Skip to content

Slightly better C++ design #15

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
alexandrefrancois merged 1 commit into from
Oct 25, 2025
Merged

Slightly better C++ design #15

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
130 changes: 52 additions & 78 deletions Sources/OscillatorsCpp/ResonatorBankVec.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,10 @@ using namespace oscillators_cpp;
constexpr float PI = 3.14159265358979323846; // PI
constexpr float twoPi = 2.0 * PI;

ResonatorBankVec::ResonatorBankVec(size_t numResonators, const std::vector<float> &frequencies, const std::vector<float> &alphas, const std::vector<float> &betas, float sampleRate)
: ResonatorBankVec(numResonators, frequencies.data(), alphas.data(), betas.data(), sampleRate) {
}

ResonatorBankVec::ResonatorBankVec(size_t numResonators, const float* frequencies, const float* alphas, const float* betas, float sampleRate)
: m_sampleRate(sampleRate), m_numResonators(numResonators), m_twoNumResonators(2*numResonators) {

Expand All @@ -39,49 +43,49 @@ ResonatorBankVec::ResonatorBankVec(size_t numResonators, const float* frequencie
constexpr float minusOne = -1.0f;

// initialize from passed frequencies
m_frequencies = new float[m_numResonators];
memcpy(m_frequencies, frequencies, m_numResonators * sizeof(float));
m_frequencies.resize(m_numResonators);
memcpy(m_frequencies.data(), frequencies, m_numResonators * sizeof(float));

// These must be 2 * numResonators size
m_alphas = new float[m_twoNumResonators];
memcpy(m_alphas, alphas, m_numResonators * sizeof(float));
memcpy(m_alphas + m_numResonators, alphas, m_numResonators * sizeof(float));
m_alphas.resize(m_twoNumResonators);
memcpy(m_alphas.data(), alphas, m_numResonators * sizeof(float));
memcpy(m_alphas.data() + m_numResonators, alphas, m_numResonators * sizeof(float));

m_omAlphas = new float[m_twoNumResonators];
vDSP_vfill(&one, m_omAlphas, 1, m_twoNumResonators);
vDSP_vsmsa(m_alphas, 1, &minusOne, &one, m_omAlphas, 1, m_twoNumResonators);
m_omAlphas.resize(m_twoNumResonators);
vDSP_vfill(&one, m_omAlphas.data(), 1, m_twoNumResonators);
vDSP_vsmsa(m_alphas.data(), 1, &minusOne, &one, m_omAlphas.data(), 1, m_twoNumResonators);

m_betas = new float[m_twoNumResonators];
memcpy(m_betas, betas, m_numResonators * sizeof(float));
memcpy(m_betas + m_numResonators, betas, m_numResonators * sizeof(float));
m_betas.resize(m_twoNumResonators);
memcpy(m_betas.data(), betas, m_numResonators * sizeof(float));
memcpy(m_betas.data() + m_numResonators, betas, m_numResonators * sizeof(float));

m_omBetas = new float[m_twoNumResonators];
vDSP_vfill(&one, m_omBetas, 1, m_twoNumResonators);
vDSP_vsmsa(m_betas, 1, &minusOne, &one, m_omBetas, 1, m_twoNumResonators);
m_omBetas.resize(m_twoNumResonators);
vDSP_vfill(&one, m_omBetas.data(), 1, m_twoNumResonators);
vDSP_vsmsa(m_betas.data(), 1, &minusOne, &one, m_omBetas.data(), 1, m_twoNumResonators);

// setup resonators
m_rPtr = new float[m_twoNumResonators];
vDSP_vfill(&zero, m_rPtr, 1, m_twoNumResonators);
m_r.resize(m_twoNumResonators);
vDSP_vfill(&zero, m_r.data(), 1, m_twoNumResonators);

m_rrPtr = new float[m_twoNumResonators];
vDSP_vfill(&zero, m_rrPtr, 1, m_twoNumResonators);
m_rr.resize(m_twoNumResonators);
vDSP_vfill(&zero, m_rr.data(), 1, m_twoNumResonators);

m_zPtr = new float[m_twoNumResonators];
vDSP_vfill(&one, m_zPtr, 1, m_numResonators);
vDSP_vfill(&zero, m_zPtr + m_numResonators, 1, m_numResonators);
m_z.resize(m_twoNumResonators);
vDSP_vfill(&one, m_z.data(), 1, m_numResonators);
vDSP_vfill(&zero, m_z.data()+ m_numResonators, 1, m_numResonators);

float twoPiOverSampleRate = twoPi / m_sampleRate;
m_wPtr = new float[m_twoNumResonators];
vDSP_vfill(&twoPiOverSampleRate, m_wPtr, 1, m_twoNumResonators);
m_w.resize(m_twoNumResonators);
vDSP_vfill(&twoPiOverSampleRate, m_w.data(), 1, m_twoNumResonators);

