During World War II, one of the Germans' major advances in magnetic recording was the use of an AC bias signal. Adding a high-frequency AC component to the recorded signal improved the linearity of the tape's direct signal response, greatly reducing distortion, but at the cost of adding hiss to what should be silent sections of the tape. In the days before modern digital signal processing, the need for linearity on portions of a signal path made tape hiss a necessary evil.

Digital signal processing, however, has obviously come a long way in the time between cassettes' popularity in the 1980s and today's resurgence. Have there been any efforts at using digitally-computed drive waveforms that would produce less hiss than conventional biasing? I wouldn't expect cassettes to achieve quite the level of performance available with compact discs, but I would expect htere's a lot of room for improvement even when using cheap oxide media.

  • If you take away the character of how recording to tape sounds, all you're left with is a medium that's harder to manipulate than digital audio in an editor. Most of the DSP work surrounding tape hiss is regarding restoration after the fact. For these types of plug-ins you can simply search for 'Noise Reduction' plug-ins on Google. – Simon Bosley Jan 23 '18 at 11:00

Uh, Dolby HX (with the HX-Pro variant presumably used in professional production environments)? It extends the headroom, thus implicitly reducing hiss. Your question sounds like you aren't actually interested in noise reduction systems with components at replay or obviously all other Dolby stuff (by far the most popular noise reduction systems) would equally qualify.

I think that digital processing would not be able to achieve all that much more since the problem with a lack of biasing is that the tape is left in an unpredictable state.

By the way, my own AIWA tapedeck from the 80s does not just have Dolby HX (and B and C as playback+record choice) but also a manual "bias" dial which I can use in connection with tape monitoring to get the most natural reproduction from the tape material currently used. A friend's tapedeck (don't know the brand) actually made test recordings in order to set some core recording values automatically.

However, on cassette tapes tape speed is so small and thus alignment so critical that for best recording quality, the most important ingredient is using the same tape drive for recording and playback because of the high alignment requirements for the heads, something that also made audio freaks poo-poo autoreverse tapedecks before tape altogether fell out of favor. So there is not a lot of leeway in improving the process of prerecording cassette tapes to a degree where it would matter in consumer devices.

| improve this answer | |
  • The problem with Dolby encodings and such is that they require that the playback system include a decoder. My interest was in the possibility of recording a tape which could be played back on ordinary equipment. Head alignment would naturally be an issue, but I would think that could be minimized by having an erase head force all domains perpendicular to the main recording direction and then using a record head that's narrower than the playback head. Another possibility I would think would be opened up with the aid of digital technology might be to follow the record head with... – supercat Jan 20 '18 at 17:43
  • ...some very narrow erase heads (again, perpendicular to the domains used for recording) which would trim the edges off the track. I would think, for example, that boosting the record level on quieter sections of the tape by about 10dB but then trimming off the left and right third of the track would yield a signal that was close to the original level but with 10dB less noise (since only the middle third of the track would have tape hiss). – supercat Jan 20 '18 at 17:46
  • @Supercat - head alignment always was the biggest issue. You can't get past that. Also, doing anything clever with splitting the tracks into thirds is more likely to increase noise. I think your best bet is to look at the technology that exists, and the possible solutions that were rejected as not workable. – Rory Alsop Jan 21 '18 at 15:33
  • @RoryAlsop: From my experience and recollection, blank tape (depending upon how it was erased) often has a lot less hiss than tape which was recorded with silence. I forgot to consider that hiss would be proportional to the square root of the recorded width, making the allowable-peak-signal vs noise ratio on the narrow recorded portion of the tape 5dB worse than when using the full width, but the overall signal level, including hiss, would be cut by 10dB. If the recorded signal were boosted by 10dB to compensate for the reduction in playback amplitude... – supercat Jan 22 '18 at 18:39
  • ...the net effect would be a 5dB improvement in SNR. Of course, that would only work for portions of the tape whose peak amplitude wasn't within 10dB of the maximum the tape could support, but the idea would be to trim the track and boost the audio selectively. Ensuring that such effects were properly balanced may be a little tricky, but I would think that digital mastering techniques should help a lot. – supercat Jan 22 '18 at 18:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.