Support for cassette equalisation and noise reduction is due for release in Stereo Lab Version 4.2.

Compact Cassettes – the world’s dominant music medium

After the gramophone record, the Compact Cassette is arguably the second-most significant analogue storage-medium for music. Invented by Philips in 1962 as a tape for speech recording for dictation, the cassette gradually improved over the next twenty years. , By the mid-1980s, cassette sales surpassed LPs worldwide. The global reach of the format was remarkable. By the 1980s, cassettes were more affordable and durable than records, so in much of the developing world, cassettes became the dominant format – sometimes well into the 21st century. Its legacy is even more significant. Home taping changed the music industry’s business model forever, and the compact cassette launched today’s personal, mobile music culture.

Cassette mechanism

The compact cassette is a miscellany of synthetic polymer materials. In the device, magnetically coated polyester (PET) tape is passed between two small spools made of polyacetal (POM) to which the leader tape is firmly attached. These spools along with small, POM tape guide-rollers (running on nickel-plated steel, or stainless steel pins) are held inside a protective polystyrene shell (shown open, above). The flangeless, POM hubs are separated from the polystyrene housing by a PTFE liner (slip sheets) to avoid edge damage to the tape.

The cassette is snapped into position inside the player such that the capstan protrudes into the shell behind the tape, and the head-and-pinch-roller assembly are pushed into the void in the body of the shell at the front. Contrary to the majority of open-reel tape, in the compact cassette, the tape is wound oxide-out on the spools. The position of the cassette in the mechanism is aligned by two pins in the baseplate which engage with the radiused square holes adjacent to the two capstan openings.

The cassette mechanism – with the cassette opened and the heads pushed into the play position – reveals the full tape path. From left to right we can see the erase head, followed by the combined record/playback head and the capstan and pinch roller on the right.

The mechanism is shown stopped. When set to play (or to record) the whole lower assembly is physically pushed into the cassette housing. The heads thereby contact the tape against the spring of the pressure pad, and the pinch roller pushes against the tape and capstan. The capstan and pinch roller mechanism is detailed in Philips’ original cassette patent (right).

The use of pressure pads to hold the tape against the record or replay heads is considered unsatisfactory in open-reel tape recorders. It increases the drag on the tape and can accelerate head wear. However, in the case of cassettes, the original patent specification incorporated the pressure pad, and it is always present. The pressure pad is necessary because there is no control of the rear tension of tape prior to the heads as there is in all open-reel recorders. So, the tape wouldn’t wrap properly around the head unless is pushed there by the pad. The pad assembly consist of a phosphor bronze, leaf spring with a felt cushion.

Philips only specified the size and pressure of the pad in their licence paperwork, (1 mm thick, density ≈ 0.25 g/cmł and a pressure of 15 to 20 grans-force on the tape), but not the felt material. Charmingly, for a format built on an assortment of plastics, the material of the felt was often natural wool fibre in early and high-quality cassettes, right up to the 1980s, although polyester felt (Terylene) became common in the 1990s. The fibre nap of the pad must be very tight so that loose fibres do not work their way between the head and the tape to cause dropouts of the audio signal.

Note the shiny mu-metal shield behind the pressure pad. This is an integral part of the cassette tape design - its role is to shield the tape head from stray magnetic fields. The signals derived from the replay head from a narrow tape running at low speed with a tiny gap are very small indeed (≈ 0.3mV RMS for a 1kHz tone with recorded flux -12dB from tape saturation), and “hum” is an ever-present menace – especially given the enormous level of amplification tape equalisation requires at low frequencies. As cassette decks improved (with better head shielding, lower field leakage, and lower impedance heads), the shielding role became less critical, and manufacturers switched to using plain steel or other cheaper alloys for the backing plate. By the 1980s, only high-quality blanks retained proper mu-metal shields.

Track arrangement

In a stereo audio cassette, four (0.6mm) audio tracks are recorded across the width of the tape: two (stereo left and right) in one direction, and two in the other. This means that the cassette is “flippable” like a record, and plays both sides, hence the symmetrical design of the cassette shell.

