minor updates

egs/libritts/ASR/prepare.sh

@@ -85,10 +85,10 @@ if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then
   # Here we shuffle and combine the train-clean-100, train-clean-360 and 
   # train-other-500 together to form the training set.
   if [ ! -f data/fbank/libritts_cuts_train-all-shuf.jsonl.gz ]; then
-    cat <(gunzip -c ./libritts_cuts_train-clean-100.jsonl.gz) \
+    cat <(gunzip -c data/fbank/libritts_cuts_train-clean-100.jsonl.gz) \
-      <(gunzip -c ./libritts_cuts_train-clean-360.jsonl.gz) \
+      <(gunzip -c data/fbank/libritts_cuts_train-clean-360.jsonl.gz) \
-      <(gunzip -c ./libritts_cuts_train-other-500.jsonl.gz) | \
+      <(gunzip -c data/fbank/libritts_cuts_train-other-500.jsonl.gz) | \
-      shuf | gzip -c > ./libritts_cuts_train-all-shuf.jsonl.gz
+      shuf | gzip -c > data/fbank/libritts_cuts_train-all-shuf.jsonl.gz
   fi
 
   if [ ! -e data/fbank/.libritts-validated.done ]; then

egs/libritts/CODEC/local/compute_spectrogram_libritts.py

@@ -46,6 +46,7 @@ from icefall.utils import get_executor
 torch.set_num_threads(1)
 torch.set_num_interop_threads(1)
 
+
 def get_args():
     parser = argparse.ArgumentParser()
 
@@ -64,12 +65,13 @@ def get_args():
     return parser.parse_args()
 
 
-def compute_spectrogram_libritts(dataset: Optional[str] = None, sampling_rate: int = 24000,):
+def compute_spectrogram_libritts(
+    dataset: Optional[str] = None, sampling_rate: int = 24000
+):
     src_dir = Path("data/manifests")
     output_dir = Path("data/spectrogram")
     num_jobs = min(32, os.cpu_count())
 
-
     frame_length = 1024 / sampling_rate  # (in second)
     frame_shift = 256 / sampling_rate  # (in second)
     use_fft_mag = True

