Measure Preserve's splice-alignment geometry on low-frequency material
A splice can only relocate by [0.75, 1.25]*window, so periods with no multiple in that interval never phase-align — at 50 ms, f < 16 Hz and 26.7-32 Hz. Harness runs by hand; too slow to gate.
This commit is contained in:
@@ -0,0 +1,700 @@
|
||||
// Measurement harness for the Preserve engine's behaviour on LOW-FREQUENCY material.
|
||||
// Reports numbers; it renders no perceptual verdict and changes no DSP.
|
||||
//
|
||||
// The question it answers: a splice relocates the read tap by a nominal `window` refined by a
|
||||
// correlation search over +/- maxLag, so the reachable relocation distances form ONE bounded
|
||||
// interval. Phase-aligning a splice needs a WHOLE NUMBER OF SOURCE PERIODS inside that
|
||||
// interval, and for some periods none exists — a geometric limit, separate from the
|
||||
// splice-CADENCE inequality. Both now sit in time_stretch.h; this is what measured them.
|
||||
//
|
||||
// Two findings shaped the sections below and are worth knowing before reading the output:
|
||||
// whether an unreachable multiple accumulates into a DETUNE or only wobbles the phase depends
|
||||
// on whether the nearest multiple misses on one side or straddles (a straddle cancels in the
|
||||
// mean); and PITCH IS THE WRONG THING TO MEASURE HERE — the fundamental usually survives, so
|
||||
// the load-bearing metric is energy outside it. Zero-crossing counting in particular reports a
|
||||
// wrong period on renders whose fundamental is provably correct, which is section C.
|
||||
//
|
||||
// Measures, at both 44.1k and 48k geometry:
|
||||
// A. the reachable relocation interval, observed rather than derived (jump/lag/frac off
|
||||
// every SpliceEvent), and the alignment-reachability predicate over frequency.
|
||||
// B. Voice-level renders at 30 Hz: the shipped default path, then transposition at rate 1.0
|
||||
// and rate 0.5/2.0 with none, then a 20-200 Hz sweep to locate the turnover.
|
||||
// C. the P=500-frame (~88 Hz) case the pitch_shift floor probe fails on, measured with three
|
||||
// independent pitch estimators to separate the two mechanisms from a measurement artifact.
|
||||
// D. a window sweep at 30 Hz — what a larger window would buy, and what it would cost.
|
||||
// E. alignable frequencies under identical conditions, without which D and B have no scale.
|
||||
|
||||
#include "../src/core/instrument/engine/pitch_shift.h"
|
||||
#include "../src/core/instrument/engine/time_stretch.h"
|
||||
#include "../src/core/instrument/engine/voice.h"
|
||||
|
||||
#include <algorithm>
|
||||
#include <cmath>
|
||||
#include <cstdint>
|
||||
#include <cstdio>
|
||||
#include <utility>
|
||||
#include <vector>
|
||||
|
||||
using namespace reasampler;
|
||||
using namespace reasampler::instrument::engine;
|
||||
|
||||
static int g_fail = 0;
|
||||
#define CHECK(cond) do { if(!(cond)) { \
|
||||
std::printf("FAIL line %d: %s\n", __LINE__, #cond); ++g_fail; } } while(0)
|
||||
|
||||
constexpr double kPi = 3.14159265358979323846;
|
||||
|
||||
// ---------------------------------------------------------------------------------------
|
||||
// Source + render helpers
|
||||
// ---------------------------------------------------------------------------------------
|
||||
|
||||
// A pure sine at `freqHz`, phase-continuous, long enough that a rate-2.0 render never
|
||||
// exhausts it (the caller sizes `frames`).
|
||||
static SampleData sineSample(double freqHz, int sampleRate, std::size_t frames,
|
||||
PitchEngine engine, double phase = 0.0) {
|
||||
SampleData s;
|
||||
s.frames.resize(frames);
|
||||
const double w = 2.0 * kPi * freqHz / static_cast<double>(sampleRate);
|
||||
for (std::size_t i = 0; i < frames; ++i) {
|
||||
s.frames[i] = static_cast<AudioSample>(std::sin(w * static_cast<double>(i) + phase));
|
||||
}
|
||||
s.sampleRate = sampleRate;
|
||||
s.rootNote = 60;
|
||||
s.play.pitchEngine = engine; // Gate, no loop, default (fully open) AHDSR
|
||||
return s;
|
||||
}
|
||||
|
||||
// One note through the REAL Voice: presize (off-thread step), start with the stretch rate,
|
||||
// then pull `outFrames` mono frames. `note - 60` is the transposition in semitones.
|
||||
static std::vector<double> renderVoice(const SampleData& s, int note, double stretchRate,
|
||||
std::int64_t window, std::size_t outFrames) {
|
||||
Voice v;
|
||||
v.presizePreserveShifters(window);
|
||||
v.start(note, 127, s, /*declickTakeover=*/false, stretchRate);
|
||||
std::vector<double> out(outFrames, 0.0);
|
||||
for (std::size_t i = 0; i < outFrames; ++i) {
|
||||
out[i] = static_cast<double>(v.renderFrame());
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------------------
|
||||
// Metrics
|
||||
// ---------------------------------------------------------------------------------------
|
||||
|
||||
// Mean spacing between positive-going zero crossings over [from, to) — the same estimator
|
||||
// test_pitch_shift.cpp uses, kept identical so the two files' numbers are comparable.
|
||||
static double periodIn(const std::vector<double>& v, std::size_t from, std::size_t to) {
|
||||
double sum = 0.0;
|
||||
std::size_t prev = 0, count = 0;
|
||||
for (std::size_t i = from + 1; i < to && i < v.size(); ++i) {
|
||||
if (v[i - 1] <= 0.0 && v[i] > 0.0) {
|
||||
if (count > 0) sum += static_cast<double>(i - prev);
|
||||
prev = i;
|
||||
++count;
|
||||
}
|
||||
}
|
||||
return count > 1 ? sum / static_cast<double>(count - 1) : 0.0;
|
||||
}
|
||||
|
||||
// Least-squares fit of a single tone at `cyclesPerFrame` over [from, from+len): returns the
|
||||
// fitted phase and writes the residual energy fraction (1 - explained), which is the
|
||||
// single-tone-purity metric — 0 = a perfect sine at that frequency, 1 = none of the energy
|
||||
// is there. Robust to amplitude but NOT to phase drift within the block, which is why the
|
||||
// caller keeps blocks near one period.