By means of a slow tape speed of 17/8 inches/second (4.76 cm/s) and very thin tape, a typical cassette runs for 30, 45, or 60 minutes of audio per side (called C60, C90, and C120 formats respectively). Cassette tape thicknesses are lower than the tapes for professional machines: 18µm for C60 cassettes, 12µm for C90 tapes, and 9µm for C120 cassettes. The C60 cassette contained 90 metres (300 feet) of tape.

The performance of the original analogue cassette was very poor, and way below the performance of contemporary gramophone-records. However, its performance was improved by Dolby B, and by the very considerable improvement and innovations in tapes themselves, as we shall see.

Cassette players were among the most widely produced consumer electronic audio devices in the late 20th century given their presence in portable radios, car stereos, “boom boxes”, portable players, and different home setups. Combined with mechanical refinements (better shells, slip sheets, and precision hubs), various innovations meant that by the early 1980s, the compact cassette rivalled LPs in fidelity for many non-critical users. And, thanks to its portability and recordability, the cassette briefly became the world’s dominant music format.

Cassette tape tabs

Cassette shell design incorporated tabs in the top rear edge of the cassette housing, one above each reel. These were to prevent accidental erasure or to prevent a new recording to be made over a previous precious recording. They are sometimes called erasure prevention or write-protect tabs.

When the tab is intact (left), a sensor pin on the record deck is depressed when the cassette is inserted, and the machine allows recording. If the tab is snapped off, the hole is exposed, the recorder’s pin penetrates the hole and disables the record control or circuit. (If you subsequently want to record again, you can “fix” it by covering the hole with a bit of adhesive tape.)

As cassette technology advanced and different tape formulations were introduced, it was considered desirable that the machine knew automatically what type of tape was to be recorded (each required different bias levels) and played back (each required different equalisation) and the tab system was extended to indicate to the machine what bias and EQ was required.

The photograph right illustrates the standardised notches for automatic tape selection: There are two notches for Type II (next to the write-protect tabs), and four for Type IV. The Type I cassette (and Type III) tape is top, Type II is in the middle, and Type IV at the bottom. The position of these notches was standardised in IEC 60094 (originally designated IEC 94).

Tape formulations

Gamma ferric oxide (γ-Fe2O3) coatings remained the baseline magnetic coating of all magnetic tapes from the 1940s to the 1970s. Prior to 1940, tape coatings (all open reel at that stage, of course) were based on Fe3O4 – the naturally occurring iron oxide named magnetite, the most magnetic of all the naturally occurring minerals. Naturally magnetised pieces of magnetite, called lodestone (illustrated), are how ancient peoples first discovered magnetism. It was an obvious choice as the basis for a magnetic layer to be applied to tape.

The switch to gamma ferric oxide (γ-Fe2O3, sometimes called maghemite) was made because maghemite can be produced in considerably finer particles than ground magnetite. The gamma in γ-Fe2O3 refers to a specific crystal structure of iron(III) oxide — one that differs from the more common α-Fe2O3 (hematite) which is only very weakly magnetic. The γ-Fe2O3 crystal structure is similar to Fe3O4 (magnetite), and the production of tape dopes involve, specially prepared magnetite being oxidised to gamma-ferric oxide.

You may be puzzled. If gamma ferric oxide has a similar crystal structure to magnetite, how is it possible to produce smaller particles? The reason is that maghemite (γ-Fe2O3) is easier to grind than magnetite (Fe3O4) because its crystal structure is full of tiny imperfections. When magnetite oxidises into maghemite, some of its iron ions are lost, leaving empty sites – or vacancies – within the lattice. These missing atoms weaken the network of iron–oxygen bonds and introduce strain that makes the crystal more brittle. Magnetite, by contrast, has a tightly ordered structure. In short, maghemite grains are more likely to fracture when ground or milled. Stick and ball models of the unit cells of magnetite and maghemite are illustrated.

