我正在尝试建立一个系统来接收来自连续发言者的演讲。
喜欢这个。
1- start recording
2- When you hear a sound start a trigger which detects -when that speech is ended-?
3- When 1-2 seconds silent happened fire that trigger and stop recording
4- analyze and recognize the recorded speech ( I have a component for that I dont do that by my self)
5- Start the system from the beginning
我尝试了一些算法，比如以100 ms的间隔测量2个secons的声级。然后将它们相加并得到一个值来检查。 但是因为它太复杂了我无法实现我的目标现在做什么？