v-jddlmZddlZddlmZmZddlZddlZddl m Z er ddl m Z ddlm Z dZd$d Z d%d&dZ d'd(d#ZdS))) annotationsN) TYPE_CHECKINGBinaryIO) AudioInfo)Path)Tensorcd}d|d}|S)N paddleaudiozxonly PCM16 WAV supported. if want support more other audio types, please manually installed (usually with `pip install zQ`). and use paddle.audio.backends.set_backend('soundfile') to set audio backend)packagewarn_msgs b/var/www/html/banglarbhumi/venv/lib/python3.11/site-packages/paddle/audio/backends/wave_backend.py_error_messager s.G V9@ V V V Ofilepathstr | BinaryIOreturnrct|dr|}nt|d} tj|}nY#tj$rG|d|t }t|wxYw|}| }| }| dz}d}|t|||||S)aGet signal information of input audio file. Args: filepath: audio path or file object. Returns: AudioInfo: info of the given audio. Example: .. code-block:: python >>> import os >>> import paddle >>> sample_rate = 16000 >>> wav_duration = 0.5 >>> num_channels = 1 >>> num_frames = sample_rate * wav_duration >>> wav_data = paddle.linspace(-1.0, 1.0, int(num_frames)) * 0.1 >>> waveform = wav_data.tile([num_channels, 1]) >>> base_dir = os.getcwd() >>> filepath = os.path.join(base_dir, "test.wav") >>> paddle.audio.save(filepath, waveform, sample_rate) >>> wav_info = paddle.audio.info(filepath) readrbrPCM_S) hasattropenwaveErrorseekcloserNotImplementedError getnchannels getframerate getnframes getsampwidthr) rfile_objfile_err_msgchannels sample_rate sample_framesbits_per_sampleencodings rinfor-+s8x  ($''+ (## :+++ a ""!'*** + !!##H$$&&K$$&&M((**Q.OH NN ]Hox   :ABT str | Path frame_offsetint num_frames normalizeboolchannels_firsttuple[Tensor, int]cBt|dr|}nt|d} tj|}nY#tj$rG|d|t }t|wxYw|}| } | } | | } |tj | tj} | tj} |r| dz }n| }tj|| |f}|dkr||||zddf}t%j|}|rt%j|ddg }|| fS) aLoad audio data from file. load the audio content start form frame_offset, and get num_frames. Args: frame_offset: from 0 to total frames, num_frames: from -1 (means total frames) or number frames which want to read, normalize: if True: return audio which norm to (-1, 1), dtype=float32 if False: return audio with raw data, dtype=int16 channels_first: if True: return audio with shape (channels, time) Return: Tuple[paddle.Tensor, int]: (audio_content, sample rate) Examples: .. code-block:: python >>> import os >>> import paddle >>> sample_rate = 16000 >>> wav_duration = 0.5 >>> num_channels = 1 >>> num_frames = sample_rate * wav_duration >>> wav_data = paddle.linspace(-1.0, 1.0, int(num_frames)) * 0.1 >>> waveform = wav_data.tile([num_channels, 1]) >>> base_dir = os.getcwd() >>> filepath = os.path.join(base_dir, "test.wav") >>> paddle.audio.save(filepath, waveform, sample_rate) >>> wav_data_read, sr = paddle.audio.load(filepath) rrr)dtyper/Nr)perm)rrrrrrrr r!r"r# readframesnp frombufferint16astypefloat32reshapepaddle to_tensor transpose)rr1r3r4r6r%r&r'r(r)frames audio_content audio_as_np16 audio_as_np32 audio_normwaveforms rloadrL_sPx  ($''+ (## :+++ a ""!'*** + !!##H$$&&K     F$$V,,M NNM-rx@@@M!((44M#"e, # z*vx&899HRL<*+DDaaaGH))H;#HAq6::: [  r.strsrcr r)r, str | Noner+ int | NoneNonecj|jdks Jd|}|rtj|}|jd}|dvrt dt |dz }|jtj kr|dz d}tj |d 5} | || || || |d d d d S#1swxYwYd S) a Save audio tensor to file. Args: filepath: saved path src: the audio tensor sample_rate: the number of samples of audio per second. channels_first: src channel information if True, means input tensor is (channels, time) if False, means input tensor is (time, channels) encoding: audio encoding format, wave_backend only support PCM16 now. bits_per_sample: bits per sample, wave_backend only support 16 bits now. Returns: None Examples: .. code-block:: python >>> import paddle >>> sample_rate = 16000 >>> wav_duration = 0.5 >>> num_channels = 1 >>> num_frames = sample_rate * wav_duration >>> wav_data = paddle.linspace(-1.0, 1.0, int(num_frames)) * 0.1 >>> waveform = wav_data.tile([num_channels, 1]) >>> filepath = "./test.wav" >>> paddle.audio.save(filepath, waveform, sample_rate) zExpected 2D tensorr)NrMz,Invalid bits_per_sample, only support 16 bitrr:zrgs #""""" ******** 1111l L!L!L!L!L!f "$ >->->->->->->-r