QR-based Watermarking in Audio Subband Using DCT

Abstract

Audio Watermarking is one of the technologies used for copyright and audio data by inserting data or digital information that contains the original information of the audio owner. In this paper, we have designed audio watermarking based on Quantization Index Modulation (QIM) with Lifting Wavelet Transform (LWT), Discrete Cosine Transform (DCT), QR Decomposition (QR). The LWT process produces several subbands of an audio. After selecting a lower frequency of audio subband, the selected subband is transformed from the time domain to the frequency domain by DCT. Each frame of the DCT coefficients is decomposed into an orthogonal matrix and a triangular matrix by QR. Using QIM, a watermark is inserted into triangular matrix by quantizing the data to a value corresponding to the quantizer contained in the watermark. In the audio watermarking experiment, we obtain robust and imperceptible watermarking with very high watermark payload. Our method performs good robustness against Low Pass Filter (LPF), Linear Speed Change (LSC), MP4 compression, resampling, and MP3 compression with payload 615.23 bps and average MOS is more than 4.