Multi-bit Data Hiding in Randomly Chosen LSB Layers of an Audio

LSB techniques generally embed data in the same LSB position of consecutive samples which helps intruders to extract secret information easily. This paper solve this problem by introducing a robust audio steganography technique where data is embedded in multiple layers of LSB chosen randomly and in non-consecutive samples. The choice of random LSB layers and non-consecutive pixels for embedding increases robustness as well as the strength of proposed steganography algorithm. It is seriously a problem that the data hiding at non-contiguous sample locations loses the capacity of stego audio. This problem is solved here by embedding three bits within a target samples. The capacity is also increased by using 6 bits ASCII representation of secret message instead of 7. The proposed technique is tested by embedding text of different payloads within the cover audio and also compared with existing techniques based on quality and capacity.