness settings as needed. As discussed in Chapter 6, “Controlling Color,” be careful with high settings under Image Effects. Excessive contrast can permanently erase detail in highlights and shadows, and excessive sharpness may create artificial edges or video grain or accentuate irregularities in skin tones. Do not set image effects unless you can judge the output on a well-calibrated monitor instead of the camcorder’s 3.5-inch LCD screen. Otherwise, leave these kinds of adjustments to postproduction if you cannot verify them on location.

Not all matching has to happen in the field. The most important considerations in the field are that each camcorder is white balanced and exposed as correctly as possible. The G10’s and XA10’s waveform monitor is a big help in fine-tuning exposure. You can precisely identify whether white areas are clipped due to excessive exposure and determine whether important shadow detail is crushed. White balance each camcorder to the actual lighting conditions. Color grading in postproduction can usually take it the rest of the way to equalize skin tones, contrast, background hues, and saturation as long as it does not have to radically correct poor white balance, irredeemable exposure errors, or degraded images.

If two or more camcorders are going to be used together repeatedly, save their setup parameters so they can be instantly matched for future shoots. (Refer to Chapter 15, “Special Features,” for more information.) Develop a strategy that exploits what each camcorder is best suited for.

Integrating with HDSLRs

HDSLR (high definition single lens reflex) still cameras are digital 35mm, APS-C, and mini cameras that have the ability to record full-motion video. HDSLRs have been hot items for the past several years because they fill a gap in a price range of $800–$12,000 that existing camcorders did not previously cover—namely, providing interchangeable prime lenses, a 35mm or near 35mm image size, and selective focus with shallow depth of field. The television series House shot one highly publicized end-of-season episode with a Canon 5D Mark II DSLR as an experiment, but they have not adopted it for regular production. However, numerous fiction and documentary feature productions, including Sundance winners, have adopted HDSLRs. (See Figure 16.2.)

As for price, a Canon 5D Mark III body alone costs more than $3,400 with a couple of batteries and two 32GB recording chips. By the time you add two or three L Series Canon lenses at $900 to $2,500 each or two or three Zeiss prime cine lenses at $3,900 each, a cine viewfinder or portable HD monitor, a separate digital audio recorder with a professional microphone (because the 5D has amateur sound), and Zacuto braces to make the camera stable, you are in the $9,000 to $24,000 range for each package. The Mark III makes sense for still photographers making the transition to video (for example, wedding photographers) who may already have an investment of $5,000–$20,000 in still-camera lenses that they can now use for video. At the low end of HDSLRs is the Canon T3i body (and its predecessor, the Canon T2i), which has the same video functionality as the middle-level Canon 7D in an APS-C size CMOS sensor. The T3i body sells for $799 with an inexpensive but slow 18–55mm zoom lens. The price plus longer-term batteries, 32GB memory chips, and a separate middle-quality prime lens totals about $1,500. Adding a second prime lens plus sound equipment will bring the price closer to $2,000. This puts the T3i squarely in the price range of the XA10 and would be a strong alternative for a student or an independent media maker.

Figure 16.2 Canon T2i HDSLR equipped for video and the Canon XA10.

The shooting style of the HDSLR lends itself to carefully set up shots, a tripod, a slider, a mini dolly or Steadicam operation, a 12-minute limit between cool-down times, shallow depths of field, and avoiding situations that may provoke skew, aliasing, and moiré artifacts. Footage from HDSLRs can get noisy from heat, and the camera may shut down due to overheating. Because of the real danger of overheating, people who shoot with HDSLRs often alternate between two HDSLRs so one can cool as the other shoots. If you mix an HDSLR with the G10/XA10, plan a strategy to get the most out of the features of each kind of camera. Because most HDSLRs have a limit of a 12-minute take, the G10/XA10 becomes the long-take camera and very likely the more mobile camera because of its dynamic image stabilization.

The main problem with using HDSLRs is working around their limitations. With HDSLRs, image stabilization is an issue. Unlike the G10/XA10, image stabilization in an HDSLR is not designed for shooting video. Image stabilization is available only on certain individual SLR telephoto lenses and is designed to take sharp still photographs, not to smooth out the bounce of walking when moving with the camera. HDSLRs will not have zebra patterns and waveform monitors. An HDSLR may have a histogram for measuring exposure, but this function is usually available only for reviewing stills.

HDSLR cameras pose enormous, but surmountable, problems for video production:

Awkward ergonomics (balance, shape, and controls appropriate for taking stills instead of moving images)

Inadequate viewfinding without additional attachments

Limited video controls

Limited recording time (usually 12 minutes)

Rolling shutter artifacts

Moiré patterns

Workflow issues

Heat issues (because they were not made for continuous video recording)

Poor sound

The fact is, HDSLR cameras are simply make-shift tools in digital video production. In spite of all these limitations, however, the images from HDSLRs are often stunning because of the high-quality lenses that can be used, the lower optical and mechanical tolerances needed for a larger format, and the greater potential for selective focus with a shallower depth of field.

