Cassette Artifacts

Semantic type: CassetteArtifacts · Category: lo-fi

Models the record and reproduce chain of a compact-cassette deck, including Dolby B/C, saturation, wow and flutter, hiss, dropouts, and azimuth error.

This type can intentionally generate output from zero input at an active setting. See Processing model.

Contract

• Seeded: yes
• Catalog sample rates: not declared; this does not mean unsupported
• Assets: none
• Catalog-declared latency: zero

EffeTune app documentation

The following section is reproduced from the English EffeTune app documentation. Its parameter names and values describe the app UI and can differ from semantic API parameters through transforms or value maps. The generated contract above is authoritative.

Cassette Artifacts

Cassette Artifacts records the music onto a modeled compact cassette and plays it back. The signal passes through the Dolby encoder, the record amplifier with the treble lift and the bounded low-frequency boost it puts onto the tape, the magnetic saturation of the coating, the treble erasure caused by the record bias, the wavelength losses of the reproduce head, local dropouts in the coating, the wow and flutter of the transport, the drifting head azimuth of the deck, the head contour of the reproduce head, and the playback curve that takes the treble lift off again, before tape hiss and modulation noise are added and the Dolby decoder runs. Use it when you want music that has been through a cassette deck rather than music with cassette-like noise placed on top of it.

How It Differs from Other Lo-Fi Effects

• Tape Artifacts models an open-reel studio machine, and the difference between the two is the difference between the formats. At their defaults, with a small signal on a 96 kHz host, the cassette is 2.0 dB down at 8 kHz, 4.4 dB at 12 kHz and 7.9 dB at 16 kHz where the open-reel machine is 0.7, 1.7 and 3.5 dB down. With noise reduction off the cassette’s background is also the louder of the two - -65.5 dBFS against the open-reel machine’s -68.5 dBFS - and with Dolby B or C it drops below it, to -73.6 and -82.8 dBFS, the same relationship the real formats have. Speed is a control there and a fixed 4.76 cm/s here, and Deck Grade, the Type I/II/IV columns, Dolby B/C, the dropouts, the head azimuth and the Dolby level error exist only here.
• Wow Flutter (Modulation) reproduces only the speed variation of a transport. Choose it when you want the wobble without tape saturation, the Type and bias behavior, the noise reduction, or the hiss.
• Saturation and Hard Clipping add nonlinearity on its own, without the frequency-dependent behavior and the transport of a tape machine.
• Vinyl Artifacts, Noise Blender and Hum Generator add a noise layer over unchanged music. Here the hiss, the modulation noise and the dropouts are generated at the correct point in the deck, so the noise reduction acts on them and they follow Tape Type and Hiss the way real cassette noise does.

