generate_wav.sh

#!/bin/bash

# Copyright 2018 Nagoya University (Tomoki Hayashi)
#  Apache 2.0  (http://www.apache.org/licenses/LICENSE-2.0)

# Begin configuration section.
nj=2
fs=22050
n_fft=1024
n_shift=256
cmd=run.pl
help_message=$(cat <<EOF
Usage:
  $0 [options] <model-path> <data-dir> [<log-dir> [<fbank-dir>] ]
Example:
  $0 ljspeech.wavenet.ns.v1/checkpoint-1000000.pkl data/train exp/wavenet_vocoder/train wav
Note:
  <log-dir> defaults to <data-dir>/log, and <fbank-dir> defaults to <data-dir>/data
Options:
  --nj <nj>             # number of parallel jobs
  --fs <fs>             # sampling rate (default=22050)
  --n_fft <n_fft>       # number of FFT points (default=1024)
  --n_shift <n_shift>   # shift size in point (default=256)
  --cmd (utils/run.pl|utils/queue.pl <queue opts>) # how to run jobs.
EOF
)
# End configuration section.

echo "$0 $*"  # Print the command line for logging

. parse_options.sh || exit 1;

if [ $# -lt 2 ] || [ $# -gt 4 ]; then
    echo "${help_message}"
    exit 1;
fi

model=$1
data=$2
if [ $# -ge 3 ]; then
  logdir=$3
else
  logdir=${data}/log
fi
if [ $# -ge 4 ]; then
  wavdir=$4
else
  wavdir=${data}/data
fi

# use "name" as part of name of the archive.
name=$(basename ${data})

mkdir -p ${wavdir} || exit 1;
mkdir -p ${logdir} || exit 1;

scp=${data}/feats.scp

split_scps=""
for n in $(seq ${nj}); do
    split_scps="$split_scps $logdir/feats.${n}.scp"
done

utils/split_scp.pl ${scp} ${split_scps} || exit 1;

${cmd} JOB=1:${nj} ${logdir}/generate_with_wavenet_${name}.JOB.log \
    generate_wav_from_fbank.py \
        --model ${model} \
        --fs ${fs} \
        --n_fft ${n_fft} \
        --n_shift ${n_shift} \
        scp:${logdir}/feats.JOB.scp \
        ${wavdir}

rm ${logdir}/feats.*.scp 2>/dev/null

echo "Succeeded creating wav for ${name}"