- Generated by
1.15.0
The real-unit vocabulary shared by every DSP block in the catalog. More...
Go to the source code of this file.
Namespaces | |
| namespace | pulp |
| namespace | pulp::signal |
| namespace | pulp::signal::units |
Functions | |
| template<typename T> | |
| T | pulp::signal::units::db_to_linear (T db) |
| Amplitude dB → linear gain. 0 dB → 1, −6.0206 dB → 0.5. | |
| template<typename T> | |
| T | pulp::signal::units::linear_to_db (T linear) |
| Linear gain → amplitude dB, floored at kDbFloorAmplitude so silence maps to a finite −200 dB rather than −inf. | |
| template<typename T> | |
| T | pulp::signal::units::midi_to_hz (T note) |
| MIDI note number → Hz (note 69 = 440 Hz). | |
| template<typename T> | |
| T | pulp::signal::units::hz_to_midi (T hz) |
| Hz → MIDI note number. | |
| template<typename T> | |
| T | pulp::signal::units::semitones_to_ratio (T semitones) |
| Semitones → frequency ratio. +12 → 2, −12 → 0.5. | |
| template<typename T> | |
| T | pulp::signal::units::ratio_to_semitones (T ratio) |
| Frequency ratio → semitones. Inverse of semitones_to_ratio. | |
| template<typename T> | |
| T | pulp::signal::units::cents_to_ratio (T cents) |
| Cents → frequency ratio. +100 → 2^(1/12), +1200 → 2. | |
| template<typename T> | |
| T | pulp::signal::units::ratio_to_cents (T ratio) |
| Frequency ratio → cents. Inverse of cents_to_ratio. | |
| template<typename T> | |
| T | pulp::signal::units::volts_to_semitones (T volts) |
| Pitch CV volts → semitones, on the 1 V/octave standard. | |
| template<typename T> | |
| T | pulp::signal::units::semitones_to_volts (T semitones) |
| Semitones → pitch CV volts, on the 1 V/octave standard. | |
| template<typename T> | |
| T | pulp::signal::units::ms_to_samples (T ms, T sample_rate) |
| Milliseconds → samples at sample_rate Hz. | |
| template<typename T> | |
| T | pulp::signal::units::samples_to_ms (T samples, T sample_rate) |
| Samples → milliseconds at sample_rate Hz. | |
| template<typename T> | |
| T | pulp::signal::units::hz_to_period_samples (T hz, T sample_rate) |
| Frequency → its period in samples. A 100 Hz tone at 48 kHz is 480 samples. | |
| template<typename T> | |
| T | pulp::signal::units::period_samples_to_hz (T samples, T sample_rate) |
| A period in samples → its frequency in Hz. | |
| template<typename T> | |
| T | pulp::signal::units::hz_to_cycles_per_sample (T hz, T sample_rate) |
| Hz → normalised frequency in cycles per sample (f / fs) — the unit a phase accumulator increments by. | |
| template<typename T> | |
| T | pulp::signal::units::hz_to_radians_per_sample (T hz, T sample_rate) |
| Hz → radians per sample (2π·f / fs) — the unit a sin/cos oscillator and most filter design formulas want. | |
| template<typename T> | |
| T | pulp::signal::units::ms_to_onepole_coef (T ms, T sample_rate) |
| A time constant in ms → the a in the one-pole smoother y += a·(x − y), i.e. | |
| template<typename T> | |
| T | pulp::signal::units::ms_to_onepole_pole (T ms, T sample_rate) |
| The complementary pole of ms_to_onepole_coef: p = 1 − a = exp(−1/(τ·fs)), the p in y = p·y + (1 − p)·x. | |
| template<typename T> | |
| T | pulp::signal::units::t60_ms_to_tau_ms (T t60_ms) |
| A 60 dB decay time in ms → the one-pole time constant τ in ms that produces it. | |
| template<typename T> | |
| T | pulp::signal::units::taper_log (T u, T lo, T hi) |
| Geometric ("log") taper: maps a normalised knob position u ∈ [0, 1] onto [lo, hi] so equal knob travel is equal RATIO — the right law for anything heard geometrically (frequency, delay time, drive). | |
| template<typename T> | |
| T | pulp::signal::units::untaper_log (T value, T lo, T hi) |
