Jones Frame Buffer Prototype
While working for the Experimental Television Center I built the first four Jones Frame Buffers with David. Pictured here is the one I built for myself. The first few buffers where built into the same card racks that were ubiquitous at the ETC at that time, later David rewired most of the buffers into newer larger boxes. This one might be the only one that stayed in the old rack.
My Jones frame buffer is the same as the other first generation buffers we built for the ETC except that it has additional second level control. Another difference between the other Jones buffers is the addition of 2v differential video inputs and output. At first the frame buffer was controlled by a Sinclair ZX81 micro computer. Plans were made to update to a Z81 big board (or Xerox 820), but when the Amiga 1000 came to market David modified the I/O and software to control the frame buffer from the Amiga. This was a great improvement, with the Sinclair simple programs were written in basic. The only way to save a program was to use audio tape. For the most part the Sinclair was used to simply poke a memory location to put the buffer in the write or read mode and control came from analog patches on the panel.
The Jones frame buffer is a unique device that offers digital processing with analog control. The internal keyer cuts between a live digital plane and the memory plane. This version had 4 memory boards with 64 frames of memory. Later Jones added a pseudo color option which is also present in this version. I and others were writing routines in Amiga Basic that were compiled to C to control the buffer, however no one ever built the libraries that were needed to blank switching, so the routines tended to be a little noisy.
(This also appears here.)
I was approached by Athena Christa Holbrook about academic research she was doing. My responses to here questions around the frame buffer follow:
Q: What drew you to the concept of using a frame buffer tool in your work?
There were several draws, including: getting to work with time, and slow down the video rate, the texture, and it operated within the context of an analog module with a patchable front panel.
At that moment, when the buffer was first being developed, what we had access to for the most part were analog tools. The Jones Frame buffer allowed one to embrace using a digital tool, but within a real time analog framework. This was a leap. For one thing, the frame buffer offered the ability to play with the frame rate and to animate. With straight analog the frame rate was locked to 60 fields/sec, 30 frames/sec. With the frame buffer one could achieve an animation feel equivalent to using 2s, 3s, 4s, etc, each buffer frame holding for some number of video frames. Simply this was the grab rate. Later it even had a tricky patch with a dongle on the Amiga that allowed switching between a fast update - the computer program determined rate, and then a hold using a square wave clock pulse. Artistically these elements created a different temporal feeling.
At the ETC, Before the Jones buffer we had a CAT frame buffer, it was also low resolution, and the CAT had flash frame artifact that made using it maddening. It had no second level control, was a single frame (or maybe 2), it was a bit primitive.
The idea of a buffer now is ubiquitous, for instance there’s one embedded in a Panasonic WJ MX-50, a small switcher from years back often used by artists. Most buffers had more resolution than the Jones, even around that time, think Failrlight with 512 x 512 pixels, but the Jones was one of the first. I recall we used the very first, and then the second generation of, memory chips being sold for the then very new IBM PC market. And later, as with the WJ MX-50, typically the resolution is indistinguishable from the analog image, or nearly so. But there is something about the low resolution quality of the old buffers, of the Jones Buffer, that was, and remains, very appealing. You were able to affect the texture of an image in ways that couldn’t be done with the analog tools we had at the time. The Jones was 4 bit, 16 shades of gray B&W capture, and had pixel resolution of 256 by 256. This is very course by today’s standards. I would probably pick as an ideal something with a bit more resolution but I still like the crude texture it created.
Later a pseudo color board was added. This allowed one to pick from a larger range of colors but still each color was being applied to one of the 16 grey level slices. I am not sure I used the pseudo color module very much in my work, though I did add the board to my buffer.
The Jones buffer also had the ability to store some number of frames, 64 in my case (four memory boards filled with early memory chips). One could capture a series of frames, which became an animation you could play back, versus a single frame store where you can sequentially grab a single frame (like the WJ MX-50). Once this was connected to the Amiga, one was able to move images between the Buffer to the Amiga, and back again to the buffer pages from the Amiga. The Jones also had a built in keyer allowing one to layer between the live image, post A to D, and the stored image.
What was most appealing though was that the Jones buffer fit well into a modular real time analog context, allowing voltage control patches to the front panel to modulate gain (contrast) and pedestal (brightness) of the incoming signal, and the key clip. I also added second level control to the normal/reverse and on/off for the keyer, a common practice with my Jones prototypes at the time.
When we first built the buffer it was controlled by a Sinclair ZX81 computer. I recall writing programs for this which needed to be saved to audio tape, there wasn't a floppy or hard drive. There was a custom Jones interface to adapt the computer to the buffer’s data bus. Later David wrote the buffer software for the Amiga platform along with a newer hardware interface. This allowed for writing software to control buffer parameters.
