Search

Reverberation, especially in large rooms, severely degrades speech recognition performance and speech intelligibility. Since direct measurement of room characteristics is usually not possible, blind estimation of reverberation-related metrics such as the reverberation time (RT) and the direct-to-reverberant energy ratio (DRR) can be valuable information to speech recognition and enhancement algorithms operating in enclosed environments. The objective of this work is to evaluate the performance of five variants of blind RT and DRR estimators based on a modulation spectrum representation of reverberant speech with single- and multi-channel speech data. These models are all based on variants of the so-called Speech-to-Reverberation Modulation Energy Ratio (SRMR). We show that these measures outperform a state-of-the-art baseline based on maximum-likelihood estimation of sound decay rates in terms of root-mean square error (RMSE), as well as Pearson correlation. Compared to the baseline, the best proposed measure, called NSRMR_k , achieves a 23% relative improvement in terms of RMSE and allows for relative correlation improvements ranging from 13% to 47% for RT prediction., Comment: In Proceedings of the ACE Chal- lenge Workshop - a satellite event of IEEE-WASPAA 2015 (arXiv:1510.00383)

A generalisation of the Shannon complex wavelet is introduced, which is related to raised cosine filters. This approach is used to derive a new family of orthogonal complex wavelets based on the Nyquist criterion for Intersymbolic Interference (ISI) elimination. An orthogonal Multiresolution Analysis (MRA) is presented, showing that the roll-off parameter should be kept below 1/3. The pass-band behaviour of the Wavelet Fourier spectrum is examined. The left and right roll-off regions are asymmetric; nevertheless the Q-constant analysis philosophy is maintained. Finally, a generalisation of the (square root) raised cosine wavelets is proposed., Comment: 8 pages, 14 figures

This paper introduces some foundations of wavelets over Galois fields. Standard orthogonal finite-field wavelets (FF-Wavelets) including FF-Haar and FF-Daubechies are derived. Non-orthogonal FF-wavelets such as B-spline over GF(p) are also considered. A few examples of multiresolution analysis over Finite fields are presented showing how to perform Laplacian pyramid filtering of finite block lengths sequences. An application of FF-wavelets to design spread-spectrum sequences is presented., Comment: 4 pages, 1 figure. conference: XIX Simposio Brasileiro de Telecomunicacoes, 2001, Fortaleza, CE, Brazil

We explore the parameter space of the minimal supersymmetric extension of the Standard Model (MSSM), allowing the soft supersymmetry-breaking masses of the Higgs multiplets, m_{1,2}, to be non-universal (NUHM). Compared with the constrained MSSM (CMSSM) in which m_{1,2} are required to be equal to the soft supersymmetry-breaking masses m_0 of the squark and slepton masses, the Higgs mixing parameter mu and the pseudoscalar Higgs mass m_A, which are calculated in the CMSSM, are free in the NUHM model. We incorporate accelerator and dark matter constraints in determining allowed regions of the (mu, m_A), (mu, M_2) and (m_{1/2}, m_0) planes for selected choices of the other NUHM parameters. In the examples studied, we find that the LSP mass cannot be reduced far below its limit in the CMSSM, whereas m_A may be as small as allowed by LEP for large tan \beta. We present in Appendices details of the calculations of neutralino-slepton, chargino-slepton and neutralino-sneutrino coannihilation needed in our exploration of the NUHM., Comment: 92 pages LaTeX, 32 eps figures, final version, some changes to figures pertaining to the b to s gamma constraint

We extend previous combinations of LEP and cosmological relic density constraints on the parameter space of the constrained MSSM, with universal input supersymmetry-breaking parameters, to large tan beta. We take account of the possibility that the lightest Higgs boson might weigh about 115 GeV, but also retain the possibility that it might be heavier. We include the most recent implementation of the b to s gamma constraint at large tan beta. We refine previous relic density calculations at large tan beta by combining a careful treatment of the direct-channel Higgs poles in annihilation of pairs of neutralinos chi with a complete treatment of chi - stau coannihilation, and discuss carefully uncertainties associated with the mass of the b quark. We find that coannihilation and pole annihilations allow the CMSSM to yield an acceptable relic density at large tan beta, but it is consistent with all the constraints only if m_chi > 140 (180) GeV for mu > 0 (mu < 0) for our default choices m_b(m_b) = 4.25 GeV, m_t = 175 GeV, and A_0 = 0., Comment: 16 pages, 4 eps figures, as accepted for Phys. Lett. B, with improved references: for a version with high-resolution figures, go to http://www.hep.umn.edu/~olive/EFGOSi.ps

We reexamine large CP-violating phases in the general Minimal Supersymmetric Standard Model, as well as more restricted models. We perform a detailed scan over parameter space to find solutions which satisfy the current experimental limits on the electric dipole moments of the electron, neutron and $^{199}$Hg atom, exploring the allowed configurations of phases and masses, and we attempt to quantify the level of tuning of the parameters necessary to populate the regions of cancellations. We then consider the measurement of CP-violating phases at a future linear collider. We find that measurements of chargino and neutralino masses and production cross-sections allow for a determination of $\phi_1$(the phase of $M_1$) to a precision of $\pi/30$, while the EDM constraints require that $\theta_\mu$ be too small to be measured. Using the EDM constraints we find that the CP-even model parameters and the phase $\phi_1$ can be determined at a Linear Collider with $400 \gev$ c.m. energy. As long as some information on the size of $|\mu|$ is included in the observables, a measurement of $\phi_1$ is guaranteed for $\phi_1 > \pi/10$. To unambiguously identify CP violation, we construct CP-odd kinematical variables at a linear collider. However, the CP asymmetries are rather small, typically about $0.1-1.5%$, and it will be challenging to experimentally observe the predicted asymmetries., Comment: 29 pages, version accepted by PRD, Figure 8 changed, minor changes in wording

Assuming that the lightest supersymmetric particle (LSP) is the lightest neutralino, we present a detailed exploration of neutralino-stau coannihilation channels, including analytical expressions and numerical results. We also include neutralino coannihilations with the selectron and smuon. We evaluate the implications of coannihilations for the cosmological relic density of the LSP, which is assumed to be stable, in the constrained minimal supersymmetric extension of the Standard Model (CMSSM), in which the soft supersymmetry-breaking parameters are universal at the supergravity GUT scale. We evaluate the changes due to coannihilations in the region of the MSSM parameter space that is consistent with the cosmological upper limit on the relic LSP density. In particular, we find that the upper limit on the mass of the neutralino is increased from about 200 GeV to about 600 GeV in the CMSSM, and estimate a qualitatively similar increase for gauginos in the general MSSM., Comment: 38 pages in LaTeX, including 25 encapsulated postscript figures. Added small clarification in the text, corrected some typos in the appendix. Figures and conclusions unchanged

The constraints on neutralino dark matter \chi obtained from accelerator searches at LEP, the Fermilab Tevatron and elsewhere are reviewed, with particular emphasis on results from LEP 1.5. These imply within the context of the minimal supersymmetric extension of the Standard Model that m_\chi \ge 21.4 GeV if universality is assumed, and yield for large tan\beta a significantly stronger bound than is obtained indirectly from Tevatron limits on the gluino mass. We update this analysis with preliminary results from the first LEP 2W run, and also preview the prospects for future sparticle searches at the LHC., Comment: Presented by J. Ellis at the Workshop on the Identification of Dark Matter, Sheffield, September, 1996. 14 pages; Latex; 12 Figs