Envelope Difference Index Research Articles

Role of Temporal Fine Structure and Envelope in Identifying Ragas: A Prospective Study for Temporal Based Raga Classifier

Background and Aim: A raga is characterized by its distinctive melodic shape. The ability to perceive intricate melodic and pitch patterns depends on the Temporal Envelope (ENV) and Fine Structure (TFS). The present study aimed to understand the importance of temporal envelope and temporal fine structure cues in identifying ragas in Indian music. Methods: Twenty‏-‏one adult’s musicians were included in the study. In experiment 1, professional musicians were involved in a raga identification task using music chimaeras. In experiment 2, the chimaeras were subjected to acoustic analysis using the envelope difference index, to understand better how the ENV and TFS changed depending on how many frequency bands were used to create the chimaeras. The subjective impression of a new group of trained musicians was then compared to these results. Friedman’s test and Wilcoxon tests were carried out. Results: Results showed that both cues are crucial in a trade-off manner; when TFS are not significantly accessible, ENV aids in raga identification. It was reflected in experiment 1 as an increase in ENV scores and a decrease in TFS scores as the number of frequency bands increased. In experiment 2, the envelope difference index for ENV increases with a number of frequency bands, and it correlates with perceptual scores for ENV. Conclusion: The current study highlights the perceptual role of temporal cues in raga identification and directs future work for a temporal-based raga classifier. Keywords: Ragas; envelope; fine structure; Indian music; music information retrieval

Temporal and spectral analysis of melody perception with different simulated cochlear implant coding strategies

This study examined the contribution of pitch and rhythm cues in melody recognition, as well as the quality from three well-known CI coding strategies, Advanced Combinational Encoder (ACE), Fundamental Frequency Modulation (F0mod), and Envelope Enhancement (EE). Firstly, we processed the 30 popular melodies of Taiwan children's songs using NCU-CI, a cochlear implant simulation software. In the pilot subjective listening test (n = 5), rhythm cues perceived better (p < 0.05) by having a higher percent correct and faster response time than pitch cues. Foremost, rhythm cues processed by ACE strategy achieved the best score, 86.80%. Then, objective tests were conducted to measure the spectral and temporal quality of processed melody using log spectral distance (LSD), intensity mismatch pattern, and envelope difference index (EDI). LSD between original and processed by the ACE, F0mod, and EE strategy were 2.10, 2.16, and 2.19, respectively. Then, average intensity mismatch patterns between original and processed by the ACE, F0mod, and EE strategy were 5.9, 6.4, and 6.0, respectively. The lower LSD and intensity mismatch pattern, the better the amplitude melody and spectral quality was preserved. Then, EDI between original and processed by the ACE, F0mod, and EE strategy were 0.11, 0.11, and 0.15, respectively. The higher EDI, the better the temporal envelope was preserved. It reveals that rhythm cues combined with the ACE strategy performed the best feature in CI melody perception.

EFFECT OF DIGITAL NOISE REDUCTION AND DIRECTIONALITY ALGORITHMS IN HEARING AIDS ON TEMPORAL ENVELOPE DISTORTION AND SPEECH RECOGNITION

Background Digital signal processing algorithms tend to alter temporal cues in the speech signal and yet individuals with hearing impairment rely strongly on these cues for speech perception. Hence, there is a need to assess the effect of these algorithms on temporal cues and speech perception. The present study aimed to quantify, in a wide dynamic range compression hearing aid, the individual and combined effects of digital noise reduction and directionality algorithms on temporal cues, syllable recognition, and sentence recognition. Temporal cues were quantified by the envelope difference index (EDI). Material and methods The study included 20 individuals (in the age range of 21 to 44 years) with mild to moderate sensorineural hearing loss. Sentence recognition, syllable recognition, and EDI were obtained at different levels in four different aided conditions – digital noise reduction algorithm only, directionality only, both digital noise reduction and directionality on, and both algorithms off. Sentences were presented in noise at +5 dB signal-to-noise ratio from 180° azimuth. Results Compared to independent activation of the algorithms, the combined algorithms significantly improved speech recognition scores at all presentation levels. The temporal changes induced by the algorithms were only mild, even though EDI was the highest when all the algorithms and the directional microphone were activated. Conclusions The noise reduction algorithms and compression induce temporal changes; however speech recognition improved when the algorithms were activated, presumably due to countervailing psychophysical factors.

