From d0635c6592a2e787ca296e15241e7371a83ca55f Mon Sep 17 00:00:00 2001
From: Hui Zhang <zhtclz@foxmail.com>
Date: Mon, 17 May 2021 19:50:29 +0800
Subject: [PATCH] using soxbinddings (#619)

---
 .gitignore                   |  3 +++
 deepspeech/frontend/audio.py | 21 +++++++++++----------
 tools/Makefile               | 25 +++++++++++++++++++------
 3 files changed, 33 insertions(+), 16 deletions(-)

diff --git a/.gitignore b/.gitignore
index 953cc28d..9b225057 100644
--- a/.gitignore
+++ b/.gitignore
@@ -9,6 +9,9 @@
 *.tar.gz
 .ipynb_checkpoints
 *.npz
+*.done
 
 tools/venv
 tools/kenlm
+tools/sox-14.4.2
+tools/soxbindings
diff --git a/deepspeech/frontend/audio.py b/deepspeech/frontend/audio.py
index a1c7df63..4488f5f2 100644
--- a/deepspeech/frontend/audio.py
+++ b/deepspeech/frontend/audio.py
@@ -21,6 +21,7 @@ import struct
 import numpy as np
 import resampy
 import soundfile
+import soxbindings as sox
 from scipy import signal
 
 
@@ -339,18 +340,18 @@ class AudioSegment(object):
             raise ValueError("speed_rate should be greater than zero.")
 
         # numpy
-        old_length = self._samples.shape[0]
-        new_length = int(old_length / speed_rate)
-        old_indices = np.arange(old_length)
-        new_indices = np.linspace(start=0, stop=old_length, num=new_length)
-        self._samples = np.interp(new_indices, old_indices, self._samples)
+        # old_length = self._samples.shape[0]
+        # new_length = int(old_length / speed_rate)
+        # old_indices = np.arange(old_length)
+        # new_indices = np.linspace(start=0, stop=old_length, num=new_length)
+        # self._samples = np.interp(new_indices, old_indices, self._samples)
 
         # sox, slow
-        # tfm = sox.Transformer()
-        # tfm.set_globals(multithread=False)
-        # tfm.speed(speed_rate)
-        # self._samples = tfm.build_array(
-        #     input_array=self._samples, sample_rate_in=self._sample_rate).copy()
+        tfm = sox.Transformer()
+        tfm.set_globals(multithread=False)
+        tfm.speed(speed_rate)
+        self._samples = tfm.build_array(
+            input_array=self._samples, sample_rate_in=self._sample_rate).copy()
 
     def normalize(self, target_db=-20, max_gain_db=300.0):
         """Normalize audio to be of the desired RMS value in decibels.
diff --git a/tools/Makefile b/tools/Makefile
index 01228271..8ae06a20 100644
--- a/tools/Makefile
+++ b/tools/Makefile
@@ -1,13 +1,18 @@
 PYTHON:= python3.8
-.PHONY: all clean kenlm
+.PHONY: all clean
 
-all: virtualenv kenlm
+all: virtualenv kenlm.done sox.done soxbindings.done
 
 virtualenv:
 	test -d venv || virtualenv -p $(PYTHON) venv
 	touch venv/bin/activate
 
-kenlm:
+clean:
+	rm -fr venv
+	find -iname "*.pyc" -delete
+	rm -rf kenlm
+
+kenlm.done:
 	# Ubuntu 16.04 透過 apt 會安裝 boost 1.58.0
 	# it seems that boost (1.54.0) requires higher version. After I switched to g++-5 it compiles normally.
 	apt install -y build-essential cmake libboost-system-dev libboost-thread-dev libboost-program-options-dev libboost-test-dev libeigen3-dev zlib1g-dev libbz2-dev liblzma-dev
@@ -15,8 +20,16 @@ kenlm:
 	test -d kenlm || wget -O - https://kheafield.com/code/kenlm.tar.gz | tar xz
 	mkdir -p kenlm/build && cd kenlm/build && cmake .. && make -j4 && make install
 	cd kenlm && python setup.py install
+	touch kenlm.done
 
-clean:
-	rm -fr venv
-	find -iname "*.pyc" -delete
+sox.done:
+	apt install -y libvorbis-dev libmp3lame-dev libmad-ocaml-dev
+	test -d sox-14.4.2 || wget https://nchc.dl.sourceforge.net/project/sox/sox/14.4.2/sox-14.4.2.tar.gz
+	tar -xvzf sox-14.4.2.tar.gz -C .
+	cd sox-14.4.2 && ./configure --prefix=/usr/ && make -j4 && make install
+	touch sox.done
 
+soxbindings.done:
+	test -d soxbindings || git clone https://github.com/pseeth/soxbindings.git
+	source venv/bin/activate; cd soxbindings && python3 setup.py install
+	touch soxbindings.done