Search

Recently, a new complementary joint sparse representations (CJSR) method was proposed for monaural speech enhancement, which utilizes the relationships among speech, noise and mixture. One of the joint sparse representations (JSRs) uses the mapping relationship between mixture and speech while the other uses the mapping relationship between mixture and noise. However, since they only use the joint information and overcomplete dictionaries, there may be some confusion components between the estimated speech and noise. In this paper, a novel model with fusion process is proposed, which further using an additional composite dictionary composed of clean speech dictionary and noise dictionary as prior knowledge. When the estimated speech and noise are obtained by JSRs, they are further sparsely represented on the composite dictionary, respectively. Because that the composite dictionary has the discriminative property, the source confusion problem of previous methods is coped well. In order to take advantage of the complementary knowledge of the two estimated signals, we propose a weighting parameter based on the residuals of sparse representation on composite dictionary. If one of the residuals is smaller, the corresponding weighting is greater, and this weighting parameter varies with noise type and speaker flexibly and effectively. Experimental results show that the proposed algorithm has superior performance compared with other tested approaches.

Maximum sound pressure levels are commonly used for environmental noise and building acoustics measurements. This paper investigates the signal processing errors due to Fast or Slow time-weighting detectors when combined with octave band filters, one-third octave band filters or an A-weighting filter. For 6th order Butterworth CPB filters the inherent time delay caused by the phase response of filters is quantified using three different approaches to establish the following rules-of-thumb: (1) time-to-gradient/amplitude matching occurs when Bt ≈ 1, (2) time-to-peak matching occurs when Bt ≈ 2 and (3) time-to-settle matching occurs when Bt ≈ 4 for octave band filters, and when Bt ≈ 3 for one-third octave band filters. Four different commercially-available sound level meters are used to quantify the variation in measured maximum levels using tone bursts, half-sine pulses, ramped noise and recorded transients. Tone bursts indicate that Slow time-weighting is inappropriate for maximum level measurements due to the large bias error. The results also show that there is more variation between sound level meters when considering Fast time-weighted maximum levels in octave bands or one-third octave bands than with A-weighted levels. To reduce the variation between measurements with different sound level meters, it is proposed that limits could be prescribed on the phase response for CPB filters and A-weighting filters.

A system is introduced with the purpose of showing how an auditory perception system may be built up to include the basic quantities on loudness domain. The quantities are the critical bands, the power law, and the weighting. The power law seems to be the most crucial basis for hypothesizing a loudness function. It has been shown that the power law could be applied as such by assuming the auditory perception system to have two essentially different stimuli: the intensity (sound pressure level) and pure pressure. These physically different quantities seem to be combined in the root of the power law, and in this study the roots are determined from equal-loudness contours. A loudness function is derived on the basis of this finding. By adding the weighting, a method has been constructed for assessing loudness. After defining the weighting, the equal-loudness contours are constructed and are seen to be virtually identical to the contours in ISO 226. It has also been found that the equations for deriving the contours in this standard and in the new ISO 226 may be incorrect, because there is no definition of a sensible loudness function. Finally, it is deduced that the derived weighting must be unequivocal for an auditory perception system (depending solely on the otologically representative group). Finally, the A-weighting (as part of an A-weighted sound pressure level) as such is reasonably similar to the weighting derived in this study. Therefore, this weighting is not the main problem when assessing sounds in respect to loudness. The A-weighting is thus chosen as the weighting for the indicator derived in the study for assessing environmental sounds.

Previously it has been found through a series of psychoacoustical experiments that the arithmetic average of sound pressure level calculated in octave bands is a good estimator of loudness for various kinds of environmental noise. Remarkably, the arithmetic average of sound pressure level in octave bands from 63 Hz to 4 kHz, Lm,1/1(63–4k), strongly correlates with the loudness level specified in ISO 532B, LL(Z), as well as with loudness assessment. To investigate this relationship further, a numerical study has been carried out based on Zwicker’s loudness model. As a result, practical expressions to estimate the loudness levels of general environmental noises from the sound pressure levels in octave bands from 63 Hz or 125 Hz to 4 kHz are proposed.

It is suggested from a previous theoretical study that an angular distribution function can be successfully used for describing the directional distribution of incident energy on the wall, instead of the field incidence assumption being used in the prediction of sound transmission loss. This study is devoted to an experimental investigation on the directionality of incidence sound field for validating the underlying assumption in the theoretical work. To this end, the directional magnitudes of incident energy on a wall in a reverberation chamber are measured by employing the sound intensimetry. The experimental results suggest that the transmission coefficient formula should be revised by introducing a proper angular distribution function as a weighting function for describing the directional energy density distribution, which replaces the conventional field incidence method. The revised formula is applied to the prediction of transmission loss for a double-leaf panel with an air cavity, in which its effectiveness is strongly supported by good agreements between measured and predicted values.