Limitations of the Envelope Difference Index as a Metric for Nonlinear Distortion in Hearing Aids.

The envelope difference index (EDI) compares the envelopes of two signals. It has been used to measure nonlinear distortion in hearing aids, but it also responds to linear processing. This article compares linear and nonlinear processing effects on the EDI. The EDI for spectral tilt and peak clipping distortion is computed to illustrate the effects of linear and nonlinear signal modifications. The EDI for wide dynamic-range compression is then compared with that obtained for linear amplification for a set of standard audiograms to show the expected range of EDI values for linear and nonlinear hearing aid processing. The EDI for hearing aid amplification and compression is also compared with a measure of time-frequency envelope modulation distortion for the same conditions. The EDI is shown to be as sensitive to linear amplification as it is to nonlinear processing. The EDI values for spectral tilt can exceed those for peak clipping, and the EDI values for linear amplification exceed those for wide dynamic-range compression for four of the nine audiograms considered. The agreement of the EDI with a nonlinear envelope distortion measure is shown to depend on the long-term spectra of the signals being compared when computing the EDI. The accuracy of the EDI as an indicator of nonlinear distortion for sentence materials can be improved by equalizing the long-term spectrum of the processed signal to match that of the unprocessed input. However, the EDI does not have a clear interpretation because of the confound between linear and nonlinear processing effects and the lack of an auditory model in calculating the signal differences.

Representation of amplified speech at cortical level in good and poor hearing aid performers

IntroductionHearing aid users reject their own hearing aid because of annoyance with background noise. The reason for dissatisfaction is located anywhere from the hearing aid microphone to the integrity of neurons along the auditory pathway. In this preview, the output of hearing aid was recorded at the level of ear canal and at auditory cortex in good and poor hearing aid users, who were classified using acceptable noise level. ObjectiveTo study the representation of amplified speech in good and poor hearing aid performers. MethodsA total of 60 participants (age ranged 15–65 years) with moderate bilateral sensorineural hearing impairment grouped into good (n=35) and poor (n=25) hearing aid performers. Gap detection test and aided SNR 50 were administered. In addition, ear canal acoustic measures and cortical auditory evoked potentials were recorded in unaided and aided conditions at 65dB SPL. ResultsHearing aid minimally alters temporal contrast of speech reflected in envelope difference index. Although having similar temporal impairment, acoustic characteristics of amplified speech sounds and SNR 50 scores from both groups, the aided cortical auditory evoked potentials surprisingly showed significant earlier latencies and higher amplitudes in good performers than poor performers. In addition, good and poor performers classified based on annoyance level was predicted by latencies of 2N1 and 2P2 components of acoustic change complex. Further, a follow-up revealed hearing aid use has relation with acceptance towards noise. ConclusionParticipants who are willing to accept noise from those who are not willing to accept noise have subtle physiological changes evident at the auditory cortex, which supports the hearing aid usage.

Effect of compression release time of a hearing aid on sentence recognition and the quality judgment of speech.

