Digital cinema cameras

On Aug. 31, Red Digital Cinema delivered its first 25 RED production digital cinema cameras. RED and other digital cinema-oriented products like ARRI's ARRIFLEX D20, DALSA's Origin, Sony's F23 and Thomson's Viper FilmStream cameras represent the high end of a spectrum that goes all the way to webcams shooting high-resolution pictures for amateur astronomy.

Breaking down each camera mentioned above:

  • The ARRIFLEX D-20 provides variable frame rates. In the data mode, it offers 2880 × 2160 at 4:3 or 16:9 with 12-bit resolution. In the HD mode, it provides a 1920 × 1080 YUV or RGB 10-bit logarithmic or ITU 709 gamma-corrected output.
  • Origin provides a single output format of 2:1 with 4096 × 2048 imaging, with 16 bits.
  • The F23 supports 4:4:4 1920 × 1080 RGB imaging and 1080/23.98p, 24p, 25p, 29.97p, 50p, 59.94p, 50i and 59.94i formats. It offers 14 bits.
  • Viper FilmStream shoots in variable aspect ratio (16:9 and 2.37:1) at up to 1080p resolution, with 10 bits.

While Sony, Thomson and others build cameras suitable to capture images for cinema output, they also offer standard broadcast interfaces (1080i/p and 720p) that make their utility different. Other factors, including bit depth, define performance as well. In a new twist on technology and marketing, RED offers both 4K and broadcast imaging, with 12 bits.

The common worldwide TV broadcast formats are 525- and 625-line based. Above that, of course, are the common HDTV formats 1080i/pand 720p. Interestingly, NHK has a research project for ultra-high-definition imaging at 8K — 7680 × 4320 pixels with a 1.77:1 aspect ratio (16:9). High-end government and astronomical cameras often have a similar capability of 7216 × 5412 (and probably much higher for the ones we don't even know about). Pixel depth varies, with 8 bits or less for many broadcast products and goes as high as 18 bits for Kodak sensors used as image acquisition devices. Some hints at capability can be garnered by looking at Kodak's line arrays, which have up to 14,404 pixels across. Imagine a scanned 16:9 image based on that with 117 million pixels per color. The data rate could easily extend to 302Gb/s (4:4:4, 72Hz, 12 bit).

Resolution independence

Why does this incredible range of capability need to exist? Simply put: Because it can. Most of today's broadcasters are familiar with downsampled images. Many facilities have been using HD cameras in SD applications with the expectation of needing HD capabilities later. Similarly, upsampling is no big deal today. When economics or technology dictate, SD 16:9 cameras are used as input sources for HD workflow.

I remember when Quantel first started touting “resolution independence.” Us TV folk didn't get it for the most part, because we couldn't figure out why you would want to do that. Now applications abound. Today, digital cinema cameras shoot a commercial from which high-quality stills are grabbed for print use. HD shots with HDV are used for SD final delivery so an HD copy can cost-effectively be kept on the shelf.

Just a few years ago, SD cameras enjoyed the lion's share of the market, but today SD cameras are a rapidly diminishing share of the total market. One major manufacturer has openly said it will cease manufacturing SD-only cameras by the end of the decade. Even replays for the NFL have converted to HD this year in part because rights holders have to deliver an HD copy of every game before a certain date.

Such firm business strategies are changing what we consider broadcast quality. Two years ago, two of the CEA giants began delivering HD consumer cameras (HDV). Prices are still high, but think about the value. Professional HD video, notably newsgathering, is being shot with cameras that cost less than an ENG lens cost a few years ago. This would not be so amazing if it weren't for the fact that the quality is amazingly good. It's not the best HD one can acquire — and certainly not the output from a DC camera — but it sustains pretty well in many applications.

To add to the confusion, there are multiple types of image sensors. IT, FIT and FT CCDs have been joined by CMOS sensors in the last few years. While CMOS offers higher sensitivity and lower power consumption, which are particularly important for the still camera industry, the differences between CCD and CMOS include tradeoffs that make neither the clear winner for motion imaging.

Critical parameters, including lower noise and higher dynamic range, make CCD attractive. CMOS offers a direct digital output, while CCD requires analog-to-digital conversion. Both appear in professional products and likely will for some time to come. As CMOS improves, in part due to significant research for consumer applications, it's likely to be used more in professional situations as well.

Consumers set the standard

This intrusion of consumer product development into the professional realm is an important and pervasive change in our industry. We might not like it much, but we are a small and unattractive industry to many companies. Small niche manufacturers will always be able to justify investing in part because they can be agile as the industry changes, and they can manufacture products without huge R&D and overhead costs. Large manufacturers face considerable uncertainty in our small marketplace in which they can choose segments to service.

Nowhere is this a bigger factor than in cameras. The engineering talent and manufacturing expertise needed to develop and then successfully market television cameras is both scarce and expensive. Time to market is long, and the size of the market is tiny compared with consumer applications. This is why HDV professional cameras are essentially upscale, high-end consumer products. Without consumers, the professional industry would not likely experience such important advances. Big ticket items, such as asset management systems that cost hundreds of thousands of dollars to install, will always be a fertile business for manufacturers. But $1500 dollar cameras, with only tens of thousands of units sold per year, cannot make a viable market. When a large company builds a big, expensive television camera, expect the price to remain high so they can recover the R&D costs and make a profit. Don't expect many small companies to get into this type of business.

Is RED the answer?

This brings me back to RED. It's a 4K digital cinema-capable camera that costs less than $20,000. It should be no surprise that the development money came from a wealthy and committed patron, who happens to be the founder of Oakley sunglasses. The camera was a hit at NAB again this year and probably shocked some more traditional players with its quality images.

Is it television? Beauty is in the eye of the beholder, or maybe it is in how the resolution-independent technology is applied. Can an upstart with a low price, but no installed base, invade a high-end, film-oriented community that embraces products from traditional vendors? And, can it capture other traditional broadcast and production applications because of the price point? Is RED better than other options? The answer may depend more on the application and less on the price.

John Luff is a broadcast technology consultant.

Send questions and comments

View the Red Digital Cinema’s RED ONE camera at NAB2008