@siteed/expo-audio-stream 1.16.0 → 2.0.0

package/src/AudioAnalysis/AudioAnalysis.types.ts

@@ -1,32 +1,56 @@
 // packages/expo-audio-stream/src/AudioAnalysis/AudioAnalysis.types.ts
 
+import { BitDepth, ConsoleLike } from '../ExpoAudioStream.types'
+
 /**
  * Represents the configuration for decoding audio data.
  */
 export interface DecodingConfig {
+    /** Target sample rate for decoded audio (Android and Web) */
     targetSampleRate?: number
+    /** Target number of channels (Android and Web) */
     targetChannels?: number
-    targetBitDepth?: number
+    /** Target bit depth (Android and Web) */
+    targetBitDepth?: BitDepth
+    /** Whether to normalize audio levels (Android and Web) */
     normalizeAudio?: boolean
 }
 
+/**
+ * Represents speech-related features extracted from audio.
+ */
+export interface SpeechFeatures {
+    isActive: boolean // Whether speech is detected in this segment
+    speakerId?: number // Optional speaker identification
+    // Could add more speech-related features here like:
+    // confidence: number
+    // language?: string
+    // sentiment?: number
+    // etc.
+}
+
 /**
  * Represents various audio features extracted from an audio signal.
  */
 export interface AudioFeatures {
-    energy: number // The infinite integral of the squared signal, representing the overall energy of the audio.
-    mfcc: number[] // Mel-frequency cepstral coefficients, describing the short-term power spectrum of a sound.
-    rms: number // Root mean square value, indicating the amplitude of the audio signal.
-    minAmplitude: number // Minimum amplitude value in the audio signal.
-    maxAmplitude: number // Maximum amplitude value in the audio signal.
-    zcr: number // Zero-crossing rate, indicating the rate at which the signal changes sign.
-    spectralCentroid: number // The center of mass of the spectrum, indicating the brightness of the sound.
-    spectralFlatness: number // Measure of the flatness of the spectrum, indicating how noise-like the signal is.
-    spectralRolloff: number // The frequency below which a specified percentage (usually 85%) of the total spectral energy lies.
-    spectralBandwidth: number // The width of the spectrum, indicating the range of frequencies present.
-    chromagram: number[] // Chromagram, representing the 12 different pitch classes of the audio.
-    tempo: number // Estimated tempo of the audio signal, measured in beats per minute (BPM).
-    hnr: number // Harmonics-to-noise ratio, indicating the proportion of harmonics to noise in the audio signal.
+    energy?: number // The infinite integral of the squared signal, representing the overall energy of the audio.
+    mfcc?: number[] // Mel-frequency cepstral coefficients, describing the short-term power spectrum of a sound.
+    rms?: number // Root mean square value, indicating the amplitude of the audio signal.
+    minAmplitude?: number // Minimum amplitude value in the audio signal.
+    maxAmplitude?: number // Maximum amplitude value in the audio signal.
+    zcr?: number // Zero-crossing rate, indicating the rate at which the signal changes sign.
+    spectralCentroid?: number // The center of mass of the spectrum, indicating the brightness of the sound.
+    spectralFlatness?: number // Measure of the flatness of the spectrum, indicating how noise-like the signal is.
+    spectralRolloff?: number // The frequency below which a specified percentage (usually 85%) of the total spectral energy lies.
+    spectralBandwidth?: number // The width of the spectrum, indicating the range of frequencies present.
+    chromagram?: number[] // Chromagram, representing the 12 different pitch classes of the audio.
+    tempo?: number // Estimated tempo of the audio signal, measured in beats per minute (BPM).
+    hnr?: number // Harmonics-to-noise ratio, indicating the proportion of harmonics to noise in the audio signal.
+    melSpectrogram?: number[] // Mel-scaled spectrogram representation of the audio.
+    spectralContrast?: number[] // Spectral contrast features representing the difference between peaks and valleys.
+    tonnetz?: number[] // Tonal network features representing harmonic relationships.
+    pitch?: number // Pitch of the audio signal, measured in Hertz (Hz).
+    crc32?: number // crc32 checksum of the audio signal, used to verify the integrity of the audio.
 }
 
