enthyp / sdr Goto Github PK

ideas:

• record with manually chosen gain (ref: Setting the RF gain suggest RTL AGC kinda sucks for non-TV)
• offset_tuning thing - figure out why capture_rate / 4 is added to capture_freq, it may help with DC spike interference
• related: tune to other frequency and shift - straight avoidance of DC spike
• de-emphasis filter (although rtl_fm rocks without it)
• simple rectangular filters instead of scipy.signal.decimate
• investigate rtl_fm code for more

use GNURadio first!

Steps:

• down shift in frequency (np.exp for now, can be PLL later)
• low-pass filter (2.4 kHz?)
• sampling (based on np.exp frequency again? i.e. no timing recovery for now, PLL/sth else later)
• PSK demodulation? (should be easy from I/Q samples we obtain above)
• differential decoding

• #3
• lock onto 19 kHz to demodulate 38 kHz subcarrier (Double sideband suppressed carrier (DSBSC) modulation)
• produce 2-channel WAV file

• right sideband of L - R is more noisy than left one (they should contain the same info) and it's the one that was used - left could be utilized instead, this requires LSB demodulation

• error correction
• frame recovery

• for stereo FM demodulation - would it improve quality?
• for RDS subcarrier recovery