Kotlin Speech Features/org.merlyn.kotlinspeechfeatures/SpeechFeatures

SpeechFeatures

class SpeechFeatures(fft: FFT = KotlinFFT())

Constructors

SpeechFeatures

fun SpeechFeatures(fft: FFT = KotlinFFT())

Functions

delta

fun delta(feat: FloatArray, n: Int): Array<FloatArray>

Compute delta features from a feature vector sequence.

fbank

fun fbank( signal: FloatArray, sampleRate: Int = 16000, winLen: Float = 0.025f, winStep: Float = 0.01f, nFilt: Int = 26, nfft: Int = 512, lowFreq: Int = 0, highFreq: Int? = null, preemph: Float = 0.97f, winFunc: FloatArray? = null): Pair<Array<FloatArray>, FloatArray>

Compute Mel-filterbank energy features from an audio signal.

getFilterBanks

fun getFilterBanks( nFilt: Int = 20, nfft: Int = 512, sampleRate: Int = 16000, lowFreq: Int = 0, highFreqIn: Int? = null): Array<FloatArray>

Compute a Mel-filterbank. The filters are stored in the rows, the columns correspond to fft bins. The filters are returned as an array of size nfilt * (nfft/2 + 1)

hz2mel

fun hz2mel(hz: Double): Double

fun hz2mel(hz: DoubleArray): DoubleArray

Convert a value in Hertz to Mels

lifter

fun lifter(cepstra: FloatArray, l: Int = 22): FloatArray

Apply a cepstral lifter the the matrix of cepstra. This has the effect of increasing the magnitude of the high frequency DCT coeffs.

logfbank

fun logfbank( signal: FloatArray, sampleRate: Int = 16000, winLen: Float = 0.025f, winStep: Float = 0.01f, nFilt: Int = 26, nfft: Int = 512, lowFreq: Int = 0, highFreq: Int? = null, preemph: Float = 0.97f, winFunc: FloatArray? = null): Array<FloatArray>

Compute log Mel-filterbank energy features from an audio signal.

mel2hz

fun mel2hz(mel: Double): Double

fun mel2hz(mel: DoubleArray): DoubleArray

Convert a value in Mels to Hertz

mfcc

fun mfcc( signal: FloatArray, sampleRate: Int = 16000, winLen: Float = 0.025f, winStep: Float = 0.01f, numCep: Int = 13, nFilt: Int = 26, nfft: Int? = 512, lowFreq: Int = 0, highFreq: Int? = null, preemph: Float = 0.97f, ceplifter: Int = 22, appendEnergy: Boolean = true, winFunc: FloatArray? = null): Array<FloatArray>

Compute MFCC features from an audio signal.

ssc

fun ssc( signal: FloatArray, sampleRate: Int = 16000, winLen: Float = 0.025f, winStep: Float = 0.01f, nFilt: Int = 26, nfft: Int = 512, lowFreq: Int = 0, highFreq: Int? = null, preemph: Float = 0.97f, winFunc: FloatArray? = null): Array<FloatArray>

Compute Spectral Subband Centroid features from an audio signal.