0:00:13 | you |
---|
0:00:14 | a fine |
---|
0:00:15 | we work |
---|
0:00:17 | uh in which he that issue |
---|
0:00:19 | oh well will be C i E |
---|
0:00:22 | oh a function of |
---|
0:00:23 | yeah |
---|
0:00:24 | mean |
---|
0:00:26 | oh what |
---|
0:00:27 | we action |
---|
0:00:28 | a a yeah for had a |
---|
0:00:31 | a function |
---|
0:00:32 | oh would you that of a function i three |
---|
0:00:36 | oh the future it actual P in a total of not |
---|
0:00:41 | it's a a a a single the issue and uh it's all those a you know because |
---|
0:00:46 | uh one thing to have a different difference |
---|
0:00:49 | oh it's was five so you know have and total |
---|
0:00:52 | um |
---|
0:00:54 | it's not yeah |
---|
0:00:55 | um can be a change just you three five regiment |
---|
0:00:58 | oh |
---|
0:00:59 | of of of a should a shot yeah |
---|
0:01:02 | i i but is only |
---|
0:01:04 | um |
---|
0:01:05 | a response and |
---|
0:01:07 | uh sum |
---|
0:01:08 | some |
---|
0:01:09 | they uh some spatial position and |
---|
0:01:12 | the whole procedure is |
---|
0:01:14 | oh we see time as you me |
---|
0:01:17 | so you want to have a but and uh |
---|
0:01:20 | can you model to recreate then |
---|
0:01:22 | in we a charge yeah |
---|
0:01:24 | oh but a higher order to C |
---|
0:01:27 | yeah |
---|
0:01:28 | i'm going to uh uh and for the use the three D five you |
---|
0:01:32 | a a continuous hrtf model |
---|
0:01:35 | uh |
---|
0:01:36 | here's the |
---|
0:01:37 | a a a a reference |
---|
0:01:38 | with with the house |
---|
0:01:40 | uh that's we can see that |
---|
0:01:42 | um in the model |
---|
0:01:44 | model of is the frequency components to from |
---|
0:01:48 | spatial component |
---|
0:01:49 | and because call it so that the you know |
---|
0:01:52 | oh function |
---|
0:01:53 | and it |
---|
0:01:54 | a a very easy to reconstruct a shot yeah |
---|
0:01:57 | at a at a a a three that they shouldn't and in the and |
---|
0:02:01 | a and G O frequency point |
---|
0:02:05 | a |
---|
0:02:05 | we have |
---|
0:02:06 | to question is for this model |
---|
0:02:08 | oh well as to how should |
---|
0:02:10 | it this model |
---|
0:02:12 | and the seconds |
---|
0:02:13 | i also and that's if you are are also if representation is that we |
---|
0:02:19 | to here |
---|
0:02:21 | oh oh one of this question |
---|
0:02:23 | oh first the we you can have a you know that he man to match to they should be and |
---|
0:02:29 | then oh we use the true |
---|
0:02:31 | a he met to evaluate the function able |
---|
0:02:36 | there are very |
---|
0:02:37 | oh we have you |
---|
0:02:39 | a random or it can be expressed as a linear combination |
---|
0:02:43 | oh on it and then no determined that |
---|
0:02:47 | deterministic all normal functions is weighted a random variables |
---|
0:02:51 | oh the um that C functions are orthonormal with |
---|
0:02:56 | read that to and in a product |
---|
0:02:59 | in fact is |
---|
0:03:00 | we use a rate be free small approximate this is |
---|
0:03:04 | using a an it's number of of the old normal function |
---|
0:03:08 | then and O |
---|
0:03:10 | the expected value of the squared norm of error of the phoenix representation |
---|
0:03:16 | can be given by these uh by prominent a two |
---|
0:03:20 | where a is the random process |
---|
0:03:22 | and has a a it |
---|
0:03:25 | representation of the random four |
---|
0:03:28 | uh the core work if they don't be |
---|
0:03:31 | also also no expansion the or |
---|
0:03:35 | so that he man |
---|
0:03:37 | a now we focus on a team for at least for illustration |
---|
0:03:42 | but we have a a random you find on them |
---|
0:03:46 | and that interval with zero mean and and in in the G |
---|
0:03:50 | you can handle |
---|
0:03:52 | um all normal compression be in a four inches rule that |
---|
0:03:59 | um |
---|
0:04:00 | the random draw that can be right yeah |
---|
0:04:03 | um |
---|
0:04:04 | and the |
---|
0:04:06 | you combination or you know |
---|
0:04:09 | uh uh another |
---|
0:04:12 | uh |
---|
0:04:13 | or in finance |
---|
0:04:14 | uh |
---|
0:04:15 | the complete |
---|
0:04:17 | a basis |
---|
0:04:18 | and way by a very well is a random variable is given by the end of all that so at |
---|
0:04:24 | the end |
---|
0:04:25 | the complete basis |
---|
0:04:28 | and the truncation |
---|
0:04:30 | um |
---|
0:04:31 | for for five |
---|
0:04:32 | or shows the arrow for uh |
---|
0:04:36 | oh of in a representation |
---|
0:04:39 | and the fact that the right along |
---|
0:04:43 | oh oh error |
---|
0:04:45 | a we a good a part of um that thing and so we can see that |
---|
0:04:50 | that was that so will change |
---|
0:04:52 | and and it you can do |
---|
0:04:55 | oh in that that is all of the |
---|
0:04:57 | a a representation |
---|
0:04:59 | so to the meeting my the that have that |
---|
0:05:02 | oh oh we need to make my |
---|
0:05:04 | uh the second time |
---|
0:05:07 | now |
---|
0:05:07 | oh you got a a a a a second term you take |
---|
0:05:12 | that second so |
---|
0:05:14 | that second so um |
---|
0:05:15 | a a you know by |
---|
0:05:17 | a by a far |
---|
0:05:19 | as as useful we a random variable is |
---|
0:05:22 | uh a given by a |
---|
0:05:24 | you know but that so at the end i T so from the calculation |
---|
0:05:29 | uh |
---|
0:05:31 | they |
---|
0:05:34 | they fire |
---|
0:05:35 | a can be right yeah |
---|
0:05:38 | a a problem that |
---|
0:05:40 | uh where yeah |
---|
0:05:41 | see the autocorrelation function of the random process |
---|
0:05:46 | and |
---|
0:05:47 | the uh a a a at the one change |
---|
0:05:50 | so |
---|
0:05:51 | we can say that |
---|
0:05:52 | a a a a a a a mean it's on the choice of the complete a base |
---|
0:05:59 | okay |
---|
0:06:00 | but to |
---|
0:06:01 | uh uh any a complete or normal because can be used |
---|
0:06:05 | so we have a question which is a that |
---|
0:06:08 | now a a a a a just five C |
---|
0:06:11 | uh uh i i think of of the for |
---|
0:06:14 | in will in creation with for no |
---|
0:06:17 | the also correlation function of the random drug that's |
---|
0:06:20 | and then we use each |
---|
0:06:23 | um i think function |
---|
0:06:24 | uh to we spend the random for that |
---|
0:06:28 | at the formal or the and and the |
---|
0:06:30 | oh random for a random variable |
---|
0:06:33 | a a is given by you inner product of that |
---|
0:06:36 | and the and the these |
---|
0:06:38 | oh i can function |
---|
0:06:41 | now |
---|
0:06:42 | oh |
---|
0:06:43 | compute |
---|
0:06:44 | a a screen play five |
---|
0:06:47 | from the calculation we can see that |
---|
0:06:50 | oh okay "'cause" i was to |
---|
0:06:53 | uh that some model |
