sandeep-krishnamurthy commented on a change in pull request #13325: [MXNET-1210 ] Gluon Audio - Example URL: https://github.com/apache/incubator-mxnet/pull/13325#discussion_r236479037
########## File path: example/gluon/urban_sounds/transforms.py ########## @@ -0,0 +1,207 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + +# coding: utf-8 +# pylint: disable= arguments-differ +"Audio transforms." + +import warnings +import numpy as np +try: + import librosa +except ImportError as e: + warnings.warn("librosa dependency could not be resolved or \ + imported, could not provide some/all transform.") + +from mxnet import ndarray as nd +from mxnet.gluon.block import Block + +class MFCC(Block): + """Extracts Mel frequency cepstrum coefficients from the audio data file + More details : https://librosa.github.io/librosa/generated/librosa.feature.mfcc.html + + Attributes + ---------- + sampling_rate: int, default 22050 + sampling rate of the input audio signal + num_mfcc: int, default 20 + number of mfccs to return + + + Inputs: + - **x**: input tensor (samples, ) shape. + + Outputs: + - **out**: output array is a scaled NDArray with (samples, ) shape. + + """ + + def __init__(self, sampling_rate=22050, num_mfcc=20): + self._sampling_rate = sampling_rate + self._num_fcc = num_mfcc + super(MFCC, self).__init__() + + def forward(self, x): + if isinstance(x, np.ndarray): + y = x + elif isinstance(x, nd.NDArray): + y = x.asnumpy() + else: + warnings.warn("MFCC - allowed datatypes mx.nd.NDArray and numpy.ndarray") + return x + + audio_tmp = np.mean(librosa.feature.mfcc(y=y, sr=self._sampling_rate, n_mfcc=self._num_fcc).T, axis=0) + return nd.array(audio_tmp) + + +class Scale(Block): + """Scale audio numpy.ndarray from a 16-bit integer to a floating point number between + -1.0 and 1.0. The 16-bit integer is the sample resolution or bit depth. + + Attributes + ---------- + scale_factor : float + The factor to scale the input tensor by. + + + Inputs: + - **x**: input tensor (samples, ) shape. + + Outputs: + - **out**: output array is a scaled NDArray with (samples, ) shape. + + Examples + -------- + >>> scale = audio.transforms.Scale(scale_factor=2) + >>> audio_samples = mx.nd.array([2,3,4]) + >>> scale(audio_samples) + [1. 1.5 2. ] + <NDArray 3 @cpu(0)> + + """ + + def __init__(self, scale_factor=2**31): + self.scale_factor = scale_factor + super(Scale, self).__init__() + + def forward(self, x): + if self.scale_factor == 0: + warnings.warn("Scale factor cannot be 0.") + return x + if isinstance(x, np.ndarray): + return nd.array(x/self.scale_factor) + return x / self.scale_factor + + +class PadTrim(Block): Review comment: This looks like a generally useful transforms. Post this PR, can you please see if this can be part of gluon transforms? ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
