diff --git a/doc/jsk_tools/scripts/audible_warning.md b/doc/jsk_tools/scripts/audible_warning.md
index 75888827f..72849c0d8 100644
--- a/doc/jsk_tools/scripts/audible_warning.md
+++ b/doc/jsk_tools/scripts/audible_warning.md
@@ -23,9 +23,9 @@ Robots using diagnostics can use this node.
 
 ## Parameter
 
-* `~speak_rate` (`Float`, default: `1.0`)
+* `~speak_rate` (`Float`, default: `1.0 / 100.0`)
 
-    Rate of speak loop.
+    Rate of speak loop. If `~wait_speak` is `True`, wait until a robot finish speaking.
 
 * `~speak_interval` (`Float`, default: `120.0`)
 
diff --git a/jsk_tools/node_scripts/audible_warning.py b/jsk_tools/node_scripts/audible_warning.py
index 7f1659cbe..6da9ef6b5 100755
--- a/jsk_tools/node_scripts/audible_warning.py
+++ b/jsk_tools/node_scripts/audible_warning.py
@@ -176,7 +176,7 @@ def run(self):
 class AudibleWarning(object):
 
     def __init__(self):
-        speak_rate = rospy.get_param("~speak_rate", 1.0)
+        speak_rate = rospy.get_param("~speak_rate", 1.0 / 100.0)
         wait_speak = rospy.get_param("~wait_speak", True)
         language = rospy.get_param('~language', '')
         seconds_to_start_speaking = rospy.get_param(