I am working with an audio sound profile and. I want to analyse the frequency of that sound and I am using wavpad sound editing software for frequency analysis of that sound. In that case, I have generated frequency versus time graph. But these sound shows multiple frequencies at a time. So I'm not able to generate a perfect graph and also not able to find a frequency range of that sound. So, can you tell me how I can analyse these frequencies ?
Your Question is not very clear. you need to be very specific in what exactly you are planning to analyse.
There are many ways sound can be analysed. Some of the basic ones
1. Spectrum: Which gives information about amplitude at every frequency.
2. Spectrogram: Which gives information how dose frequency changes over time.
The first question you have to ask yourself is: do the frequencies of my signal change over time, or they are always the same.
If you signal is stationary, then you can answer your questions by looking at its Fourier transform (or spectrum).
If your signal is not stationary, then it can be a bit more complicated. The Short-Time Fourier Transform could be a good tool for the analysis.
Hello,
The most popular way of frequency analysis is by spectrum, but the disadvantage is that we can't do time-frequency analysis. You can use Short Time Fourier Transform (STFT) or Wavelet Ttransform for time frequency analysis. Wavelet transform can provide you better time and frequency resolution because we don't fix the envelop as we do it in STFT. Depending on the type of problem and the results you are looking for you can choose either of these methods.
Hi,
A time-frequency representation of the sound always give you different frequencies at each time, except in the case you have a perfect sinusoidal signal. These frequencies that appear are the frequency content of the signal that characterizes the timbre of the sound.
It's important to know what do you mean with "perfect graph". If you are looking for the fundamental frequency of the sound you will need an algorithm that computes this fundamental frequency from the frequency spectrum of your sound.
If what you want is do pitch tracking (get a graph that shows you what note is being played at a particular time) you need something like melodyne. All frequency representations mentioned above will give you all the frequency components that are present on the signal, which is what I think you're getting with wavpad.
I believe that the basic theory was already covered in the previous answers, but if you need a stool to do that, I'd suggest the open-sources http://web.audacityteam.org or (even better) http://www.sonicvisualiser.org Both can plot the spectrogram of a given audio file but (I think) the last one even let you export the spectrum data.
As Jose Ramon or Sandra already said, it is normal to find multiple frequency components in an audio sound profile (e.g. speech). If you are looking for a frequency range for that sound just define an amplitude of a noise floor in your spectrum and each frequency component that exceeds that threshold belongs to your frequency range. You might be able to define a lower frequency border when first peaks exceed that threshold and you might find an upper bound when the last peak exceeds that threshold.
- Badges
- Science topic