egs/libritts/CODEC/local/display_manifest_statistics.py

@@ -0,0 +1,341 @@
+#!/usr/bin/env python3
+# Copyright    2023  Xiaomi Corp.             (authors: Zengwei Yao)
+#              2024  The Chinese Univ. of HK  (authors: Zengrui Jin)
+#
+# See ../../../../LICENSE for clarification regarding multiple authors
+#
+# Licensed 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.
+
+"""
+This file displays duration statistics of utterances in a manifest.
+You can use the displayed value to choose minimum/maximum duration
+to remove short and long utterances during the training.
+"""
+
+
+from lhotse import load_manifest_lazy
+
+
+def main():
+    paths = [
+        "./data/spectrogram/libritts_cuts_train-clean-100.jsonl.gz",
+        "./data/spectrogram/libritts_cuts_train-clean-360.jsonl.gz",
+        "./data/spectrogram/libritts_cuts_train-other-500.jsonl.gz",
+        "./data/spectrogram/libritts_cuts_dev-clean.jsonl.gz",
+        "./data/spectrogram/libritts_cuts_dev-other.jsonl.gz",
+        "./data/spectrogram/libritts_cuts_test-clean.jsonl.gz",
+        "./data/spectrogram/libritts_cuts_test-other.jsonl.gz",
+    ]
+    for path in paths:
+        cuts = load_manifest_lazy(path)
+        cuts.describe()
+
+
+if __name__ == "__main__":
+    main()
+
+"""
+./data/spectrogram/libritts_cuts_train-clean-100.jsonl.gz statistics:
+________________________________________
+_ Cuts count:               _ 33236    _
+________________________________________
+_ Total duration (hh:mm:ss) _ 53:47:18 _
+________________________________________
+_ mean                      _ 5.8      _
+________________________________________
+_ std                       _ 4.6      _
+________________________________________
+_ min                       _ 0.2      _
+________________________________________
+_ 25%                       _ 2.4      _
+________________________________________
+_ 50%                       _ 4.5      _
+________________________________________
+_ 75%                       _ 7.9      _
+________________________________________
+_ 99%                       _ 21.4     _
+________________________________________
+_ 99.5%                     _ 23.7     _
+________________________________________
+_ 99.9%                     _ 27.8     _
+________________________________________
+_ max                       _ 33.2     _
+________________________________________
+_ Recordings available:     _ 33236    _
+________________________________________
+_ Features available:       _ 33236    _
+________________________________________
+_ Supervisions available:   _ 33236    _
+________________________________________
+SUPERVISION custom fields:
+Speech duration statistics:
+__________________________________________________________________
+_ Total speech duration        _ 53:47:18 _ 100.00% of recording _
+__________________________________________________________________
+_ Total speaking time duration _ 53:47:18 _ 100.00% of recording _
+__________________________________________________________________
+_ Total silence duration       _ 00:00:01 _ 0.00% of recording   _
+__________________________________________________________________
+
+./data/spectrogram/libritts_cuts_train-clean-360.jsonl.gz statistics:
+_________________________________________
+_ Cuts count:               _ 116500    _
+_________________________________________
+_ Total duration (hh:mm:ss) _ 191:17:42 _
+_________________________________________
+_ mean                      _ 5.9       _
+_________________________________________
+_ std                       _ 4.6       _
+_________________________________________
+_ min                       _ 0.1       _
+_________________________________________
+_ 25%                       _ 2.4       _
+_________________________________________
+_ 50%                       _ 4.6       _
+_________________________________________
+_ 75%                       _ 8.1       _
+_________________________________________
+_ 99%                       _ 21.3      _
+_________________________________________
+_ 99.5%                     _ 23.4      _
+_________________________________________
+_ 99.9%                     _ 27.4      _
+_________________________________________
+_ max                       _ 40.4      _
+_________________________________________
+_ Recordings available:     _ 116500    _
+_________________________________________
+_ Features available:       _ 116500    _
+_________________________________________
+_ Supervisions available:   _ 116500    _
+_________________________________________
+SUPERVISION custom fields:
+Speech duration statistics:
+___________________________________________________________________
+_ Total speech duration        _ 191:17:42 _ 100.00% of recording _
+___________________________________________________________________
+_ Total speaking time duration _ 191:17:42 _ 100.00% of recording _
+___________________________________________________________________
+_ Total silence duration       _ 00:00:01  _ 0.00% of recording   _
+___________________________________________________________________
+
+./data/spectrogram/libritts_cuts_train-other-500.jsonl.gz statistics:
+_________________________________________
+_ Cuts count:               _ 205043    _
+_________________________________________
+_ Total duration (hh:mm:ss) _ 310:04:36 _
+_________________________________________
+_ mean                      _ 5.4       _
+_________________________________________
+_ std                       _ 4.4       _
+_________________________________________
+_ min                       _ 0.1       _
+_________________________________________
+_ 25%                       _ 2.3       _
+_________________________________________
+_ 50%                       _ 4.2       _
+_________________________________________
+_ 75%                       _ 7.3       _
+_________________________________________
+_ 99%                       _ 20.6      _
+_________________________________________
+_ 99.5%                     _ 22.8      _
+_________________________________________
+_ 99.9%                     _ 27.4      _
+_________________________________________
+_ max                       _ 43.9      _
+_________________________________________
+_ Recordings available:     _ 205043    _
+_________________________________________
+_ Features available:       _ 205043    _
+_________________________________________
+_ Supervisions available:   _ 205043    _
+_________________________________________
+SUPERVISION custom fields:
+Speech duration statistics:
+___________________________________________________________________
+_ Total speech duration        _ 310:04:36 _ 100.00% of recording _
+___________________________________________________________________
+_ Total speaking time duration _ 310:04:36 _ 100.00% of recording _
+___________________________________________________________________
+_ Total silence duration       _ 00:00:01  _ 0.00% of recording   _
+___________________________________________________________________
+
+./data/spectrogram/libritts_cuts_dev-clean.jsonl.gz statistics:
+________________________________________
+_ Cuts count:               _ 5736     _
+________________________________________
+_ Total duration (hh:mm:ss) _ 08:58:13 _
+________________________________________
+_ mean                      _ 5.6      _
+________________________________________
+_ std                       _ 4.3      _
+________________________________________
+_ min                       _ 0.3      _
+________________________________________
+_ 25%                       _ 2.4      _
+________________________________________
+_ 50%                       _ 4.4      _
+________________________________________
+_ 75%                       _ 7.8      _
+________________________________________
+_ 99%                       _ 19.9     _
+________________________________________
+_ 99.5%                     _ 21.9     _
+________________________________________
+_ 99.9%                     _ 26.3     _
+________________________________________
+_ max                       _ 30.1     _
+________________________________________
+_ Recordings available:     _ 5736     _
+________________________________________
+_ Features available:       _ 5736     _
+________________________________________
+_ Supervisions available:   _ 5736     _
+________________________________________
+SUPERVISION custom fields:
+Speech duration statistics:
+__________________________________________________________________
+_ Total speech duration        _ 08:58:13 _ 100.00% of recording _
+__________________________________________________________________
+_ Total speaking time duration _ 08:58:13 _ 100.00% of recording _
+__________________________________________________________________
+_ Total silence duration       _ 00:00:01 _ 0.00% of recording   _
+__________________________________________________________________
+
+./data/spectrogram/libritts_cuts_dev-other.jsonl.gz statistics:
+________________________________________
+_ Cuts count:               _ 4613     _
+________________________________________
+_ Total duration (hh:mm:ss) _ 06:25:52 _
+________________________________________
+_ mean                      _ 5.0      _
+________________________________________
+_ std                       _ 4.1      _
+________________________________________
+_ min                       _ 0.3      _
+________________________________________
+_ 25%                       _ 2.2      _
+________________________________________
+_ 50%                       _ 3.8      _
+________________________________________
+_ 75%                       _ 6.5      _
+________________________________________
+_ 99%                       _ 19.7     _
+________________________________________
+_ 99.5%                     _ 24.5     _
+________________________________________
+_ 99.9%                     _ 31.0     _
+________________________________________
+_ max                       _ 32.6     _
+________________________________________
+_ Recordings available:     _ 4613     _
+________________________________________
+_ Features available:       _ 4613     _
+________________________________________
+_ Supervisions available:   _ 4613     _
+________________________________________
+SUPERVISION custom fields:
+Speech duration statistics:
+__________________________________________________________________
+_ Total speech duration        _ 06:25:52 _ 100.00% of recording _
+__________________________________________________________________
+_ Total speaking time duration _ 06:25:52 _ 100.00% of recording _
+__________________________________________________________________
+_ Total silence duration       _ 00:00:01 _ 0.00% of recording   _
+__________________________________________________________________
+
+./data/spectrogram/libritts_cuts_test-clean.jsonl.gz statistics:
+________________________________________
+_ Cuts count:               _ 4837     _
+________________________________________
+_ Total duration (hh:mm:ss) _ 08:34:09 _
+________________________________________
+_ mean                      _ 6.4      _
+________________________________________
+_ std                       _ 5.1      _
+________________________________________
+_ min                       _ 0.3      _
+________________________________________
+_ 25%                       _ 2.4      _
+________________________________________
+_ 50%                       _ 4.8      _
+________________________________________
+_ 75%                       _ 8.9      _
+________________________________________
+_ 99%                       _ 22.6     _
+________________________________________
+_ 99.5%                     _ 24.4     _
+________________________________________
+_ 99.9%                     _ 29.6     _
+________________________________________
+_ max                       _ 36.7     _
+________________________________________
+_ Recordings available:     _ 4837     _
+________________________________________
+_ Features available:       _ 4837     _
+________________________________________
+_ Supervisions available:   _ 4837     _
+________________________________________
+SUPERVISION custom fields:
+Speech duration statistics:
+__________________________________________________________________
+_ Total speech duration        _ 08:34:09 _ 100.00% of recording _
+__________________________________________________________________
+_ Total speaking time duration _ 08:34:09 _ 100.00% of recording _
+__________________________________________________________________
+_ Total silence duration       _ 00:00:01 _ 0.00% of recording   _
+__________________________________________________________________
+
+./data/spectrogram/libritts_cuts_test-other.jsonl.gz statistics:
+________________________________________
+_ Cuts count:               _ 5120     _
+________________________________________
+_ Total duration (hh:mm:ss) _ 06:41:31 _
+________________________________________
+_ mean                      _ 4.7      _
+________________________________________
+_ std                       _ 3.8      _
+________________________________________
+_ min                       _ 0.3      _
+________________________________________
+_ 25%                       _ 1.8      _
+________________________________________
+_ 50%                       _ 3.6      _
+________________________________________
+_ 75%                       _ 6.5      _
+________________________________________
+_ 99%                       _ 17.8     _
+________________________________________
+_ 99.5%                     _ 20.4     _
+________________________________________
+_ 99.9%                     _ 23.8     _
+________________________________________
+_ max                       _ 27.3     _
+________________________________________
+_ Recordings available:     _ 5120     _
+________________________________________
+_ Features available:       _ 5120     _
+________________________________________
+_ Supervisions available:   _ 5120     _
+________________________________________
+SUPERVISION custom fields:
+Speech duration statistics:
+__________________________________________________________________
+_ Total speech duration        _ 06:41:31 _ 100.00% of recording _
+__________________________________________________________________
+_ Total speaking time duration _ 06:41:31 _ 100.00% of recording _
+__________________________________________________________________
+_ Total silence duration       _ 00:00:01 _ 0.00% of recording   _
+__________________________________________________________________
+"""