DSPSplitComplex W = {m_wPtr, m_wPtr + m_numResonators};
DSPSplitComplex W = {m_w.data(), m_w.data() + m_numResonators};
// multiply 2 * PI / sampleRate by frequency for each resonator
vDSP_vmul(W.realp, 1,
m_frequencies, 1,
m_frequencies.data(), 1,
W.realp, 1,
m_numResonators);
vDSP_vmul(W.imagp, 1,
m_frequencies, 1,
m_frequencies.data(), 1,
W.imagp, 1,
m_numResonators);

Expand All @@ -90,24 +94,9 @@ ResonatorBankVec::ResonatorBankVec(size_t numResonators, const float* frequencie
vvcosf(W.realp, W.realp, &count);
vvsinf(W.imagp, W.imagp, &count);

m_alphasSample = new float[m_twoNumResonators];
m_smPtr = new float[m_numResonators];
m_rsqrtPtr = new float[m_numResonators];
}

ResonatorBankVec::~ResonatorBankVec() {
delete [] m_frequencies;
delete [] m_alphas;
delete [] m_omAlphas;
delete [] m_betas;
delete [] m_omBetas;
delete [] m_rPtr;
delete [] m_rrPtr;
delete [] m_zPtr;
delete [] m_wPtr;
delete [] m_alphasSample;
delete [] m_smPtr;
delete [] m_rsqrtPtr;
m_alphasSample.resize(m_twoNumResonators);
m_sm.resize(m_numResonators);
m_rsqrt.resize(m_numResonators);
}

float ResonatorBankVec::frequencyValue(size_t index) {
Expand Down Expand Up @@ -136,7 +125,7 @@ void ResonatorBankVec::getPowers(float *dest, size_t size) {
{
throw std::out_of_range("Buffer passed to getPowers() is not large enough");
}
DSPSplitComplex R = {m_rrPtr, m_rrPtr + m_numResonators};
DSPSplitComplex R = {m_rr.data(), m_rr.data() + m_numResonators};
vDSP_zvmags(&R, 1, dest, 1, m_numResonators);
}

Expand All @@ -145,34 +134,34 @@ void ResonatorBankVec::getAmplitudes(float *dest, size_t size) {
{
throw std::out_of_range("Buffer passed to getAmplitudes() is not large enough");
}
DSPSplitComplex R = {m_rrPtr, m_rrPtr + m_numResonators};
DSPSplitComplex R = {m_rr.data(), m_rr.data() + m_numResonators};
vDSP_zvmags(&R, 1, dest, 1, m_numResonators);
int count = static_cast<int>(m_numResonators);
vvsqrtf(dest, dest, &count);
}

void ResonatorBankVec::update(const float sample) {
vDSP_vsmul(m_alphas, 1, &sample, m_alphasSample, 1, m_twoNumResonators);
vDSP_vsmul(m_alphas.data(), 1, &sample, m_alphasSample.data(), 1, m_twoNumResonators);

// resonator
vDSP_vmma(m_rPtr, 1,
m_omAlphas, 1,
m_zPtr, 1,
m_alphasSample, 1,
m_rPtr, 1,
vDSP_vmma(m_r.data(), 1,
m_omAlphas.data(), 1,
m_z.data(), 1,
m_alphasSample.data(), 1,
m_r.data(), 1,
m_twoNumResonators);

// Smoothing with betas
vDSP_vmma(m_rrPtr, 1,
m_omBetas, 1,
m_rPtr, 1,
m_betas, 1,
m_rrPtr, 1,
vDSP_vmma(m_rr.data(), 1,
m_omBetas.data(), 1,
m_r.data(), 1,
m_betas.data(), 1,
m_rr.data(), 1,
m_twoNumResonators);

// phasor
DSPSplitComplex Z = {m_zPtr, m_zPtr + m_numResonators};
DSPSplitComplex W = {m_wPtr, m_wPtr + m_numResonators};
DSPSplitComplex Z = {m_z.data(), m_z.data() + m_numResonators};
DSPSplitComplex W = {m_w.data(), m_w.data() + m_numResonators};
vDSP_zvmul(&Z, 1,
&W, 1,
&Z, 1,
Expand All @@ -185,11 +174,6 @@ void ResonatorBankVec::update(const std::vector<float> &samples) {
update(sample);
}
stabilize(); // this is overkill but necessary
// compute amplitudes
// DSPSplitComplex R = {m_rrPtr, m_rrPtr + m_numResonators};
// vDSP_zvmags(&R, 1, m_powers, 1, m_numResonators);
// int count = static_cast<int>(m_numResonators);
// vvsqrtf(m_amplitudes, m_powers, &count);
}

/// Process a frame of samples.
Expand All @@ -200,11 +184,6 @@ void ResonatorBankVec::update(const float *frameData, size_t frameLength, size_t
update(frameData[i]);
}
stabilize(); // this is overkill but necessary
// compute amplitudes
// DSPSplitComplex R = {m_rrPtr, m_rrPtr + m_numResonators};
// vDSP_zvmags(&R, 1, m_powers, 1, m_numResonators);
// int count = static_cast<int>(m_numResonators);
// vvsqrtf(m_amplitudes, m_powers, &count);
}

/// Process a frame of samples.
Expand All @@ -215,20 +194,15 @@ void ResonatorBankVec::update(const float *frameData, size_t frameLength, size_t
update(frameData[i]);
}
stabilize(); // this is overkill but necessary
// compute amplitudes
// DSPSplitComplex R = {m_rrPtr, m_rrPtr + m_numResonators};
// vDSP_zvmags(&R, 1, powers, 1, m_numResonators);
// int count = static_cast<int>(m_numResonators);
// vvsqrtf(amplitudes, powers, &count);
}

/// Apply norm correction to phasor.
/// This can be done every few hundreds (?) of iterations
void ResonatorBankVec::stabilize() {
DSPSplitComplex Z = {m_zPtr, m_zPtr + m_numResonators};
vDSP_zvmags(&Z, 1, m_smPtr, 1, m_numResonators);
DSPSplitComplex Z = {m_z.data(), m_z.data() + m_numResonators};
vDSP_zvmags(&Z, 1, m_sm.data(), 1, m_numResonators);
// use reciprocal square root
int count = static_cast<int>(m_numResonators);
vvrsqrtf(m_rsqrtPtr, m_smPtr, &count);
vDSP_zrvmul(&Z, 1, m_rsqrtPtr, 1, &Z, 1, m_numResonators);
vvrsqrtf(m_rsqrt.data(), m_sm.data(), &count);
vDSP_zrvmul(&Z, 1, m_rsqrt.data(), 1, &Z, 1, m_numResonators);
}
27 changes: 14 additions & 13 deletions Sources/OscillatorsCpp/ResonatorBankVec.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,39 +33,40 @@ class ResonatorBankVec {
private:
float m_sampleRate;
size_t m_numResonators;
float* m_frequencies;
float* m_alphas;
float* m_omAlphas;
float* m_betas;
float* m_omBetas;

std::vector<float> m_frequencies;
std::vector<float> m_alphas;
std::vector<float> m_omAlphas;
std::vector<float> m_betas;
std::vector<float> m_omBetas;

size_t m_twoNumResonators;

/// Accumulated resonance values, non-interlaced real (cos) | imaginary (sin) parts
float* m_rPtr;
std::vector<float> m_r;
/// Smoothed accumulated resonance values, non-interlaced real (cos) | imaginary (sin) parts
float* m_rrPtr;
std::vector<float> m_rr;

/// Phasors
float* m_zPtr;
std::vector<float> m_z;
/// Phasor multipliers
float* m_wPtr;
std::vector<float> m_w;

/// hold sample value * alphas
float* m_alphasSample;
std::vector<float> m_alphasSample;

/// Squared magnitudes buffer (ntermediate calculations)
float* m_smPtr;
std::vector<float> m_sm;
/// Reverse square root buffer (intermediate calculations)
float* m_rsqrtPtr;
std::vector<float> m_rsqrt;


public:
ResonatorBankVec & operator=(const ResonatorBankVec&) = delete;
ResonatorBankVec(const ResonatorBankVec&) = delete;

ResonatorBankVec(size_t numResonators, const std::vector<float> &frequencies, const std::vector<float> &alphas, const std::vector<float> &betas, float sampleRate);
ResonatorBankVec(size_t numResonators, const float* frequencies, const float* alphas, const float* betas, float sampleRate);
~ResonatorBankVec();

float sampleRate() { return m_sampleRate; }
size_t numResonators() { return m_numResonators; }
Expand Down
Loading