Echoes can make speech harder to understand, and turning off echoes in an audio recording is a notoriously difficult engineering problem. However, the human brain appears to successfully solve the problem by separating the sound in direct speech and its echo, according to a study published February 15 in the open access journal. More biology by Jiaxin Gao of Zhejiang University, China, and colleagues.
Audio signals in poorly designed online meetings and auditoriums often echo at least 100 milliseconds behind the original speech. These echoes greatly distort speech, interfering with the slowly varying sound characteristics that are most important for understanding conversations; however, people still reliably understand echoic speech. To better understand how the brain enables this, the authors used magnetoencephalography (MEG) to record neural activity while human participants listened to a story with and without echo. They compared the neural signals with two computational models: one that simulates the brain adaptation to the echo, and another simulating the brain pulling away the echo of the original speech.
Participants understood the story with more than 95% accuracy, regardless of the echo. The researchers observed that cortical activity follows the energy changes related to direct speech, despite strong echo interference. Simulating neural adaptation only partially captured the brain response they observed: The neural activity was best explained by a model that split the original speech and its echo into separate processing streams. This remained true even when participants were asked to direct their attention to a silent film and ignore the story, suggesting that top-down attention is not required to mentally separate direct speech and its echo. The researchers say that auditory stream segregation may be important both for identifying a specific speaker in a crowded environment and for clearly understanding an individual speaker in a reverberant space.
The authors add: “Echoes strongly distort the sound characteristics of speech and create a challenge for automatic speech recognition. However, the human brain can segregate speech from its echo and achieve reliable recognition of echoic speech.”