[audio utils] fix fft_bin_width computation (#1274)

Author: xenova

src/models/audio_spectrogram_transformer/feature_extraction_audio_spectrogram_transformer.js

@@ -10,7 +10,7 @@ export class ASTFeatureExtractor extends FeatureExtractor {
 
         const sampling_rate = this.config.sampling_rate;
         const mel_filters = mel_filter_bank(
-            256, // num_frequency_bins
+            257, // num_frequency_bins
             this.config.num_mel_bins, // num_mel_filters
             20, // min_frequency
             Math.floor(sampling_rate / 2), // max_frequency
@@ -19,11 +19,6 @@ export class ASTFeatureExtractor extends FeatureExtractor {
             "kaldi", // mel_scale
             true, // triangularize_in_mel_space
         );
-
-        // Do padding:
-        for (let i = 0; i < mel_filters.length; ++i) {
-            mel_filters[i].push(0);
-        }
         this.mel_filters = mel_filters;
 
         this.window = window_function(400, 'hann', {

src/models/seamless_m4t/feature_extraction_seamless_m4t.js

@@ -9,7 +9,7 @@ export class SeamlessM4TFeatureExtractor extends FeatureExtractor {
 
         const sampling_rate = this.config.sampling_rate;
         const mel_filters = mel_filter_bank(
-            256, // num_frequency_bins
+            257, // num_frequency_bins
             this.config.num_mel_bins, // num_mel_filters
             20, // min_frequency
             Math.floor(sampling_rate / 2), // max_frequency
@@ -18,11 +18,6 @@ export class SeamlessM4TFeatureExtractor extends FeatureExtractor {
             "kaldi", // mel_scale
             true, // triangularize_in_mel_space
         );
-
-        // Do padding:
-        for (let i = 0; i < mel_filters.length; ++i) {
-            mel_filters[i].push(0);
-        }
         this.mel_filters = mel_filters;
 
         this.window = window_function(400, 'povey', {

src/models/wespeaker/feature_extraction_wespeaker.js

@@ -10,7 +10,7 @@ export class WeSpeakerFeatureExtractor extends FeatureExtractor {
 
         const sampling_rate = this.config.sampling_rate;
         const mel_filters = mel_filter_bank(
-            256, // num_frequency_bins
+            257, // num_frequency_bins
             this.config.num_mel_bins, // num_mel_filters
             20, // min_frequency
             Math.floor(sampling_rate / 2), // max_frequency
@@ -19,11 +19,6 @@ export class WeSpeakerFeatureExtractor extends FeatureExtractor {
             "kaldi", // mel_scale
             true, // triangularize_in_mel_space
         );
-
-        // Do padding:
-        for (let i = 0; i < mel_filters.length; ++i) {
-            mel_filters[i].push(0);
-        }
         this.mel_filters = mel_filters;
 
         this.window = window_function(400, 'hamming', {

src/utils/audio.js

@@ -235,7 +235,8 @@ function linspace(start, end, num) {
  * various implementation exist, which differ in the number of filters, the shape of the filters, the way the filters
  * are spaced, the bandwidth of the filters, and the manner in which the spectrum is warped. The goal of these
  * features is to approximate the non-linear human perception of the variation in pitch with respect to the frequency.
- * @param {number} num_frequency_bins Number of frequencies used to compute the spectrogram (should be the same as in `stft`).
+ * @param {number} num_frequency_bins Number of frequency bins (should be the same as `n_fft // 2 + 1`
+ * where `n_fft` is the size of the Fourier Transform used to compute the spectrogram).
  * @param {number} num_mel_filters Number of mel filters to generate.
  * @param {number} min_frequency Lowest frequency of interest in Hz.
  * @param {number} max_frequency Highest frequency of interest in Hz. This should not exceed `sampling_rate / 2`.
@@ -261,6 +262,14 @@ export function mel_filter_bank(
         throw new Error('norm must be one of null or "slaney"');
     }
 
+    if (num_frequency_bins < 2) {
+        throw new Error(`Require num_frequency_bins: ${num_frequency_bins} >= 2`);
+    }
+
+    if (min_frequency > max_frequency) {
+        throw new Error(`Require min_frequency: ${min_frequency} <= max_frequency: ${max_frequency}`);
+    }
+
     const mel_min = hertz_to_mel(min_frequency, mel_scale);
     const mel_max = hertz_to_mel(max_frequency, mel_scale);
     const mel_freqs = linspace(mel_min, mel_max, num_mel_filters + 2);
@@ -269,7 +278,7 @@ export function mel_filter_bank(
     let fft_freqs; // frequencies of FFT bins in Hz
 
     if (triangularize_in_mel_space) {
-        const fft_bin_width = sampling_rate / (num_frequency_bins * 2);
+        const fft_bin_width = sampling_rate / ((num_frequency_bins - 1) * 2);
         fft_freqs = hertz_to_mel(Float64Array.from({ length: num_frequency_bins }, (_, i) => i * fft_bin_width), mel_scale);
         filter_freqs = mel_freqs;
     } else {