I recently found 2 book covers. First a new one about KI depicting an IBM 1401 panel (I first hadn’t recognized it although I have an emulator. Very interesting is the display showing the different instruction length. More about the emulator on an upcoming thread).
Then I found one or 2 covers with lineprinter print outs. Maybe from a 1401 as well. I had to think of the lineprinter arts. Most famous one was probably Madonna with child (first? published in Best of Creative Computing Vol 1, 1976, page 158, there are also some more including a cat and Mr. Spock on the cover). I remember having printed it myself on a C64 printer, I don’t remember my source. (Click to enlarge)
Printing a text file is easy, but I wonder how a photo or painting got into a computer/mainframe. This kind of art was famous in the 80s with the rise of the home computers and printers for everyone. According wikipedia there were kiosks in malls where someone could have print it. I assume this was in the early 80s or late 70s. But I think the first images were printed in ~1961.
Unless if they were created manually (what I don’t think), there must be some hardware like a scanner. I assume that the source image was the same size than the print, so here ~30x80 cm. First I thought that the scanner must be the same size, but one line would be enough and then pulling the image upwards.
But they also needed a software, deciding what character to print.
Does anybody know on what computer it was done first, who did it, what was the scanner (patent?) and what the software (in the 60s and 80s maybe different). And who remembers the kiosks? I don’t remember any in Germany.
There is also (modern) typewriter art, which is even more impressive. With many characters printed between lines and over another. But that is usually be done by an artist and not a computer. Would be interesting if it could be done by one. An early one is this
I’m pretty sure they were typed in manually. The nose area rendered with only M and C looks manual rather than anything algorithmic. If it were an algorithm, you’d expect the range of characters to be more varied.
You know how “photo-realistic” painters can paint something without a scanner? Same thing.
However, it’s also possible they used some assistance from a ruler, or an overlay with regularly spaced hole punches.
Note that this skill is basically pixel art, which goes back thousands of years (invention of weaving, loom, etc). In particular, punch card technology was, at the time, commercially used for pixel art textile mass production. So there would have been plenty of people skilled with it, including professional artists.
Also, typewriter art was a popular hobby at the time, albeit not something of particular value as far as Capitalism would measure. Nevertheless, there would be any number of people who could type that stuff “free-hand”.
There were scanners dating back to the 1960s and before, but they were analogue (for making duplicator stencils or transmitting images for facsimile reproduction). They also tended to be helical scan: the photodetector was driven by a lead screw, while the image to be scanned was attached to a rotating drum. It may have been possible to adapt this technology to produce low-density greyscale images. High resolution drum scanners came much later.
The image density of printed characters is fairly well known (or easy to work out). pbmtoascii has been doing it since 1988, which is almost retro. Typewriter artists such as Bob Neill (who I think we’ve mentioned here before) worked it out manually.
Linotype artists were also doing this in the early part of the 20th century:
Apologies for not linking to the actual Inland Printer pages. While they are all on the Internet Archive, each volume is thousands of pages thick and rather hard to search.
I remember (comparably) cheap scanner heads for home computer printers. This was basically a simple sensor to be mounted on the print head. The printer would feed the image line per line as it would usually feed the paper and the head mechanism stepped the sensor over each line.
I guess, if you were creative, you could have come up with a similar concept using a teletype. Moreover, you’d get a rather precise per character rasterization for free. There is a certain probability that these printer scan heads didn’t emerge out of the blue and may have had a predecessor in a hypothetical setup like this.
Edit: There’s evidence for rather sophisticated scanners at Lawrence Livermore National Laboratory (PDP-1 EYEBALL) and MIT (TX-2 EYE), compare the images at the bottom of this page. Even video signals were grabbed successfully on a PDP-1 by the early/mid 1960s.
The madonna is quite simple. But the Spock looks very photo realistic.
Yes, everything is possible to do manually. But it would be much work, especially when making many images.
Now searching for teleprinter art on Google images, I found the spock image and the history, a scanner was used indeed (1973)
I remember the top left part of that Spock image was still knocking about in the computer science department at University of Strathclyde c. 1998. I can’t remember if it was in simple ASCII art or over-printed like with typewriter art. The original scan (thanks for the article!) used overprinting.
The one I know of was Luna 9, the Russian probe that sent the first pictures from the moon’s surface. It didn’t use ASCII art but transmitted analogue radiofax data. This signal was recognized by listeners at the Jodrell Bank station (UK) and they brought in a radiofax printer from the Daily Express newspaper to publish the images
