|
|
|
|
|
|
|
American Ingenuity: Tinkering with "DV"
BRAINTRUSTdv interviews Dan Vance
|
|
|
|
|
|
|
Dan Vance (the other "DV") is a celluloid filmmaker turned engineer turned digital filmmaker. As a filmmaker, he has made short films and industrial videos and has written four feature-length screenplays. As an engineer, he has designed innumerable electronic and mechanical devices and holds seven patents.
Vance is currently shooting his first feature film, Cold Day in Hell, using the VC25P, a tapeless camera system he designed and built by himself.
|
|
|
|
|
|
BRAINTRUSTdv: What inspired you to design a digital video camera of your own?
Dan Vance: I really only became aware of the possibility of using DV for a feature film about eighteen months ago. I immediately began investigating all the available camcorders and started coming up to speed on the technology. Almost as quickly, I was disappointed to find that the "affordable" (under $5000) camcorders all used CCDs that were smaller than an 8mm film frame, with about the same resolution. The thought of blowing that up to theater screen size was worrisome to me, to say the least.
My fears were confirmed a little later, when I saw the results of Steven Soderberg's use of a Canon XL1S in his film Full Frontal. Though he was clearly going for a specific effect, it was hard to watch. I decided that I would have to have at least 1/2" CCDs. Still pretty low resolution, but a big improvement over the ubiquitous 1/3" CCDs. I began writing a script for a no-budget film and I purchased a JVC DV500 (1/2" CCD). I did some test shooting and determined right away that normal interlaced video was not going to work dramatically. So I did some more research and found that the major contributor to the "video look" was the interlace scanning process. At that time, Panasonic's DVX100 had not come out yet, and the least expensive camcorder with progressive scan was over $60,000. That's when I realized that there might be an opportunity to have a slight advantage over the thousands of other no-budget DV filmmakers by having a unique camera with a higher quality image. And I set out to see if I could achieve that.
BTdv: How many functional prototypes did you build before arriving at the final VC25P model?
| DV: Three. Since I wasn't sure the concept was really workable at first, I wanted to minimize the expense of trying it out, so I started with an NTSC camera (Hitachi HV-C10A) and a Super VHS recorder. Once I had proven the concept (30P), I invested in a PAL camera (Hitachi HV-C20E/K), which I intended to dock to a Sony Betacam SP PAL recorder as prototype number two. But there were problems with the Betacam SP approach and I decided to go with the hard disk recording system instead. The existing model is the third version, which is essentially the same as the second, but assembled into a single enclosure. |
|
|
|
|
|
|
|
|
|
|
|
|
|
BTdv: What was the problem with the Betacam SP prototype?
DV: There were two problems that I hadn't anticipated. One is that the input connector on the unit was obsolete and would have required either fabricating a custom connector or accessing the internal wiring, neither of which was an attractive option. The other problem was that the VTR had no component output, which I later found out is a standard design practice, so I would have had to purchase a second deck, a studio type deck, to transfer from the Betacam SP tape to the analog-to-DV converter. Even a used deck would have cost as much as the entire camera, so I decided it would make more sense to go straight to a hard drive. That also results in a marginally better signal than Betacam SP.
BTdv: What are the major components used by the VC25P and how did you acquire them?
DV: The major components are camera, lens, shutter assembly, shutter driver/sync board, hard disk recorder, and analog-to-DV converter. The hard disk and analog-to-DV converter were purchased new, the camera and lens were bought on eBay, and the shutter and shutter drive/sync board were built from scratch.
BTdv: How is the camera powered?
DV: There's an integral holder for a standard NP-1 12V NiCad battery. There's also an external power jack and an AC supply which I use whenever power is available. Power consumption is higher than many camcorders, due mainly to the hard disk and analog converter.
BTdv: Which common video artifacts are eliminated by the VC25P system?
DV: The shutter system eliminates the interlace artifact, which is the primary giveaway that something was shot on video. There's a huge ongoing debate about the "film look" and which elements really define it: progressive scan, contrast, lighting, camera technique, depth of field, production values, etc. For me it's simple, because you can shoot a scene on film that has "video-style" lighting, lousy camera technique, infinite depth of field, high contrast, no production valuesand it still looks like film. Conversely, you can do everything completely "film-style" on video, and it still looks like video, unless you shoot progressive scan. My perception is that ninety-percent-plus of the "film look" is achieved through progressive scan. Five percent is latitude (contrast ratio, Gamma), and the other five percent is everything else.
