Discuss the history, art and science of Pixelvision. Detail the technical and marketing issues that gave rise to both its creation and apparent demise, and discuss the unexpected place in which Pixelvision found itself in the hands of alternative mediamakers. This essay will put Pixelvision in a historical, artistic and technical context and will posit the idea that Pixelvision can still be a viable image acquisition tool despite the ability of digital imaging to "ape" its signature look.

"Film will only become art when the materials are as inexpensive as pencil and paper" -Jean Cocteau

"Some little fat girl in Ohio is going to make a beautiful movie with her father's camcorder...." -Francis Ford Coppola, Hearts of Darkness.

While I don't agree with the Cocteau quote, many people feel that the PXL-2000, the formal name for the camera that produces the Pixelvision format, has been the camera that most closely fulfilled Cocteau's thesis. Articles have been written that cite the Coppola quote and Sadie Benning as the little girl that fulfills that idea.


Let's begin with the technical and scientific issues surrounding the PXL-2000 and Pixelvision.

The format of Pixelvision has several unique features, none of which is individually responsible for the "look" of the format.

1. B+W

2. 15 fps
Standard NTSC video is 30 fps, 2 fields per frame, for a screen refresh rate of 60 times per second. This only refreshes the screen 15 times per second.

3. Less than 100 lines of resolution with the image area. (I counted 90) (traditional NTSC video has at most 500 lines of resolution)

4. Frame within a frame
This reduces the information that is needed to be recorded. This is great when the video is projected, because the black around the pixelvision image fades insensibly into the darkness around the edge of the screen. The frame within a frame is probably the first thing one notices about Pixelvision.

The above 4 aspects were built into the design of the Pixelvision image so as to reduce the amount of information that would be needed to be encoded onto (what still boggles my technical mind) a thin piece of audio tape. Granted a 45 minute tape will hold 10 minutes of material because the speed at which the tape travels is 4 times faster than that of normal audio cassettes, but the data stream necessary for accurate pixel reproduction is so narrow that information can be recorded on both sides of the tape.

5.Shutter speed. The camera does very well in low light. If standard NTSC video refreshes the screen every 60th of a second (remember there are two fields for every thirtieth of a second frame, making the two fields happen within that frame, thus resulting in new information every 60th of a second), then the lowest possible shutter speed would need to be 60th of a second. Since the PXL-2000 is only 15 FPS, the lowest shutter it will have is 15th of a second, effectively quadrupling the light gathering capacity of this lowly $100 camera. There is an automatic aperture that adjusts the shutter speed so that in bright light, the images are more discrete, and appear to have a more staccato nature, while in lower light conditions, the images are more blurred.

6. Records both sound and image on audiotape.
We often forget that ALL (or at least most) media is magnetic and therefore if the appropriate recording mechanism is manufactured around a standard piece of media (like recording 8 track digital audio onto an SVHS tape), any information can be recorded onto it, even if it is not the media originally intended for that format.

7. Supports subpixel events
While anything with a total of 100 lines of resolution sounds totally unusable, Pixelvision succeeds because of it's ability to support subpixel events. That is, the dithering that occurs between pixels is not limited to the pixels themselves and when viewed, this dithering has a counter intuitive effect. It is the same effect that anti-aliasing has on fonts on a computer screen. The fonts look best (and are perceived to be sharper or of a higher quality) when there is a small amount of blurring that takes place, instead of seeing the jagged edge of each pixel that makes up the letter.
One example of sub-pixel imaging is when a film is transferred to video. Motion picture film has a much greater resolution than video can support. When we watch a film on video, it would be like going to the movie theater and watching the movie screen through a screen door. After a very short time, one forgets that the window screen is there at all. This is because our brain is looking beyond the window screen and connecting what we are seeing on the movie screen.
Events and imaging of a sub-pixel nature are of paramount importance to making video look less like video and as a by-product more like film, though that is not necessarily the point. While attending the recent Bill Viola retrospective at the Whitney, I was struck by how much video is changing at the crux of the millennium. Viola and his work seemed to be firmly fixed in video as video, when video was so new that no matter what one did, it was considered groundbreaking. It was as if no one had ever heard music before and along comes this guy who says listen to this and he just plays a scale and everyone is oohing and ahing because they've never heard anything like it. But essentially it is without soul.
I would like to offer this new terminology to reflect the growing shift of makers towards dealing with sub-pixel events and imaging, instead of dealing with pre-millennial video technologies, work strategies and concerns. This shift is from a pre-millennial attitude towards a sub-pixellist viewpoint.

