I have an application which records audio from the phones mic and uploads the audio to a server. However, to reduce the data charges incurred I only want to upload audio if it contains speech (basically any audio which contains more than plain background noise).
Is there an easy way to do this?
I presume you could analyse the amr file to determine the "frequency" of recorded segments to determine the presence of greater than normal noise... However, if this is a solution I dont have a clue how I would go about coding it.
Any help is greatly appreciated!