egs/libritts/CODEC/local/validate_manifest.py

@@ -0,0 +1 @@
+../../../ljspeech/TTS/local/validate_manifest.py

egs/libritts/CODEC/prepare.sh

@@ -0,0 +1,87 @@
+#!/usr/bin/env bash
+
+# fix segmentation fault reported in https://github.com/k2-fsa/icefall/issues/674
+export PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python
+
+set -eou pipefail
+
+stage=0
+stop_stage=100
+sampling_rate=24000
+nj=32
+
+dl_dir=$PWD/download
+
+. shared/parse_options.sh || exit 1
+
+# All files generated by this script are saved in "data".
+# You can safely remove "data" and rerun this script to regenerate it.
+mkdir -p data
+
+log() {
+  # This function is from espnet
+  local fname=${BASH_SOURCE[1]##*/}
+  echo -e "$(date '+%Y-%m-%d %H:%M:%S') (${fname}:${BASH_LINENO[0]}:${FUNCNAME[1]}) $*"
+}
+
+log "dl_dir: $dl_dir"
+
+if [ $stage -le 0 ] && [ $stop_stage -ge 0 ]; then
+  log "Stage 0: Download data"
+
+  # If you have pre-downloaded it to /path/to/LibriTTS,
+  # you can create a symlink
+  #
+  #   ln -sfv /path/to/LibriTTS $dl_dir/LibriTTS
+  #
+  if [ ! -d $dl_dir/LibriTTS ]; then
+    lhotse download libritts $dl_dir
+  fi
+
+  # If you have pre-downloaded it to /path/to/musan,
+  # you can create a symlink
+  #
+  #   ln -sfv /path/to/musan $dl_dir/musan
+  #
+  if [ ! -d $dl_dir/musan ]; then
+    lhotse download musan $dl_dir
+  fi
+fi
+
+if [ $stage -le 1 ] && [ $stop_stage -ge 1 ]; then
+  log "Stage 1: Prepare LibriTTS manifest"
+  # We assume that you have downloaded the LibriTTS corpus
+  # to $dl_dir/LibriTTS
+  mkdir -p data/manifests
+  if [ ! -e data/manifests/.libritts.done ]; then
+    lhotse prepare libritts --num-jobs 32 $dl_dir/LibriTTS data/manifests
+    touch data/manifests/.libritts.done
+  fi
+fi
+
+
+if [ $stage -le 2 ] && [ $stop_stage -ge 2 ]; then
+  log "Stage 2: Compute Spectrogram for LibriTTS"
+  mkdir -p data/spectrogram
+  if [ ! -e data/spectrogram/.libritts.done ]; then
+    ./local/compute_spectrogram_libritts.py --sampling-rate $sampling_rate 
+    touch data/spectrogram/.libritts.done
+  fi
+
+  # Here we shuffle and combine the train-clean-100, train-clean-360 and 
+  # train-other-500 together to form the training set.
+  if [ ! -f data/spectrogram/libritts_cuts_train-all-shuf.jsonl.gz ]; then
+    cat <(gunzip -c data/spectrogram/libritts_cuts_train-clean-100.jsonl.gz) \
+      <(gunzip -c data/spectrogram/libritts_cuts_train-clean-360.jsonl.gz) \
+      <(gunzip -c /data/spectrogramlibritts_cuts_train-other-500.jsonl.gz) | \
+      shuf | gzip -c > data/spectrogram/libritts_cuts_train-all-shuf.jsonl.gz
+  fi
+
+  if [ ! -e data/spectrogram/.libritts-validated.done ]; then
+    log "Validating data/spectrogram for LibriTTS"
+    ./local/validate_manifest.py \
+      data/spectrogram/libritts_cuts_train-all-shuf.jsonl.gz
+    touch data/spectrogram/.libritts-validated.done
+  fi
+fi
+

egs/libritts/CODEC/shared

@@ -0,0 +1 @@
+../../../icefall/shared/