|
||||
static double toneFit(const std::vector<double>& v, std::size_t from, std::size_t len,
|
||||
double cyclesPerFrame, double* residFraction) {
|
||||
double sc = 0.0, ss = 0.0, cc = 0.0, s2 = 0.0, cs = 0.0, e = 0.0;
|
||||
for (std::size_t k = 0; k < len && from + k < v.size(); ++k) {
|
||||
const double t = 2.0 * kPi * cyclesPerFrame * static_cast<double>(k);
|
||||
const double c = std::cos(t), s = std::sin(t);
|
||||
const double x = v[from + k];
|
||||
sc += x * c; ss += x * s; cc += c * c; s2 += s * s; cs += c * s; e += x * x;
|
||||
}
|
||||
const double det = cc * s2 - cs * cs;
|
||||
double a = 0.0, b = 0.0;
|
||||
if (std::fabs(det) > 1e-12) {
|
||||
a = (sc * s2 - ss * cs) / det;
|
||||
b = (ss * cc - sc * cs) / det;
|
||||
}
|
||||
const double explained = a * sc + b * ss; // energy captured by the fit
|
||||
if (residFraction != nullptr) *residFraction = e > 0.0 ? 1.0 - explained / e : 0.0;
|
||||
return std::atan2(b, a);
|
||||
}
|
||||
|
||||
// Total unwrapped phase drift (in CYCLES) of the render relative to an ideal tone at
|
||||
// `cyclesPerFrame`, measured across [from, to) in one-period blocks. This is the direct
|
||||
// observable behind "the rendered pitch is wrong": a nonzero drift IS a frequency error.
|
||||
static double phaseDriftCycles(const std::vector<double>& v, std::size_t from, std::size_t to,
|
||||
double cyclesPerFrame, double* worstStepCycles) {
|
||||
const std::size_t blk = static_cast<std::size_t>(1.0 / cyclesPerFrame);
|
||||
double total = 0.0, prev = 0.0, worst = 0.0;
|
||||
bool first = true;
|
||||
for (std::size_t p = from; p + blk <= to && p + blk < v.size(); p += blk) {
|
||||
const double ph = toneFit(v, p, blk, cyclesPerFrame, nullptr);
|
||||
if (!first) {
|
||||
double d = ph - prev;
|
||||
while (d > kPi) d -= 2.0 * kPi;
|
||||
while (d < -kPi) d += 2.0 * kPi;
|
||||
total += d / (2.0 * kPi);
|
||||
if (std::fabs(d) / (2.0 * kPi) > worst) worst = std::fabs(d) / (2.0 * kPi);
|
||||
}
|
||||
prev = ph;
|
||||
first = false;
|
||||
}
|
||||
if (worstStepCycles != nullptr) *worstStepCycles = worst;
|
||||
return total;
|
||||
}
|
||||
|
||||
// Median single-tone residual fraction over the render, in one-period blocks.
|
||||
static double medianResidual(const std::vector<double>& v, std::size_t from, std::size_t to,
|
||||
double cyclesPerFrame) {
|
||||
const std::size_t blk = static_cast<std::size_t>(1.0 / cyclesPerFrame);
|
||||
std::vector<double> r;
|
||||
for (std::size_t p = from; p + blk <= to && p + blk < v.size(); p += blk) {
|
||||
double resid = 0.0;
|
||||
toneFit(v, p, blk, cyclesPerFrame, &resid);
|
||||
r.push_back(resid);
|
||||
}
|
||||
if (r.empty()) return 0.0;
|
||||
std::size_t mid = r.size() / 2;
|
||||
std::nth_element(r.begin(), r.begin() + static_cast<std::ptrdiff_t>(mid), r.end());
|
||||
return r[mid];
|
||||
}
|
||||
|
||||
// Period of the highest normalized-autocorrelation peak over [minLag, maxLag] — a pitch
|
||||
// estimator that, unlike zero-crossing counting, is not fooled by a low-level fast component
|
||||
// adding spurious crossings. The two disagreeing is itself the diagnosis.
|
||||
static double autocorrPeriod(const std::vector<double>& v, std::size_t from, std::size_t len,
|
||||
std::int64_t minLag, std::int64_t maxLag) {
|
||||
double e0 = 0.0;
|
||||
for (std::size_t k = 0; k < len && from + k < v.size(); ++k) e0 += v[from + k] * v[from + k];
|
||||
if (e0 <= 0.0) return 0.0;
|
||||
double best = -1e18; std::int64_t bestLag = 0;
|
||||
std::vector<double> score(static_cast<std::size_t>(maxLag - minLag + 1), 0.0);
|
||||
for (std::int64_t lag = minLag; lag <= maxLag; ++lag) {
|
||||
double s = 0.0, e = 0.0;
|
||||
for (std::size_t k = 0; k < len && from + k + static_cast<std::size_t>(lag) < v.size();
|
||||
++k) {
|
||||
const double b = v[from + k + static_cast<std::size_t>(lag)];
|
||||
s += v[from + k] * b;
|
||||
e += b * b;
|
||||
}
|
||||
const double r = e > 0.0 ? s / std::sqrt(e0 * e) : 0.0;
|
||||
score[static_cast<std::size_t>(lag - minLag)] = r;
|
||||
if (r > best) { best = r; bestLag = lag; }
|
||||
}
|
||||
// Parabolic refinement so the estimate isn't quantized to whole frames.
