Hacker News new | past | comments | ask | show | jobs | submit login

I used my 16-bit 192 kHz sound card to determine how much data can fit on a cassette by "prolonging" the length of each bit until I got an acceptable error rate. I found that the absolute best you can do with a decent old Technics tape deck is 8-bit at 12kHz which means for a 90-min Tape you can fit about 60 MB on one tape.

I noticed however that switching between bits produced a "ringing artifact" oscillating at about 60 kHz. If this is caused by my particular setup I don't know. With a better modulation than just PCM this could be compensated and you could pull probably more data out of the stream.




That 60 kHz tone could be the tape bias frequency.

https://en.m.wikipedia.org/wiki/Tape_bias


Thanks for the hint. The next thing to try would be to connect the tape head directly to the IO-Pins of a Raspberry PI and then use a simple manchester encoded signal ramping up the frequency. A Raspberry PI can push up to 1 MHz and has an internal Schmitt-Trigger which would be perfect for reading the data back.


You’re going to have to deal with the nonlinearity of tape yourself, if you go this route.


Holy crap that's way more data than I could ever have imagined being possible on regular cassette.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: