Companding to improve cochlear-implant speech recognition in speech-shaped noise

Nonlinear sensory and neural processing mechanisms have been exploited to enhance spectral contrast for improvement of speech understanding in noise. The "companding" algorithm employs both two-tone suppression and adaptive gain mechanisms to achieve spectral enhancement. This study implemented a 50-channel companding strategy and evaluated its efficiency as a front-end noise suppression technique in cochlear implants. The key parameters were identified and evaluated to optimize the companding performance. Both normal-hearing (NH) listeners and cochlear-implant (CI) users performed phoneme and sentence recognition tests in quiet and in steady-state speech-shaped noise. Data from the NH listeners showed that for noise conditions, the implemented strategy improved vowel perception but not consonant and sentence perception. However, the CI users showed significant improvements in both phoneme and sentence perception in noise. Maximum average improvement for vowel recognition was 21.3 percentage points (p<0.05) at 0 dB signal-to-noise ratio (SNR), followed by 17.7 percentage points (p<0.05) at 5 dB SNR for sentence recognition and 12.1 percentage points (p<0.05) at 5 dB SNR for consonant recognition. While the observed results could be attributed to the enhanced spectral contrast, it is likely that the corresponding temporal changes caused by companding also played a significant role and should be addressed by future studies.     

Across-site patterns of modulation detection: relation to speech recognition

The aim of this study was to identify across-site patterns of modulation detection thresholds (MDTs) in subjects with cochlear implants and to determine if removal of sites with the poorest MDTs from speech processor programs would result in improved speech recognition. Five hundred millisecond trains of symmetric-biphasic pulses were modulated sinusoidally at 10 Hz and presented at a rate of 900 pps using monopolar stimulation. Subjects were asked to discriminate a modulated pulse train from an unmodulated pulse train for all electrodes in quiet and in the presence of an interleaved unmodulated masker presented on the adjacent site. Across-site patterns of masked MDTs were then used to construct two 10-channel MAPs such that one MAP consisted of sites with the best masked MDTs and the other MAP consisted of sites with the worst masked MDTs. Subjects' speech recognition skills were compared when they used these two different MAPs. Results showed that MDTs were variable across sites and were elevated in the presence of a masker by various amounts across sites. Better speech recognition was observed when the processor MAP consisted of sites with best masked MDTs, suggesting that temporal modulation sensitivity has important contributions to speech recognition with a cochlear implant.     

Excess noise in intracavity laser frequency modulation

The response to perturbations and to stochastic noise of a laser below threshold subjected to an intracavity periodic frequency modulation is theoretically studied. It is shown that, when the modulation frequency is close to the cavity axial mode separation but yet detuned from exact resonance, the laser exhibits a strongly enhanced sensitivity to external noise, which includes large transient energy amplification of perturbations in the deterministic case and enhancement of field fluctuations in presence of a continuous stochastic noise. This large excess noise is due to the nonorthogonality of Floquet laser modes which makes it possible continuous energy transfer from the forcing noise to transient growing perturbations.     

Benefits of amplification for speech recognition in background noise

The purpose of the present study was to examine the benefits of providing audible speech to listeners with sensorineural hearing loss when the speech is presented in a background noise. Previous studies have shown that when listeners have a severe hearing loss in the higher frequencies, providing audible speech (in a quiet background) to these higher frequencies usually results in no improvement in speech recognition. In the present experiments, speech was presented in a background of multitalker babble to listeners with various severities of hearing loss. The signal was low-pass filtered at numerous cutoff frequencies and speech recognition was measured as additional high-frequency speech information was provided to the hearing-impaired listeners. It was found in all cases, regardless of hearing loss or frequency range, that providing audible speech resulted in an increase in recognition score. The change in recognition as the cutoff frequency was increased, along with the amount of audible speech information in each condition (articulation index), was used to calculate the "efficiency" of providing audible speech. Efficiencies were positive for all degrees of hearing loss. However, the gains in recognition were small, and the maximum score obtained by an listener was low, due to the noise background. An analysis of error patterns showed that due to the limited speech audibility in a noise background, even severely impaired listeners used additional speech audibility in the high frequencies to improve their perception of the "easier" features of speech including voicing.     

语音识别中信道和噪音的联合补偿

频谱和倒谱的联合调整方法,用于对语音识别中信道差异和背景噪音的存在进行联合补偿。该方法根据干净语音的最大似然准则在频域和倒谱域分别对噪音和信道进行补偿,避免了对噪音和信道影响模型进行简化所带来的误差影响,且实现时间复杂度较低。在信噪比由10dB到20dB的含有信道和加性噪音的汉语数字串识别实验中,该方法使平均音节错误率相对下降了50.44%。实验表明频谱和倒谱的联合调整方法可以快速的补偿信道差异和背景噪音。     

Improvement of speech recognition by nonlinear noise reduction

The success of nonlinear noise reduction applied to a single channel recording of human voice is measured in terms of the recognition rate of a commercial speech recognition program in comparison to the optimal linear filter. The overall performance of the nonlinear method is shown to be superior. We hence demonstrate that an algorithm that has its roots in the theory of nonlinear deterministic dynamics possesses a large potential in a realistic application.     

Dichotic speech recognition in noise using reduced spectral cues

It is generally accepted that the fusion of two speech signals presented dichotically is affected by the relative onset time. This study investigated the hypothesis that spectral resolution might be an additional factor influencing spectral fusion when the spectral information is split and presented dichotically to the two ears. To produce speech with varying degrees of spectral resolution, speech materials embedded in +5 dB S/N speech-shaped noise were processed through 6-12 channels and synthesized as a sum of sine waves. Two different methods of splitting the spectral information were investigated. In the first method, the odd-index channels were presented to one ear and the even-index channels to the other ear. In the second method the lower frequency channels were presented to one ear and the high-frequency channels to the other ear. Results indicated that spectral resolution did affect spectral fusion, and the effect differed across speech materials, with the sentences being affected the most. Sentences, processed through six or eight channels and presented dichotically in the low-high frequency condition were not fused as accurately as when presented monaurally. Sentences presented dichotically in the odd-even frequency condition were identified more accurately than when presented in the low-high condition.     

Contribution of low-frequency acoustic information to Chinese speech recognition in cochlear implant simulations

Chinese sentence recognition strongly relates to the reception of tonal information. For cochlear implant (CI) users with residual acoustic hearing, tonal information may be enhanced by restoring low-frequency acoustic cues in the nonimplanted ear. The present study investigated the contribution of low-frequency acoustic information to Chinese speech recognition in Mandarin-speaking normal-hearing subjects listening to acoustic simulations of bilaterally combined electric and acoustic hearing. Subjects listened to a 6-channel CI simulation in one ear and low-pass filtered speech in the other ear. Chinese tone, phoneme, and sentence recognition were measured in steady-state, speech-shaped noise, as a function of the cutoff frequency for low-pass filtered speech. Results showed that low-frequency acoustic information below 500 Hz contributed most strongly to tone recognition, while low-frequency acoustic information above 500 Hz contributed most strongly to phoneme recognition. For Chinese sentences, speech reception thresholds (SRTs) improved with increasing amounts of low-frequency acoustic information, and significantly improved when low-frequency acoustic information above 500 Hz was preserved. SRTs were not significantly affected by the degree of spectral overlap between the CI simulation and low-pass filtered speech. These results suggest that, for CI patients with residual acoustic hearing, preserving low-frequency acoustic information can improve Chinese speech recognition in noise.     

Monosyllabic word recognition at higher-than-normal speech and noise levels

The effects of intensity on monosyllabic word recognition were studied in adults with normal hearing and mild-to-moderate sensorineural hearing loss. The stimuli were bandlimited NU#6 word lists presented in quiet and talker-spectrum-matched noise. Speech levels ranged from 64 to 99 dB SPL and S/N ratios from 28 to -4 dB. In quiet, the performance of normal-hearing subjects remained essentially constant in noise, at a fixed S/N ratio, it decreased as a linear function of speech level. Hearing-impaired subjects performed like normal-hearing subjects tested in noise when the data were corrected for the effects of audibility loss. From these and other results, it was concluded that: (1) speech intelligibility in noise decreases when speech levels exceed 69 dB SPL and the S/N ratio remains constant; (2) the effects of speech and noise level are synergistic; (3) the deterioration in intelligibility can be modeled as a relative increase in the effective masking level; (4) normal-hearing and hearing-impaired subjects are affected similarly by increased signal level when differences in speech audibility are considered; (5) the negative effects of increasing speech and noise levels on speech recognition are similar for all adult subjects, at least up to 80 years; and (6) the effective dynamic range of speech may be larger than the commonly assumed value of 30 dB.     

Channel selection in the modulation domain for improved speech intelligibility in noise

Background noise reduces the depth of the low-frequency envelope modulations known to be important for speech intelligibility. The relative strength of the target and masker envelope modulations can be quantified using a modulation signal-to-noise ratio, (S/N)(mod), measure. Such a measure can be used in noise-suppression algorithms to extract target-relevant modulations from the corrupted (target + masker) envelopes for potential improvement in speech intelligibility. In the present study, envelopes are decomposed in the modulation spectral domain into a number of channels spanning the range of 0-30 Hz. Target-dominant modulations are identified and retained in each channel based on the (S/N)(mod) selection criterion, while modulations which potentially interfere with perception of the target (i.e., those dominated by the masker) are discarded. The impact of modulation-selective processing on the speech-reception threshold for sentences in noise is assessed with normal-hearing listeners. Results indicate that the intelligibility of noise-masked speech can be improved by as much as 13 dB when preserving target-dominant modulations, present up to a modulation frequency of 18 Hz, while discarding masker-dominant modulations from the mixture envelopes.     

Fundamental frequency is critical to speech perception in noise in combined acoustic and electric hearing

Cochlear implant (CI) users have been shown to benefit from residual low-frequency hearing, specifically in pitch related tasks. It remains unclear whether this benefit is dependent on fundamental frequency (F0) or other acoustic cues. Three experiments were conducted to determine the role of F0, as well as its frequency modulated (FM) and amplitude modulated (AM) components, in speech recognition with a competing voice. In simulated CI listeners, the signal-to-noise ratio was varied to estimate the 50% correct response. Simulation results showed that the F0 cue contributes to a significant proportion of the benefit seen with combined acoustic and electric hearing, and additionally that this benefit is due to the FM rather than the AM component. In actual CI users, sentence recognition scores were collected with either the full F0 cue containing both the FM and AM components or the 500-Hz low-pass speech cue containing the F0 and additional harmonics. The F0 cue provided a benefit similar to the low-pass cue for speech in noise, but not in quiet. Poorer CI users benefited more from the F0 cue than better users. These findings suggest that F0 is critical to improving speech perception in noise in combined acoustic and electric hearing.     

Effects of age and mild hearing loss on speech recognition in noise

Using an adaptive strategy, the effects of mild sensorineural hearing loss and adult listeners' chronological age on speech recognition in babble were evaluated. The signal-to-babble ratio required to achieve 50% recognition was measured for three speech materials presented at soft to loud conversational speech levels. Four groups of subjects were tested: (1) normal-hearing listeners less than 44 years of age, (2) subjects less than 44 years old with mild sensorineural hearing loss and excellent speech recognition in quiet, (3) normal-hearing listeners greater than 65 with normal hearing, and (4) subjects greater than 65 years old with mild hearing loss and excellent performance in quiet. Groups 1 and 3, and groups 2 and 4 were matched on the basis of pure-tone thresholds, and thresholds for each of the three speech materials presented in quiet. In addition, groups 1 and 2 were similar in terms of mean age and age range, as were groups 3 and 4. Differences in performance in noise as a function of age were observed for both normal-hearing and hearing-impaired listeners despite equivalent performance in quiet. Subjects with mild hearing loss performed significantly worse than their normal-hearing counterparts. These results and their implications are discussed.     

