diff --git a/audio/tests/audiotools/test_audiotools.sh b/audio/tests/audiotools/test_audiotools.sh
index e1856a8ed..387059d51 100644
--- a/audio/tests/audiotools/test_audiotools.sh
+++ b/audio/tests/audiotools/test_audiotools.sh
@@ -1,5 +1,5 @@
 python -m pip install -r ../../audiotools/requirements.txt
-export PYTHONPATH=$PYTHONPATH:$(realpath ../../..)
+export PYTHONPATH=$PYTHONPATH:$(realpath ../../..) # this is root path of `PaddleSpeech`
 wget  https://paddlespeech.bj.bcebos.com/PaddleAudio/audio_tools/audio.tar.gz
 wget  https://paddlespeech.bj.bcebos.com/PaddleAudio/audio_tools/regression.tar.gz
 tar -zxvf audio.tar.gz