Context:The sentence recognition score and quality of speech differ when hearing aid compression is set at different release times at different signal to noise ratios (SNRs) for the normal and compressed rate of sentences.Aims:To investigate the effect of amplitude-compression release time of a hearing aid on sentence recognition and quality judgment: (1) for normal rate and time-compressed sentences (2) in quiet and noisy conditions.Settings and Design:A post-test repeated measures design.Methods and Material:We recruited fifteen adult participants with bilateral sensorineural hearing loss in each group, the younger (35–45 years), and the older ones (60–70 years). A gap detection test assessed temporal processing ability. We used three compression settings, fast-acting, slow-acting, and linear. Sentence recognition and quality and envelope difference index in normal and altered rates were assessed in quiet and in noise at these three compression settings.Statistical analysis used:A repeated measure ANOVA.Results:We found a significant improvement in recognition of sentences at a normal rate in slow compression release time, compared to fast and linear gain settings at each SNR. Similar results were observed for sentences compressed at the rate of 35% in quiet and +10 dB SNR. Further, the participants preferred the quality of speech in quiet with the hearing aid set to slow compared to fast compression release time. The benefit from the slow compression release time was higher than either linear or fast compression release time on sentence recognition. Further, we saw that there was a negative impact on sentence recognition at 3 dB SNR (normal-rate) and in quiet (35% compression rate) in older adults.Conclusions:The slow compression release time in a hearing aid is superior to the fast one in noisy conditions and also with higher subjective ratings of speech quality in quiet.

Effects of WDRC release time and number of channels on output SNR and speech recognition.

The purpose of this study was to investigate the joint effects that wide dynamic range compression (WDRC) release time (RT) and number of channels have on recognition of sentences in the presence of steady and modulated maskers at different signal-to-noise ratios (SNRs). How the different combinations of WDRC parameters affect output SNR and the role this plays in the observed findings were also investigated. Twenty-four listeners with mild to moderate sensorineural hearing loss identified sentences mixed with steady or modulated maskers at three SNRs (-5, 0, and +5 dB) that had been processed using a hearing aid simulator with six combinations of RT (40 and 640 msec) and number of channels (4, 8, and 16). Compression parameters were set using the Desired Sensation Level v5.0a prescriptive fitting method. For each condition, amplified speech and masker levels and the resultant long-term output SNR were measured. Speech recognition with WDRC depended on the combination of RT and number of channels, with the greatest effects observed at 0 dB input SNR, in which mean speech recognition scores varied by 10 to 12% across WDRC manipulations. Overall, effect sizes were generally small. Across both masker types and the three SNRs tested, the best speech recognition was obtained with eight channels, regardless of RT. Increased speech levels, which favor audibility, were associated with the short RT and with an increase in the number of channels. These same conditions also increased masker levels by an even greater amount, for a net decrease in the long-term output SNR. Changes in long-term SNR across WDRC conditions were found to be strongly associated with changes in the temporal envelope shape as quantified by the Envelope Difference Index; however, neither of these factors fully explained the observed differences in speech recognition. A primary finding of this study was that the number of channels had a modest effect when analyzed at each level of RT, with results suggesting that selecting eight channels for a given RT might be the safest choice. Effects were smaller for RT, with results suggesting that short RT was slightly better when only 4 channels were used and that long RT was better when 16 channels were used. Individual differences in how listeners were influenced by audibility, output SNR, temporal distortion, and spectral distortion may have contributed to the size of the effects found in this study. Because only general suppositions could made for how each of these factors may have influenced the overall results of this study, future research would benefit from exploring the predictive value of these and other factors in selecting the processing parameters that maximize speech recognition for individuals.

Effect of Compression, Digital Noise Reduction and Directionality on Envelope Difference Index, Log-Likelihood Ratio and Perceived Quality.

The aim of the present study was to evaluate the use of the envelope difference index (EDI) and log-likelihood ratio (LLR) to quantify the independent and interactive effects of wide dynamic range compression, digital noise reduction and directionality, and to carry out self-rated quality measures. A recorded sentence embedded in speech spectrum noise at +5 dB signal to noise ratio was presented to a four channel digital hearing aid and the output was recorded with different combinations of algorithms at 30, 45 and 70 dB HL levels of presentation through a 2 cc coupler. EDI and LLR were obtained in comparison with the original signal using MATLAB software. In addition, thirty participants with normal hearing sensitivity rated the output on the loudness and clarity parameters of quality. The results revealed that the temporal changes happening at the output is independent of the number of algorithms activated together in a hearing aid. However, at a higher level of presentation, temporal cues are better preserved if all of these algorithms are deactivated. The spectral components speech tend to get affected by the presentation level. The results also indicate the importance of quality rating as this helps in considering whether the spectral and/or temporal deviations created in the hearing aid are desirable or not.