BTdv: How is progressive scan technology important to someone working in digital video?
DV: Progressive scan is an accessible option for any type of video production, but my only interest in it is in using it to achieve the classic dramatic look of film. Progressive scan and digital video are separate subjects that are united right now simply because digital video is the video at the moment.
|
VC25P: Hitachi camera and wiring
|
BTdv: How do the Canon XL1S and the Panasonic DVX100 fail to provide adequate progressive scan performance?
DV: The XL1S uses an interpolation method to generate an image that is like a progressive scan image, but isn't really. This method inherently loses some information, thus some resolution. Since it's only a 1/3" CCD camera to start with, the resolution loss is a high price to pay.
The DVX100 provides true progressive scan. Its limitations are in its CCD size, the incorporation of consumerish features that make it difficult to get professional results, its non-removable lens with a professionally useless optical image stabilizer, and its 2:3:3:2 internal pulldown recording method, which can cause problems with some editing software.
BTdv: Why are 1/3" CCDs in prosumer cameras undesirable for the professional videographer?
DV: The two inherent drawbacks of 1/3" CCDs are that the small image size equates to huge depth of field, so you lose the control over depth of field that you would have with larger formats, and smaller CCDs mean less light-gathering capability. In addition, smaller CCDs usually have fewer pixels and less resolution than their larger counterparts, though there are exceptions. Another drawback of available 1/3" cameras is that, except for the Canon XL1S, all current models have fixed lenses. Since these cameras are at the low end of the price spectrum, they tend to have marginal lenses, but since the lens is fixed, the user is stuck with it.
BTdv: What is frame integration mode and what sort of cameras have this feature?
DV: Frame integration mode is a CCD timing technique that serves two main purposes: in scientific applications where a camera is used to track small particles, it increases the effective resolution of the camera, which allows for the tracking of smaller particles; in surveillance applications, it increases the low-light sensitivity of the camera. In frame integration mode, the exposure time is an even multiple of the field rate, so any moving objects exhibit pronounced motion blur. In the VC25P, the external shutter excludes half the light, eliminating the motion blur while preserving the image from two identical fields, resulting in a progressive scan image. Many industrial/medical cameras have this optional mode.
BTdv: On your Web site, you mention having contacted engineers at JVC. What led to that?
DV: After figuring out the progressive scan technique using frame integration mode and the external shutter, I was researching the various cameras that had this mode. Then I came across the new (at that time) JVC DV5000, which unlike the DV500, had slow shutter speed capability. I wanted to find out what method they used to achieve the slow shutter speed, because one possibility was that they did it with frame integration mode. If that was true, it would greatly simplify the whole project: I could just add the shutter and eliminate having to interface and assemble all the discrete components! The published specifications were no help, and the service manual was not yet available, so I called the west coast engineer and asked him.
BTdv: Was he helpful?
DV: He just gave me a complete run-around, something like "Well, the DV5000 uses complex CCD timing algorithms..." I wasn't sure if he just didn't know, or if he didn't want to facilitate the enabling of progressive scan on the DV5000. So I tried the east coast engineer and got almost exactly the same line. Then I decided to purchase the DV5000 and try itand I found that it does not use frame integration.
I found out much more recently that the DV5000 (and many other camcorders) probably can be made to do frame integration, but it would require significant internal modifications to the CCD driver circuit. At some point I'll probably try that, but not until I'm done shooting my current film.
BTdv: Explain the shutter design on the VC25P.
DV: The shutter itself is a five-inch semi-circular disk of thin polymerized cellulose...Okay, it's black railroad board spray-coated with polyurethane to make it waterproof! I made it out of cardboard just for the prototype, intending to make the real one out of aluminum, but it works so well, I just can't justify making a new one.
The shutter motor is from a four-inch cooling fan. I cut off the blades and bonded the shutter disk to the hub of the motor. It's a 12VDC brushless motor, which is ideal for this application because it's easy to control and it's quiet.
