Filter Slope Research Articles

Optical narrow band filter without resonance's

This paper introduces an optical wave filter, which uses gratings at 45? or 135? inclined grating lines that avoid any resonance's. Therefore, many more options to form the filter shape exist. In general, the filter design can be traced to that of transversal filters (finite impulse response filter, FIR filter). Such an integrated optical wave filter is characterized by steep filter slopes and a narrow pass band (less then 01nm) combined with a high stop band attenuation (more than 40dB) and a linear phase response in the pass band. Compared to conventional Bragg grating filters, the inclined grating line filters can have a flatter pass band and steeper filter skirts related to the width of the pass band. In general, the filter's realization is possible using any optical material. In view of the excellent optical properties the semiconductor material system InP/InGaAsP is used for manufacturing the filter.

An Optical Wave Filter with 45° Inclined Grating Lines

This letter presents a novel integrated optical wave filter which can be realized in InP/InGaAsP, for example, in this device a definite shape of the filter function can be adjusted. The shape is characterized by steep filter slopes and a narrow pass band (down to 0.1 nm) combined with a high stop band attenuation (more than 40 dB), and a linear phase response in the pass band. Compared to conventional Bragg grating filters the inclined grating line filters can have a flatter pass band and steeper filter skirts related to the width of the pass band.

Factors affecting sediment trapping in vegetated filter strips: simulation study using VFSMOD

AbstractSoil and water conservation practices have been promoted for a long time, in order to sustain agricultural activities and prevent environmental pollution. Vegetated filter strips (VFS) have been used to reduce sediment pollution into water bodies at or near the pollutant source. However, factors effecting VFS performance under natural conditions have not been well understood owing to the physical, time and financial limitations of field experiments. The use of well‐validated simulation models to understand the performance of VFS and factors affecting sediment deposition is highly justified. The objective of this research is to investigate sediment trapping in VFS and to study various factors affecting VFS performance using the simulation model VFSMOD, which was developed by researchers at University of North Carolina. Recently, VFSMOD has been validated successfully by using 21 filters with varying length, slope and vegetated cover. A wide range of five parameters was selected for the simulations, namely filter length, filter slope, manning roughness coefficient, soil type and characteristics of incoming sediment from adjacent fields. Computer simulations revealed that the length of filter is the most significant factor affecting sediment trapping in VFS. The relative increase in trapping efficiencies was not linearly related to an increase in filter length. Inflow sediment class also has a major influence on sediment trapping in VFS. The trapping efficiency of clay sediments in a 15 m length VFS was 47% compared with 92% for silt from incoming sediment. Manning roughness coefficient had a moderate effect on sediment trapping and was more significant in short filters. Land slope and soil type of VFS had a minor influence on the performance of VFS. Copyright © 2001 John Wiley & Sons, Ltd.

Intelligibility of bandpass speech: effects of truncation or removal of transition bands.

An intelligibility of over 90% was reported for keywords in "everyday" 1/3-octave sentences centered on 1500 Hz and having steep transition band slopes of 100 dB/octave [Warren et al., Percept. Psychophys. 57, 175-182 (1995)]. A subsequent study by Warren and Bashford [J. Acoust. Soc. Am. 106, L47-L52 (1999)] found that it was not the 1/3-octave passband, but the transition bands that were chiefly responsible for this high intelligibility: When the passband and transition bands were segregated using filter slopes of 1000 dB/octave, the isolated passband had an intelligibility score of only 24%, while the pair of transition bands had a score of over 80%. In the present study, experiment 1 examined the distribution of information along the transition bands' slopes by truncation at graded downpoints: Truncation at downpoints of 40 dB or more produced no significant change in intelligibility. Experiment 2 closed the gap separating the transition bands so that their slopes intersected at 1500 Hz. This triangular band had a negligible passband (as defined conventionally by 3-dB downpoints) and an intelligibility score of 60%; truncation at downpoints of 50 dB or more produced no significant change in intelligibility. Experiment 3 determined the intelligibilities of rectangular bands (1000-dB/octave slopes) centered on 1500 Hz. Their bandwidths ranged from 3 to 12 semitones in 1-semitone steps, resulting in intelligibility scores increasing monotonically from 14% to 94%. Calculations based upon experiments 2 and 3 showed that the triangular band truncated at 30-dB downpoints had half the intelligibility of a rectangular band having the same frequency range.

