A localization algorithm based on head-related transfer functions *The Journal of the Acoustical Society of America*, Vol. 123, No. 6. (June 2008), pp. 4290-4296 by Justin A. Macdonald

@article{macdonald-2008, abstract = {Two sound localization algorithms based on the head-related transfer function were developed. Each of them uses the interaural time delay, interaural level difference, and monaural spectral cues to estimate the location of a sound source. Given that most localization algorithms will be required to function in background noise, the localization performance of one of the algorithms was tested at signal-to-noise ratios ({SNRs}) from 40 to -40 {dB}. Stimuli included ten real-world, broadband sounds located at 5 degrees intervals in azimuth and at 0 degrees elevation. Both two- and four-microphone versions of the algorithm were implemented to localize sounds to 5 degrees precision. The two-microphone version of the algorithm exhibited less than 2 degrees mean localization error at {SNRs} of 20 {dB} and greater, and the four-microphone version committed approximately 1 degrees mean error at {SNRs} of 10 {dB} or greater. Potential enhancements and applications of the algorithm are discussed.}, author = {MacDonald, Justin A.}, citeulike-article-id = {13445817}, citeulike-linkout-0 = {http://view.ncbi.nlm.nih.gov/pubmed/18537380}, citeulike-linkout-1 = {http://www.hubmed.org/display.cgi?uids=18537380}, issn = {1520-8524}, journal = {The Journal of the Acoustical Society of America}, keywords = {hrtf, ssl}, month = jun, number = {6}, pages = {4290--4296}, pmid = {18537380}, posted-at = {2014-11-26 13:35:52}, priority = {2}, title = {A localization algorithm based on head-related transfer functions}, url = {http://view.ncbi.nlm.nih.gov/pubmed/18537380}, volume = {123}, year = {2008} }

See the CiteULike entry for more info, PDF links, BibTex etc.

The way sound is shaped by the head and body before reaching the ears of a listener is described by a head-related transfer function (HRTF). There is a different HRTF for every angle of incidence.⇒

MacDonald proposes two methods for sound source localization based on head-related transfer functions (actually the HRIR, their representation in the time domain).⇒

The first method for SSL proposed by MacDonald applies the **inverse** of the HRIR $F^{(i,\theta)}$ to the signal recorded by $i$ For each microphone $i$ and every candidate angle $\theta$.
It then uses the Pearson correlation coefficient to compare the resultant signals.
Only for the correct angle $\theta$ should the signals match.⇒

The second method for (binaural) SSL proposed by MacDonald applies the HRIR $F^{(o,\theta)}$ to the signals recorded by the left and right microphones every candidate angle θ, where $F^{(o,\theta)}$ is the respective opposite microphone. It then uses the Pearson correlation coefficient to compare the resultant signals. Only for the correct angle θ should the signals match.⇒

The binaural sound-source localization methods proposed by MacDonald can be extended to larger arrays of microphones.⇒