The fast Hartley transform (FHT) is emerging as an alternative to the more popular fast Fourier transform (FFT) in the processing of real signals. In this paper, a fixed-point error analysis has been carried out. The basic differences between the FFT and FHT flowcharts from the point of view of error analysis are first discussed. A suitable scaling model is chosen for the FHT and the average output signal-to-noise ratio is calculated both for decimation-in-frequency and decimation-in-time cases. The results are then compared with the FFT error-analysis results. © 1990.

K. M.Muraleedhara Prabhu

Siva Bala Narayanan

MATHEMATICAL TECHNIQUES - ERROR ANALYSIS

Signal theory

DISCRETE HARTLEY TRANSFORM

FAST HARTLEY TRANSFORM

FIXED-POINT ERROR ANALYSIS

HARTLEY TRANSFORM

Mathematical transformations

IIT Madras is a public technical and research university located in Chennai, Tamil Nadu. Founded in 1959, it is recognised as an Institute of National Importance.

IIT Madras has been ranked as the top engineering institute in India for four years in a row by the National Institutional Ranking Framework of the MHRD

It currently offers undergraduate, postgraduate and research degrees across 16 disciplines in Engineering, Sciences, Humanities and Management. About 596 faculty belonging to science and engineering departments and centres of the Institute are engaged in teaching, research and industrial consultancy.

IIT Madras

Fixed-point error analysis of fast Hartley transform

Signal Processing

The Wigner-Ville distribution (WVD) has been a powerful signal processing tool for time-frequency signal analysis. Consequently, many algorithms have been proposed in the literature for computing the WVD in real-time applications [2, 8, 7, 5]. Recently, Pei and Yang proposed an efficient algorithm which uses discrete Hartley transform (DHT) to evaluate the WVD entirely in the real domain [5]. In this paper, a fixed-point error analysis of this algorithm has been carried out. An expression for output noise-to-signal ratio (NSR) is derived. The results indicate that this algorithm has better noise properties when compared to the FFT-based algorithm, even though the FFT has better noise properties than the FHT.

Some results on fixed-point error analysis of Wigner-Ville distribution

Of late, the fast Hartley transform (FHT) has attracted considerable research interest as an alternative tool to the fast Fourier transform (FFT). In this paper, efficient implementation of the FHT on different DSP processors is considered. Instead of counting the required arithmetic operations, the necessary number of instruction cycles for an implementation of FHT is used as a measure. Also, different scaling schemes, which are available in the literature, have been considered to increase the signal-to-noise ratio (SNR) of the algorithm. Finally, the details of the programs on ADSP2101, TMS320C25 and TMS320C30 have been presented.

Microprocessors and Microsystems

Fast Hartley transform implementation on DSP chips

IEEE Transactions on Acoustics, Speech, and Signal Processing

Improved Fourier and Hartley transform algorithms: Application to cyclic convolution of real data

Journal	Signal Processing
ISSN	01651684
Open Access	No