Relative contributions of passband and filter skirts to intelligibility of bandpass speech: Some effects of amplitude and context

Recently, Warren and Bashford [J. Acoust. Soc. Am. 106, L47–L52 (1999)] reported over 90% intelligibility for everyday sentences reduced to a 1/3-octave band (center frequency 1500 Hz, slopes 100 dB/octave, peak levels 75 dB). Surprisingly, when the bandpass sentences were partitioned using 2000-order FIR filtering, the rectangular 1/3-octave passbands had only 24% intelligibility, while the filter skirts separated by a 1/3-octave notch had an intelligibility of 83%, despite their severe spectral tilts. The present study found that when monosyllabic word lists were substituted for the sentences, the whole-band intelligibility was 26%, the passband 4%, and the filter skirts 16%. Intelligibility also was measured for 1/3-octave sentences with peak levels ranging from 85 down to 35 dB: The whole band intelligibilities ranged from 90% to 68%, and the filter skirt pairs each had three to four times the passbands’ intelligibility. These findings demonstrate the ability to compensate for extreme distortions of the spectral profile of speech. Hence, it is necessary to consider the contribution of frequencies outside the nominal passband as defined by the 3 dB downpoints along the filter slopes when dealing with filtered speech. [Work supported by NIH.]

On loudness and pitch of ‘‘complex noise’’ in relation to the factors extracted from the autocorrelation function

In a previous study, it was shown that the loudness of bandpass noise within the critical band changes increasing the effective duration of the envelope of the normalized autocorrelation function (ACF) τe changing the filter slope of bandpass noise and the subsequent reverberation time Tsub of sound field [Y. Ando, Architectural Acoustics—Blending Sound Sources, Sound Fields, and Listeners (AIP Press/Springer-Verlag, New York, 1998), Chap. 6]. In this study, the loudness of the complex signals including bandpass noises with different frequency (complex noise) is examined. The complex noises consist of the bandpass noises whose center frequencies are the harmonics of the fundamental frequency. The bandwidth of each component is changed within the critical band by use of a sharp filter of more than 1000 dB/octave. Subjects judge the loudness of complex noise by comparing the tone of its perceived pitch frequency, the ‘‘missing fundamental.’’ The results are discussed with the four factors extracted from the autocorrelation function, which are τe, the energy represented at the origin of the delay Φ(0), and the amplitude and the delay time of the first peak of ACF φ1 and τ1. [Work supported by ACT-JST and Grant-in-aid for Scientific Research, JSPS, Japan.]

Preattentive Control of Serial Auditory Processing in Dichotic Listening

Dichotic listening performance was examined in an auditory selective attention task where subjects responded to occasional consonant-vowel (CV) or shaped broadband noise-burst (NB) targets in rapid serial auditory presentation (RASP). Trial types were randomized and included monaural CVs and NBs as well as dichotic CV-CV and CV-NB pairings. CVs were spoken by two different voices (male and female), and the two NB stimuli differed in their filter slopes at higher frequencies. The target was designated by stimulus category (/ba/, /da/, /ga/, or NB) and voice (e.g., “female /ba/”). Performance was compared for targets in the left and right ears on monaural and dichotic trials using accuracy and reaction time (RT) measures. Right ear advantages (REAs) were present for CV targets with either CV or NB distractors, but not for monaural CVs. The REA found for monaural NB targets was eliminated by CV distractors, yielding a left ear advantage (LEA) for the distractor effect of CVs on NB targets. The pattern of results suggests initial preprocessing of speech stimuli through phonetic feature analysis, followed by serial attentional processing of the objects in the auditory field. REAs are attributed to a rightward asymmetry in the preattentive control of auditory attention similar to that found in visual search.

