Hi,
I am keen to know the basics to remove echo knowing FarEnd Signal and NearEnd Signal. I have read research papers in which its commonly said:
y(n) = hTx
where hT is the room impulse response and echo signal estimated is y(n).
Now, what I really want to understand is the computation of room impulse response. Given two-time frame arrays of PCM 16 bit. Let's say:
Near_end_sig = [x1,x2,x3,x4,....,xN] (one time frame)
Far_end_sig = [y1,y2,y3,....,yN] (one time frame)
Here I want to know in this case working on digital acoustic signals (PCM-16 bit arrays of size let's assume 1024). How do I compute room impulse response h and how do I estimate echo signal? My main concern is the computation of room impulse response and the echo signal estimation based on the time frames.
Regards,
Khubaib Ahmad
Please have a look at the following useful link:
http://www.diva-portal.org/smash/get/diva2:23654/FULLTEXT01.pdf
https://smartech.gatech.edu/bitstream/handle/1853/44851/wada_ted_s_201208_phd.pdf
hello,
the room impulse response h = [h(1) h(2) ....h(N)] is modelled as a digital filter with a L taps-finite impulse response (FIR). In room acoustics, to measure this impulse response, a white noise (or pseudo-white noise as MLS sequence) x[n] n=0,1,... is presented at the input of a lousdspeaker, and the corresponding received signal y[n], n=0,1,.... is recorded at the output of a microphone that is placed in the same acoustical environment.
This room impulse response is then estimated by computing the intercorrelation function betwenen the received and transmited signal :
Ryx[k] = E( y[n] x[n+k] )
= h[n] * Rxx[k]
= h [n]
since the autocorrelation function of the input sequence is Rxx[k]= Dirac_Impulse[k] (as we have used a (pseudo)-white noise).
So by computing the intercorrelation function between the known transmitted (i.e; x[n]) and received (i.e. y[n]) sequences, we have directly access to an estimate of the L-taps of the room impulse response h .
I hope this will help.
Pascal.
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers