0:00:18i'm going to have proposed here with my only collectively from if you k
0:00:23it's rental
0:00:24and basically score multichannel i-vector combination for robust speaker verification from the most environments
0:00:32basically we were well a couple of years ago in a in a project about
0:00:35a automation application basically was glued inner product and the main characteristics of the project
0:00:44was uttered below a kind of or
0:00:49speech interface in order to talk to activate the windows the doors and so on
0:00:54for people with physical impairments
0:00:56i either they're
0:00:58them as you mentioned above the main that this is was that the use or
0:01:01try to users like a multiple microphones are not was from the big microphones with
0:01:06of the system always listening and of course the use was supposed to be able
0:01:09to say comments from anywhere in any position in the room
0:01:14so it we of course that some speaker tasks and some speaker services we were
0:01:21there's the in and basically the conference we were expecting in this kind of applications
0:01:29not it was
0:01:30of course of the speaker can be anyone so we expect to have a huge
0:01:34mismatch of in between enrollment the spaces
0:01:37even if you have fixed the problem and or even clean enrolment with a and
0:01:42i from application about as more from location
0:01:44and of course we need to the models that were able to cope with this
0:01:47problem a typically wouldn't five number of microphones the solution that
0:01:52you can find
0:01:54in the literature are trying to play something of the speech enhancement even microphone array
0:02:00beamforming and someone usually you need but it will clearly that devices for that and
0:02:05sensors you can do some common combination the post processed with the to describe it
0:02:12and to play every with i-vectors or to combine different channels in the sense different
0:02:17channels that are recorded in a at the same utterance so we have
0:02:22the same
0:02:24different samples of the same utterance i for that we would basically use this apartment
0:02:29with we may sure many impulse responses with a from many locations of positions and
0:02:35which relate the database
0:02:37i'm basically we will that the scenario we will focus on some have been a
0:02:41fixed position for enrollment and you can be anywhere for speaker verification we are explained
0:02:48every for the combined effect but we can try and we can talk more an