Effect of stimulus rise time and high-pass masking on early auditory evoked potentials

Problems of frequency-specific objective assessment of hearing threshold by means of auditory brainstem response (ABR) have been discussed recently. While a number of workers have recommended methods of selective masking to improve the frequency specificity, others believe that frequency-specific potentials can also be obtained without masking. In this context, the effects of rise-decay time and high-pass masking on ABRs were investigated. ABRs were recorded in normal-hearing subjects and patients with high and low frequency hearing loss by means of surface electrodes between the vertex and the ipsilateral mastoid. The frequency of the stimulus was 1 kHz, and the rise-decay time 1 ms (1-0-1) or 2 ms (2-0-2). High-pass filtered noise (cutoff frequency 1.5 kHz; filter slope 250 dB/octave) was employed for masking. Particular attention was paid to the problem of efficient masking. In normal-hearing subjects under the influence of high-pass masking compared to non-masked ABRs, longer mean latencies and diminished means of the amplitudes of wave V were found, with differences in the near-threshold domain being less pronounced. Similar results were observed in patients with high frequency hearing loss. In patients with low frequency hearing loss, the influence of high-pass masking was especially marked distinctly near to threshold. Furthermore, latency and amplitude differences of wave V of the 1-0-1 and the 2-0-2 stimuli were determined from the ABRs obtained with and without high-pass masking. The differences between the latency differences of both stimuli in the suprathreshold range (70 dB nHL) only were statistically significant. The results are suggestive of an inadequate frequency specificity of unmasked stimuli in the suprathreshold range. Evaluation of the latencies revealed for both rise-decay times a similar frequency specificity near the threshold and a higher frequency specificity of the longer stimulus in the suprathreshold range.

Auditory suppression and frequency selectivity in older and younger adults

Age differences in auditory suppression were examined by comparing auditory-filter shapes obtained with simultaneous and forward masking at 2 kHz in young and elderly normal-hearing listeners. To compensate for the decay of forward masking, growth of masking functions were used to transform thresholds obtained with a notched-noise masker to the level of a continuous noise band that would give the same threshold values. Although both age groups exhibited smaller equivalent rectangular bandwidths (ERBs) when the filters derived from transformed thresholds were obtained with forward masking, the change from simultaneous to nonsimultaneous masking was significantly greater for young adults. Measures of the low- (Pl) and high- (Pu) frequency sides of the filters for young listeners indicated that the slopes of both sides increased following a change from simultaneous to forward masking but that the high-frequency side exhibited significantly greater sharpening. Filter slopes (both upper and lower) for older adults, in contrast, did not differ between the two masking procedures. The findings from the study are discussed as reflecting possible age differences in auditory suppression. However, it is also noted that conclusions regarding differences between filter shapes derived with simultaneous and forward masking are limited to filter parameters determined with transformed (as described above) thresholds.

Equalization of nonuniform EDFA gain using a fiber-loop mirror

We have experimentally demonstrated cascaded EDFA gain equalization using a fiber-loop mirror (FLM) acting as a linear wavelength filter. By changing the intra-loop polarization, the FLM passband center wavelength can be tuned. Additionally, there is some ability to change the bandwidth and filter slope of the FLM. The FLM is used to equalize the nonuniform gain of a cascade of EDFAs. After 1500 km, the power differential is reduced from 30 to 2 dB for three WDM channels covering a 9-nm wavelength range.

Perceived location of bars and edges in one-dimensional images: computational models and human vision.

Observers used a cursor to mark the location and polarity of all the bar and edge features seen in compound (f + 3f) gratings of moderate frequency and contrast. They almost always reported six bars and six edges per cycle of the fundamental frequency (f = 0.4 c/deg, contrast 32%), for all phases of the third harmonic (3f = 1.2 c/deg, contrast 10.7%). This general pattern of features was predicted by the positions of peaks and troughs in the outputs of even and odd filters applied to the stimulus waveform, but not by peaks of "local energy" since there were only two energy peaks per cycle. We considered a family of filters whose amplitude spectrum has slope p on a log-log plot. The best-fitting filter slope was determined for bars (even filter) and edges (odd filter) in conjunction with a classification rule in which all peaks and troughs in the response profile are counted as features. If bars were seen at luminance peaks, and edges seen at gradient peaks (zero-crossings in the second derivative) we should have found p = 0 for bars and p = 1 for edges. In fact, for both bars and edges the best-fitting slope was about p = 0.5. For edges, this is consistent with the use of a smoothed (Gaussian) derivative operator. The filters form a quadrature pair, as in the energy model, but features are not constrained to lie at energy peaks. A compressive transducer preceding the filters improved the goodness-of-fit for predicted edge locations, but did not affect the estimate of filter slopes, nor the goodness-of-fit for bar locations. In an experiment with single blurred edges we confirmed that the perceived location of edges is shifted towards the darker side of the edge in direct proportion to the contrast of the edge. This was well predicted by adding a compressive transducer to the filter model.