The shutter driver/sync circuit brings the motor up to the correct speed (1500 RPM), then locks it to the vertical sync pulses from the camera. Because it's a 4-pole motor, the shutter can lock in one of two positions, either the correct position or 180° out of phase. For this reason, the motor/shutter assembly is attached to a rotatable mount, which allows the assembly to be rotated 180° to achieve the proper phase. (Alternatively, you can just restart the motorit will come up and lock in the correct position fifty percent of the time.) The phase match is necessary for transfer to film. If the phase is wrong then the film transfer will reintroduce the interlace artifact.
|
BTdv: Why is PAL the best format for video-to-film transfers?
DV: Since PAL at 25fps is so close to film's 24fps, the transfer can be done frame-for-frame, by slowing the PAL footage by four percent. For progressive scan imaging, PAL and true 24P are the only standard formats that can be successfully transferred to film. That is, if you shoot in 30P (NTSC), there is no way to convert that to 24fps without noticeable jerkiness because of the mismatched frame rates. NTSC to film only works for 60i (interlaced), via the 3:2 pulldown scheme.
BTdv: How is audio adjusted from a frame rate of 25 to a frame rate of 24?
|
|
|
|
|
|
|
|
|
|
|
|
DV: Some filmmakers have just slowed down the audio and lived with the pitch change. It's not as noticeable for dialogue, but pretty awful for music. I've yet to do a film transfer, but I understand that the slow-down and pitch correction is a fairly straightforward process with off-the-shelf software. For my own music cutting, I plan to speed up the audio by four percent and cut the scenes to that. Then when the slow-down is done, the music will match.
BTdv: Would the VC25P benefit someone interested only in Internet streaming and DVD distribution, or is it designed only for video-to-film?
DV: 25P would actually be a detriment to DVD distribution because it would require conversion for NTSC players. As for Internet streaming, I don't know anything about that!
BTdv: Why did you need to build a preamp for the microphone?
DV: I didn't really have to build my own preamp. There are a couple of off-the-shelf ones available, but I wanted an analog VU meter, specifically a little round analog VU meter like those found on most of the professional and broadcast camcorders. And in the context of the VanceCam project, a preamp is a pretty simple add-on, so I went ahead and built my own.
BTdv: How many balanced XLR microphones can be mixed into the camera?
DV: On the latest version of the camera, I opted for a single XLR input, but there is a second line input that can also be used for a microphone, with an external preamp.
BTdv: How does the video signal become digital in the VC25P?
DV: There are a couple of devices (DataVideo and ADS) that take analog video and audio inputs and convert them to DV. I used the ADS Pyro A/V Link because it has component inputs and is fairly small. In the second prototype, I simply bolted it to the side of the camera, but in the current version, I removed it from its housing and mounted the circuit board inside the enclosure. This saves space and looks better, too!
BTdv: In recording directly to a hard disk, do you have instant playback capability?
DV: I can connect a monitor directly to the camera output, or from the analog-to-digital converter. If I'm running on batteries, I would usually connect from the camera. That way I can turn off the hard drive and A-to-D and just run the camera to set up and frame the scene. That saves a lot of power until I'm ready for a take.
I can play back instantly from the hard drive, but I don't. Since I come from a film background, I'm used to not seeing what I've shot until days later. If I need to look at a shot, I can, but it's a crutch I try to do without.
BTdv: What is the storage capacity of the on-board hard disk?
DV: The CAPDIV hard drive holds 40GB, which equates to about 3 hours of DV. For editing, I transfer the files via Firewire to an external 120GB hard drive.
BTdv: What editing software is best suited to the files generated by the VC25P?
DV: The DV format is so universalone of the many really nice things about itthat any of the popular editing packages can be used. I am using Adobe Premiere, mainly because it works with my operating system, Windows 98SE. It's an excellent program, but if I were running Windows XP Pro, I would probably switch to Avid, because it is the software for film-style DV. (Avid won't run on Windows 98SE.)
BTdv: What modifications do you plan to incorporate in subsequent DV cameras?
DV: There may be a window of opportunity to go to the next level, which would be to incorporate progressive scan on a 2/3" CCD 16:9 format camera. However, the cost savings would not be as dramatic as the VC25P, because the 2/3" cameras are quite a bit more expensive to start with. Also, the price of true 24P cameras is coming down, while features and performance are rising, so the window of opportunity is closing fast. After I finish my indy feature (my first), I will definitely want to shoot my next one in HD 24P. Whether that means building or buying remains to be seen.
|
VC25P: Analog-to-digital board
|
BTdv: What do you think of the innumerable minor DV festivals which have emerged in the past few years? Do you think the traditional film festival model can benefit DV artists and storytellers?