8. Pinhole lens for amazing depth of field
I recently met Jim Denault, the cinematographer for Nadja and Another girl, another planet both feature films in which Pixelvision is used heavily and wholly, respectively. He was very excited about the fact that he had modified his PXL-2000 so that he could use standard "C" mount lens with it instead of the tiny plastic bubble of a lens that is standard. While this is certainly a more professional lens, I believe that the charm of Pixelvision is derived heavily from the extreme depth of field which is achieved through that little crappy lens. With the adjustment detailed in number eight, one can actually put pieces of paper ON the lens and have them be in focus (if the paper is translucent enough and the light source behind it is strong enough).
Playing with this extreme depth of field is one of the things that Another girl, another planet started to do but that Nadja totally abandoned. When first using the PXL-2000, the camera seemingly cries out for extremely foregrounding and backgrounding characters, settings and props. Mounting anything but the factory issued lens, robs Pixelvision of this ability.

9. Focusable with user adjustment
As far as I know, from viewing many pixelvision pieces, I am the only maker that has taken advantage of this adjustment. Adjusting the lens for focus necessitates taking off the housing of the camera, which may explain others not wanting to mess with their collectors item or not wanting to jeopardize their working investment.

10. Output is RF and will drive channel 3 or 4 (the jack , though, is RCA and is traditionally used for line level signals and can be confusing when the output jack is the same as the input jack on one's VCR. When you hook them together they don't work, because one needs to purchase a simple adapter from radio shack that doesn't change the signal, but just the jack so that one can plug into the antenna/cable in on the VCR.)

Released in 1987.
Taken off the market in 1989.
Rumor: Many breakdowns
Real reason: Economic issue. The profit margin was shrinking because of an increase in the price of key parts, most notably the recording heads.
Also not marketed well. It seems to have hit a target market with adults who like to do weird videos and not with kids at all. Commercials using the Pixelvision image were produced but never aired.
Cheap! At the time (and probably still) it was the cheapest self contained camcorder for kids or adults. Retailed for like $100. This may have been too expensive for most kids, though it seems like video games are more expensive than that.
Probably the most famous Pixelvision user is of course Sadie Benning. James Benning gave his daughter a PXL-2000 and she used this to document her nascent lesbianism. Youngest person ever to receive a MaCarthur genius grant.

Because of Sadie Benning, Pixelvision initially became a lesbian and gay mediamaker's tool. Festivals sprung up to address the burgeoning use of this quirky camera to tape the confessions of every sad dyke, gay boy and transgenderedvestite.

It wasn't long after the camera was discontinued that Gerry Fialka (PXL this festival) set up his festival devoted to anything made with a PXL 2000. I believe it is in its 10th year if he's still doing it. I've lost touch with festivals and the rag that promotes them, the Independent. I feel I get more useful information out of DV magazine and that is free because I lie about how many computers I have and the purchasing budget for which I am responsible.

But seriously, the Independent always seemed to be begging for scraps from the big media controllers, whereas with the advent of DV we can all make our stuff and not have to worry about where we're going to get the money to pay for this transfer, or that service or renting this thing.

Timeline of selected Pixelvision productions
(please note: this is a very biased list. It is primarily based on the pieces I've heard about from people who are familiar with Pixelvision have mentioned.)

Jollies (1989) - Sadie Benning
A Place Called Lovely (1990) - Sadie Benning
Elegy (1991)- Joe Gibbons
Don from Lakewood (1991)- Erik Saks
Glass Jaw (1992) Michael O'Reilly
People thought I was the victim of a gay bashing simply because of the gay/PXL-2000 connection.
Another Girl, Another Planet (1992) - Michael Almereyda Pixelvision feature with actual distribution deal.
It Wasn't Love (1992) Sadie Benning
Orion Climbs (1994)- Michael O'Reilly Colorized Pixelvision
Strange Weather (1994) Peggy Ahwesh/Margie Strosser
Nadja (1996) Michael Almereyda - One of the few (and possibly only) feature films that mixes standard motion picture film with images acquired in Pixelvision.


The camera is the confessional - One could make the argument that the PXL-2000 was highly influential in the many confessional and diaristic pieces that were made at the end of the 1980's. Even to those unfamiliar with Pixelvision specifically and mediamaking in general, there is a tendency to get right up into the lens. Perhaps this is because the camera is often hooked to a TV and the participant can see their own image and they move in close because that is what looks good, the minutiae.

But maybe it is because it's a toy and is less threatening. The camera is housed in a rugged plastic case and the lens itself looks like it was meant for one of the total toy camera's made by Fisher-Price. Whatever the case the PXL-2000 made Sadie Benning famous. Some say she is the embodiment of the quote by Francis Ford Coppola at the beginning of this essay. And others say that she would never have been so wildly successful had her father, noted experimental filmmaker James Benning, not given her the camera and then promoted and helped her edit the videos.

