audio_utils.py

import re
from functools import lru_cache
from subprocess import CalledProcessError, run

import numpy as np
import librosa

flg_ffmpeg = False


# hard-coded audio hyperparameters
SAMPLE_RATE = 16000
N_FFT = 400
N_MELS = 80
HOP_LENGTH = 160
CHUNK_LENGTH = 30
N_SAMPLES = CHUNK_LENGTH * SAMPLE_RATE  # 480000 samples in a 30-second chunk


def get_T_after_cnn(L_in, dilation=1):
    for padding, kernel_size, stride in eval("[(1,3,1)] + [(1,3,2)] "):
        L_out = L_in + 2 * padding - dilation * (kernel_size - 1) - 1
        L_out = 1 + L_out // stride
        L_in = L_out
    return L_out


def load_audio(file: str, sr: int = SAMPLE_RATE):
    """
    Open an audio file and read as mono waveform, resampling as necessary
    """

    if flg_ffmpeg:
        # This launches a subprocess to decode audio while down-mixing
        # and resampling as necessary.  Requires the ffmpeg CLI in PATH.
        # fmt: off
        cmd = [
            "ffmpeg",
            "-nostdin",
            "-threads", "0",
            "-i", file,
            "-f", "s16le",
            "-ac", "1",
            "-acodec", "pcm_s16le",
            "-ar", str(sr),
            "-"
        ]
        # fmt: on
        try:
            out = run(cmd, capture_output=True, check=True).stdout
        except CalledProcessError as e:
            raise RuntimeError(f"Failed to load audio: {e.stderr.decode()}") from e

        return np.frombuffer(out, np.int16).flatten().astype(np.float32) / 32768.0
    else:
        # prepare input data
        audio, _ = librosa.load(file, sr=sr, mono=True, dtype=np.float32)
        return audio


def pad_or_trim(array, length: int = N_SAMPLES, *, axis: int = -1):
    """
    Pad or trim the audio array to N_SAMPLES, as expected by the encoder.
    """
    if array.shape[axis] > length:
        array = array.take(indices=range(length), axis=axis)

    if array.shape[axis] < length:
        pad_widths = [(0, 0)] * array.ndim
        pad_widths[axis] = (0, length - array.shape[axis])
        array = np.pad(array, pad_widths)

    return array


@lru_cache(maxsize=None)
def mel_filters(n_mels: int = N_MELS):
    """
    the mel filterbank matrix for projecting STFT into a Mel spectrogram.
    """
    filters = librosa.filters.mel(sr=SAMPLE_RATE, n_fft=N_FFT, n_mels=n_mels)

    return filters


def log_mel_spectrogram(
    audio: np.ndarray,
    n_mels: int = N_MELS,
    padding: int = 0,
):
    """
    Compute the log-Mel spectrogram of

    Parameters
    ----------
    audio: np.ndarray, shape = (*)
        The path to audio or either a NumPy array or Tensor containing the audio waveform in 16 kHz

    n_mels: int
        The number of Mel-frequency filters, only 80 is supported

    padding: int
        Number of zero samples to pad to the right

    device: Optional[Union[str, torch.device]]
        If given, the audio tensor is moved to this device before STFT

    Returns
    -------
    np.ndarray, shape = (80, n_frames)
        A Tensor that contains the Mel spectrogram
    """
    if padding > 0:
        audio = np.pad(audio, (0, padding))
    stft = librosa.stft(
        y=audio,
        n_fft=N_FFT,
        hop_length=HOP_LENGTH,
        window="hann",
        pad_mode="reflect",
    )
    magnitudes = np.abs(stft[:, :-1]) ** 2

    filters = mel_filters(n_mels)
    mel_spec = filters @ magnitudes

    log_spec = np.log10(np.clip(mel_spec, 1e-10, None))
    log_spec = np.maximum(log_spec, np.max(log_spec) - 8.0)
    log_spec = (log_spec + 4.0) / 4.0

    return log_spec


def process_audio(content):
    pattern = r"<audio>(.*?)</audio>"
    audio_urls = re.findall(pattern, content)
    if len(audio_urls) == 0:
        return None

    audios, audio_lens, audio_span_tokens = [], [], []
    for audio_path in audio_urls:
        cache = getattr(process_audio, "cache", {})
        if audio_path in cache:
            mel, audio_len, audio_token_num = cache[audio_path]
            audios.append(mel)
            audio_lens.append(audio_len)
            audio_span_tokens.append(audio_token_num + 2)
            continue

        audio = load_audio(audio_path)
        L = audio.shape[0] if audio.shape[0] <= 480000 else 480000  # max_length < 30s
        mel_len = L // 160
        audio = pad_or_trim(audio.flatten())
        mel = log_mel_spectrogram(audio)
        audio_len_after_cnn = get_T_after_cnn(mel_len)
        audio_token_num = (audio_len_after_cnn - 2) // 2 + 1
        audio_len = [audio_len_after_cnn, audio_token_num]
        audios.append(mel)
        audio_lens.append(audio_len)
        audio_span_tokens.append(audio_token_num + 2)  # add audio bos eos

        cache[audio_path] = (mel, audio_len, audio_token_num)
        process_audio.cache = cache

    input_audio_lengths = np.array(audio_lens)
    input_audios = np.stack(audios, axis=0)

    return {
        "input_audios": input_audios,
        "input_audio_lengths": input_audio_lengths,
        "audio_span_tokens": audio_span_tokens,
        "audio_urls": audio_urls,
    }