Create FM stereo from Matlab.
in [DSP]
|
Prev: conventional wisdom how to upsample very large arrays, accurately?
Next: FIR quantization effect ( implementing in fpga)
From: Robert Orban on 6 Jan 2010 21:15 In article <6748e7fe-9cbe-408f-8fa9- ff1105bf8260(a)g22g2000prf.googlegroups.com>, disney_snoopy(a)hotmail.com says... > > >Hi all, > >I would appreciate if anyone here could help me. >I was thinking to create a FM stereo with the RDS and L+R and L-R >channel in Matlab. >But i doesnt know how to do on that... >Can anyone here help me plz? > >Thankz (use a monospaced font for the text below) The FM stereo encoding equation, where w = 2 pi 38000 is: #1: comp(L, R) := (L + R) + (L - R)�COS(w�t) + 0.09�COS(w/2�t) The decoding equations using synchronous demodulation are: #2: Lout = comp(L, R)�(0.5 + COS(w�t)) #3: Rout = comp(L, R)�(0.5 - COS(w�t)) Substituting #1 into #2 and #3 and simplifying gives: ( 3�w�t ) ( w�t ) 9�COS(-------) 9�COS|-----| #4: ( 2 ) ( 2 ) (L - R)�COS(2�w�t) Lout = ---------------- + -------------- + -------------------- + 200 100 2 (3�L + R)�COS(w�t) -------------------- + L 2 ( 3�w�t ) 9�COS|-------| #5: ( 2 ) (R - L)�COS(2�w�t) Rout = - ---------------- + -------------------- - 200 2 (L + 3�R)�COS(w�t) -------------------- + R 2 This shows that L and R appear in the outputs of the synchronous demodulator as desired, but there are a number of other terms as well. Assuming that L and R are spectrally limited to less than 19 kHz, L and R can be separated from the remaining terms by applying lowpass filters to #4 and #5 because in this case, the lowest spectral component of the "other terms" is 19 kHz. Typically, the lopwass filters used have a passband to 15 kHz and a stopband beginning at 19 kHz, but nothing in the physics of the system prevents audio response to 19 kHz - (some very small delta). As for the MATLAB implementation of the above, this is an "exercise for the reader." |