China
Africa
Explore chapters and articles related to this topic
Objective Quality and Intelligibility Measures
Published in Philipos C. Loizou, Speech Enhancement, 2013
A series of studies [82] were published validating the AI culminating with the creation of the ANSI S3.5-1969 standard [69]. These studies investigated among other things the impact of low-pass, high-pass, and bandpass filtering of speech in various noise masking conditions [82]. The standard was later revised in 1997. In the ANSI S3.5-1997 standard [70], the name of the measure changed from AI to SII. The computation of the audibility function (SNR) was modified to take into account the effects of spread of masking and vocal effort. Masking becomes an issue when higher energy vowels make lower energy consonants inaudible. The revised standard takes into consideration the fact that SI can decrease at extremely high sound pressure levels, something known as “roll-over” effect [68,83]. In the presence of high background noise levels, a talker is likely to raise the voice level (vocal effort) due to the Lombard effect. Increased vocal effort is associated with variation in the amplitude spectrum causing possible changes to intelligibility. The audibility function was also modified in the revised standard to accommodate individuals with conductive hearing loss. Hearing threshold levels may be used as additional input to the SII computation.
Practical Techniques for Improving Speech Recognition Performance
Published in John Holmes, Wendy Holmes, Speech Synthesis and Recognition, 2002
The difficulties of dealing with noise and other imposed signal disturbances are exacerbated by the tendency for talkers to modify the way they speak, and in particular to increase their vocal effort, when the acoustic environment worsens. This phenomenon is known as the Lombard effect, named after Etienne Lombard who first described it (Lombard, 1911). As environmental noise level increases, people’s natural response is to talk more loudly and often with a more exaggerated style of articulation. The consequence for the acoustic signal is an increase in overall level, but also, perhaps more significantly, changes in spectrum shape. One effect that has been observed is a change in spectral tilt. This change can be regarded as being due to another type of convolutional disturbance, the characteristics of which are dependent upon other (external) causes of corruption. There may also be changes in formant frequencies and in the durations of many of the speech sounds, due to factors such as more precise articulation, increased muscular tension and reduced speaking rate.
Signal Types
Published in Eddy B. Brixen, Audio Metering, 2020
When we speak, and background noise gets higher, we tend to raise the level (and the pitch) of the voice. We do this involuntarily to enhance the audibility of the voice. This special behavior is called the Lombard effect or the Lombard reflex (named after the French otolaryngologist Étienne Lombard, who discovered the phenomenon in 1909).
A study of human vocal effort in response to the architectural auditory environment
Published in Architectural Science Review, 2020
Pantea Alambeigi, Jane Burry, Sipei Zhao, Eva Cheng
The discovery of the Lombard effect by a French otolaryngologist named Étienne Lombard in 1911, demonstrates the propensity to elevate sound level in the presence of high background noise. Kryter (2013) termed the noise influence on vocal effort as an axiom. Elevating the speech level in higher background noise implies that speakers audit the sound level output actively to ensure a constant level of understanding in their communications (Summers et al. 1988). More background noise will lead to an unconscious increase in the voice level to maintain the intelligibility of the conversation (Bradley and Gover 2011; Lazarus 1987).