### Table 2. Performance of the proposed multiple-substream UEP algorithm for the FOOTBALL sequence. Frequency domain parti- tioning was used for these experiments. a109 is the mean packet loss

2002

Cited by 1

### Table 1: Four types of signals and their properties in the time domain. Type Time Domain Frequency Domain Transform

"... In PAGE 1: ... INTRODUCTION There are four types of signals often used in signal processing, and to analyze these signals, there are four types of Fourier trans- forms. In Table1 we list the four types of signals along with their properties and the appropriate Fourier transform. Since the Fourier transform is linear, the discrete Fourier transforms are samples of the continuous Fourier transform under the appropriate sampling conditions.... In PAGE 1: ... The Cohen class of time-frequency distributions [1, 2] was originally formulated for type I signals. Recently, this class has been extended to the three types of discrete signals in Table1 us- ing both an axiomatic approach [3, 4] and an operator theory ap- proach [5, 6, 7].... ..."

### Table 3: Bayesian estimation of a long-memory stochastic volatility model where d = 0:25, QMLE is the approximate frequency domain, quasi-maximum likelihood estimator, and GPH is the log-periodogram regression using [T1=2] frequencies.

"... In PAGE 15: ...) of Tables 2 - 4 indicate that our MCMC algorithm mixes well and produces near independent draws from the posterior distribution. All of the ine ciency measures are below ve with the largest (4:4087) being found in Table3 for 2 when d = 0:25, T = 4096, and 2 = 0:5. The long-memory parameters ine ciency mea- sures are all less than four.... In PAGE 19: ...Table3 where d = 0:25. It is important to also point out that empirically the standard error of the QMLE is not available and hence, construction of the theoretical 95% con dence interval is generally not possible.... ..."

### Table 4: Bayesian estimation of a long-memory stochastic volatility model where d = 0:4, QMLE is the approximate frequency domain, quasi-maximum likelihood estimator, and GPH is the log-periodogram regression using [T1=2] frequencies.

"... In PAGE 18: ... In all but three cases, the 95% Bayesian con dence interval constructed from the draws of djW(y) contain the true parameter value. In the three cases where the con dence interval does not (all found in Table4 where d = 0:4) the 97.5 percentiles of d are 0:3983 (T = 4096; 2 = 1), 0:3809 (T = 4096; 2 = 0:5), and 0:3998 (T = 8192; 2 = 0:5); i.... ..."

### Table 2: Adjusted Periodogram Regression for four countries. Results for increasing range, with zero padding and frequency domain seasonal adjustment. Estimates of d.

"... In PAGE 4: ...by omitting all periodogram ordinates at the seasonal frequencies. In Table2 we present the results of this estimation procedure for the four \seasonal quot; in ation series analyzed in Hassler and Wolters (1995), which are now also reliable for m gt; 95 as well. We also show outcomes for the asymptotically e cient approximate frequency domain ML estimator for the simple fractionally integrated process applied in Boes et al.... In PAGE 4: ... The results of the periodogram regression and the two approximate ML estimators are now close, see the last rows of Table 2. Table2 around here One might be tempted to use seasonal adjustment for quot;stochastic seasonality quot; like Cen- sus X-11 or ARIMA model based methods as an alternative way to avoid the singularities. That is not an option.... ..."

### TABLE V MPEG-7 BASIC AND BASIC SPECTRAL DESCRIPTORS PROVIDE A BASIC TIME DOMAIN ANALYSIS AND A BASIC FREQUENCY DOMAIN ANALYSIS.

in 5 envelope

### TABLE V MPEG-7 BASIC AND BASIC SPECTRAL DESCRIPTORS PROVIDE A BASIC TIME DOMAIN ANALYSIS AND A BASIC FREQUENCY DOMAIN ANALYSIS.

in dist frdb

2004