- Compute the average amplitude for each of fifty 0.1 second snippets of the five seconds
- Iterate from each end of the clip until a given number of snippets have amplitude greater than some threshold.
- Tune (a) the number of high-amplitude snippets required to trigger the beginning and end of the audio and (b) the threshold amplitude to achieve your desired accuracy.
Word Prediction in Matlab?
1 次查看(过去 30 天)
显示 更早的评论
Hello I have a tricky situation ?
- I need to develop a code which can record user voice and play back?-> the code is ready and running fine.
- Now the real problem, I need to identify the position of the spoken word.
In detail: say for example i have defined 5 seconds time duration for recording. during this time i speak a sentence "how are you ", say this whole sentence completes in 3 seconds (ideally) and the remaining 2 seconds are vacant and there are no information available.
Now, I have plotted the recorded file also. now i need to determine the position of either of the three words, that is the location of the words in whole 5 seconds duration?
could you please help me in determining code?
0 个评论
回答(1 个)
Luke Halberstadt
2016-8-17
Assuming that the sound amplitude is lower when the words are not being spoken, the following solution should roughly give you a start and end time:
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Speech Recognition 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!