seek-tune/server/shazam/spectrogram.go
Chigozirim Igweamaka e3a35ef1eb fix(spectrogram): correct STFT and peak extraction algorithm
- Fix frame calculation with proper sliding window iteration
- Change hop size to windowSize/2 for 50% overlap
- Return magnitude spectrum instead of complex values
- Fix Peak time/frequency calculations using proper frame-based indexing
- Add Hz conversion using frequency resolution
- Remove incorrect frequency-based time calculations
2025-11-19 16:47:01 +01:00

192 lines
4.8 KiB
Go

package shazam
import (
"errors"
"fmt"
"math"
"math/cmplx"
)
const (
dspRatio = 4
windowSize = 1024
maxFreq = 5000.0 // 5kHz
hopSize = windowSize / 2 // 50% overlap for better time-frequency resolution
windowType = "hanning" // choices: "hanning" or "hamming"
)
func Spectrogram(sample []float64, sampleRate int) ([][]float64, error) {
filteredSample := LowPassFilter(maxFreq, float64(sampleRate), sample)
downsampledSample, err := Downsample(filteredSample, sampleRate, sampleRate/dspRatio)
if err != nil {
return nil, fmt.Errorf("couldn't downsample audio sample: %v", err)
}
window := make([]float64, windowSize)
for i := range window {
theta := 2 * math.Pi * float64(i) / float64(windowSize-1)
switch windowType {
case "hamming":
window[i] = 0.54 - 0.46*math.Cos(theta)
default: // Hanning window
window[i] = 0.5 - 0.5*math.Cos(theta)
}
}
// Initialize spectrogram slice
spectrogram := make([][]float64, 0)
// Perform STFT
for start := 0; start+windowSize <= len(downsampledSample); start += hopSize {
end := start + windowSize
frame := make([]float64, windowSize)
copy(frame, downsampledSample[start:end])
// Apply window
for j := range window {
frame[j] *= window[j]
}
// Perform FFT
fftResult := FFT(frame)
// Convert complex spectrum to magnitude spectrum
magnitude := make([]float64, len(fftResult)/2)
for j := range magnitude {
magnitude[j] = cmplx.Abs(fftResult[j])
}
spectrogram = append(spectrogram, magnitude)
}
return spectrogram, nil
}
// LowPassFilter is a first-order low-pass filter that attenuates high
// frequencies above the cutoffFrequency.
// It uses the transfer function H(s) = 1 / (1 + sRC), where RC is the time constant.
func LowPassFilter(cutoffFrequency, sampleRate float64, input []float64) []float64 {
rc := 1.0 / (2 * math.Pi * cutoffFrequency)
dt := 1.0 / sampleRate
alpha := dt / (rc + dt)
filteredSignal := make([]float64, len(input))
var prevOutput float64 = 0
for i, x := range input {
if i == 0 {
filteredSignal[i] = x * alpha
} else {
filteredSignal[i] = alpha*x + (1-alpha)*prevOutput
}
prevOutput = filteredSignal[i]
}
return filteredSignal
}
// Downsample downsamples the input audio from originalSampleRate to targetSampleRate
func Downsample(input []float64, originalSampleRate, targetSampleRate int) ([]float64, error) {
if targetSampleRate <= 0 || originalSampleRate <= 0 {
return nil, errors.New("sample rates must be positive")
}
if targetSampleRate > originalSampleRate {
return nil, errors.New("target sample rate must be less than or equal to original sample rate")
}
ratio := originalSampleRate / targetSampleRate
if ratio <= 0 {
return nil, errors.New("invalid ratio calculated from sample rates")
}
var resampled []float64
for i := 0; i < len(input); i += ratio {
end := i + ratio
if end > len(input) {
end = len(input)
}
sum := 0.0
for j := i; j < end; j++ {
sum += input[j]
}
avg := sum / float64(end-i)
resampled = append(resampled, avg)
}
return resampled, nil
}
// Peak represents a significant point in the spectrogram.
type Peak struct {
Freq float64 // Frequency in Hz
Time float64 // Time in seconds
}
// ExtractPeaks analyzes a spectrogram and extracts significant peaks in the frequency domain over time.
func ExtractPeaks(spectrogram [][]float64, audioDuration float64, sampleRate int) []Peak {
if len(spectrogram) < 1 {
return []Peak{}
}
type maxies struct {
maxMag float64
freqIdx int
}
bands := []struct{ min, max int }{
{0, 10}, {10, 20}, {20, 40}, {40, 80}, {80, 160}, {160, 512},
}
var peaks []Peak
frameDuration := audioDuration / float64(len(spectrogram))
// Calculate frequency resolution (Hz per bin)
effectiveSampleRate := float64(sampleRate) / float64(dspRatio)
freqResolution := effectiveSampleRate / float64(windowSize)
for frameIdx, frame := range spectrogram {
var maxMags []float64
var freqIndices []int
binBandMaxies := []maxies{}
for _, band := range bands {
var maxx maxies
var maxMag float64
for idx, mag := range frame[band.min:band.max] {
if mag > maxMag {
maxMag = mag
freqIdx := band.min + idx
maxx = maxies{mag, freqIdx}
}
}
binBandMaxies = append(binBandMaxies, maxx)
}
for _, value := range binBandMaxies {
maxMags = append(maxMags, value.maxMag)
freqIndices = append(freqIndices, value.freqIdx)
}
// Calculate the average magnitude
var maxMagsSum float64
for _, max := range maxMags {
maxMagsSum += max
}
avg := maxMagsSum / float64(len(maxMags))
// Add peaks that exceed the average magnitude
for i, value := range maxMags {
if value > avg {
peakTime := float64(frameIdx) * frameDuration
peakFreq := float64(freqIndices[i]) * freqResolution
peaks = append(peaks, Peak{Time: peakTime, Freq: peakFreq})
}
}
}
return peaks
}