Dear Dick,
Not directly related to source separation as you queried in your original message, but these may be also be of interest nonetheless:
1) I used a simple statistical model of the precedence effect as part of my PhD in 2004. The statistical model was published in AAuA in 2006:
Hacihabiboglu, H., and Murtagh, F., (2006), "An observational study of the precedence effect", Acta Acustica united with Acustica, 90(3), pp. 440-456, May/June 2006.
We have later used that model to simplify ISM-based binaural room auralization by a perceptually informed clustering of image sources:
Hacihabiboglu, H., and Murtagh, F., (2008), "Perceptual simplification for model-based binaural room auralisation", Applied Acoustics, 69(8), pp. 715-727, August 2008.
2) In an earlier work published in IEEE WASPAA in 2003 I used onset dominance to reduce the complexity of the binaural synthesis of early reflections. A graceful degradation approach was used to keep the filter size fixed while processing not only the direct sound but also the reflections.
Hacihabiboglu H., (2003) “A fixed-cost variable-length auralization filter model utilizing the precedence effect”, Proc. IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA-03), pp. 1-4, New York, USA, 19-22 October, 2003.
I hope that these are also relevant.
Best,
Hüseyin
--
Dr Huseyin Hacihabiboglu
Assistant Professor
Head of Department of Modelling and Simulation
Informatics Institute, METU, 06800, Ankara, Turkey
tel: +90 312 2107889 | e-mail: hhuseyin@xxxxxxxxxxx | web: http://www.hacihabiboglu.org | VoIP: husshho (skype)
On 24 Sep 2014, at 07:54, Richard F. Lyon <dicklyon@xxxxxxx> wrote:
> Axel, yes, those are good ones; thanks for your paper, and for reminding me of Palomäki et al 2004, which I have.
> I will mention these as precedence-effect application successes in my binaural chapter.
>
> Dick
>
>
> On Tue, Sep 23, 2014 at 12:22 PM, Axel Plinge <axel.plinge@xxxxxxxxxxxxxx> wrote:
> Hello Dick!
>
> I am not sure if this is what you are looking for, however I am using a simple precedence effect in my 'machine hearing' model for speaker localization by subtracting a time-shifted average from the signal in each band, as introduced in:
>
> Plinge, A., Hennecke, M. H., & Fink, G. A. "Robust Neuro-Fuzzy Speaker Localization Using a Circular Microphone Array", in 12th International Workshop on Acoustic Echo and Noise Control; Tel Aviv, Israel, 2010 http://patrec.cs.tu-dortmund.de//pubs/abstracts/Plinge2010-RNF.htm
>
> A similar idea was used in:
> K. J. Palomäki, G. J. Brown, and D. L.Wang, “A binaural processor for missing data speech recognition in the presence of noise and small-room reverberation,” in Speech Communication, 2004, vol. 43 (4), pp. 273–398.
> Kind regards,
>
> Axel
> --
>
> http://patrec.cs.tu-dortmund.de/cms/en/home/People/Staff/plinge.html
>
> TU Dortmund, Department of Computer Science, LS XII,
> Pattern Recognition, Otto-Hahn-Str. 16, 44227 Dortmund, Germany
>
>