|
||||
const std::size_t i = static_cast<std::size_t>(bestLag - minLag);
|
||||
double frac = 0.0;
|
||||
if (i > 0 && i + 1 < score.size()) {
|
||||
const double den = score[i - 1] - 2.0 * score[i] + score[i + 1];
|
||||
if (den < 0.0) frac = 0.5 * (score[i - 1] - score[i + 1]) / den;
|
||||
}
|
||||
return static_cast<double>(bestLag) + frac;
|
||||
}
|
||||
|
||||
// The five strongest spectral peaks over a Hann-windowed segment, scanned on a fine period
|
||||
// grid (Goertzel-style direct evaluation, no FFT-bin quantization). Prints period in frames
|
||||
// and magnitude relative to the strongest — the decisive "is the rendered pitch wrong, or is
|
||||
// there a second component fooling the zero-crossing count" measurement.
|
||||
static void reportSpectrum(const char* label, const std::vector<double>& v, std::size_t from,
|
||||
std::size_t len, double wantPeriod) {
|
||||
const int kGrid = 2000;
|
||||
const double pLo = 30.0, pHi = 8000.0;
|
||||
std::vector<double> mag(static_cast<std::size_t>(kGrid), 0.0);
|
||||
std::vector<double> per(static_cast<std::size_t>(kGrid), 0.0);
|
||||
for (int g = 0; g < kGrid; ++g) {
|
||||
// Geometric grid: constant relative resolution across three octaves of period.
|
||||
const double p = pLo * std::pow(pHi / pLo, static_cast<double>(g) / (kGrid - 1));
|
||||
per[static_cast<std::size_t>(g)] = p;
|
||||
double re = 0.0, im = 0.0;
|
||||
const double w = 2.0 * kPi / p;
|
||||
for (std::size_t k = 0; k < len && from + k < v.size(); ++k) {
|
||||
const double hann = 0.5 * (1.0 - std::cos(2.0 * kPi * static_cast<double>(k) /
|
||||
static_cast<double>(len)));
|
||||
const double x = v[from + k] * hann;
|
||||
re += x * std::cos(w * static_cast<double>(k));
|
||||
im += x * std::sin(w * static_cast<double>(k));
|
||||
}
|
||||
mag[static_cast<std::size_t>(g)] = std::sqrt(re * re + im * im);
|
||||
}
|
||||
double top = 0.0;
|
||||
for (double m : mag) top = std::max(top, m);
|
||||
// Local maxima, ranked by MAGNITUDE (not by grid order) so the fundamental cannot be
|
||||
// pushed off the list by low-level debris at a shorter period.
|
||||
std::vector<std::pair<double, double>> peaks; // (magnitude, period)
|
||||
for (int g = 1; g + 1 < kGrid; ++g) {
|
||||
const std::size_t i = static_cast<std::size_t>(g);
|
||||
if (mag[i] <= mag[i - 1] || mag[i] < mag[i + 1]) continue;
|
||||
if (mag[i] < 0.02 * top) continue;
|
||||
peaks.emplace_back(mag[i], per[i]);
|
||||
}
|
||||
std::sort(peaks.begin(), peaks.end(),
|
||||
[](const std::pair<double, double>& a, const std::pair<double, double>& b) {
|
||||
return a.first > b.first;
|
||||
});
|
||||
// Energy fraction OUTSIDE the fundamental's mainlobe — the honest "how much of this render
|
||||
// is not the wanted tone" number, since a peak list alone can hide broadband debris.
|
||||
double eTotal = 0.0, eFund = 0.0;
|
||||
for (int g = 0; g < kGrid; ++g) {
|
||||
const std::size_t i = static_cast<std::size_t>(g);
|
||||
const double e = mag[i] * mag[i];
|
||||
eTotal += e;
|
||||
if (std::fabs(per[i] - wantPeriod) / wantPeriod < 0.06) eFund += e;
|
||||
}
|
||||
std::printf(" %s spectrum (want period %.1f fr); strongest peaks >2%% of max:\n", label,
|
||||
wantPeriod);
|
||||
for (std::size_t k = 0; k < peaks.size() && k < 8; ++k) {
|
||||
std::printf(" period %8.1f fr rel %.4f%s\n", peaks[k].second,
|
||||
peaks[k].first / top,
|
||||
std::fabs(peaks[k].second - wantPeriod) / wantPeriod < 0.03
|
||||
? " <-- the wanted tone" : "");
|
||||
}
|
||||
std::printf(" energy outside the wanted tone's mainlobe: %.2f%%\n",
|
||||
eTotal > 0.0 ? 100.0 * (1.0 - eFund / eTotal) : 0.0);
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------------------
|
||||
// A. Splice geometry, observed off the shifter's own SpliceEvent stream
|
||||
// ---------------------------------------------------------------------------------------
|
||||
|
||||
struct SpliceStats {
|
||||
long long count = 0;
|
||||
double minReloc = 1e18, maxReloc = -1e18;
|
||||
std::int64_t minLag = 1LL << 40, maxLag = -(1LL << 40);
|
||||
double meanInterval = 0.0;
|
||||
bool jumpAlwaysNominal = true; // |jump| == window on every splice (steady state)
|
||||
};
|
||||
|
||||
// Drives a bare PitchShifter over the same feed schedule Voice uses, recording every splice.
|
||||
// The audio is not kept — this measures the DECISIONS, not the sound.