Using the PXL-2000 and editing, processing and mixing Pixelvision with video and film formats has taught me some life long lessons about making things, art and media.

Economics. Make everything do more than one thing. (I am happiest when one thing does 4 jobs)
Work Backwards. What will it look/sound like when it reaches its final destination exhibition mode? Instead of writing a script first and then composing images to fit the words, compose the words to reflect what is on the screen. And don't stick to one of these methods, use both.
Out of the frame and format. Thinking about sound and image delivery systems and final destination exhibition modes rather than just film or just video.
Inspiration can sneak up on you. Tools can serve many purposes and don't have to ONLY do what they were designed for. If you get a better sound from the water cooler in the office of a nuclear power plant than the plant itself, record it and use it. (see cheat and steal).
In the seams. Sometimes the most interesting and revealing things happen in the seams. Perhaps the Pixelvision image recorded on the audiotape produces tape playback anomalies, that are cooler than the live signal being outputted to your production house friend's Betacam SP rig. Sometimes the broken or malfunctioning contribute things that are more interesting than the properly working, and sometimes this is just annoying. (This doesn't have to apply just to equipment)
Cheat and steal. If the Pixelvision image recorded on the audiotape is not acceptable due to tape playback anomalies, then borrow your production house friend's Betacam SP rig and output to that.
Good sound can rescue poor image, but not vice versa.
Recognize the counter-intuitive when it happens and embrace and examine it.

THOUGHTS ON PIXELVISION I think Gerry White, in writing about my use of Pixelvision for a profile in the Independent (Dec 95), summed it up best when writing about my desire to use formats other than just Pixelvision he paraphrased, "A carpenter wouldn't use his favorite saw to pound a nail".

I think Pixelvision is an amazing tool that has opened up opportunities for myself, but one that I see myself using with more and different image acquisition formats, image processing techniques and more and different forms of narratives.

At one point, I used it exclusively because I couldn't afford any other kind of camera and I thought it looked cool. I still think it looks cool and evokes a certain kind of feeling, and will continue to use it in new and different ways. I could very easily take any moving images I've acquired and process them so they would very closely resemble those of Pixelvision.

I think Pixelvision is entirely counter-intuitive. The general thinking among those that are responsible for designing moving image devices is that more is better. More resolution makes for a better picture. The picture is certainly more robust, but the lessons I've learned from Pixelvision tell me that less is more. It is easier for your eye/mind to digest, more visceral and just cooler looking when you are looking at 15 or even 30 frames per second instead of having new visual information thrown at you every 60th of a second.

Pixelvision can still be a viable image acquisition tool despite the ability of digital imaging to "ape" its signature look

1. The camera lends itself to revealing things in a confessional nature.
2. Video on the web is blurry, small and with a slow refresh rate due to the exceedingly narrow "pipe" through which the information needs to fit. The one HUGE difference between the web and Pixelvision is that with Pixelvision, what you record and see in the monitor (even though it is a small amount of information like that transmitted by the web) is what is going to tape, so the maker is ACTIVELY composing the best possible lighting and composition. With video for the web, makers are creating images from a variety of sources, but aren't working with their images at web resolution. Rather the encoding for the web is a passive process that takes place after the maker is finished creating their piece.

For web video to succeed either one of two things must happen:

1. The bandwidth must increase (like with widespread cable modem use) OR
2. A camera must be developed that could mimic the different delivery speeds of the web so that you could immediately see what your image would look like on the web and make decisions based on that.

The one thing that is being missed here is that the process of acquiring images with Pixelvision is an active one. That is, I am always looking at a monitor when I am shooting, and I adjust lighting and mise en scene to exploit the strengths and minimize the weaknesses of Pixelvision. I make sure the depth of field is appropriate to what I'm trying to say with the shot, I make sure any movement is appropriate for the shutter speed, and above all, I make sure what we are seeing is "reading" in the definition deprived frame of Pixelvision. This is the main reason that images acquired and processed to look like Pixelvision invariably fall down. They were not acquired with a mind to how they were going look ultimately. So footage that relies on it's amount of definition in the acquisition format to be "read" (say, the direction in which eyes are turned to indicate another character's position), will ultimately fail to convey any meaning if translated to Pixelvision. This is why Pixelvision can still be seen as a viable image acquisition format.

That is if you can still find one that is working.


All of the material in this essay is from personal experience and memory. It is highly subjective and biased.