Typical oxidation conditions consist of “slow cooking” the magnetite at a temperature of 200–300°C in a rotary kiln for many hours in an atmosphere of moist air or a carefully mixed flow of air and inert gas, like N2. Nitrogen acts as a diluent and atmosphere moderator, controlling how much oxygen the magnetite actually “sees,” and therefore how fast the oxidation proceeds. It also provides a safety margin — it reduces the risk of local overheating or combustion of fine powders (which are pyrophoric when partially oxidised). The 200–300 °C range used for converting Fe3O4 to γ-Fe2O3 is chosen because it is deliberately below the temperatures that causes the phase transformation to α-Fe2O3 (hematite). That happens above ≈ 350°C.The dimensions of the resulting iron oxide particles are in the region of 0.1 - 0.75µm with length-to-diameter ratios from 3:1 to 10:1.

In the professional audio world (studios, broadcasting, and mastering), gamma ferric oxide tapes for open-reel machines were overwhelmingly dominant, right up until the advent of digital recording technology. But the extreme limitations of the cassette tape medium drove substantial improvements in tape formulations for this market.

Why cassettes drove innovation

Magnetic tape’s intrinsic noise arises because of the essentially granular, or particulate, nature of the magnetic material in the tape coating. The phenomenon is essentially a statistical problem, related to the distribution of the individual signal elements. This is best appreciated by comparing a fast tape-speed, reel-to-reel recording with a recording on a standard cassette (assuming the same tape formulation). A professional recorder recording a 2.5mm track on a machine running at 15ips (38cm/s) encounters thirty times more grains (and thereby magnetic domains) per second than the cassette player, with a track width of 0.6mm running at 4.76cm/s.

Noise scales inversely with the square root of the number of domains interacting with the head at any instant, so the cassette tape is six (√33.6) times, or 15dB, more noisy than the same tape on the open-reel machine. Now, 15dB is a very severe loss of dynamic range – remember professional tape machines were limited (without Dolby A) to a dynamic range of about 60dB at this time. Clearly, the compact cassette was destined to remain a very limited quality medium without innovation on the tape front.

On the other hand, the format was becoming increasingly popular with consumers. In the heyday of the cassette, the market for cassette tape was truly enormous. A Washington Post article states the worldwide market for blank cassette tapes was over one billion units/year in 1977. The economic incentive to develop new tape formulations to improve the performance of the format was thereby enormously significant.

It is to the chemists, as much as to the precision mechanical engineers, that we owe the development of the compact cassette medium from a format for office dictation to the hi-fi medium it became by the 1980s.

Cassettes and Stereo Lab

The Stereo Lab app’ offers tools to maximise the quality of the recordings.

Upper time constant

Cassettes only used two upper time constants: 120µs for IEC Type I (ferric) tapes, and 70µs for Type II and Type IV cassettes. Because of all the phenomena that prejudice against high frequency information from the cassette format (low tape speed, head-gap loss, azimuth, HF saturation), we advise recording all tapes (of all types) with 120µs equalisation applied. This maximises high frequencies off tape. If your deck does not allow EQ selection manually, you will have to block the equalisation slots on the cassette tape to stop the machine applying 70µs equalisation automatically. If the tape requires 70µs EQ, then use TEQ E equalisation to apply this equalisation to your head-drop recording in Stereo Lab.

For similar reasons, don’t select Dolby noise reduction on the machine: this is better applied in the app’ (see below). The difference between the transfer functions of 120µs and 70µs equalisation (normalised at low frequency) are illustrated.

Lower time constant

When Philips introduced their Compact Cassette in 1963, the format specification called for 120µs playback equalisation time constant (treble roll-off) and 1590µs low-frequency turnover time constant (100 Hz). The 1590 µs pole provided bass roll-off to limit head contour effects and keep hum low.

By the mid-1970s, as tape formulations improved and head design advanced, the cassette became more capable at low frequencies. So, to extend bass response, the IEC redefined the low-frequency playback equalization constant to 3180 µs (50 Hz). This change became part of IEC Publication 94 (first issued in 1976). After that, most quality decks aligned to the new standard. By the early 1980s, 3180µs was universal in practice, and 1590µs was obsolete. All the more surprising then that the National Association of Broadcasters should have adopted the 1950µs long-time constant for cassette equalization in 1976, at exactly the same time the commercial standards (IEC94) switched to 3180µs.