|
||||
static SpliceStats spliceGeometry(const std::vector<AudioSample>& src, std::int64_t window,
|
||||
double rate, double shift, std::size_t outFrames,
|
||||
std::vector<double>* audio = nullptr) {
|
||||
PitchShifter ps;
|
||||
ps.configure(window);
|
||||
ps.prime(src.data(), window);
|
||||
ps.setShiftRatio(shift);
|
||||
ps.setFeedRate(rate);
|
||||
StretchCursor cur;
|
||||
cur.start(window);
|
||||
loop::ResolvedLoop lp{}; // inactive: the source is long enough to run straight through
|
||||
|
||||
SpliceStats st;
|
||||
if (audio != nullptr) audio->assign(outFrames, 0.0);
|
||||
std::size_t lastSpliceAt = 0;
|
||||
double intervalSum = 0.0;
|
||||
long long intervals = 0;
|
||||
for (std::size_t i = 0; i < outFrames; ++i) {
|
||||
const std::int64_t due = cur.due(rate);
|
||||
AudioSample last = 0.0f;
|
||||
bool fed = false;
|
||||
for (std::int64_t k = 0; k < due; ++k) {
|
||||
if (fed) ps.writeFrame(last);
|
||||
const std::int64_t q = cur.next(lp);
|
||||
last = (q >= 0 && static_cast<std::size_t>(q) < src.size())
|
||||
? src[static_cast<std::size_t>(q)] : 0.0f;
|
||||
fed = true;
|
||||
}
|
||||
const AudioSample o = fed ? ps.process(last) : ps.processNoInput();
|
||||
if (audio != nullptr) (*audio)[i] = static_cast<double>(o);
|
||||
const SpliceEvent& ev = ps.lastSplice();
|
||||
if (!ev.fired) continue;
|
||||
++st.count;
|
||||
const double reloc = std::fabs(static_cast<double>(ev.jump) -
|
||||
static_cast<double>(ev.lag) - ev.frac);
|
||||
if (reloc < st.minReloc) st.minReloc = reloc;
|
||||
if (reloc > st.maxReloc) st.maxReloc = reloc;
|
||||
if (ev.lag < st.minLag) st.minLag = ev.lag;
|
||||
if (ev.lag > st.maxLag) st.maxLag = ev.lag;
|
||||
if (std::llabs(ev.jump) != window) st.jumpAlwaysNominal = false;
|
||||
if (lastSpliceAt != 0) { intervalSum += static_cast<double>(i - lastSpliceAt); ++intervals; }
|
||||
lastSpliceAt = i;
|
||||
}
|
||||
st.meanInterval = intervals > 0 ? intervalSum / static_cast<double>(intervals) : 0.0;
|
||||
return st;
|
||||
}
|
||||
|
||||
// Is there a whole number of source periods inside the reachable relocation interval?
|
||||
static bool alignmentReachable(double periodFrames, double lo, double hi, int* whichN) {
|
||||
for (int n = 1; n <= 64; ++n) {
|
||||
const double m = periodFrames * n;
|
||||
if (m > hi) break;
|
||||
if (m >= lo) { if (whichN != nullptr) *whichN = n; return true; }
|
||||
}
|
||||
if (whichN != nullptr) *whichN = 0;
|
||||
return false;
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------------------
|
||||
// 1. The reachable relocation interval, measured
|
||||
// ---------------------------------------------------------------------------------------
|
||||
|
||||
static void reportReachableInterval() {
|
||||
std::printf("\n=== A. Reachable splice relocation interval (measured) ===\n");
|
||||
for (const auto& g : {std::pair<int, std::int64_t>{44100, 2205},
|
||||
std::pair<int, std::int64_t>{48000, 2400}}) {
|
||||
const int sr = g.first;
|
||||
const std::int64_t w = g.second;
|
||||
// Broadband noise: every lag is a plausible candidate, so the search's own limits —
|
||||
// not the source's periodicity — set the observed extremes.
|
||||
std::vector<AudioSample> noise(600000);
|
||||
std::uint32_t rng = 12345u;
|
||||
for (auto& x : noise) {
|
||||
rng = rng * 1664525u + 1013904223u;
|
||||
x = static_cast<AudioSample>((static_cast<double>(rng >> 8) / 8388608.0) - 1.0);
|
||||
}
|
||||
SpliceStats up = spliceGeometry(noise, w, 1.0, 1.5, 120000); // up-shift: jump +w
|
||||
SpliceStats dn = spliceGeometry(noise, w, 1.0, 0.7, 120000); // down-shift: jump -w
|
||||
const double lo = std::min(up.minReloc, dn.minReloc);
|
||||
const double hi = std::max(up.maxReloc, dn.maxReloc);
|
||||
std::printf(" %d Hz, window %lld frames (%.1f ms):\n", sr, static_cast<long long>(w),
|
||||
1000.0 * static_cast<double>(w) / sr);
|
||||
std::printf(" up-shift splices %lld, lag [%lld, %lld], reloc [%.2f, %.2f]\n",
|
||||
up.count, static_cast<long long>(up.minLag),
|
||||
static_cast<long long>(up.maxLag), up.minReloc, up.maxReloc);
|
||||
std::printf(" down-shift splices %lld, lag [%lld, %lld], reloc [%.2f, %.2f]\n",
|
||||
dn.count, static_cast<long long>(dn.minLag),
|
||||
static_cast<long long>(dn.maxLag), dn.minReloc, dn.maxReloc);
|
||||
std::printf(" observed reachable relocation interval: [%.2f, %.2f] frames "
|
||||
"= [%.2f, %.2f] ms\n", lo, hi, 1000.0 * lo / sr, 1000.0 * hi / sr);
|
||||
std::printf(" structural bound (window +/- window/4): [%lld, %lld]\n",
|
||||
static_cast<long long>(w - w / 4), static_cast<long long>(w + w / 4));
|
||||
// The jump is nominal and the lag is inside +/- window/4 — the two facts the
|
||||
// reachable interval is derived from.
|
||||
CHECK(up.jumpAlwaysNominal && dn.jumpAlwaysNominal);
|
||||
CHECK(up.minLag >= -(w / 4) && up.maxLag <= w / 4);
|
||||
CHECK(dn.minLag >= -(w / 4) && dn.maxLag <= w / 4);
|
||||
}
|
||||
}
|
||||
|
||||
// The reachability predicate over frequency, at both geometries. Pure arithmetic over the
|
||||
// interval measured above — no render, stated as such.
