diff --git a/tools/AP_BWE_main/24kto48k/readme.txt b/tools/AP_BWE_main/24kto48k/readme.txt
index de118a9..21cef4a 100644
--- a/tools/AP_BWE_main/24kto48k/readme.txt
+++ b/tools/AP_BWE_main/24kto48k/readme.txt
@@ -1,3 +1,6 @@
+For the inference of the v3 model, if you find that the generated audio sounds somewhat muffled, you can try using this audio super-resolution model.
+对于v3模型的推理，如果你发现生成的音频比较闷，可以尝试这个音频超分模型。
+
 put g_24kto48k.zip and config.json in this folder
 把g_24kto48k.zip and config.json下到这个文件夹