---|
0:06:54 | a a a a and i i is the eigenvalue |
---|
0:06:58 | and we know that |
---|
0:07:00 | a a a that was to i can function account |
---|
0:07:03 | i well as i my of error a uh variability of the random process at was able |
---|
0:07:10 | so we |
---|
0:07:11 | we say that |
---|
0:07:12 | we a a "'cause" i have three |
---|
0:07:15 | you or greater than that you buy out the |
---|
0:07:18 | a a complete basis |
---|
0:07:22 | and |
---|
0:07:23 | then all with a a a a that can be a a a a similar if then a long uh |
---|
0:07:29 | i can function |
---|
0:07:31 | but |
---|
0:07:32 | quite well and the distribution of property in the random i thought |
---|
0:07:37 | i a i don't know all too complicated to be described as think they |
---|
0:07:42 | all in close to form |
---|
0:07:44 | so at to read about it |
---|
0:07:46 | extremely you can go to find a |
---|
0:07:48 | oh okay L expression for the random problem |
---|
0:07:53 | i |
---|
0:07:54 | uh |
---|
0:07:54 | given a a a a a set of real i they shouldn't of of the random process |
---|
0:07:59 | oh |
---|
0:08:00 | and it's and i three are all normal be we can calculate |
---|
0:08:05 | uh the random error |
---|
0:08:06 | oh variable able for each realisation |
---|
0:08:09 | and then we apply that there |
---|
0:08:11 | random variable and use |
---|
0:08:13 | a a of various to the them that so |
---|
0:08:17 | the that the energy of the random or |
---|
0:08:20 | and then the ah |
---|
0:08:22 | a a a a a a hell of a a a is a random variable so are part |
---|
0:08:27 | uh |
---|
0:08:27 | screen a five in |
---|
0:08:29 | we know that it's of ones that are then |
---|
0:08:32 | a a sign which is given by a |
---|
0:08:34 | oh of to more expansion |
---|
0:08:38 | so we |
---|
0:08:39 | for a T max |
---|
0:08:40 | oh which is that um well |
---|
0:08:43 | and i think the sample variance |
---|
0:08:45 | are are among random variable |
---|
0:08:47 | oh which is |
---|
0:08:48 | involved in the model |
---|
0:08:52 | now still we use the |
---|
0:08:53 | i um i don't five three made to evaluate |
---|
0:08:57 | the function model |
---|
0:09:00 | i |
---|
0:09:00 | oh the kl expansion is unavailable |
---|
0:09:04 | oh |
---|
0:09:05 | there are two seconds to is are and a consideration |
---|
0:09:09 | oh the |
---|
0:09:10 | uh the |
---|
0:09:11 | representation is not offer more |
---|
0:09:14 | redundant you must at this to the in the mall then |
---|
0:09:18 | how do we measure the redundancy or equivalence is the H E |
---|
0:09:24 | i we know that is small so in the |
---|
0:09:27 | oh model uh imply that error of the representation is a there is a trade-off efficient the end and ri |
---|
0:09:36 | uh |
---|
0:09:38 | you can uh uh uh and six fashion |
---|
0:09:40 | oh we can calculate a very is for and coefficient |
---|
0:09:45 | then |
---|
0:09:46 | we find that |
---|
0:09:47 | only only |
---|
0:09:48 | uh |
---|
0:09:51 | or i coefficients are a very small which used |
---|
0:09:56 | the corresponding sounds to be removed them from the model |
---|
0:10:00 | that's in a large than the error |
---|
0:10:02 | and i |
---|
0:10:03 | so only uh and i a uh so that the model |
---|
0:10:08 | and uh |
---|
0:10:11 | a so uh we can you to me |
---|
0:10:14 | the efficiency of the model in terms of that so |
---|
0:10:18 | which are given by a and the it |
