add new modif

local/format_lms.sh

@@ -43,7 +43,8 @@ trap "rm -r $tmpdir" EXIT
 
 mkdir -p $tmpdir
 #for lm_suffix in tgsmall tgmed tglarge fglarge; do
-for lm_suffix in fglarge tglarge tgsphinx french-small mixed; do
+#for lm_suffix in fglarge tglarge tgsphinx french-small mixed; do
+for lm_suffix in sg; do
 #for lm_suffix in fglarge; do
 #for lm_suffix in mixed; do
 #for lm_suffix in linto1; do

local/lm/train_lm.sh

@@ -39,7 +39,7 @@ lm_dir=$4
 
 [[ -d "$corpus_dir" ]] || { echo "No such directory '$corpus_dir'"; exit 1; }
 # Normalize data
-normjobs=2
+normjobs=32
 split_prefix=$tmp_dir/split
 
 if [ "$stage" -le 1 ]; then

local/parseESTERSyncV2.py

@@ -371,7 +371,8 @@ if __name__=="__main__":
         if spk[0] in Spk_that_contribute_to_meeting:
             spk_id = str(basename)+'_%s-%03d' % (str(re.sub('\d+','',spk[0])),int(re.sub('[a-zA-Z]','',spk[0])))
             spk2gender.write(spk_id+" "+spk[1]+"\n")
-    wav_scp.write(basename+" sox "+os.path.dirname(file_trs) + '/' + basename + '.wav'+" -t wav -r 16000 -c 1 - |\n")
+    wav_scp.write(basename+" ffmpeg-normalize "+os.path.dirname(file_trs) + '/' + basename + '.wav' + " && sox "+os.path.dirname(file_trs) + '/normalized-' + basename + '.wav'+" -t wav -r 16000 -c 1 - |\n")
+#    wav_scp.write(basename+" sox --norm "+os.path.dirname(file_trs) + '/normalized-' + basename + '.wav'+" -t wav -r 16000 -c 1 - |\n")
     segments_file.close()
     utt2spk_file.close()
     text_file.close()