雑音下語音認知に及ぼす低音域減弱の影響

Many types of hearing aids apply low-cut filtering for noise reduction. This method is based on the fact that background noises are mainly low frequency. The advantages of low-cut filtering for speech recognition with noise have been contentious. We evaluated the effectiveness of the low-cut filtering method for improving the recognition of Japanese sentences with noise.Speech recognition tests were undertaken by 56 patients with sensorineural hearing loss and 14 volunteers with normal hearing. Multitalker noise was added to sounds of four Japanese sentences by a male announcer at an S/N ratio of 6 dB. Two of the four sentences were processed through a low-cut filter (1000 Hz cutoff frequency, -12 dB per octave filter slope). In 48 of the patients with sensorineural hearing loss, the two filtered sentences were not understood better than the unfiltered sentences. In 34 of these patients, speech recognition deteriorated as a result of low-cut filtering. These results indicated that low-cut filtering is not effective for speech recognition with noise.

Electrosensory stimulus-intensity thresholds in the weakly electric knifefish Eigenmannia: reduced sensitivity at harmonics of its own organ discharge

The South American knifefish Eigenmannia sp. can detect the electric organ discharges (EODs; 250­600 Hz) of conspecifics when they are superimposed over its own EOD. This study investigates the minimum frequency difference necessary for such signal perception, using the application of sine-wave stimuli. Electrosensory stimulus-intensity thresholds were determined for trained fish using stimuli associated with food rewards. These sine-wave stimuli were 'clamped' to the EOD frequency of the fish. Electrosensory thresholds were also determined for the spontaneous jamming avoidance response (JAR; a change in EOD frequency evoked by a stimulus of sufficiently similar frequency), in this case using unclamped stimuli. Over the wide frequency range investigated (0.3­3.01 times EOD frequency), the lowest stimulus-intensity thresholds of 0.6 µV cm-1 (peak-to-peak) (0 dB) at a water conductivity of 100 µS cm-1 were found close to (but not exactly at) the EOD fundamental frequency. At exact frequency identity between the EOD and the stimulus, the stimulus-intensity response threshold rose abruptly by more than 10 dB compared with slightly higher or lower stimulus frequencies. A similar 'needle-like' threshold increase was found at exactly two and three times the EOD frequency, but neither at harmonic ratios between stimulus and EOD frequency that represent fractions (e.g. at 5:4=1.25, 4:3=1.33, 3:2=1.5 or 5:3=1.67 times EOD frequency) nor at subharmonics such as half or two-thirds of the EOD frequency. The steepest increase of stimulus-intensity response threshold was in the range 0.998­1.002 times EOD frequency, corresponding to a threshold change, or electrosensory filter slope, of 5000 dB per octave. For the spontaneous JAR, a similar stimulus-intensity threshold increase was observed when EOD frequency equalled stimulus frequency. Because of the longer rise time for the stimulus amplitude (400 ms rather than 35 ms) the stimulus intensity threshold was higher (up to 32 dB; mean, 20 dB) than in the other experiments (up to 15 dB; mean, 13 dB). A difference in frequency between the EOD and the applied stimulus as small as 1 Hz (that is, 0.2 % of the EOD frequency) was sufficient for good signal perception in Eigenmannia sp. The JAR appears to be useful in avoiding insensitivity at exact integer harmonics of the EOD frequency.

The development of frequency resolution in humans as revealed by the auditory brain-stem response recorded with notched-noise masking