|
||||
static void reportReachabilityByFrequency() {
|
||||
std::printf("\n=== A2. Alignment reachability by frequency (arithmetic, not rendered) ===\n");
|
||||
const double freqs[] = {12, 14, 16, 18, 20, 22, 24, 26, 26.6, 28, 30, 31, 31.9, 32,
|
||||
34, 36, 40, 50, 60, 80, 88.2, 100, 140, 200};
|
||||
for (const auto& g : {std::pair<int, std::int64_t>{44100, 2205},
|
||||
std::pair<int, std::int64_t>{48000, 2400}}) {
|
||||
const int sr = g.first;
|
||||
const std::int64_t w = g.second;
|
||||
const double lo = static_cast<double>(w - w / 4), hi = static_cast<double>(w + w / 4);
|
||||
std::printf(" %d Hz / window %lld, interval [%.0f, %.0f] frames:\n", sr,
|
||||
static_cast<long long>(w), lo, hi);
|
||||
for (double f : freqs) {
|
||||
const double P = static_cast<double>(sr) / f;
|
||||
int n = 0;
|
||||
const bool ok = alignmentReachable(P, lo, hi, &n);
|
||||
if (ok) {
|
||||
std::printf(" %6.1f Hz P=%8.1f ALIGNABLE (n=%d, n*P=%.1f)\n", f, P, n,
|
||||
n * P);
|
||||
} else {
|
||||
// How far the nearest multiple sits outside the interval, and the phase error
|
||||
// that residual forces at every splice.
|
||||
double best = 1e18; double bestM = 0.0;
|
||||
for (int k = 1; k <= 64; ++k) {
|
||||
const double m = P * k;
|
||||
const double d = m < lo ? lo - m : (m > hi ? m - hi : 0.0);
|
||||
if (d < best) { best = d; bestM = m; }
|
||||
}
|
||||
std::printf(" %6.1f Hz P=%8.1f UNALIGNABLE (nearest n*P=%.1f, off by "
|
||||
"%.1f frames = %.1f deg of phase)\n",
|
||||
f, P, bestM, best, 360.0 * best / P);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------------------
|
||||
// 2. Voice-level renders at 30 Hz
|
||||
// ---------------------------------------------------------------------------------------
|
||||
|
||||
// The reassurance case: the SHIPPED default path. 30 Hz played at its root, rate 1.0, no
|
||||
// transposition. The shift is exactly 1.0, so the tap's delay never drifts and no splice can
|
||||
// fire; a primed shifter at unity is a bit-exact pass-through. Baseline is the SAME source
|
||||
// through Varispeed at the root, which is a straight readPos_ += 1.0 read of the PCM — i.e.
|
||||
// the unprocessed sample. This is NOT a comparison against a pre-change binary; it is the
|
||||
// stronger claim that the path is transparent.
|
||||
static void testRootRateUnityIsBitIdenticalToTheDirectRead() {
|
||||
std::printf("\n=== B1. 30 Hz, root note, rate 1.0, no transposition ===\n");
|
||||
const int sr = 44100;
|
||||
const std::int64_t w = 2205;
|
||||
const std::size_t frames = 300000, outFrames = 250000;
|
||||
const SampleData pres = sineSample(30.0, sr, frames, PitchEngine::Preserve);
|
||||
const SampleData vari = sineSample(30.0, sr, frames, PitchEngine::Varispeed);
|
||||
const std::vector<double> p = renderVoice(pres, 60, 1.0, w, outFrames);
|
||||
const std::vector<double> v = renderVoice(vari, 60, 1.0, w, outFrames);
|
||||
std::size_t firstDiff = outFrames;
|
||||
for (std::size_t i = 0; i < outFrames; ++i) {
|
||||
if (p[i] != v[i]) { firstDiff = i; break; }
|
||||
}
|
||||
std::printf(" Preserve vs Varispeed at root, %zu frames: %s\n", outFrames,
|
||||
firstDiff == outFrames ? "BIT-IDENTICAL"
|
||||
: "differ (first at frame ?)");
|
||||
if (firstDiff != outFrames) {
|
||||
std::printf(" first difference at frame %zu (%.9f vs %.9f)\n", firstDiff, p[firstDiff],
|
||||
v[firstDiff]);
|
||||
}
|
||||
CHECK(firstDiff == outFrames);
|
||||
// And the same claim at the 48k geometry.
|
||||
const SampleData pres48 = sineSample(30.0, 48000, frames, PitchEngine::Preserve);
|
||||
const SampleData vari48 = sineSample(30.0, 48000, frames, PitchEngine::Varispeed);
|
||||
const std::vector<double> p48 = renderVoice(pres48, 60, 1.0, 2400, outFrames);
|
||||
const std::vector<double> v48 = renderVoice(vari48, 60, 1.0, 2400, outFrames);
|
||||
bool same48 = true;
|
||||
for (std::size_t i = 0; i < outFrames && same48; ++i) if (p48[i] != v48[i]) same48 = false;
|
||||
std::printf(" same at 48k / window 2400: %s\n", same48 ? "BIT-IDENTICAL" : "DIFFER");
|
||||
CHECK(same48);
|
||||
// Splice count on the same conditions, read off the shifter directly.
|
||||
std::vector<AudioSample> src(pres.frames.begin(), pres.frames.end());
|
||||
const SpliceStats st = spliceGeometry(src, w, 1.0, 1.0, outFrames);
|
||||
std::printf(" splices fired over %zu frames at shift 1.0, rate 1.0: %lld\n", outFrames,
|
||||
st.count);
|
||||
CHECK(st.count == 0);
|
||||
|
||||
// Onset at a MUCH larger window — the cost side of any window-resize option. prime()
|
||||
// parks the tap on src[0] whatever the window, so frame 0 must still be source frame 0.
|
||||
// Started at quarter-phase so src[0] is FULL SCALE, not the zero a sine would give: a
|
||||
// frame-0 match against 0.0 would also pass on a voice that produced silence.
