"Real" music has frequencies much higher than most human voice, but that's not going to help you, because the frequency range of the telephone network, at least in North America, is about 300 to 3400 Hz. We don't realize the missing frequencies because we're all so used to hearing poor quality sound from telephones, that our brains automatically fill in the missing frequencies.

Unfortunately for you, you still get that same low quality frequency range when hearing music over the telephone, so there isn't an easy way to tell music from speech.

Steve

On Mon, Mar 17, 2014 at 5:55 PM, Muhammad Shahzad <shaheryarkh@gmail.com> wrote:

nah, i just want to skip the music, not the human voice even if s/he is singing. I am not an expert in this area but i think usually male's voice is around between 300 - 3000 Hz and females have a little higher frequency, while most of musical instruments are over 5000 Hz. So, if we can see audio is less then say 5000 Hz then record else, skip. Just a rough idea, i am sure professionals like you can devise something better or improve this.

Thank you.