# signal's spectrum energy

I have a signal with bandwidth between 0-2KHz. If i would like to know the signal's frequency spectrum energy, then I use Parseval's theorem and calculate it over time. Now it's my question: What if I would like to know only part of the spectrum energy, let's say 1.5Khz-2Khz, Is there a low cost way to do it, instead of FFT the signal and then sum the relevant frequency band?

• You could band pass filter the data in the desired frequency range, and then compute the energy of the output signal in the time domain. Jun 8, 2015 at 12:09
• Thanks, i have thought about it, but I wondered if there is another way, maybe something that can give me more insight.
– Udi
Jun 8, 2015 at 12:14
• I guess it's either time domain (i.e., filtering) or frequency domain (i.e., FFT). I don't see many more options. I'm not sure what kind of insight you're expecting. Jun 8, 2015 at 12:17
• I agree with @MattL. Consider a similar question: You have the signal's spectrum, and need to find its energy over a time interval. It can't be done without manipulating the signal in time.
– MBaz
Jun 8, 2015 at 13:50
• What is low cost? FFT could cost with todays technology for your needs under \$5.
– Moti
Jun 10, 2015 at 6:19