|
||||
const SampleData cosPhase =
|
||||
sineSample(30.0, sr, 300000, PitchEngine::Preserve, kPi / 2.0);
|
||||
CHECK(cosPhase.frames[0] == 1.0f);
|
||||
for (std::int64_t big : {std::int64_t{2205}, std::int64_t{8820}}) {
|
||||
const std::vector<double> up = renderVoice(cosPhase, 67, 1.0, big, 64); // +7 st
|
||||
std::printf(" window %5lld, +7 st: out[0]=%.9f (src[0]=%.9f), |out| over frames 1..63 "
|
||||
"min %.6f\n", static_cast<long long>(big), up[0],
|
||||
static_cast<double>(cosPhase.frames[0]),
|
||||
*std::min_element(up.begin() + 1, up.end(),
|
||||
[](double a, double b) { return std::fabs(a) < std::fabs(b); }));
|
||||
CHECK(up[0] == static_cast<double>(cosPhase.frames[0])); // zero added latency
|
||||
}
|
||||
// A sample SHORTER than the window: start() primes the whole playable span and freezes
|
||||
// the writer immediately. A larger window moves that threshold, so check it still speaks
|
||||
// on frame 0 at the largest window swept below.
|
||||
const SampleData shortSample =
|
||||
sineSample(30.0, sr, 3000, PitchEngine::Preserve, kPi / 2.0);
|
||||
const std::vector<double> shortOut = renderVoice(shortSample, 67, 1.0, 8820, 64);
|
||||
std::printf(" 3000-frame sample under a 8820-frame window, +7 st: out[0]=%.9f src[0]=%.9f\n",
|
||||
shortOut[0], static_cast<double>(shortSample.frames[0]));
|
||||
CHECK(shortOut[0] == static_cast<double>(shortSample.frames[0]));
|
||||
}
|
||||
|
||||
// One measured row: render through the Voice and report every metric for that condition.
|
||||
static void measureRow(const char* label, double freqHz, int sr, std::int64_t window,
|
||||
int note, double rate) {
|
||||
const std::size_t frames = 900000;
|
||||
const std::size_t outFrames = 300000;
|
||||
const SampleData s = sineSample(freqHz, sr, frames, PitchEngine::Preserve);
|
||||
const double shift = std::pow(2.0, (note - 60) / 12.0);
|
||||
const std::vector<double> out = renderVoice(s, note, rate, window, outFrames);
|
||||
|
||||
bool finite = true;
|
||||
double peak = 0.0;
|
||||
for (double x : out) { if (!std::isfinite(x)) finite = false; peak = std::max(peak, std::fabs(x)); }
|
||||
|
||||
const double srcPeriod = static_cast<double>(sr) / freqHz;
|
||||
const double wantPeriod = srcPeriod / shift; // pitch is the TAP's, not the feed's
|
||||
const double wantCpf = 1.0 / wantPeriod;
|
||||
const std::size_t from = 40000, to = 280000;
|
||||
const double gotPeriod = periodIn(out, from, to);
|
||||
double worstStep = 0.0;
|
||||
const double drift = phaseDriftCycles(out, from, to, wantCpf, &worstStep);
|
||||
const double resid = medianResidual(out, from, to, wantCpf);
|
||||
|
||||
std::vector<AudioSample> src(s.frames.begin(), s.frames.end());
|
||||
const SpliceStats st = spliceGeometry(src, window, rate, shift, outFrames);
|
||||
const double lo = static_cast<double>(window - window / 4);
|
||||
const double hi = static_cast<double>(window + window / 4);
|
||||
int n = 0;
|
||||
const bool reach = alignmentReachable(srcPeriod, lo, hi, &n);
|
||||
|
||||
// Effective frequency error implied by the drift, and the phase step it works out to per
|
||||
// splice — the number that says whether a splice is stepping the phase or not.
|
||||
const double driftPerSplice = st.count > 0 ? drift / static_cast<double>(st.count) : 0.0;
|
||||
std::printf(" %-26s f=%6.1f Hz shift=%.4f rate=%.2f | period got %8.2f want %8.2f "
|
||||
"(%+.2f%%) | splices %4lld every %7.0f fr | phase drift %+8.3f cyc "
|
||||
"(%+7.1f deg/splice, worst step %.1f deg) | resid %.4f | peak %.3f | "
|
||||
"align %s%s\n",
|
||||
label, freqHz, shift, rate, gotPeriod, wantPeriod,
|
||||
wantPeriod > 0.0 ? 100.0 * (gotPeriod - wantPeriod) / wantPeriod : 0.0,
|
||||
st.count, st.meanInterval, drift, 360.0 * driftPerSplice, 360.0 * worstStep,
|
||||
resid, peak, reach ? "YES" : "NO",
|
||||
reach ? "" : " <-- no whole period in the reachable interval");
|
||||
CHECK(finite);
|
||||
}
|
||||
|
||||
// Three independent pitch estimators plus the spectrum, on one condition. Where the
|
||||
// zero-crossing count and the autocorrelation disagree, the render is not simply detuned —
|
||||
// something else is crossing zero.
|
||||
static void deepDive(const char* label, double freqHz, int sr, std::int64_t window, int note,
|
||||
double rate) {
|
||||
const SampleData s = sineSample(freqHz, sr, 900000, PitchEngine::Preserve);
|
||||
const double shift = std::pow(2.0, (note - 60) / 12.0);
|
||||
const std::vector<double> out = renderVoice(s, note, rate, window, 300000);
|
||||
const double wantPeriod = (static_cast<double>(sr) / freqHz) / shift;
|
||||
const double zc = periodIn(out, 40000, 280000);
|
||||
// Search bounded to [0.5, 1.7] x the wanted period: a pure sine autocorrelates equally at
|
||||
// EVERY multiple of its period, so an unbounded search reports 2P about half the time.
