View Full Version : Cassette Tape - Max Data Rate

October 4th, 2011, 07:53 AM
Just wondering on the above,

the fastest commercial implementation, 'back-in-the-day', and
the theoretical data rate possible, assuming a basic portable deck only


October 4th, 2011, 03:47 PM
I believe the highest data rates for cassettes were in the vicinity of 1200bps. Although I may be wrong about this..
I would think the theoretical maximum rate for cassette storage would be dependent on the frequency response of audio tape. I imagine the data rates for cassette were chosen to accomodate the average cassette - You could probably squeeze more bps out of a high quality audio cassette.

October 4th, 2011, 03:53 PM
1200 baud Kansas City standard only requires a maximum frequency of 2400 Hz, though. Even cheap consumer cassette equipment can't be that poor, can it? I'd think it was probably more oriented towards the speed at which typical CPUs could decode...

October 4th, 2011, 04:06 PM
Specialized drives used audio cassettes and could go much faster because they used recording more akin to a floppy drive (i.e. saturation recording). I had a Techtran drive that did 2400 bps with no problem.

October 4th, 2011, 10:45 PM
Specialized drives used audio cassettes and could go much faster because they used recording more akin to a floppy drive (i.e. saturation recording). I had a Techtran drive that did 2400 bps with no problem.I also recall the Techtran drives, and at least one other that was usually in a dual-drive cabinet and could go even faster IIRC; I'll have to see if I still have some literature on it.

If we're going to talk about specialized drives my Burroughs cassette drives normally read and write at 8000 bps and can read in fast seek mode at 24000 bps; they use two tracks, one for clock and one for data. They did use special digital cassettes, but I never had any trouble backing up and restoring to/from ordinary cassettes, although they couldn't be used as is because the drives used BOT/EOT holes in the tape.

It also took a long time before modems broke the 2400baud barrier, although with clever encoding methods that could be almost ten times as many bits per second. Unlike modems though, if you had an error reading a cassette there was no way to retransmit the bad block without manually rewinding and starting from scratch, so the speeds and protocols had to be pretty conservative and reliable, especially considering the quality of the average portable cassette player.

October 4th, 2011, 11:05 PM
Probably related--meant to be a replacement for punched tape--2 drives with high-speed record search capability. 2400 is what I recall using it at. The switch on the back went to 9600 bps. But plain audio cassettes, nothing special about them.

I think NCR made a similar drive.

The problem with using a portable audio drive is that you're limited to about 3Khz bandwidth and it's not distortion-free. So much about 1Kbps is very dicey.

October 5th, 2011, 02:18 AM
Sinclair used a system where a '0' was encoded as a short pulse, and a '1' as a longer pulse. The standard ROM load/save routine used a data rate that stored approx 2000 '0's or 1000 '1's per second - so roughly 1500bps if we assume an even mix of zeros and ones.

Speed-loaders that used their own loading routine rather than ROM probably came close to doubling that (2500-3000bps iirc). And I believe it was CPU speed rather than tape quality that stopped it from being faster.

October 6th, 2011, 03:00 AM
Many thanks for the replies on this one. @FishFinger - interesting, thanks - the IBM PC interface is the same as that.

So maybe 8000bps should be achievable? I tried a 3-bit 8kHz AM encoding..., needless to say the noise floor was WAY too high, and I retrieved no content whatsoever. Going to try an RLL 2,7 next.

October 6th, 2011, 05:20 AM
If the medium does not have to be tape then youtube has several videos of high speed loading from CD or MP3 player, I think a speed of 12000BPS was mentioned on one video.
The Amstrad CPC has a high speed rate of 2000 baud from cassette tape with 1000 baud as "safe" speed.
Speed is one thing but tape quality is much more difficult to gauge.