 /**
@@ -44,6 +68,11 @@ export interface AudioFeaturesOptions {
     chromagram?: boolean
     tempo?: boolean
     hnr?: boolean
+    melSpectrogram?: boolean
+    spectralContrast?: boolean
+    tonnetz?: boolean
+    pitch?: boolean
+    crc32?: boolean
 }
 
 /**
@@ -51,11 +80,12 @@ export interface AudioFeaturesOptions {
  */
 export interface DataPoint {
     id: number
-    amplitude: number
-    activeSpeech?: boolean
-    dB?: number
-    silent?: boolean
+    amplitude: number // Peak amplitude for the segment
+    rms: number // Root mean square value
+    dB: number // Always computed
+    silent: boolean // Always computed
     features?: AudioFeatures
+    speech?: SpeechFeatures
     startTime?: number
     endTime?: number
     // start / end position in bytes
@@ -63,33 +93,38 @@ export interface DataPoint {
     endPosition?: number
     // number of audio samples for this point (samples size depends on bit depth)
     samples?: number
-    // TODO: speaker detection
-    speaker?: number
 }
 
-export type AmplitudeAlgorithm = 'peak' | 'rms'
-
 /**
  * Represents the complete data from the audio analysis.
  */
 export interface AudioAnalysis {
-    pointsPerSecond: number // How many consolidated value per second
+    segmentDurationMs: number // Duration of each segment in milliseconds
     durationMs: number // Duration of the audio in milliseconds
     bitDepth: number // Bit depth of the audio
     samples: number // Size of the audio in bytes
     numberOfChannels: number // Number of audio channels
     sampleRate: number // Sample rate of the audio
     dataPoints: DataPoint[] // Array of data points from the analysis.
-    amplitudeAlgorithm: AmplitudeAlgorithm // Algorithm used to calculate amplitude values.
     amplitudeRange: {
         min: number
         max: number
     }
-    // TODO: speaker detection
-    speakerChanges?: {
-        timestamp: number // Timestamp of the speaker change in milliseconds.
-        speaker: number // Speaker identifier.
-    }[]
+    rmsRange: {
+        min: number
+        max: number
+    }
+    // TODO: speaker changes into a broader speech analysis section
+    speechAnalysis?: {
+        speakerChanges: {
+            timestamp: number
+            speakerId: number
+        }[]
+        // Could add more speech analysis data here like:
+        // dominantSpeaker?: number
+        // totalSpeechDuration?: number
+        // speakerStats?: { [speakerId: number]: { duration: number, segments: number } }
+    }
 }
 
 /**
@@ -97,9 +132,9 @@ export interface AudioAnalysis {
  */
 export interface AudioRangeOptions {
     /** Start time in milliseconds */
-    startTime?: number
+    startTimeMs?: number
     /** End time in milliseconds */
-    endTime?: number
+    endTimeMs?: number
 }
 
 /**
@@ -115,10 +150,9 @@ export interface PreviewOptions extends AudioRangeOptions {
      */
     numberOfPoints?: number
     /**
-     * Algorithm used to calculate amplitude values
-     * @default "rms"
+     * Optional logger for debugging.
      */
-    algorithm?: AmplitudeAlgorithm
+    logger?: ConsoleLike
     /**
      * Optional configuration for decoding the audio file.
      * Defaults to:
@@ -129,21 +163,3 @@ export interface PreviewOptions extends AudioRangeOptions {
      */
     decodingOptions?: DecodingConfig
 }
-
-/**
- * Represents a simplified preview of audio waveform,
- * optimized for quick visualization.
- */
-export interface AudioPreview {
-    /** Number of data points per second */
-    pointsPerSecond: number
-    /** Duration of the audio in milliseconds */
-    durationMs: number
-    /** Range of amplitude values in the preview */
-    amplitudeRange: {
-        min: number
-        max: number
-    }
-    /** Array of data points representing the waveform */
-    dataPoints: DataPoint[]
-}