Depth of field will be one of the biggest visual differences between the G10/XA10 and a larger-format HDSLR. You do not want to intercut similar shots from different formats together in the same scene unless there is an aesthetic reason to do so because the depth of field will radically shift with each cut. Instead of cutting between shallow and deep space arbitrarily, save the HDSLR for moments that should be shallow or devise a scheme that justifies the change of appearance. It might be possible to shoot with the G10/XA10 in medium to telephoto at f/1.8 to f/2.8 to achieve a shallower depth of field. This may match with a DSLR in normal or mild wide angle and at higher f/stops, where its shallow field is less pronounced. But instead of a strategy of making the spatial representation of each camera match, you more likely will want to exploit the unique look of each camera: the G10/XA10 in wide angle at an infinite depth of field and the large-format camera taking close-ups with a shallow depth of field. The intercutting will be between extremes of scale and depth of field, which can work if there is a consistency and purpose for shallow and deep (like the Sergio Leone spaghetti westerns of the 1960s). Alternatively, allocate the shots that have extensive movement to the G10/XA10 and assign individually set up shots that need to be carefully focused to the HDSLR.

The HDSLR is designed and optimized for taking still pictures and has been retrofitted for video as an afterthought. Excessively fine lines and textures below the resolution threshold of a digital camera create aliasing artifacts as the digital circuit misreads irresolvable detail. To improve performance, an optical low-pass filter is used to better match the threshold of fine detail with the camera’s pixel size and the potential number of pixels. However, what is required for a still photographic image on an 18-megapixel T3i or 22.3-megapixel 5D Mark III is completely different from the requirements of the 2-megapixel resolution of 1,920×1,080 video on the same camera. This produces aliasing, false detail, and moiré when a camera designed for still photography is used for video.

When a high-megapixel HDSLR creates a 1,920×1,080 video image, it simply skips lines of in-between pixels that would have been available for a still photograph. This produces stair-casing and intensifies other artifacts. It would be a much richer image if the HDSLR interpolated the in-between information into calculating the 1,920×1,080 video signal. But an HDSLR’s processor cannot do this in real time, and is already verging on overheating by taking 24 or more pictures a second when it was designed for simply one picture at a time.

Before the HDSLR revolution, the lowest-priced package that could provide large-sensor capabilities dedicated to video was the RED, starting at $45,000 with professional cine lenses and monitor. Canon’s EOS C300 PL is one of the new generation of large-sensor camcorders designed specifically for video. To solve the artifact and heating problems in large-format video, the price tag for the EOS C300 camera body alone is $16,000. The Canon T3i body at $800, the Canon 7D at $1,700, or the Canon 5D Mark III at $3,400 provide a good approximation of the large-format image at a fraction of the price if you can work around the HDSLR shortcomings. The newest camera from RED, the Scarlet, costs $11,700 for a stripped-down body, and about $17,000 for a fully operational camera system that still needs expensive lenses. Competing large-format bodies dedicated to video like the BlackMagic Cinema Camera have come on the market at about $3,000 and will still need another $3,000 to $7,000 in lenses and accessories. Most HDSLRs are still a less expensive alternative for large-format video, but the gap is narrowing.

The Canon T3i has the same video capabilities as the middle-level Canon 7D at about half the price. Both have an APS-C sensor that shoots 24p/25p/30p at 1,920×1,080; and 50p/60p at 1,280×720 which can be used for slow-motion work. Compared to the T3i, the Canon 7D has a heavier-duty waterproofed body and some additional still-photography capabilities. The 7D has HDMI-out during 1080i recording for high-definition monitoring, whereas the T3i reverts to standard-definition monitoring while the camera is recording.

When mixing a G10/XA10 camcorder with an HDSLR on a multi-camera shoot, the ideal is to assign each to what it can do best. You can pre-adjust each to look similar in white balance, exposure, camera profile, and filtration, but the formats will still look different. Figure out a form-and-content strategy that will exploit their different looks. You might assign the camcorder and HDSLR quite different types of shots or completely different subject matter so that matched action and direct shot-by-shot comparisons do not occur. HDSLRs not only look good in shallow focus, to some degree they need shallow focus to look good. Aliasing and moiré disappear in shallow focus. Shallow depth of field is probably the look that you want to exploit with HDSLRs.

An L Series Canon lens on an HDSLR can cost as much as the entire XA10 camcorder. If lens quality is a major comparative difference, then lock the G10/XA10 in Av mode to f/4 for its optimum aperture. Do tests from matching cameras and find the ideal lighting ratio, ISO and gain sensitivity settings, and camera presets. The more expensive camera and lens may be assigned to shoot the primary footage and the G10/XA10 relegated to shooting B-roll. But in certain situations, the opposite strategy might work. If the main subject needs very long takes, run-and-gun mobility, or a deep depth of field, the G10/XA10 excels at that, and a large-format HDSLR might be dedicated to composing and taking exquisite establishing shots and cutaways as B-roll.

When testing and comparing the images of a G10/XA10 and an HDSLR, you might compare the HDSLR’s standard camera profile versus neutral, as well as hacked profiles and setups available on the Internet. To further match the HDSLR to video, you might also try a Caprock 2.0 anti-moiré filter to cut the hyper-real look of HDSLR video and its artifacts.