Hi Paul,<br /><br />Thanks for the prompt answer.<br /><br />I am trying to make a freeware software for identification of bird calls in noisy conditions. One of critical points in the task is to correctly identify "audio feature" (e.g. a bird call). As you can imagine length of a call is not fixed and distance from onset to onset in not necessarily call length. That is why I am looking for offset detection mechanism. <br /><br />If I need to retrieve both onset and offset, should I then use aubionotes instead of aubioonset?<br /><br /><br />Cheers,<br />Lucas<br /><br /><span>On 04/01/14, <b class="name">Paul Brossier </b> <piem@piem.org> wrote:</span><blockquote cite="mid:52C80DA1.6000402@piem.org" class="iwcQuote" style="border-left: 1px solid #00F; padding-left: 13px; margin-left: 0;" type="cite"><div class="mimepart text plain">On 04/01/2014 06:00, Lukasz Tracewski wrote:<br />>Hi,<br />><br />>Aubio has a plethora of great onset detection methods. How about offset<br />>detection? I can imagine that running algorithm on reverse-ordered data<br />>should provide me with offsets, but maybe there is a smarter approach<br />>that I am missing.<br /><br />Hi Lukasz,<br /><br />There is no offset detection method in aubio (yet). The onset detection<br />functions will sometimes peak during the transient sound created by the<br />release of a note, but these peaks are expected be small compared to the<br />ones produced by note onsets.<br /><br />So far, offsets times are assumed to be where the level of the signal<br />drops under a given threshold. aubionotes for instance uses a simple<br />silence gate to determine the end of notes.<br /><br />What kind of offsets are you looking at? Do you have a specific<br />algorithm in mind?<br /><br />There is the work of Emmanouil Benetos and Simon Dixon, who use a Hidden<br />Markov Model to determine whether a pitch track is active or not:<br /><br /><a href="http://www.eecs.qmul.ac.uk/~simond/pub/2011/Benetos-Dixon-ICASSP2011.pdf" target="1">http://www.eecs.qmul.ac.uk/~simond/pub/2011/Benetos-Dixon-ICASSP2011.pdf</a><br /><br />Best, Paul<br /><br /><br />><br />>Thanks,<br />>Lucas<br />><br />><br />>_______________________________________________<br />>aubio-user mailing list<br />>aubio-user@aubio.org<br />><a href="https://lists.aubio.org/listinfo/aubio-user" target="1">https://lists.aubio.org/listinfo/aubio-user</a><br />><br /></div></blockquote>