The Consonant-Weighted Envelope Difference Index (cEDI): A Proposed Technique for Quantifying Envelope Distortion

The benefits of amplitude compression in hearing aids may be limited by distortion resulting from rapid gain adjustment. To evaluate this, it is convenient to quantify distortion by using a metric that is sensitive to the changes in the processed signal that decrease consonant recognition, such as the Envelope Difference Index (EDI; Fortune, Woodruff, & Preves, 1994). However, the EDI relies on the entire duration of the signal, including portions irrelevant to consonant recognition. This note describes a computationally efficient method of automatically segmenting speech in time according to the syllable structure. Our technique uses the 1st derivative of the envelope as a basis. Peaks located in the derivative were used to generate a weighting function for the computation of a metric of signal distortion. The weighting function significantly improved the variance explained in consonant recognition scores over previous methods. However, only 3.2% of the variance was explained in the revised model. This technique was effective in focusing the analysis of distortion on specific segments of the signal. Use of the technique has implications for speech analysis. The difference in the amplitude envelope of consonants is not a robust model of the effect of hearing aid compression on consonant recognition.

Application of the Envelope Difference Index to Spectrally Sparse Speech

Amplitude compression is a common hearing aid processing strategy that can improve speech audibility and loudness comfort but also has the potential to alter important cues carried by the speech envelope. In previous work, a measure of envelope change, the Envelope Difference Index (EDI; Fortune, Woodruff, & Preves, 1994), was moderately related to recognition of spectrally robust consonants. This follow-up study investigated the relationship between the EDI and recognition of spectrally sparse consonants. Stimuli were vowel-consonant-vowel tokens processed to reduce spectral cues. Compression parameters were chosen to achieve a range of EDI values. Recognition was measured for 20 listeners with normal hearing. Both overall recognition and perception of consonant features were reduced at higher EDI values. Similar effects were noted with noise-vocoded and sine-vocoded processing and regardless of whether periodicity cues were available. The data provide information about the acceptable limits of envelope distortion under constrained conditions. These limits can be used to consider the impact of envelope distortions in situations where other cues are available to varying extents.

A comparison of modulation indices for describing amplitude-compressed speech.

This presentation explores two approaches to characterizing changes to speech with amplitude-compression amplification. In the first, we considered changes to waveform morphology using the envelope difference index (EDI); [Fortune et al. (1994)] to quantify the effects on amplitude envelope in the time domain. In the second, the spectral correlation index (SCI); [Gallun & Souza (2008)] was used to compare signal envelopes in the frequency domain. In the SCI, similarity is represented by the correlation of the average modulation spectra at six octave-spaced amplitude modulation frequencies (1–32 Hz) in six octave-spaced carrier frequency regions (250–8000 Hz) for a pair of phonemes. Test stimuli were a set of 16 vowel-consonant-vowel (VCV) syllables spoken by four talkers. The VCVs were digitally processed to limit spectral information using a signal-correlated noise technique [Schroeder (1968)] then submitted to a compression amplification simulation in which parameters (compression ratio and release time) were selected to achieve a range of temporal alteration. Consonant identification scores for 10 normal-hearing listeners were obtained and related to the modulation indices. Although the SCI was more sensitive to changes in specific compression parameters, the EDI was more closely related to recognition. [Work supported by R01 DC0060014 and VA RR&D CDA-2 C4963W.]