Sound Character Guide

• Deck Grade sets the class of machine: it moves the band edges and the short-term stability together, and nothing else. Measured at the reference bias with a small signal on a 96 kHz host, the -3 dB points run 13.6 Hz to 18.0 kHz on Reference, 16.7 Hz to 14.0 kHz on Hi-Fi, 19.9 Hz to 10.0 kHz on Consumer and 26.1 Hz to 6.5 kHz on Portable. At 16 kHz that is 2.4, 4.0, 7.8 and 16.7 dB down. The azimuth wobble grows with the same order - none at all on Reference, which is the class of deck that carries an azimuth servo, and largest on Portable, where the top end visibly breathes.
• Record Level is the operating point, not a gain: it states how hard a 0 dBFS peak drives the tape, and the output level does not move with it. What moves is everything else. At the default +9.0 dB, a full-scale 1 kHz tone comes out 3.6 dB rounded off with about 6% third harmonic, and on a dense modern master the top 6 dB of the material comes out as about 4.2 dB. At +0.0 dB the same material comes out almost uncompressed (the top 6 dB stays 5.9 dB) and the tape is never used; by +15.0 dB the top 6 dB has collapsed to about 1.9 dB. The background follows it one decibel for one decibel in the other direction, so raising Record Level buys signal-to-noise and spends dynamic range.
• Bass hits the ceiling first: the record side has to boost everything below 50 Hz to get it onto the tape, and that boost has a ceiling set by Deck Grade, so deep bass reaches saturation before the midrange does. On a Consumer deck at Record Level +12.0 dB, a full-scale tone comes out 7.8 dB down at 20 Hz and 5.4 dB down at 40 Hz against 3.3 dB at 315 Hz. The same asymmetry is why the low end rolls off at all.
• Tape Type is not a volume switch: a small 1 kHz signal comes out within 0.01 dB of the same level on all three. What changes is headroom and noise. Type IV has the most high-frequency headroom - its 10 kHz saturation output is 6.5 dB above Type II and its 315 Hz maximum output level 1 dB above - and at the default Record Level it keeps about 4.9 dB more at 10 kHz than Type II on a -6 dBFS tone, and about 7.5 dB more on a full-scale one; its own noise floor, though, is 2.5 dB worse than Type II. Type I is the noisiest of the three, 4 dB above Type II, and colors small signals exactly as Type II does, so most of what you hear from it is the extra background.
• Noise reduction is a matched round trip: the same sliding-band compander encodes before the tape and expands after it, so under ideal conditions the music comes back out. The quieting is the measured, effective figure rather than the nominal 10 and 20 dB: about 8 dB for Dolby B and about 17 dB for Dolby C, A-weighted, at the default settings. The price is treble mistracking at high levels, where the decoder reads tape-compressed high frequencies as a quieter signal and turns them down further; Dolby C’s anti-saturation shelves reduce it, and at the default Record Level C keeps about 3.9 dB more than B on a 10 kHz tone at -6 dBFS, and about 8.4 dB more on a full-scale one.
• The top octave is the deck’s limit: at the reference bias, with a small signal on a 96 kHz host, the default Consumer deck measures +1.1 dB at 50 Hz (the head contour), 0.0 dB at 1 kHz, -0.8 dB at 5 kHz, -3.0 dB at 10 kHz and -7.8 dB at 16 kHz, and -2.9 dB at 20 Hz. That gentle low-frequency lift over a falling low end and that falling top are the deck’s own response, before anything is driven hard.
• The transport is slower and larger than an open reel’s: a capstan wow at 6.9 Hz, a hub rotation at 0.42 Hz and broadband flutter between 1 and 40 Hz. At the default 0.200% the pitch moves about 9 cents peak to peak, right at the 5 to 10 cents where frequency modulation starts to be heard as such, so the wobble is audible on held notes and masked in dense program; 0.040% is the figure a reference deck publishes and moves only about 2 cents, and the 1.000% top of the range gives about 46 cents, a strong warble.
• A living background: hiss in the gaps, plus modulation noise that rides on the music itself, about 48, 50 and 52 dB below the signal on Type I, Type II and Type IV. Take Hiss all the way down to -92.0 dB re 250 nWb/m when you want a silent background.
• Dropouts are losses, not clicks: each event is a smooth raised-cosine dip of 2.1 to 21 ms and 3 to 30 dB rather than a gate, and the depth draw is weighted toward the shallow end, so what you usually hear is the music briefly ducking rather than a tick. The default 2.0 events/min is about one every half minute on any single track.
• Azimuth and Dolby Level Error are the compatibility axes: both are signed, and their sign is the point. Azimuth darkens the top on both channels and puts a lag between them, and which channel leads follows its sign. A Dolby Level Error above zero makes the decoder cut too little, so the result is brighter and hissier; below zero it cuts too much and the result is duller.

Parameters

Speed is stated rather than selected: the compact cassette runs at 4.76 cm/s (1⅞ ips) by definition, so it is not a control and the status line at the bottom of the panel names it once, as part of the Wow/Flutter readout.

