Intro msg: Teraks, emulators, reviving cassette data

I'd like to introduce myself to the list. I maintain the somewhat imaginary "Jefferson Computer Museum" at <http://www.threedee.com/jcm>, which has info about my several Terak systems. The Terak was a PDP-11/03-based graphics workstation circa 1978-85. My "Terak Museum" web page is the proud recipient of the "Geek Site of the Day" award for October 16, 1996. I also have historical info about the UCSD P-System, including an emulator, source code and very early Pascal compilers. In the months to come, I will add info about other systems I have, including a Zilog MCZ Z-80 development system, CBM PET and NEC systems. I've also been given permission by Claus Giloi to distribute the C source to his Windows-based Altair and IMSAI emulators. If you haven't seen this, it's a nifty graphical recreation, letting you click on the toggle switches to drive the emulator and watch the LEDs blink. In the years to come, I'd like to enhance it to include virtual peripherals, or with inexpensive recreations like an opto paper tape reader. Recently, I've been researching the possibility of reviving old audio cassette tape storage of computer data. With today's PC audio digitizers and a little software, it should be possible to decode and synthesize tapes in formats such as Tarbell tapes for S-100 systems, 88-ACR, Commodore PET, VIC and C-64, Bell 103 recordings, etc. A software approach would have several advantages: you don't need the original hardware, and it has a better chance of restoring out-of-spec data. A little digging revealed the "soundmodem", a driver for Linux and DOS that is a software-based FSK modem that can handle 300 to 9600 baud in real-time using a SoundBlaster as the digitizer / DSP. It is used by ham radio operators. To experiment, I digitized an old wobbly CBM PET tape and did a bit of post-processing in contemporary sound software and it normalized the volume quite well. I suspect with commercial audio software, one could even invoke filters to remove print-through. At 22 or 44 kHz mono 8-bit samples, there's certainly enough headroom to distinguish these relatively slow-speed signals. I'm sure with some formats, just watching the zero-crossing timing would work. I wonder if this technique could be used to rescue old N-track reel tapes that have become unreadable over the years, by intercepting, digitizing, and post-processing the tape-head signal. What would help the most is to get specifications for the old standards. I don't have any documentation, although I'm digging through the basement to find my old Kilobauds. I sent an e-mail to Don Tarbell, who is apparently still on the Internet, but no response so far. It would also help to see more samples! If you have any old cassettes, please consider digitizing them and sending them to me. In another area of the JCM, I've begun to collect ancient ASCII art program that converts teletype-style overstrike art into Adobe Illustrator documents, which are easy to re-size and print on today's laser printers. I have dozens of pictures from DECUS tapes, but I'm always looking for more, and I'd like to record personal anecdotes about the creation of these old artworks. I'd also like to get an actual print sample of the entire printable font from an ASR-33 teletype, in order to scan and convert it into an authentic Postscript bitmap font. - John