---|
0:10:23 | then uh |
---|
0:10:24 | that i is one is |
---|
0:10:26 | oh there are many a point is all orthonormal basis |
---|
0:10:31 | so will which one is a bass |
---|
0:10:34 | oh |
---|
0:10:36 | and uh |
---|
0:10:37 | oh we know that a human |
---|
0:10:39 | uh well i think that all of normal basis we can calculate a |
---|
0:10:43 | oh but and all coefficients |
---|
0:10:46 | and |
---|
0:10:47 | a i |
---|
0:10:48 | screen a a |
---|
0:10:49 | i friends |
---|
0:10:51 | so we compare them at all three play |
---|
0:10:55 | then the lack of that bad has so um |
---|
0:10:58 | the well known be up by thing a like to create a a a a is the best choice |
---|
0:11:07 | okay |
---|
0:11:08 | oh |
---|
0:11:09 | we use the uh |
---|
0:11:11 | proposed a map to be evaluate to each you yeah no |
---|
0:11:15 | for the efficiency in |
---|
0:11:18 | a a spatial component is there's |
---|
0:11:22 | in |
---|
0:11:23 | a a a there is an um |
---|
0:11:25 | each each they show the spatial components it |
---|
0:11:28 | a a a a a a a very low the harmonic |
---|
0:11:32 | and uh a a i i the |
---|
0:11:34 | coefficients still we hold |
---|
0:11:37 | spherical harmonic |
---|
0:11:38 | uh coefficients |
---|
0:11:39 | which is a given by a |
---|
0:11:41 | of the you know what are so it not yeah a and a very money |
---|
0:11:46 | and uh |
---|
0:11:48 | we might not be way and have to a a and is given by a |
---|
0:11:53 | uh |
---|
0:11:55 | uh right |
---|
0:11:56 | um |
---|
0:11:57 | and |
---|
0:11:58 | it is a need to me by a a a a wave number of K |
---|
0:12:04 | okay and not a |
---|
0:12:06 | oh |
---|
0:12:08 | oh of for that model we have |
---|
0:12:10 | oh have they have a |
---|
0:12:12 | capital and coefficient |
---|
0:12:15 | then oh we can do |
---|
0:12:17 | the are there the whole |
---|
0:12:19 | all these there will have money |
---|
0:12:21 | oh oh coefficient |
---|
0:12:23 | oh we find that |
---|
0:12:25 | and right oh okay efficient |
---|
0:12:28 | a a a a a nine one nine then |
---|
0:12:31 | of the total energy |
---|
0:12:33 | so uh for me looking and |
---|
0:12:36 | uh we define the H I T N as the initial efficient the a a and the whole key |
---|
0:12:44 | oh i shows |
---|
0:12:46 | um um |
---|
0:12:47 | the uh very is all all their a harmonic |
---|
0:12:51 | coefficients for the frequency and |
---|
0:12:53 | for a you know that |
---|
0:12:54 | and with C |
---|
0:12:56 | that |
---|
0:12:57 | the was the C coefficients of hand |
---|
0:13:00 | oh most of the energy |
---|
0:13:03 | and then uh |
---|
0:13:04 | think of be shows that the speaking a at |
---|
0:13:09 | okay so and and the capital and wrong |
---|
0:13:12 | calculate and in these oh to try to go and O |
---|
0:13:17 | a a and one the red |
---|
0:13:19 | a down |
---|
0:13:21 | and a problem if we should both he we can |
---|
0:13:25 | uh as you can be a the of they show of component expansion is a wrong |
---|
0:13:31 | seventeen then |
---|
0:13:33 | and then |
---|
0:13:34 | a a a a think good shows a migration |
---|
0:13:37 | a model the man |
---|
0:13:38 | a H I yeah the blue one |
---|
0:13:41 | and then |
---|
0:13:42 | a a a a a a a one in a a a a we constructed |
---|
0:13:46 | each each yeah using |
---|
0:13:48 | and and for a capital and to themselves |
---|
0:13:50 | spherical a a harmonic |
