AudioPreprocessing

org.llm4s.speech.processing.AudioPreprocessing

Functional audio preprocessing utilities. These are pure transformations described as functions that return either errors or processed audio.

Attributes

Graph
Supertypes
class Object
trait Matchable
class Any
Self type

Members list

Value members

Concrete methods

def resamplePcm16(bytes: Array[Byte], source: AudioMeta, targetRate: Int): Result[(Array[Byte], AudioMeta)]

Resample PCM16 little-endian bytes to target sample rate using Java Sound.

Resample PCM16 little-endian bytes to target sample rate using Java Sound.

Attributes

def standardizeForSTT(bytes: Array[Byte], meta: AudioMeta, targetRate: Int): Result[(Array[Byte], AudioMeta)]

Compose multiple steps functionally

Compose multiple steps functionally

Attributes

def toMono(bytes: Array[Byte], meta: AudioMeta): Result[(Array[Byte], AudioMeta)]

Convert to mono by averaging channels (PCM16 little-endian).

Convert to mono by averaging channels (PCM16 little-endian).

Attributes

def trimSilence(bytes: Array[Byte], meta: AudioMeta, threshold: Int): Result[(Array[Byte], AudioMeta)]

Trim leading and trailing silence using a simple amplitude threshold on PCM16.

Trim leading and trailing silence using a simple amplitude threshold on PCM16.

Attributes

def wrap(bytes: Array[Byte], meta: AudioMeta, format: AudioFormat): GeneratedAudio