An acoustical environment survey was carried out in six similar air conditioned landscaped offices in which human speech was considered to be the major source of complaint. Both questionnaire and physical noise measurement were administrated. Physical noise measurement data reveal higher sound energy within the human speech frequency range and high speech intelligibility level in the offices. Results also show that the preferred noise criterion gives good correlation with the auditory sensation comments of the office workers. The noise climate, the equivalent sound pressure level and the noise pollution level are also found to be useful in assessing the acoustical environment inside the surveyed offices.

In the present study, noise measurements were carried out in 26 air-conditioned landscaped offices during lunch and office hours. The noise spectra show that, in general, the noise from office activities is of a frequency higher than 500 Hz. A-, B- and C-weighted sound levels are also calculated and it is found that linear relationships exist between these weighted levels. In addition, simple formulae are proposed for the estimations of noise criterion (NC) and noise rating (NR) from the A-weighted sound level which is also known as the equivalent sound pressure level. The air-conditioning systems affect the noise levels, NC and NR, but not the basic relationships between weighted sound levels, NC and NR.

The present technique for measuring the airborne sound insulation of walls and floors, involving measurements in 16 one-third octave bands, is tedious and expensive. The method provides more information than is needed for most purposes, and is more suited to research. Several investigators have proposed the measurement of the overall A-weighted sound level difference using a sound level meter, with a broad band source of white or pink noise. Consistent results have been obtained but their relation to accepted rating methods such as STC is rather empirical. The reference curves used for airborne sound insulation, i.e. STC and HPGW are very similar to the A weighting curve, and if the latter were adopted as the reference curve, there would be a firm theoretical basis for measurement with a sound level meter. Measurement of the difference between the linear sound level of a source of white noise, and the A weighted received level, would in practice be a test of the conformity to the A weighting curve of the transmission loss curve of the partition. Adverse deviations would show as a higher received level. Favourable deviations would have little effect. A study of practical walls and floors, taken from National Building Studies Research Paper 33, showed that there was good correlation between the sound level difference calculated as if it had been measured directly with a sound level meter, and a proposed rating method similar to ISO R717, but using the A weighting curve as the reference curve. Ninety-six per cent of results were within ±1 dB. The practical difficulties of achieving a reasonably flat transmitted spectrum, and of correcting for room absorption will reduce this precision, but bearing in mind the practical success of other short tests, the proposed test should provide a rapid test which is adequate for approval purposes.

Sound pressure level measurements in cars travelling at motorway speeds have shown that, in many cases, the overall level is very high in relation to the dB(A) and octave band levels, suggesting that much of the sound energy is in the low frequency and infrasonic regions. A technique has been developed to extend accurate octave band measurements down to the octave centred on 2 Hz. The system uses a calibrated sound level meter feeding a frequency modulation tape-recorder to record noise below 64 Hz, and an octave band analysis system to analyse the resultant tape recordings. Typical results are presented for a number of vehicles and it is found that sound pressure levels as high as 120 dB can be found in the octave bands between 2 and 16 Hz.

The paper discusses the merits of using a single dB(A) reading in preference to the more complicated and time-consuming process of deriving an NR number. It is shown that, for commonly-encountered spectra, there is a small numerical difference between the dB(A) and NR values. This difference depends upon spectral shape and the overall level of the noise. Octave bands above 1 kHz tend to make the dB(A) level less than the NR number, while the octave bands below 1 kHz generally have the opposite effect.

A simple method of measuring sound power is described. Noise is added from a second calibrated source until the room sound pressure level is increased by 3 dB . The noise output of the device is then equal to the added power. A simple loudspeaker, modified to reduce the effect of the environment on the sound power output, is used as the calibrated source of sound power. Checks show that the sound power readings are substantially independent of the room in which they are made.

A panel of subjects assessed the noisiness of six cars at four constant speeds. The sound pressure levels existing in the vehicles were measured in one-third octave bands from 25 to 10,000 Hz, and also using the A and B weighting networks. From the band levels estimates were made of single figure parameters to describe the noise environment, i.e. Zwicker Phons, Stevens Phons, Noise Criterion, Noise Rating Number and Speech Interference Level. Good correlation was found between subjective assessment and Zwicker Phons, Stevens Phons, dB(A) and dB(B). Less good correlation was obtained between the subjective assessment and Noise Criterion, Noise Rating Number and Speech Interference Level. The experiment indicates that subjective response to internal car noise can adequately be measured using the A weighting network.

The ‘A’ weighting curve, drawn from the 56 dB datum, forms a very good compromise between the British House Party Wall Grade curve, and the International Standard Organisation R 717 reference curve. Since the ‘A’ weighting curve is already internationally standardised for noise measurements, its adoption as a reference curve for sound insulation measurements would make for a unification and rationalisation of acoustic criteria. The adoption of the ‘A’ weighting curve would also provide a theoretical basis for a quick test for sound transmission using a sound level meter to measure the overall levels on each side of a wall, of a white noise source. When the received level is measured on the ‘A’ weighting network, any serious adverse deviation of the partition below the reference curve would be shown by the higher overall received level. Suggestions for the revision of ISO R 717 to incorporate this proposal are included. Good correlation is shown between the sound level difference, as measured by a meter, and the revised airborne sound insulation index measured by the ISO R 140 procedure, on one sample test.