Studies of tuning in infants have reported that auditory brain-stem response (ABR) tuning curves generated using low-frequency probes are adultlike by 3 months of age while high-frequency tuning curves remain immature [Folsom and Wynne, J. Acoust. Soc. Am. 81, 412-417 (1987)]. Behavioral studies have similarly reported adultlike low-frequency psychoacoustic tuning curves by 3 months with high-frequency tuning curves immature until approximately 6 months of age [L. Olsho, Infant Behav. Dev. 8, 371-384 (1985); Spetner and Olsho, Child Dev. 61, 632-652 (1990); Schneider et al., J. Exp. Psych.: Human Percept. Perform. 16, 642-652 (1990)]. Prior to this experiment, there have been no ABR studies of the development of frequency resolution for infants older than 3 months. In this study, notched-noise tuning functions were constructed from wave-V amplitude data for 3-month-old, 6-month-old, and adult subjects. Tone-pip stimuli at 1000, 4000, and 8000 Hz (50 dB nHL) were presented simultaneously with notched-noise masking centered at frequencies related to the tone-pip frequency (1/3-oct intervals above and below the probe frequency). By plotting wave-V amplitude across notched-noise center frequency, isointensity tuning functions were generated for the three subject groups at the three probe frequencies. Auditory filter width (Q) and slope (dB/oct) were measured from each notched-noise tuning function in order to qualify degree of tuning. Consistent with previous studies, results showed that 3-month-old infants do not have adultlike tuning for high-frequency stimulation (8000 Hz). In contrast, by 6 months of age, tuning-function width (Q) is adultlike for both high- and low-frequency probes. These results, combined with previously reported evidence that the human cochlea is fully tuned at birth [Abdala et al., submitted to Hear. Res. (1995); Bargones and Burns, J. Acoust. Soc. Am. 83, 1809-1816 (1988)], suggest that immaturities in the auditory-neural system contribute to the broad high-frequency tuning consistently observed in 3-month-old human infants.

Spectral redundancy: intelligibility of sentences heard through narrow spectral slits.

The intelligibility of word lists subjected to various types of spectral filtering has been studied extensively. Although words used for communication are usually present in sentences rather than lists, there has been no systematic report of the intelligibility of lexical components of narrowband sentences. In the present study, we found that surprisingly little spectral information is required to identify component words when sentences are heard through narrow spectral slits. Four hundred twenty listeners (21 groups of 20 subjects) were each presented with 100 bandpass filtered CID ("everyday speech") sentences; separate groups received center frequencies of 370, 530, 750, 1100, 1500, 2100, 3000, 4200, and 6000 Hz at 70 dBA SPL. In Experiment 1, intelligibility of single 1/3-octave bands with steep filter slopes (96 dB/octave) averaged more than 95% for sentences centered at 1100, 1500, and 2100 Hz. In Experiment 2, we used the same center frequencies with extremely narrow bands (slopes of 115 dB/octave intersecting at the center frequency, resulting in a nominal bandwidth of 1/20 octave). Despite the severe spectral tilt for all frequencies of this impoverished spectrum, intelligibility remained relatively high for most bands, with the greatest intelligibility (77%) at 1500 Hz. In Experiments 1 and 2, the bands centered at 370 and 6000 Hz provided little useful information when presented individually, but in each experiment they interacted synergistically when combined. The present findings demonstrate the adaptive flexibility of mechanisms used for speech perception and are discussed in the context of the LAME model of opportunistic multilevel processing.

Level-dependent critical bandwidth for phase discrimination

Monaural phase discrimination was evaluated at 1000 Hz in six normal-hearing listeners as a function of the frequency difference between components in three-tone complexes at 40, 60, 80, and 100 dB SPL. The phase of the center component of 100% sinusoidally amplitude-modulated (SAM) waveforms was shifted by 90 degrees to produce quasi-frequency-modulated (QFM) waveforms that had identical long-term power spectra to the SAM waveforms but with different amplitude envelopes and temporal fine structure. At low modulation frequencies, where spectral components were close together and presumably all well within a single auditory filter, normal-hearing listeners could easily discriminate QFM from SAM waveforms. As modulation frequency increased, a point was reached where listeners could no longer distinguish QFM from SAM waveforms, referred to here as the critical bandwidth for phase discrimination (CBphs). Discrimination performance (d') was measured as a function of modulation frequency to yield a psychometric function for phase discrimination. From the psychometric function, CBphs was defined as the modulation frequency corresponding to d' = 1.0. At a carrier frequency of 1000 Hz, CBphs increased with level between 40 and 80 dB SPL according to the relation: CBphs = 34.I(0.136). Above 80 dB SPL, very little change in CBphs with level was seen. The increase with level of the geometric mean CBphs was predicted from level-dependent auditory filter slopes inferred from forward-masked tuning curves, as was the tendency to reach an asymptote above 80 dB SPL. Comparisons with previous work indicate that CBphs values from the best performing subjects were well within the audibility region for cubic difference tones. It is proposed that internally generated cubic difference tones interact with externally generated acoustic components, both limited by a level-dependent auditory filter, to produce an internal excitation envelope that is the basis for discriminating between SAM and QFM waveforms. It is also suggested that individual differences at low levels may be due to internal phase ambiguities.