|
||||
const double ac = autocorrPeriod(out, 60000, 60000,
|
||||
std::max<std::int64_t>(40,
|
||||
static_cast<std::int64_t>(wantPeriod * 0.5)),
|
||||
static_cast<std::int64_t>(wantPeriod * 1.7));
|
||||
std::printf(" %s (f=%.1f Hz, shift %.4f, rate %.2f, want period %.1f fr):\n", label, freqHz,
|
||||
shift, rate, wantPeriod);
|
||||
std::printf(" zero-crossing period %.2f (%+.2f%%) | autocorrelation period %.2f "
|
||||
"(%+.2f%%)\n", zc, 100.0 * (zc - wantPeriod) / wantPeriod, ac,
|
||||
100.0 * (ac - wantPeriod) / wantPeriod);
|
||||
reportSpectrum(label, out, 60000, 131072, wantPeriod);
|
||||
}
|
||||
|
||||
static void reportTransposedAt30Hz() {
|
||||
std::printf("\n=== B2. 30 Hz transposed, rate 1.0 (44.1k / window 2205) ===\n");
|
||||
measureRow("30 Hz +2 st", 30.0, 44100, 2205, 62, 1.0);
|
||||
measureRow("30 Hz +7 st", 30.0, 44100, 2205, 67, 1.0);
|
||||
measureRow("30 Hz -7 st", 30.0, 44100, 2205, 53, 1.0);
|
||||
std::printf("\n=== B3. 30 Hz stretched, no transposition (44.1k / window 2205) ===\n");
|
||||
measureRow("30 Hz rate 0.5", 30.0, 44100, 2205, 60, 0.5);
|
||||
measureRow("30 Hz rate 2.0", 30.0, 44100, 2205, 60, 2.0);
|
||||
std::printf("\n=== B4. the same six at 48k / window 2400 ===\n");
|
||||
measureRow("30 Hz +2 st @48k", 30.0, 48000, 2400, 62, 1.0);
|
||||
measureRow("30 Hz +7 st @48k", 30.0, 48000, 2400, 67, 1.0);
|
||||
measureRow("30 Hz rate 2.0 @48k", 30.0, 48000, 2400, 60, 2.0);
|
||||
}
|
||||
|
||||
// Where does the behaviour actually turn over? Swept at a fixed, modest transposition so the
|
||||
// only thing changing is the source period against the reachable interval.
|
||||
static void reportFrequencySweep() {
|
||||
std::printf("\n=== B5. Frequency sweep, +2 st, rate 1.0 (44.1k / window 2205) ===\n");
|
||||
const double freqs[] = {20, 22, 24, 25, 26, 26.5, 27, 28, 29, 30, 31, 31.5, 32, 33,
|
||||
34, 36, 40, 45, 50, 60, 70, 80, 88.2, 100, 120, 140, 170, 200};
|
||||
for (double f : freqs) measureRow("sweep +2 st", f, 44100, 2205, 62, 1.0);
|
||||
std::printf("\n=== B6. Same sweep at rate 2.0, NO transposition ===\n");
|
||||
for (double f : freqs) measureRow("sweep rate 2.0", f, 44100, 2205, 60, 2.0);
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------------------
|
||||
// 3. The P=500 case: geometric, or cadence?
|
||||
// ---------------------------------------------------------------------------------------
|
||||
|
||||
// pitch_shift_tests' floor probe fails at source period 500 frames, rate 2.0, shift 0.25 and
|
||||
// holds at 600/700. Under the geometric claim, alignment needs a whole number of source
|
||||
// periods in [0.75w, 1.25w] = [1654, 2756]. This reports whether that is satisfied for each of
|
||||
// the three periods — separating "no aligned landing point exists" (geometric) from "an
|
||||
// aligned landing point exists but the cadence is too fast to use it" (the inequality already
|
||||
// in time_stretch.h).
|
||||
static void reportFloorProbeMechanism() {
|
||||
std::printf("\n=== C. The P=500/600/700 floor probe: which mechanism? ===\n");
|
||||
const std::int64_t w = 2205;
|
||||
const int sr = 44100;
|
||||
const double rate = 2.0;
|
||||
const double shift = std::pow(2.0, -24.0 / 12.0); // 0.25
|
||||
const double lo = static_cast<double>(w - w / 4), hi = static_cast<double>(w + w / 4);
|
||||
for (double period : {500.0, 600.0, 700.0}) {
|
||||
int n = 0;
|
||||
const bool reach = alignmentReachable(period, lo, hi, &n);
|
||||
const double freq = static_cast<double>(sr) / period;
|
||||
// The cadence inequality from time_stretch.h, evaluated for this row.
|
||||
const double cadence = static_cast<double>(w) / std::fabs(rate - shift);
|
||||
const double outPeriod = period / shift;
|
||||
std::printf(" P=%5.0f (%.1f Hz): alignable in [%.0f,%.0f]? %s%s | cadence %.0f fr vs "
|
||||
"output period %.0f fr -> %s\n",
|
||||
period, freq, lo, hi, reach ? "YES" : "NO",
|
||||
reach ? "" : " (geometric failure)",
|
||||
cadence, outPeriod,
|
||||
cadence < outPeriod ? "SPLICE INSIDE A CYCLE (cadence failure)" : "ok");
|
||||
measureRow("floor probe", freq, sr, w, 60 - 24, rate);
|
||||
}
|
||||
// The probe's OWN signal, reproduced exactly: a bare PitchShifter fed by the same
|
||||
// schedule test_pitch_shift.cpp's runStretch uses, not the Voice. Its zero-crossing
|
||||
// number is the one that is currently RED, so it is the one that has to be explained.
