Yoshikazu Yokotani

Improved integer transforms for lossless audio coding

Lifting scheme based integer transforms are very powerful tools to construct lossless coding schemes. These transforms such as the integer fast fourier transform (IntFFT) and the integer modified discrete cosine transform (IntMDCT) are integer approximations of the original floatingpoint transforms, and hence there is an approximation error in the transform domain. This paper will propose structures for improved integer transforms in terms of improved approximation accuracy and computational efficiency. Experimental results will show that clear improvements in these two points are achieved in lossless audio coding.

Audio Data Hiding with High Data Rates Based on Intmdct

This paper describes high data-rate audio data hiding using the IntMDCT. The IntMDCT is an integer approximation of the MDCT with perfect reconstruction. Based on this transform, we describe a straight-forward way to embed data and extract it in a bit-exact manner while perceptual transparency is maintained. Since the IntMDCT spectrum can be used to obtain a closer approximation of the masking threshold compared to time domain approaches, it is possible to achieve higher data rates. In a simple experimental implementation, we found we could embed data at rates up to about 140 kb/s without introducing audible distortions

Improved lossless audio coding using the noise-shaped IntMDCT

This paper discusses approximation noise shaping to improve the efficiency of the integer modified discrete cosine transform (IntMDCT)-based lossless audio codec. The scheme is applied to rounding operations associated with lifting steps to shape the noise spectrum towards the low frequency bands. In this paper, constraints on the noise shaping filter and a design procedure with the constraints are discussed. Several noise shaping filters are designed and experimental results showing the improvement are presented.

Improved integer transforms using multi-dimensional lifting [audio coding examples]

Recently lifting-based integer transforms have received much attention, especially in the area of lossless audio and image coding. The usual approach is to apply the lifting scheme to each Givens rotation. Especially in the case of long transform sizes in audio coding applications, this leads to a considerable approximation error in the frequency domain. This paper presents a multidimensional lifting approach for reducing this approximation error. In this approach, large parts of the transform are calculated without rounding operations, only the output is rounded and added. The new approach is applied and evaluated for both the integer modified discrete cosine transform (IntMDCT) and the integer fast Fourier transform (IntFFT).