The transition from the NAB 1590µs standard to the IEC 3180 µs standard in broadcasting was gradual and varied across different regions and organisations. In North America, the NAB 1590µs equalisation remained the standard for cassette tapes well into the 1980s. The adoption of the IEC 3180µs standard was more prevalent in Europe and other regions.

This divergence leads to compatibility issues when tapes recorded under different standards are played back on machines set to the wrong equalisation curve. To address this, some tape recorders were designed with switchable or jumper-selectable EQ settings, allowing users to choose between NAB and IEC equalization to match the source tape. It is not known if any of these models survived: we know of none.

TEQ F equalisation correction in Stereo Lab allows 1590µs equalisation to be applied to recordings which have been made on a machine with 3180µs equalisation – which basically means, every machine manufactured since 1976.

Noise reduction

The popular analogue cassette noise reduction is implemented in software in the Stereo Lab app’. We implemented an algorithm which is based on the Wireless World Dolby Noise Reducer.

One of the problems with hardware Dolby B decoding is that, in order to work properly, the audio must be decoded at the correct reference level. This is often not the case in hardware tape recorders - especially with old cassettes. This accounts for the poor performance of many hardware implementations. The result is usually an unacceptable degree of treble-loss when the noise reduction is engaged.

The Wireless World Dolby decoder accomplished this with the inclusion of meters and line-up tones. In Stereo Lab, we take advantage of modern digital signal processing to ensure that the audio is first of all normalised so that it is at a reference level and with therefore be decoded properly. If you use TEQ E or TEQ F equalisation correction, be sure to apply these filters prior to Cassette noise reduction because incorrect EQ upsets the Dolby normalisation.

B-yond noise reduction

B-yond noise reduction is designed to decode Dolby C encoded material. Here again, we take advantage of modern digital signal processing to ensure that the audio is first of all normalised so that it is at a reference level and with therefore be decoded properly. If you use TEQ E or TEQ F equalisation correction, be sure to apply these filters prior to B-yond noise reduction because incorrect EQ upsets the Dolby normalisation.

CATHODE

CATHODE stands for CAssette Tape Hiss-Oriented Dynamic Expander – a system developed by Pspatial Audio to enhance digital recordings made from cassette replays. It is a single-ended system, so it may be used as a tape noise reduction system when no noise-reduction encoding has taken place.

Azimuth Dragon

The Nakamichi Dragon cassette deck is justifiably venerated for many reasons, but most of all for its automatic azimuth correction mechanism, which Nakamichi named NAAC (Nakamichi Auto Azimuth Correction).

NAAC uses a servo mechanism, driven by a DC motor and a complex gear train, to move a pivoting playback head to correct for azimuth misalignments. If the azimuth error exceeds a certain threshold, this is detected by phase-detection circuitry and drives the servo to adjust the head's position to minimise the error. The process typically takes between 1 to 5 seconds, but it is generally inaudible to the listener, who is simply aware of an improved treble response. NAAC is capable of correcting azimuth errors up to 12 arc minutes (0.2 degrees) – enough to cause a significant loss in high-frequency response.

Pspatial Audio’s Azimuth Dragon is a software version of Nakamichi’s hardware system. It is included in Stereo Lab from version 4.n. It too is capable of correcting automatically for azimuth errors of up to 0.2°. Azimuth Dragon is selected as a special EQ setting in the tape processes menu.

If you want to go on reading about cassettes, different tape formulations, even a DIY cassette deck for archiving, there's much more these subjects (and a lot of other things too!) in the new edition of the Needle-drop Handbook due out in early 2026.



Links

Pspatial Audio Home page

For all support issues, go here.

For Pspatial Audio sales, email: sales@pspatialaudio.com



© Pspatial Audio 2025. All rights reserved. Apple Certified Developer. Stereo Lab, Aria 51, Aria 20, Head Space, Groove Sleuth, iLOOP and FRANCINSTIEN T-Sym are trademarks of Pspatial Audio. FRANCINSTIEN and Bride of FRANCINSTIEN (BoF) are trademarks of Phaedrus Audio. Macintosh and the Mac logo are trademarks of Apple Computer, Inc.