o .wi@s~ddlmZddlZddlZddlmZddlmZddlm Z m Z ddgiZ  dd ed ed e d e dede defddZdS))AnyN)Tensor)_check_same_shape)_MULTIPROCESSING_AVAILABLE_PESQ_AVAILABLE)$perceptual_evaluation_speech_qualitypesqFpredstargetfsmodekeep_same_device n_processesreturnc Cststdddl}dtdtfdd}t|}|dvr#td ||d vr.td |t|||j d krR||| | |} t | } nq|d |jd  } |d |jd  } tr|d kr|j|| | ||d} t| } n)tj| jdd} t| jdD]} ||| | ddf| | ddf|| | <qt | || tj} | t| } |r| |jS| S)aCalculate `Perceptual Evaluation of Speech Quality`_ (PESQ). It's a recognized industry standard for audio quality that takes into considerations characteristics such as: audio sharpness, call volume, background noise, clipping, audio interference etc. PESQ returns a score between -0.5 and 4.5 with the higher scores indicating a better quality. This metric is a wrapper for the `pesq package`_. Note that input will be moved to `cpu` to perform the metric calculation. .. hint:: Usingsing this metrics requires you to have ``pesq`` install. Either install as ``pip install torchmetrics[audio]`` or ``pip install pesq``. Note that ``pesq`` will compile with your currently installed version of numpy, meaning that if you upgrade numpy at some point in the future you will most likely have to reinstall ``pesq``. Args: preds: float tensor with shape ``(...,time)`` target: float tensor with shape ``(...,time)`` fs: sampling frequency, should be 16000 or 8000 (Hz) mode: ``'wb'`` (wide-band) or ``'nb'`` (narrow-band) keep_same_device: whether to move the pesq value to the device of preds n_processes: integer specifying the number of processes to run in parallel for the metric calculation. Only applies to batches of data and if ``multiprocessing`` package is installed. Returns: Float tensor with shape ``(...,)`` of PESQ values per sample Raises: ModuleNotFoundError: If ``pesq`` package is not installed ValueError: If ``fs`` is not either ``8000`` or ``16000`` ValueError: If ``mode`` is not either ``"wb"`` or ``"nb"`` RuntimeError: If ``preds`` and ``target`` do not have the same shape Example: >>> from torch import randn >>> from torchmetrics.functional.audio.pesq import perceptual_evaluation_speech_quality >>> preds = randn(8000) >>> target = randn(8000) >>> perceptual_evaluation_speech_quality(preds, target, 8000, 'nb') tensor(2.2885) >>> perceptual_evaluation_speech_quality(preds, target, 16000, 'wb') tensor(1.6805) zwPESQ metric requires that pesq is installed. Either install as `pip install torchmetrics[audio]` or `pip install pesq`.rNxrcSstt|tjS)N)np issubdtypetypenumber)rr_/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/torchmetrics/functional/audio/pesq.py_issubtype_numberYsz?perceptual_evaluation_speech_quality.._issubtype_number)i@i>z:Expected argument `fs` to either be 8000 or 16000 but got )wbnbz;Expected argument `mode` to either be 'wb' or 'nb' but got r ) n_processor)shape)rModuleNotFoundErrorrrboolr vectorize ValueErrorrndimdetachcpunumpytorchtensorreshaperr pesq_batcharrayemptyrange from_numpyastypefloat32lentodevice)r r r r rr pesq_backendr_filter_error_msg pesq_val_nppesq_valpreds_np target_npbrrrrs88   (   . r)Fr )typingrr%rr&rtorchmetrics.utilities.checksrtorchmetrics.utilities.importsrr__doctest_requires__intstrrrrrrrs0