Novel findings concerning intelligibility of bandpass speech.

Very high intelligibility was found for a number of narrow-band filtering conditions. Four hundred listeners (20 groups of 20 subjects) were presented with bandpass filtered CTD sentences (‘‘everyday speech’’) and monosyllabic words. Separate groups received center frequencies of 370, 530, 750, 1100, 1500, 2100, 3000, 4200, or 6000 Hz at 70 dBA SPL. In experiment 1, intelligibility of single 1/3-octave bands with steep filter slopes (96 dB/oct) averaged more than 95% for sentences centered at 1100, 1500, and 2100 Hz, and more than 50% for monosyllabic words centered at 1500 and 2100 Hz. Experiment 2 used the same center frequencies with extremely narrow bands (slopes of 115 dB/oct intersecting at the center frequency)−intelligibility remained relatively high for most bands, with greatest intelligibility at 1500 Hz (77% for sentences, 18% for words). In experiment 3, 1/3-octave bands (96 dB/oct) centered at 370 and 6000 Hz were presented simultaneously either diotically or dichotically (when presented separately in experiment 1, intelligibility of these bands did not differ significantly, and averaged 23% for sentences and 3% for monosyllabic words). When the bands were combined, diotic and dichotic presentations were equivalent. Intelligibility rose to an average of 77% for sentences and 34% for words. [Work supported by NIH Grant No. DC00208.]

Rate and filter effects on the developing middle-latency response.

Auditory middle-latency responses (MLRs) were obtained from 71 unanesthetized gerbils ranging in age from 10 to greater than 90 days. Effects of age, stimulation rate, high- and lowpass filter settings and filter slope were examined. MLR amplitude decreased significantly with increased stimulation rate at all ages, at rates up to 40/s. The detection of MLR waves (presence or absence) varied inversely with the rate of stimulation only in immature subjects. The amplitude of waves B (15 ms) and C (25 ms) was significantly larger with a highpass filter setting of 3 Hz as compared to 10 and 30 Hz. This effect was significantly more pronounced in developing animals as compared to adults. MLR amplitude was greater with a filter slope of 6 dB/octave as compared to 48 dB/octave, (10-2,000 Hz) and this effect was also significantly greater in developing animals than in adults. There was no interaction between lowpass filter setting and age (100 vs. 2,000 Hz). A prominent positive wave occurring at approximately 50 ms was present in the 48 dB/octave condition although it was not observed with 6 dB/octave filtering. The clinical use of the MLR requires a better understanding of the effects of stimulus and recording procedures on the response, and how they vary as a function of subject age.

Analog and zero phase-shift digital filtering of the auditory brain stem response waveform.

Auditory brain stem response waveforms, derived from normal ears, were analyzed with a computer-based digital filtering program as a means of systematically evaluating the effects of phase-shift distortion. Off-line analysis involved the stimulated responses of a Butterworth-type filter at five high-pass and four low-pass frequencies, three filter slopes, and standard versus zero phase-shifting. Phase-shift, as seen with analog filters, proved to be a major cause of waveform distortion including significant latency and amplitude changes. A discussion on the need and use of digital filtering as well as suggested analog filter parameter settings is included.

Perceptual compensation for effects of transmission channel characteristics on vowel quality

The phonetic quality of front vowels along a synthetic /I/‐/e/ FI continuum can be influenced by passing the vowels through a filter that has a sloping characteristic in the FI region. The shift in phoneme boundary that occurs for different filter slopes was measured and whether the influence of various conditions listeners can compensate for the characteristics of the transmission channel was examined. These conditions include placing the vowel in a consonantal context and preceding the vowel with a carrier phrase that has been subjected to the same filtering as the vowel. Experiments have also been performed on the preception of natural vowels in isolation and in a consonantal context. [Work supported by SERC.]