DV: I think the proliferation of small DV festivals was inevitable given the accessibility of DV technology.
Film festivals have begun to accept DV entries, so eventually what we think of as "film festivals" will become format-neutral venues, and we won't have the film/DV distinction.
BTdv: Terminology has been a point of conflict among DV artists and storytellers. "Film," "filming," "filmmaking," and "footage" are problematic. "Taping," "capturing," "rendering," "acquisition medium," "electronic cinema"these are some of the entries in the DV lexicon. What do you think is an ideal solution to these semantic inconsistencies? Is there a particular system of terms you favor?
DV: Well, we can always fall back on the most enduring, if simplistic, all-encompassing "movie"! And of course there's the slightly more formal "motion picture." We've finally reached a period where technology is giving us a viable alternative to the hundred-year-old film format. So it's natural that we try to define the new techniques, processes, and imagery in terms of what they are replacing, hence the use of terms like "electronic cinema," "electronic film production," and "on-line editing." It's probably not the best idea, but historically it's a moving, fully-loaded freight train, so it's best not to stand on the tracks trying to stop it with an outstretched hand. "Cine" and "cinema" are actually format-neutral terms that have become synonymous with "film" simply because they were associated for many years with no rival formats to confuse the issue. So eventually we may drop "electronic," and "cinema" may become acceptable on its own to refer to both celluloid and electronic motion pictures. On the recording end, "capture" seems to be an apt description for any recording format. If I were building my ultimate ideal camera (and I may!), I would simply call it a motion picture camera, regardless of the technology used.
BTdv: What do you think about non-commercial Web-hosted DV festivals or commercial sites featuring DV shorts?
DV: I think they are like car-enthusiast ralliesfun and diverting. But don't expect to find a Ford recruiter there wanting to hire you.
BTdv: How does the Internet figure into your work? Do you find it frustrating or convenient, hegemonic or indispensable?
DV: It is an awesome, indispensable communications tool. Right now it's still in its infancy, and the control and abuse issues (spam, fraud) have to be dealt with, but it's overwhelmingly a positive, useful tool. And of course, without the Internet, there would be no eBay, and without eBay, it's quite likely there would be no VanceCam!
BTdv: What do you think of the recent appropriation of DV technology by mainstream filmmakers?
DV: I don't think "appropriation" is an accurate characterization. Since the dawn of motion pictures, filmmakers have always been on the leading edge of applying current technology to the filmmaking process. The move to DV by some well-known filmmakers is just a high profile example of this process. It was inevitable, and was being used by lesser known filmmakers for a long time before Lucas started using it. He just brought it into the limelight. It's all good.
BTdv: In 2003, Forrester Research, Inc., reported that Internet downloads would soon replace DVD purchases. The International Recording Media Association has contradicted the Forrester report, saying, "From the current annual level of 2.7 billion discs produced worldwide, by 2008, annual factory shipments of DVDs globally will approach 7 billion units. Adams Media Research projects DVD players in 91% of U.S. homes by 2008." What do you think of this debate?
DV: Right now, buying a DVD is much more convenient than downloading. It will probably stay that way for at least the next few years.
BTdv: What do you think of marketing minor digital video endeavors on DVD for the home market without pursuing traditional exhibition?
DV: I guess that's a viable alternative for some producers. Generally though, there's a reason why an endeavor didn't go the traditional distribution route, and often it's simply because it's nothing anyone wants to see.
BTdv: No matter how far-fetched, tell us what you see as the ultimate long-term goal for digital video technology in terms of distribution, exhibition, and public reception.
DV: First of all, "digital video" is the technology of the moment, so to project long term, you have to be willing to give up on that term to allow for whatever's next. While it will almost certainly retain digital aspects, there will also be new technologies/methodologies that are just unknown right now. I suppose we're looking at completely wireless, fully electronic distribution and exhibition soon, but I hope that there will always be an equivalent to the motion picture theater experiencesitting in a dark room with a large number of strangers, watching a drama unveil on a huge screen. There's no substitute for that.
|
|
|
|
|
|
|
|
|