Spectral models of additive and modulation noise in speech and phonatory excitation signals

The article presents spectral models of additive and modulation noise in speech. The purpose is to learn about the causes of noise in the spectra of normal and disordered voices and to gauge whether the spectral properties of the perturbations of the phonatory excitation signal can be inferred from the spectral properties of the speech signal. The approach to modeling consists of deducing the Fourier series of the perturbed speech, assuming that the Fourier series of the noise and of the clean monocycle-periodic excitation are known. The models explain published data, take into account the effects of supraglottal tremor, demonstrate the modulation distortion owing to vocal tract filtering, establish conditions under which noise cues of different speech signals may be compared, and predict the impossibility of inferring the spectral properties of the frequency modulating noise from the spectral properties of the frequency modulation noise (e.g., phonatory jitter and frequency tremor). The general conclusion is that only phonatory frequency modulation noise is spectrally relevant. Other types of noise in speech are either epiphenomenal, or their spectral effects are masked by the spectral effects of frequency modulation noise.     

Notionally steady background noise acts primarily as a modulation masker of speech

Stone et al. [J. Acoust. Soc Am. 130, 2874-2881 (2011)], using vocoder processing, showed that the envelope modulations of a notionally steady noise were more effective than the envelope energy as a masker of speech. Here the same effect is demonstrated using non-vocoded signals. Speech was filtered into 28 channels. A masker centered on each channel was added to the channel signal at a target-to-background ratio of -5 or -10 dB. Maskers were sinusoids or noise bands with bandwidth 1/3 or 1 ERB(N) (ERB(N) being the bandwidth of "normal" auditory filters), synthesized with Gaussian (GN) or low-noise (LNN) statistics. To minimize peripheral interactions between maskers, odd-numbered channels were presented to one ear and even to the other. Speech intelligibility was assessed in the presence of each "steady" masker and that masker 100% sinusoidally amplitude modulated (SAM) at 8 Hz. Intelligibility decreased with increasing envelope fluctuation of the maskers. Masking release, the difference in intelligibility between the SAM and its "steady" counterpart, increased with bandwidth from near-zero to around 50 percentage points for the 1-ERB(N) GN. It is concluded that the sinusoidal and GN maskers behaved primarily as energetic and modulation maskers, respectively.     

Use of intonation contours for speech recognition in noise by cochlear implant recipients

The corruption of intonation contours has detrimental effects on sentence-based speech recognition in normal-hearing listeners Binns and Culling [(2007). J. Acoust. Soc. Am. 122, 1765-1776]. This paper examines whether this finding also applies to cochlear implant (CI) recipients. The subjects' F0-discrimination and speech perception in the presence of noise were measured, using sentences with regular and inverted F0-contours. The results revealed that speech recognition for regular contours was significantly better than for inverted contours. This difference was related to the subjects' F0-discrimination providing further evidence that the perception of intonation patterns is important for the CI-mediated speech recognition in noise.     

Improved speech recognition in noise in simulated binaurally combined acoustic and electric stimulation

