| 摘要 |
The instantaneous amplitude of an analog audio signal is periodically sampled to be converted into multibit binary digital words representing the sampled amplitude. The bits of successive binary words are reproduced serially with appropriate word and frame synchronizing pulses in quaternary coding to pulse code modulate the FM carrier signal used in standard video recording. The recorded signal is demodulated on playback to reproduce successive digital words that are converted into discrete analog values which are low pass filtered to reproduce the original audio signal. The reproduced digital values are also stored in successive address locations of a random access memory. Upon detecting data drop-out, such as by monitoring the FM carrier in the reproduced signal, the memory address sequence is reversed to read out the previously entered values backwards, thus approximating the roughly symmetrical waveform envelope of normal high frequency musical signals so as to avoid discernable discontinuity in the audio output. At the same time, timing pulses are delivered to decrease gradually the bandwidth of the voltage controlled low pass filter to smooth the sample outputs from the memory. When actual data is reacquired, timing pulses used to increment the memory address gradually restore the maximum bandwidth of the voltage controlled low pass filter. |