• Deck Grade (Reference, Hi-Fi, Consumer or Portable) - Selects the class of the deck. It governs only the mechanisms that have no control of their own: the record equalization budget that flattens the head’s wavelength loss, the record amplifier’s bandwidth, the ceiling on the low-frequency record boost, the size of the azimuth wobble and the shape of the head contour. It never touches Wow/Flutter, Hiss or Dropouts, so changing it can never discard your own settings. The band edges, measured with a small signal on a 96 kHz host, are 13.6 Hz to 18.0 kHz (Reference), 16.7 Hz to 14.0 kHz (Hi-Fi), 19.9 Hz to 10.0 kHz (Consumer) and 26.1 Hz to 6.5 kHz (Portable), and the azimuth wobble runs 0, 1, 2 and 4 arcmin of standard deviation over the same four. Reference has no wobble at all, because a deck of that class carries an azimuth servo. The default Consumer is an ordinary domestic machine.
• Tape Type (Type I, Type II or Type IV) - Selects the tape formulation: ferric, high-bias and metal. This is a headroom and noise profile, not an equalizer preset and not a level control - a small 1 kHz signal comes out within 0.01 dB of the same level on all three. Type II is the reference column: Type I sits 4.0 dB noisier and Type IV 2.5 dB noisier, while Type IV has 6.5 dB more high-frequency headroom and 1 dB more low-frequency headroom than Type II. Each Type also has its own recommended bias point, so Bias 0 dB means a correctly aligned deck whichever one is selected.
• Noise Reduction (Off, Dolby B or Dolby C) - Selects the companding noise reduction. This is always a matched encode/decode round trip - the same sliding-band law records and plays back - so it quiets the tape without changing the music under ideal conditions. Dolby B is a single sliding band and Dolby C two staggered ones with anti-saturation shelves; the quieting they deliver here is measured rather than nominal, about 8 dB for B and about 17 dB for C, and it depends on Tape Type, Hiss, Dolby Level Error and the host sample rate, which is why the status line reports the figure for the current settings. Both also mistrack loud treble the way a real deck does, and Dolby C mistracks less than Dolby B.
• Bias (-6.0 to +6.0 dB) - Sets the recording bias relative to the recommended point for the selected Tape Type. 0 dB is the correctly aligned deck: it sits 2.5 dB (Type I), 3.0 dB (Type II) or 2.0 dB (Type IV) over the peak of the 10 kHz sensitivity curve, which is where a deck is aligned. Higher (over-biased) settings are cleaner in the low and mid range and duller on top: at +2.0 dB the 10 kHz sensitivity falls 1.67, 1.81 and 1.52 dB on the three Types, and at +6.0 dB it falls 5.31, 5.71 and 4.86 dB. Lower (under-biased) settings are brighter and more distorted, in the way a misaligned deck is, but only down to that peak - about -3.6 dB on Type I, -3.9 dB on Type II and -3.2 dB on Type IV, worth about +2.5, +3.0 and +2.0 dB at 10 kHz - and below it the treble darkens again while the distortion keeps rising, so -6.0 dB is already less bright than -4.0 dB. At 1 kHz the whole range moves less than 0.2 dB, so Bias is not a volume control.
• Record Level (-12.0 to +18.0 dB) - Sets how hard the deck records. The number is the tape level a 0 dBFS peak reaches, in dB above the 250 nWb/m reference flux, and the status line states that convention. The control adds no gain of its own: as long as the tape is not saturating, the same signal comes out at the same level wherever Record Level is set, so what it changes is the tape, not the volume. The default +9.0 dB is a normally recorded cassette, where a full-scale 1 kHz tone comes out 3.6 dB rounded off with about 6% third harmonic and the deep bass is already into the ceiling. Lower settings back off toward a transfer that never uses the tape - at +0.0 dB a full-scale tone loses only 0.5 dB - and raise the background one decibel for every decibel, since the hiss is on the tape and the tape is now further below the peak. Higher settings compress harder and quiet the background by the same rule; past about +15.0 dB the dynamic range stops widening and only the compression keeps growing.
• Wow/Flutter (0 to 1%) - Sets the speed variation of the transport, as a DIN 45507 peak-weighted deviation in percent at the fixed 4.76 cm/s. 0% is a perfectly steady transport. The default 0.200% sits in the middle of the 0.15 to 0.25% window ordinary cassette decks publish and moves the pitch about 9 cents peak to peak, right at the 5 to 10 cents where frequency modulation begins to be heard as such. 0.040% is the weighted peak a reference deck publishes and moves only about 2 cents; the 1.000% top of the range gives about 46 cents, a strong warble. The movement is slower here than on an open-reel machine, because the hub rotation at 0.42 Hz dominates it.
