I am trying to identify whether signals measured via accelerometer on the ankle bone, while someone is talking, is actually derived from his or her own voice transmitting through bone conduction. I initially used FFT to spectral analyze both the vocal speech and the bone vibration. I want to prove that both signals are derived from the same source. The harmonics in FFT is a mathematical derived value, whereas the standard voice analysis LPC is a measured value. They are not the same. Is LCP a better tool to use in this case?
Dear Abu Dzarr,
why don't you use a coherency analysis? It can be used to estimate the causality between the input and output.
See for example: http://en.wikipedia.org/wiki/Coherence_(signal_processing)
Dear Prof Pannert,
I will like to try using coherency analysis on my test data, and see what comes out of it. Unfortunately I am at the moment only familiar with Labview programming. There is no native vi for coherency analysis available and going by the forum discussion, it appeared that I will require additional mathscript coding. I will like to avoid that, as I have to purchase additional toolkit. I notice that coherency analysis is available in Matlab as the function ‘cohere’. At the moment I am not familiar with Matlab. I tried to twiddle gnu octave (downloaded free) but it appeared to me that, it is better for me to try scaling the learning curve with Matlab than octave. There appeared to be more learning material/support available for the former. I will try to get myself a student version of Matlab next week. It’s affordable. I wonder If I am on the right track?
Dear Abu Dzarr, I wish if you can give a try SCILAB... it look more like MatLab than gnu octave... and has good documentation for beginners... like some workshops for numerical analysis...
- Badges
- Science method