Temporal Envelope Changes of Compression and Speech Rate: Combined Effects on Recognition for Older Adults

When understanding speech in complex listening situations, older adults with hearing loss face the double challenge of cochlear hearing loss and deficits of the aging auditory system. Wide-dynamic range compression (WDRC) is used in hearing aids as remediation for the loss of audibility associated with hearing loss. WDRC processing has the additional effect of altering the acoustics of the speech signal, particularly the temporal envelope. Older listeners are negatively affected by other types of temporal distortions, but this has not been found for the distortion of WDRC processing for simple signals. The purpose of this research was to determine the circumstances under which older adults might be negatively affected by WDRC processing and what compensatory mechanisms those listeners might be using for the listening conditions when speech recognition performance is not affected. Two groups of adults with mild to moderate hearing loss were tested: (a) young-old (62-74 years, n=11) and (b) old-old (75-88 years, n=14). The groups did not differ in hearing loss, cognition, working memory, or self-reported health status. Participants heard low-predictability sentences compressed at each of 4 compression settings. The effect of compression on the temporal envelope was quantified by the envelope difference index (EDI; T. W. Fortune, B. D. Woodruff, & D. A. Preves, 1994). The sentences were presented at three rates: (a) normal rate, (b) 50% time compressed, and (c) time restored. There was no difference in performance between age groups, or any interactions involving age. There was a significant interaction between speech rate and EDI value; as the EDI value increased, representing higher amounts of temporal envelope distortion, speech recognition was significantly reduced. At the highest EDI value, this reduction was greater for the time-compressed than the normal rate condition. When time was restored to the time-compressed signals, speech recognition did not improve. Temporal envelope changes were detrimental to recognition of low-context speech for older listeners once a certain threshold of distortion was reached, particularly for rapid rate speech. For this sample tested, the effect was not age related within the age range tested here. The results of the time-restored condition suggested that listeners were using acoustic redundancy to compensate for the negative effects of WDRC distortion in the normal rate condition.

Quantifying the Effect of Compression Hearing Aid Release Time on Speech Acoustics and Intelligibility

Compression hearing aids have the inherent, and often adjustable, feature of release time from compression. Research to date does not provide a consensus on how to choose or set release time. The current study had 2 purposes: (a) a comprehensive evaluation of the acoustic effects of release time for a single-channel compression system in quiet and (b) an evaluation of the relation between the acoustic changes and speech recognition. The release times under study were 12, 100, and 800 ms. All of the stimuli were VC syllables from the Nonsense Syllable Task spoken by a female talker. The stimuli were processed through a hearing aid simulator at 3 input levels. Two acoustic measures were made on individual syllables: the envelope-difference index and CV ratio. These measurements allowed for quantification of the short-term amplitude characteristics of the speech signal and the changes to these amplitude characteristics caused by compression. The acoustic analyses revealed statistically significant effects among the 3 release times. The size of the effect was dependent on characteristics of the phoneme. Twelve listeners with moderate sensorineural hearing loss were tested for their speech recognition for the same stimuli. Although release time for this single-channel, 3:1 compression ratio system did not directly predict overall intelligibility for these nonsense syllables in quiet, the acoustic measurements reflecting the changes due to release time were significant predictors of phoneme recognition. Increased temporal-envelope distortion was predictive of reduced recognition for some individual phonemes, which is consistent with previous research on the importance of relative amplitude as a cue to syllable recognition for some phonemes.

Quantitative measures of envelope cues in speech recognition

A quantitative procedure is derived to evaluate the relative contribution of envelope cues to speech recognition. Recognition data of 16 consonants in the /aCa/ form were collected using signal‐correlated noise stimuli in seven normal‐hearing listeners. Several distance measures were calculated directly from duration and amplitude of the acoustic envelope. One amplitude distance measure was the Euclidean distance which was computed from the squared difference of the sample‐by‐sample amplitudes. The second measure was the envelope difference index (EDI) [Fortune et al., Ear Hear. 15, 93–95 (1994)] which was computed from the absolute value of the difference of the sample‐by‐sample amplitudes. A multidimensional scaling analysis was used to convert the perceptual confusion matrix into a distance matrix and to normalize the different distance measures. Correlation coefficients were computed between the different distance measures and the perceptual data. Preliminary analysis of data from six stop consonants ...