• Hiss (-92.0 to -42.0 dB re 250 nWb/m) - Sets the level of the tape hiss and the modulation noise together, as the A-weighted no-signal flux of Type II with noise reduction off, referred to the 250 nWb/m reference. This is the tape’s own datasheet figure rather than a level at the output: the noise is recorded on the tape, so what it measures at the output depends on Record Level. -92.0 dB re 250 nWb/m switches both off completely. The default -60.5 dB re 250 nWb/m is the bias noise the manufacturer publishes for a Type II tape. Type I sits 4.0 dB above that column and Type IV 2.5 dB above it, Record Level moves the whole thing one decibel for one decibel, and the Dolby decoder then removes its own measured amount, so what you hear in the gaps is not this number - the status line states what it comes to. All of it is ahead of Output, so a meter placed after Output reads it lifted by whatever Output is set to. While the music plays, what this control mostly adds is the modulation noise riding on the signal.
• Dropouts (0 to 20 events/min) - Sets the average rate of coating dropouts, counted per track: whichever single channel you meter, it sees this many events per minute. Half of them are tape-wide and affect every channel together, half are local to one track. Each event is a smooth raised-cosine dip of 2.1 to 21 ms and 3 to 30 dB rather than a gate, so it sounds like a brief loss of signal, not a click. The default 2.0 events/min is a cassette in ordinary service, about one event every half minute on any one track; 0 is a flawless tape and adds nothing at all, and the 20 events/min top is three times the quality-control bound a premium cassette publishes, which is clearly degraded-tape territory.
• Azimuth (-6.0 to +6.0 arcmin) - Sets the head alignment error between the deck that recorded the tape and the deck playing it. It is not a quality grade but the alignment state of that particular pair of machines, which is why it is signed and independent of Deck Grade. Any error costs treble on both channels: at the default +2.0 arcmin a 10 kHz tone loses 0.25 dB and a 16 kHz tone 0.60 dB against a perfectly aligned pair, and at ±6.0 arcmin that becomes 1.03 and 2.26 dB. It also puts a lag of 11.0 µs between the channels at +2.0 arcmin, and the sign decides which channel leads, so a mono sum of correlated material loses a further 0.8 dB at 8 kHz, 1.8 dB at 12 kHz and 3.2 dB at 16 kHz; uncorrelated material shows no such comb. Deck Grade adds a slow drift on top of this setting, so Azimuth is the centre the drift wanders around rather than a frozen value.
• Dolby Level Error (-3.0 to +3.0 dB) - Sets how far the playback deck’s Dolby reference sits from the recording deck’s. It only means anything with Noise Reduction on, and its sign is the point: above zero the decoder reads the tape as hotter than it is, cuts too little, and the result is brighter and hissier; below zero it cuts too much and the result is duller. At the default deck a mid-level tone moves about 2.4 dB down at 5 kHz and 1.0 dB down at 10 kHz at -3.0 dB, and about 1.9 dB up at 5 kHz and 2.2 dB up at 10 kHz at +3.0 dB. 0.0 dB is two decks calibrated to each other. The mistracking on loud treble is there at every setting, because the tape itself changes the signal between encode and decode; what this control opens is the bright side, which a matched pair cannot reach.
• Output (-24.0 to +24.0 dB) - Adjusts the level after the whole chain. Use it to match loudness when you compare with bypass, or to bring back the loudness a high Record Level setting has cost.
• Mix (0 to 100%) - Blends the cassette signal with the original. 100% is full cassette playback. The dry signal is delay-aligned with the tape path, so the midrange blends cleanly - 1 kHz stays within 0.06 dB of unity at 50% - but the top octave does not, because dry and tape no longer share the same phase up there and partly cancel. That cancellation is not the same on both channels, because the azimuth error puts a lag between them: at 50% at the default settings on a 96 kHz host the left channel comes out 1.7 dB down at 8 kHz, 3.6 dB at 12 kHz, 5.3 dB at 16 kHz and 6.2 dB at 20 kHz, while the right channel is 4.4, 8.9, 9.0 and 7.0 dB down at the same frequencies. At 0% the input passes through completely unchanged and the effect adds no latency; at any other setting it adds 165 samples (3.741 ms) on a 44.1 kHz host, 179 (3.729 ms) at 48 kHz, 347 (3.615 ms) at 96 kHz and 683 (3.557 ms) at 192 kHz.