| Inverse of taper_log: a value in [lo, hi] → its knob position in [0, 1]. | |
| template<typename T> | |
| T | pulp::signal::units::taper_linear (T u, T lo, T hi) |
| Linear taper — the arithmetic counterpart of taper_log, present so a parameter table can name its law instead of open-coding a lerp. | |
| template<typename T> | |
| T | pulp::signal::units::untaper_linear (T value, T lo, T hi) |
| Inverse of taper_linear. | |
| template<typename T> | |
| T | pulp::signal::units::taper_log_inverse (T value, T lo, T hi) |
| Compatibility spelling used by the shared modulation toolkit. | |
| template<typename T> | |
| T | pulp::signal::units::t60_to_per_sample_gain (T t60_seconds, T sample_rate) |
| Per-sample gain for a feedback path that decays by 60 dB in t60_seconds. | |
| template<typename T> | |
| T | pulp::signal::units::beats_to_seconds (T beats, T bpm) |
| template<typename T> | |
| T | pulp::signal::units::seconds_to_beats (T seconds, T bpm) |
| constexpr timebase::BeatDivision | pulp::signal::units::to_beat_division (Division division) noexcept |
| Converts the legacy signal spelling to the canonical timebase vocabulary. | |
| constexpr float | pulp::signal::units::division_to_beats (Division division) |
| constexpr float | pulp::signal::units::division_to_beats (int index) |
| float | pulp::signal::units::division_to_seconds (Division division, float bpm) |
| double | pulp::signal::units::division_to_samples (Division division, float bpm, double sample_rate) |
| float | pulp::signal::units::division_to_hz (Division division, float bpm) |
Variables | |
| constexpr double | pulp::signal::units::kDbFloorAmplitude = 1e-10 |
| Amplitude floor used when converting to dB. | |
| constexpr double | pulp::signal::units::kMidiA4 = 69.0 |
| MIDI note number of concert A. Fixed by the MIDI specification. | |
| constexpr double | pulp::signal::units::kA4Hz = 440.0 |
| Frequency of concert A in Hz — the tuning reference for midi_to_hz. | |
| constexpr double | pulp::signal::units::kSemitonesPerOctave = 12.0 |
| Semitones per octave, and cents per semitone. Fixed by 12-TET notation. | |
| constexpr double | pulp::signal::units::kCentsPerSemitone = 100.0 |
| constexpr double | pulp::signal::units::kSemitonesPerVolt = kSemitonesPerOctave |
| Semitones per volt on the 1 V/octave pitch-CV standard. | |
| constexpr int | pulp::signal::units::kDivisionCount = static_cast<int>(Division::count) |
The real-unit vocabulary shared by every DSP block in the catalog.
Series law 3 says parameters are expressed in real units — Hz, ms, s, dB, semitones, cents — because those are what a musician reads on a knob and what a spec can state a range in. That law is only enforceable if there is exactly ONE conversion from each real unit into the internal quantity a filter or oscillator actually wants. Two hand-rolled pow(10, db/20) call sites are two chances to disagree about the dB floor; two hand-rolled one-pole coefficients are two chances to disagree about whether "20 ms" means a time constant or a settling time. Everything converts here.
Conventions this header fixes, once, for the whole library:
Everything is a template over the sample type so a double block does not silently round its coefficients through float.
RT contract: every function here is a pure, stateless, branch-light scalar computation. Nothing allocates, locks, or performs I/O, so all of it is safe to call per sample on the audio thread. Several use std::pow / std::log2 / std::exp, which are not free — the conversions that appear in a hot path (db_to_linear, semitones_to_ratio) are cheap enough to use per sample, but a block that converts a control value that only changes per block should convert per block.
1.15.0