28
\$\begingroup\$

I know that cassettes store an analog signal, but is the signal Amplitude-Modulated (AM) or Frequency Modulated (FM)?

\$\endgroup\$
  • \$\begingroup\$ Shouldn't it be on DSP SE? \$\endgroup\$ – Failed Scientist Oct 21 '18 at 10:47
  • 1
    \$\begingroup\$ @FailedScientist there's nothing digital about your standard audio cassette (though there are digital tape formats too.) \$\endgroup\$ – topo morto Oct 21 '18 at 16:49
  • \$\begingroup\$ @topomorto True true! I missed pretty basic thing here. \$\endgroup\$ – Failed Scientist Oct 21 '18 at 17:38
70
\$\begingroup\$

Neither, there is no modulation involved. The magnetization on the media is directly (and hopefully fairly linearly) related to the waveform amplitude.

There is a high frequency bias signal added to the audio signal, to get the resulting signal on the tape to a linear range of the magnetization curve, but the signal at the head is a sum of the two, not a modulation.

enter image description here

That image is from here, a good description of the process. This is stressing my memory of tape decks I had 40 years ago... The one instance I recall of amplitude modulation (kinda) was Dolby HX, which changed to amplitude of the bias in response to high amplitude audio signals, to keep the resulting signal from going into saturation. https://en.wikipedia.org/wiki/Dolby_noise-reduction_system#Dolby_HX/HX-Pro

\$\endgroup\$
  • 1
    \$\begingroup\$ Thank you! Stupid me didn't even consider that... \$\endgroup\$ – Ian Rehwinkel Oct 18 '18 at 14:00
  • 15
    \$\begingroup\$ That said, although what ends up on the tape is linear, it is recorded onto the tape using some bias signal to get the resulting signal in a linear region of the magnetization curve, and a (usually) AC bias signal is also included in the signal going to the head. There shoudn't be any intermodulation between the two though. en.wikipedia.org/wiki/Tape_bias \$\endgroup\$ – Phil G Oct 18 '18 at 14:07
  • 1
    \$\begingroup\$ It might be appropriate to distinguish between the recording signal and the recorded signal. The signal applied to the tape (the recording signal) by the write head is AM. The signal recovered by the read head (the recorded signal) has lost the bias component and is not modulated (at least not in the sense we're using here). So labelling the process as AM or not is a matter for argument. \$\endgroup\$ – WhatRoughBeast Oct 18 '18 at 22:53
  • \$\begingroup\$ @ WhatRoughBeast I added some details about the bias. \$\endgroup\$ – Phil G Oct 18 '18 at 23:08
  • \$\begingroup\$ Nifty thing is, HX didn't require decoding, so it wasn't Dolby in the normal sense like B/C. It was just a refinement of recording. \$\endgroup\$ – Harper Oct 21 '18 at 22:59
28
\$\begingroup\$

It is neither, the signal is recorded as a magnetized pattern without any form of modulation. This can give an impression of how that works:

enter image description here

I got that from here.

See, no modulation needed, the amount of magnetic flux is proportional to the actual audio signal.

Some HiFi Stereo Video recorders do use FM to record the audio though.

\$\endgroup\$
  • 1
    \$\begingroup\$ That page is not really an accurate description. It is not the particles that are realigned during recording, it is the magnetic domains within the particles (depending on the technology, each particle can have one or more domains, in rough sense they define the analog resolution of the material). They way that is presented on the slide is misleading IMO. The domains are groups of magnetic moments that share a common orientation. It is akin to a crystallization pattern when the material is brought below its Curie temperature. If they kept the random orientation it would have been closer. \$\endgroup\$ – isdi Oct 18 '18 at 15:01
  • \$\begingroup\$ how long does it take for those moments to decay? :o \$\endgroup\$ – Daniel Oct 18 '18 at 19:16
  • \$\begingroup\$ @Daniel If the material is stored at room temperature or lower, the "magnetic decay times" are tens or hundreds of years. In practice, the plastic tape in old recordings deteriorates far quicker, and that often leads to some of the magnetic material being physically stripped off the tape every time it is played. That obviously can produce a rapid decay in the sound reproduction quality. \$\endgroup\$ – alephzero Oct 19 '18 at 18:27
11
\$\begingroup\$

Analog tape recording for audio cassette doesnt require modulation. But since magnetization ability of the particles on the tape does not have linear response at low signal levels of the audio to be recorded, poor response will result during playback.

To prevent this, technique called biasing is applied. This creates very linear response of magnetization for the entire dynamic range required for high fidelity audio.

When recording, magnetic tape has a nonlinear response as determined by its coercivity. Without bias, this response results in poor performance especially at low signal levels. A recording signal which generates a magnetic field strength less than tape's coercivity is unable to magnetise the tape and produces little playback signal. Bias increases the signal quality of most audio recordings significantly by pushing the signal into more linear zones of the tape's magnetic transfer function.

– Wikipedia

So, you may have confused between this high frequency AC bias used in audio recording with modulation.

\$\endgroup\$
  • 4
    \$\begingroup\$ @isdi: To quote wikipedia: "modulation is the process of varying one or more properties of a periodic waveform, called the carrier signal, with a modulating signal that typically contains information to be transmitted". The direct addition of two signals is not modulation under this definition. \$\endgroup\$ – mbrig Oct 18 '18 at 17:18
  • 3
    \$\begingroup\$ To further clarify, a modulated signal will have a spectrum centred on the carrier frequency, with bandwidth/sidebands/etc determined by the modulation. A hypothetical tape recording of a pure 5kHz tone and a 100kHz AC bias with have have a spectrum with content at exactly 5kHz and 100kHz, nowhere else (barring noise). \$\endgroup\$ – mbrig Oct 18 '18 at 17:20
  • 1
    \$\begingroup\$ Biasing looks more like dithering than modulation \$\endgroup\$ – Nayuki Oct 18 '18 at 19:57
  • 1
    \$\begingroup\$ When using high frequency A.C. bias, audio frequency will ride on it without affecting its frequency or its peak to peak amplitude. Only its absolute value above zero will keep on changing according to the audio. This is just arithmetic addition of two waves, not modulation. \$\endgroup\$ – soosai steven Oct 18 '18 at 22:53
  • 1
    \$\begingroup\$ Although the signal sent to the record head merely has a mix of high-frequency and audio frequency signal sent to it, the effect on the tape is very similar to pulse-width modulation. \$\endgroup\$ – supercat Oct 19 '18 at 19:59

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.