Reading the Status Line

The line under the controls states the Record Level convention and reports what the two Base settings come to on the deck as it is currently configured, in the form Record Level +9.0 dB → tape peak +9.0 dB re 250 nWb/m at 0 dBFS in · Wow/Flutter Base 0.200% → 0.200% at 4.76 cm/s (1⅞ ips) · Hiss Base -60.5 dB re 250 nWb/m → -73.6 dBFS, Type I, Dolby B.

• Record Level restates the control as what it means on the tape: the flux a 0 dBFS peak reaches. It is a statement of the convention, not a meter - there is none, and a quieter master lands correspondingly lower on the tape.
• Wow/Flutter states the Base value and the effective one. The speed is fixed, so the two are always the same figure; the line is there to name the measurement convention the percentage belongs to, and it is the one place the transport speed is stated.
• Hiss states the Base value and the no-signal A-weighted floor it becomes at the output, after the Tape Type column, Record Level and the Dolby decoder, before Output. With Hiss at -92.0 dB re 250 nWb/m the whole noise family is off and the line reads → off.

• The effective floor is measured, not derived from the nominal 10 and 20 dB of Dolby B and C, and it depends on Tape Type, noise reduction, Hiss, Dolby Level Error, Record Level and the host sample rate together. With Hiss at its -60.5 dB re 250 nWb/m default and Record Level at +9.0 dB, on a 96 kHz host, it comes to:

  Tape Type	NR Off	Dolby B	Dolby C
  Type I	-65.5 dBFS	-73.6 dBFS	-82.8 dBFS
  Type II	-69.5 dBFS	-77.7 dBFS	-87.0 dBFS
  Type IV	-67.0 dBFS	-75.2 dBFS	-84.4 dBFS

  Record Level shifts the whole table one decibel for one decibel: at +12.0 dB every figure is 3 dB lower, at +6.0 dB every figure is 3 dB higher. How much the Dolby decoder removes does not change with it, because the tape floor and the decoder’s own reference move together.

• Each combination is measured once and then remembered, so the figure appears immediately for any setting you have already visited. While you drag Hiss or Dolby Level Error through combinations that have not been measured yet, the line reads measuring… and fills in the number once the control stops - announcing a settled-looking figure that is several decibels out would be worse than announcing nothing.