Speech recognition in noise improves with combined acoustic and electric stimulation compared to electric stimulation alone [Kong et al., J. Acoust. Soc. Am. 117, 1351-1361 (2005)]. Here the contribution of fundamental frequency (F0) and low-frequency phonetic cues to speech recognition in combined hearing was investigated. Normal-hearing listeners heard vocoded speech in one ear and low-pass (LP) filtered speech in the other. Three listening conditions (vocode-alone, LP-alone, combined) were investigated. Target speech (average F0=120 Hz) was mixed with a time-reversed masker (average F0=172 Hz) at three signal-to-noise ratios (SNRs). LP speech aided performance at all SNRs. Low-frequency phonetic cues were then removed by replacing the LP speech with a LP equal-amplitude harmonic complex, frequency and amplitude modulated by the F0 and temporal envelope of voiced segments of the target. The combined hearing advantage disappeared at 10 and 15 dB SNR, but persisted at 5 dB SNR. A similar finding occurred when, additionally, F0 contour cues were removed. These results are consistent with a role for low-frequency phonetic cues, but not with a combination of F0 information between the two ears. The enhanced performance at 5 dB SNR with F0 contour cues absent suggests that voicing or glimpsing cues may be responsible for the combined hearing benefit.     

Spectro-temporal modulation subspace-spanning filter bank features for robust automatic speech recognition

In an attempt to increase the robustness of automatic speech recognition (ASR) systems, a feature extraction scheme is proposed that takes spectro-temporal modulation frequencies (MF) into account. This physiologically inspired approach uses a two-dimensional filter bank based on Gabor filters, which limits the redundant information between feature components, and also results in physically interpretable features. Robustness against extrinsic variation (different types of additive noise) and intrinsic variability (arising from changes in speaking rate, effort, and style) is quantified in a series of recognition experiments. The results are compared to reference ASR systems using Mel-frequency cepstral coefficients (MFCCs), MFCCs with cepstral mean subtraction (CMS) and RASTA-PLP features, respectively. Gabor features are shown to be more robust against extrinsic variation than the baseline systems without CMS, with relative improvements of 28% and 16% for two training conditions (using only clean training samples or a mixture of noisy and clean utterances, respectively). When used in a state-of-the-art system, improvements of 14% are observed when spectro-temporal features are concatenated with MFCCs, indicating the complementarity of those feature types. An analysis of the importance of specific MF shows that temporal MF up to 25 Hz and spectral MF up to 0.25 cycles/channel are beneficial for ASR.     

Laser frequency modulation noise measurement by recirculating delayed self-heterodyne method

I propose and demonstrate the use of the recirculating delayed self-heterodyne (DSH) method for measuring FM noise power spectral densities (PSDs), which are the most fundamental measure characterizing the spectral purity of laser sources. By analyzing the DSH beat signals with 1, 10, and 160?km delays, the FM noise PSD of a narrow-linewidth fiber laser is evaluated for Fourier frequency range between 10?Hz and 100?kHz, which exhibits flicker noise as the dominant contribution.     

The effect of infrasonic and very low frequency noise on speech intelligibility

An experiment was performed in which a noise containing frequencies from 10 Hz to 47 Hz was used to mask speech. The behaviour of speech intelligibility with speech presentation level and masking noise level was examined briefly.The infrasonic and low frequency masking noise did reduce the intelligibility of speech. The effect only became significant when the masking noise level was present at levels of 115 dB OASPL or above.     

The ability of cochlear implant users to use temporal envelope cues recovered from speech frequency modulation

Previous studies have demonstrated that normal-hearing listeners can understand speech using the recovered "temporal envelopes," i.e., amplitude modulation (AM) cues from frequency modulation (FM). This study evaluated this mechanism in cochlear implant (CI) users for consonant identification. Stimuli containing only FM cues were created using 1, 2, 4, and 8-band FM-vocoders to determine if consonant identification performance would improve as the recovered AM cues become more available. A consistent improvement was observed as the band number decreased from 8 to 1, supporting the hypothesis that (1) the CI sound processor generates recovered AM cues from broadband FM, and (2) CI users can use the recovered AM cues to recognize speech. The correlation between the intact and the recovered AM components at the output of the sound processor was also generally higher when the band number was low, supporting the consonant identification results. Moreover, CI subjects who were better at using recovered AM cues from broadband FM cues showed better identification performance with intact (unprocessed) speech stimuli. This suggests that speech perception performance variability in CI users may be partly caused by differences in their ability to use AM cues recovered from FM speech cues.