The application of convolutional error correcting codes to digital audio is examined. One implementation (R 2/3, 30 taps) requiring one extra bit for each two information bits, has been simulated and would perform about one miscorrection per hour at a mean bit error rate of 10-3. The decoding hardware would consist of 20 Kbits of memory and a comparatively small amount of logic. Implementation as an integrated circuit would be feasible.
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.