Recommended Settings

1. Ordinary Cassette Deck (default)
   • Deck Grade: Consumer, Tape Type: Type I, Noise Reduction: Dolby B, Bias: 0.0 dB, Record Level: +9.0 dB
   • Wow/Flutter: 0.200%, Hiss: -60.5 dB re 250 nWb/m, Dropouts: 2.0 events/min, Azimuth: +2.0 arcmin, Dolby Level Error: 0.0 dB, Output: 0.0 dB, Mix: 100%
   • The everyday cassette sound, and the plugin’s own default: the top softened by 7.9 dB at 16 kHz, a 1.1 dB lift around 50 Hz over a low end that is already 2.9 dB down at 20 Hz, about 6% third harmonic and 3.6 dB of rounding on a full-scale tone, a -73.6 dBFS background on a 96 kHz host, pitch movement of about 9 cents, and a dropout about every half minute on each track.
2. Reference Deck, Metal Tape with Dolby C
   • Deck Grade: Reference, Tape Type: Type IV, Noise Reduction: Dolby C, Bias: 0.0 dB, Record Level: +9.0 dB
   • Wow/Flutter: 0.040%, Hiss: -60.5 dB re 250 nWb/m, Dropouts: 0.0 events/min, Azimuth: 0.0 arcmin, Dolby Level Error: 0.0 dB, Output: 0.0 dB, Mix: 100%
   • The most capable combination the format offers: a Reference deck reaches -3 dB at 18.0 kHz and does not drift at all, Type IV brings 6.5 dB more high-frequency headroom than Type II, and the background sits at -84.4 dBFS on a 96 kHz host. Type IV’s own tape floor is 2.5 dB worse than Type II’s - it is Dolby C that makes this the quietest setting here, and it is also the setting that mistracks loud treble least. With the wobble, the hiss, the dropouts and the azimuth drift all switched off, this deck is entirely deterministic.
3. Ferric Tape, No Noise Reduction
   • Deck Grade: Consumer, Tape Type: Type I, Noise Reduction: Off, Bias: 0.0 dB, Record Level: +9.0 dB
   • Wow/Flutter: 0.200%, Hiss: -60.5 dB re 250 nWb/m, Dropouts: 2.0 events/min, Azimuth: +2.0 arcmin, Dolby Level Error: 0.0 dB, Output: 0.0 dB, Mix: 100%
   • A plain ferric tape recorded without noise reduction: the background sits at -65.5 dBFS on a 96 kHz host, 8.1 dB louder than the default, and nothing removes it, so hiss is part of the sound in every gap. The tone is exactly the default’s for small signals - the difference between the Types is noise and headroom, not color - and nothing mistracks, because there is no decoder to mistrack.
4. Home Deck, Slightly Over-Biased
   • Deck Grade: Consumer, Tape Type: Type I, Noise Reduction: Dolby B, Bias: +2.0 dB, Record Level: +12.0 dB
   • Wow/Flutter: 0.300%, Hiss: -58.0 dB re 250 nWb/m, Dropouts: 4.0 events/min, Azimuth: +3.0 arcmin, Dolby Level Error: -1.0 dB, Output: +0.5 dB, Mix: 100%
   • An ordinary domestic deck running generic tape, and a tape recorded on a different one: the bias set a little high, so the top is about 1.7 dB darker at 10 kHz than an aligned deck and the low and mid range a little cleaner, a less steady transport, a raised tape floor, a dropout every quarter minute or so per track, a wider azimuth error, and a decoder calibrated 1 dB low, which darkens it further. Record Level is 3 dB above the default, so the top 6 dB of a dense master comes out as about 3.1 dB and Output goes slightly up to pay for the compression. The status line reports what the raised Hiss setting comes to after Type I, Dolby B and the Record Level.
5. Portable, Worn Tape
   • Deck Grade: Portable, Tape Type: Type I, Noise Reduction: Off, Bias: -2.0 dB, Record Level: +12.0 dB
   • Wow/Flutter: 0.480%, Hiss: -54.0 dB re 250 nWb/m, Dropouts: 8.0 events/min, Azimuth: +4.0 arcmin, Dolby Level Error: 0.0 dB, Output: +1.0 dB, Mix: 100%
   • An intentionally degraded lo-fi effect. A Portable deck reaches -3 dB at 6.5 kHz and 26 Hz and drifts twice as much as the default one; the bias is under the aligned point, which brightens 10 kHz by about 1.6 dB and adds distortion with it; the transport is well past the point where the wobble is obvious; the tape is loud and noisy; the azimuth is badly out; and dropouts arrive several times a minute on every track. Record Level is high enough that the bass is firmly into the ceiling, and Output brings the loudness back.

Model Notes

The effect models one recording and playback pass on a deck running at the fixed 4.76 cm/s of the compact cassette. The record side lifts the treble before the tape and the playback side takes exactly the same lift off again, rather than following a published reproduce standard; the low-frequency half of the curve is deliberately not symmetric, because the record-side boost a real deck applies below 50 Hz has a ceiling, and that ceiling is what produces both the low-frequency rolloff and the bass reaching saturation first. Deck Grade governs only the mechanisms that have no control of their own, so it never changes Wow/Flutter, Hiss or Dropouts. The azimuth wobble is a bounded process that is pulled back to the Azimuth setting rather than a random walk, and Reference has none at all, because a deck of that class carries an azimuth servo; with Wow/Flutter at 0, Hiss off, Dropouts at 0 and Deck Grade at Reference, nothing random runs at all. Dolby B and Dolby C are modeled as matched sliding-band companders and always run as a complete encode and decode round trip; there is no encode-only or decode-only operation, and no claim of conformance to, or certification against, any noise-reduction specification. Dolby Level Error offsets the decoder’s reference only, which is the calibration difference between two decks, not a second processing stage. Type III tapes, microcassette and Elcaset formats, other tape speeds, pitch control, auto-reverse, print-through, splice noise, motor hum, shell and mechanism noise, and bleed from the opposite side are outside this model. No public per-tape dropout statistics exist, so the dropout rate range, durations and depths are a calibrated model constrained by a published quality-control bound rather than a transcription of measured data. The tape path carries 165 samples (3.741 ms) of transport and processing delay on a 44.1 kHz host, falling to 683 samples (3.557 ms) on a 192 kHz one; at Mix 0% the input is passed through bit-exactly with no delay at all. The tone figures quoted above are measured on a 96 kHz host at the reference 0.0 dB Bias. This effect costs about one and a half times as much as Tape Artifacts.

Back to all effects