niels_round
Niels Henrik Pontoppidan

Research Area Manager/ Research Engineer

> Learn more

mail@eriksholm.com

This study investigated the benefits experienced by normal-hearing listeners from TFS in speech recognition with three spatially separated speakers. The combination of tone-vocoding and simulated spatial conditions enabled us to isolate the benefits of monaural and binaural TFS cues. The results revealed that both monaural and binaural TFS cues can significantly improve speech recognition.
l-r

More realistic testing

Previous demonstrations of the TFS benefit were obtained in monaural listening conditions over headphones. This study measured the importance of TFS in a more realistic scenario with three spatially separated speakers. The spatial condition in Figure 1 was simulated with generic head-related transfer functions (HRTFs) over headphones.

Figure 1: Spatial speech-in-speech test setup. The target (T) was placed in front and the two competing talkers (M1 and M2) 50 degrees off to the elft and right.

binaural-vs-monaural

Binaural TFS (bTFS) and Monaural TFS (mTFS)

Speech recognition can benefit from Binaural processing of TFS (bTFS) in two ways: 1) sensing the same TFS in both ears could be an indication that the sound is originating from a single source, and 2) the time difference between the two TFS signals could help the listener to localise the source. The ordering of the tone-vocoding and the spatial simulation allowed us to isolate the bTFS and mTFS contributions in the three conditions shown in Figure 2.

When the tone-vocoding takes place prior to applying the HRTFs, the signal reaching the ears has artificial TFS, but normal spatial properties (vocBefore). When the tone-vocoding takes place after applying the HRTFs, the signal reaching the ears has artificial TFS and ITD's. Both conditions maintain the interaural level differences. Therefore the contrast between unprocessed and vocBefore simulates mTFS removal, and the contrast between vocBefore and vocAfter simulates bTFS removal.

Figure 2: Illustration of how the sequence of tone-vocoding and spatial simulation isolates monaural TFS and binaural TFS

srt

Results

The participants were eight normal-hearing listeners. The results showed a significant effect of the mTFS and the bTFS on p<0.05 level. Altogether the results indicate that TFS enhances both monaural and binaural processing in the auditory system. Furthermore, the 6.9 dB benefit in SRT that emerges through access to mTFS is of similar magnitude as that found in the simular monaural TFS speech test.


Further reading

Neher T, Behrens T, Carlile S, Jin C, Kragelund L, Petersen AS, van Schaik A (2009). Benefit from spatial separation of multiple talkers in bilateral hearing-aid users: Effects of hearing loss, age, and cognition. International Journal of Audiology, 48(11), p.758-774.

Learn more

  • learnmore-poster

    Poster

    The benefit of temporal fine structure in spatial release from masking for older hearing-impaired listeners

  • learnmore-poster

    Poster

    Effect of binaural tone vocoding on recognising target speech