The next level of manipulation was to write programs that could peek and poke addresses to change the buffer look up table. For instance I wrote a program to animate pulsing from positive to negative. Even compiled in C though the transitions were a bit glitchy as no one ever got around to writing the libraries needed to confine switching to the vertical, or horizontal, interval, and the rate was a bit limited.
Q: What were your experiences working with Dave on building and using these buffers (early prototypes and later iterations)?
This was the most complex device I recall building with David. I taped up the boards using actual tape on mylar or acetate, I forget what the material was. This was all pre-CAD. The complexity made it different. With analog devices even if I didn’t participate in any of the design I had a very good sense of what was going on. Here so much was in PAL chips hidden to all but David. I built all of the early prototypes, one for David, one for the ETC, one for Ralph Hocking, one for Peer Bode and my own. The boards were originally designed for a rack that Ralph Hocking had bought a good many of from a surplus source. Later David migrated the boards to a larger better looking case. As far as I know, my buffer is the only one still in the original footprint.
My version of the buffer, and all my Jones prototypes, were a little different. I was using a differential 2vpp video standard which placed video patching on the front panels. David was really good about working with me on my unique modifications for my version of all the hardware. Well, of course this was his idea, it’s just that I was the only one that ended up building with this format instead of using standard single ended video signals. This was forward thinking of David. Patching single ended video was difficult, while the standard I used made for video patch panels paralleling the +-5 analog patching.
The buffer was really the most annoying to build as the boards were sandwiched together in a way that made it hard to make changes and debug problems. But bottom line, as I wrote above, the buffer was the most opaque of the Jones devices I worked with and it was the most complicated and dense, and, I suspect this had a lot to do with the how early the technology was that we were using.
Q: How many/which works of yours were created, entirely or in part, using any iteration of the frame buffer, or even earlier developments such as the A-to-D/D-to-A converter, or the line buffer?
Any additional details, memories, or technological descriptions you might have of creating these works and working with the buffers would be wonderful!
Most all my work included use of the buffer. I didn’t often just use the buffer, and then the colorizer, like many others did, sometimes, but I usually, often, had an analog layer on top, colorized, and then the buffer layer behind, often black and white. As common as this was for me it did vary. Notice the end of Bug-Eyed Ramrod reversed the layering. I can’t count the works. I never showed all my recordings. They number in the hundreds. I still use the buffer in the same way with new work. Adding a course layer in the background pulsing and slow, as a backdrop.
I didn’t have access to the simpler A-to-D/ D-to-A, basically a buffer without the memory. The early one David created with Gary Hill had the ability to flip the bit relationships. I recall watching a taped letter David sent to Ralph and Sherry from Woodstock, it must have been mid 70s. I was greatly impressed by this tape, amazed by this new trick. Some of the programs I wrote for the buffer animated something like this bit flipping. Alas I can’t find a working floppy that has any of these programs, and now I would probably not like the crudeness, one can achieve better with a Max Jitter Patch these days.
I liked to add a bit of high frequency oscillation into the buffer pedestal input for added texture, and when I can get it to work, the amiga dongle was a nice trick.
Q: What, in your mind, has been the most influential aspects of the buffer over the years (on your work, on others' work, and/or on the history of video processing)?
I prefer the term the history of image processing. Has that term fell out of use? With it I mean a movement within the making of video art characterized by the work coming out of the ETC, the Vasulka's studio, those associated with Dan Sandin, Gary Hill’s work, at least early work, etc. As I wrote above, the buffer took one foot and planted it into a digital realm. Now we were able to use hybrid analog digital systems to image process in a real time context. Time manipulation was probably the biggest piece. Texture another. Before these early digital tools understand you were stuck with the onslaught of 60 fields a second. How else did one break out of this? I used to shoot on film and then use an analytic projector to control the speed and re-photograph and image process, all in real time. Re-scan was the the other alternative. Think of works using reel to reel playback re-scanned. Some years later, students of mine, fighting the same limitations, would play back in slow motion from a VHS deck onto a monitor and then re-scan the monitor. This way they could play with the texture through the re-scan and play with time using the primitive controls of early VHS slow motion. Honestly, these works could be less course and more interesting chromatically than the buffer. But the buffer was a digital leap. I see so many works that were just buffer, buffer and colorizer.
Q: Do you still have any iteration or prototype of the buffer?
Yes I still have my frame buffer. (I gave here a link to this page). I have added a nice clean front panel since the picture shown on this page.