|
||||
std::printf("\n --- the probe's exact signal (bare PitchShifter, runStretch schedule) ---\n");
|
||||
for (double period : {500.0, 600.0, 700.0}) {
|
||||
const std::size_t srcLen = 400000;
|
||||
std::vector<AudioSample> src(srcLen);
|
||||
for (std::size_t i = 0; i < srcLen; ++i) {
|
||||
src[i] = static_cast<AudioSample>(
|
||||
std::sin(2.0 * kPi * static_cast<double>(i) / period));
|
||||
}
|
||||
std::vector<double> out;
|
||||
const SpliceStats st = spliceGeometry(src, w, rate, shift, 60000, &out);
|
||||
const double want = period / shift;
|
||||
const double zc = periodIn(out, 20000, 50000);
|
||||
const double ac = autocorrPeriod(out, 20000, 20000,
|
||||
static_cast<std::int64_t>(want * 0.5),
|
||||
static_cast<std::int64_t>(want * 1.7));
|
||||
std::printf(" P=%.0f: zero-crossing %.2f (%+.2f%%) | autocorrelation %.2f (%+.2f%%) "
|
||||
"| splices %lld every %.0f fr\n", period, zc, 100.0 * (zc - want) / want,
|
||||
ac, 100.0 * (ac - want) / want, st.count, st.meanInterval);
|
||||
reportSpectrum("probe", out, 20000, 32768, want);
|
||||
}
|
||||
|
||||
std::printf("\n --- independent pitch estimators on the same three (through the Voice) ---\n");
|
||||
deepDive("P=500", 44100.0 / 500.0, sr, w, 36, rate);
|
||||
deepDive("P=600", 44100.0 / 600.0, sr, w, 36, rate);
|
||||
deepDive("P=700", 44100.0 / 700.0, sr, w, 36, rate);
|
||||
std::printf("\n --- and on the geometric cases, for contrast ---\n");
|
||||
deepDive("30 Hz +2 st rate 1.0", 30.0, sr, w, 62, 1.0);
|
||||
deepDive("30 Hz rate 2.0", 30.0, sr, w, 60, 2.0);
|
||||
deepDive("29 Hz rate 2.0", 29.0, sr, w, 60, 2.0);
|
||||
}
|
||||
|
||||
// What would a bigger window buy? The reachable interval is [0.75w, 1.25w], so it contains a
|
||||
// whole number of source periods for EVERY period P <= 0.5w — i.e. a window of at least TWO
|
||||
// source periods makes alignment reachable unconditionally. This sweeps 30 Hz across windows
|
||||
// spanning that threshold (1470 * 2 = 2940 frames = 66.7 ms at 44.1k) and reports what
|
||||
// actually changes. The window is an ARGUMENT to configure(); nothing shipped is altered.
|
||||
static void reportWindowSweep() {
|
||||
std::printf("\n=== D. Window sweep at 30 Hz — what a larger window would buy ===\n");
|
||||
const int sr = 44100;
|
||||
const double P = static_cast<double>(sr) / 30.0;
|
||||
std::printf(" source period %.1f frames; alignment is unconditional once window >= 2P = "
|
||||
"%.0f frames (%.1f ms)\n", P, 2.0 * P, 2000.0 * P / sr);
|
||||
for (std::int64_t w : {std::int64_t{2205}, std::int64_t{2646}, std::int64_t{2940},
|
||||
std::int64_t{3528}, std::int64_t{4410}, std::int64_t{8820}}) {
|
||||
const double lo = static_cast<double>(w - w / 4), hi = static_cast<double>(w + w / 4);
|
||||
int n = 0;
|
||||
const bool reach = alignmentReachable(P, lo, hi, &n);
|
||||
std::printf("\n window %lld fr (%.1f ms), interval [%.0f, %.0f]: %s\n",
|
||||
static_cast<long long>(w), 1000.0 * static_cast<double>(w) / sr, lo, hi,
|
||||
reach ? "ALIGNABLE" : "unalignable");
|
||||
// Per-voice Preserve state: two shifter rings of 2*window floats (L/R) plus the
|
||||
// window-sized prime scratch = 5*window floats (voice.cpp presizePreserveShifters).
|
||||
const double bytes = 5.0 * static_cast<double>(w) * 4.0;
|
||||
std::printf(" per-voice Preserve state %.1f KB; at the 32-voice ceiling %.2f MB\n",
|
||||
bytes / 1024.0, 32.0 * bytes / (1024.0 * 1024.0));
|
||||
measureRow(" 30 Hz +2 st", 30.0, sr, w, 62, 1.0);
|
||||
measureRow(" 30 Hz rate 2.0", 30.0, sr, w, 60, 2.0);
|
||||
deepDive(" +2 st", 30.0, sr, w, 62, 1.0);
|
||||
deepDive(" rate 2.0", 30.0, sr, w, 60, 2.0);
|
||||
}
|
||||
}
|
||||
|
||||
// Alignable neighbours under identical conditions — without these the out-of-band-energy
|
||||
// numbers above have no scale.
|
||||
static void reportAlignableControls() {
|
||||
std::printf("\n=== E. Alignable controls (same conditions, a frequency that CAN align) ===\n");
|
||||
deepDive("34 Hz +2 st rate 1.0", 34.0, 44100, 2205, 62, 1.0);
|
||||
deepDive("34 Hz rate 2.0", 34.0, 44100, 2205, 60, 2.0);
|
||||
deepDive("20 Hz +2 st rate 1.0", 20.0, 44100, 2205, 62, 1.0);
|
||||
deepDive("220 Hz +2 st rate 1.0", 220.0, 44100, 2205, 62, 1.0);
|
||||
deepDive("220 Hz rate 2.0", 220.0, 44100, 2205, 60, 2.0);
|
||||
}
|
||||
|
||||
int main() {
|
||||
reportReachableInterval();
|
||||
reportReachabilityByFrequency();
|
||||
testRootRateUnityIsBitIdenticalToTheDirectRead();
|
||||
reportTransposedAt30Hz();
|
||||
reportFrequencySweep();
|
||||
reportFloorProbeMechanism();
|
||||
reportAlignableControls();
|
||||
reportWindowSweep();
|
||||
|
||||
if (g_fail == 0) {
|
||||
std::printf("\nall preserve_low_frequency measurements completed\n");
|
||||
return 0;
|
||||
}
|
||||
std::printf("\n%d preserve_low_frequency check(s) failed\n", g_fail);
|
||||
return 1;
|
||||
}
|
||||
Reference in New Issue
Block a user