---|
0:13:53 | and that really one still |
---|
0:13:55 | uh |
---|
0:13:56 | the reconstructed hrtf |
---|
0:13:59 | oh using and trying in terms of spherical |
---|
0:14:02 | harmonic |
---|
0:14:04 | that we want to different this speaking writing one and right one is |
---|
0:14:08 | oh or wrong to one percent |
---|
0:14:12 | i |
---|
0:14:13 | oh let's say which are normal with select uh a base it is that a frequency components there is |
---|
0:14:22 | a a a a a a in it is a three D uh hrtf model |
---|
0:14:28 | we use |
---|
0:14:28 | uh |
---|
0:14:29 | spherical be L to expend the of frequency component |
---|
0:14:34 | and uh |
---|
0:14:36 | i |
---|
0:14:37 | i is that there will be a more efficient of which is given by be formula |
---|
0:14:44 | i still we have a other classes |
---|
0:14:47 | that's a |
---|
0:14:47 | a complex exponentials |
---|
0:14:50 | oh and agenda probably and T V share on a because |
---|
0:14:54 | uh are they are all you can have but it's in expanding of frequency components use or on the in |
---|
0:15:01 | term |
---|
0:15:03 | and uh |
---|
0:15:04 | collocation can be given by |
---|
0:15:06 | oh these three |
---|
0:15:08 | you cliche |
---|
0:15:09 | respectively |
---|
0:15:10 | then |
---|
0:15:11 | a here it would treat the and the |
---|
0:15:14 | a property of H T of components |
---|
0:15:17 | oh models using |
---|
0:15:18 | these four sets of orthonormal basis |
---|
0:15:21 | and |
---|
0:15:24 | but that and so on i shows the screen play |
---|
0:15:29 | and we can see that |
---|
0:15:31 | a a a a new value |
---|
0:15:32 | for for a a a a a uh |
---|
0:15:34 | according to compress C spun that actually is quite close |
---|
0:15:38 | to that there will be a function |
---|
0:15:40 | but still we model the |
---|
0:15:42 | note that |
---|
0:15:44 | uh i think he said all |
---|
0:15:46 | a a will be out in a sample |
---|
0:15:49 | and i with and still of the total energy of we're |
---|
0:15:53 | oh while uh i think in a comp simple and now shows only a have for three presents no the |
---|
0:16:01 | oh |
---|
0:16:02 | and G |
---|
0:16:03 | so |
---|
0:16:03 | with i |
---|
0:16:04 | here are |
---|
0:16:06 | spread will be found at most and is the passage voice C |
---|
0:16:10 | it's yeah frequency expansion |
---|
0:16:14 | okay |
---|
0:16:15 | uh motion |
---|
0:16:16 | uh and i right know we a a a a set the question |
---|
0:16:19 | how to we if they really the efficiency of a and |
---|
0:16:23 | oh |
---|
0:16:24 | a function of hrtf model |
---|
0:16:26 | and oh we use a |
---|
0:16:29 | we use the some of the |
---|
0:16:31 | where of random variable in in the model to uh evaluate the efficiency and the read that still |
---|
0:16:39 | efficiency of the model |
---|
0:16:41 | oh a spatial component expansion is around seventy percent and the |
---|
0:16:45 | but what is still |
---|
0:16:47 | uh a frequency component |
---|
0:16:49 | pension is the spherical bessel don't function |
---|
0:16:53 | thanks for attention |
---|
0:16:55 | you |
---|
0:17:01 | i |
---|
0:17:02 | i |
---|
0:17:17 | i |
---|
0:17:19 | okay |
---|
0:17:23 | i |
---|
0:17:24 | and |
---|
0:17:30 | all |
---|
0:17:32 | i |
---|
0:17:34 | i |
---|
0:17:34 | the that |
---|
0:17:35 | do that but um |
---|
0:17:37 | i think it it can be |
---|
0:17:39 | uh |
---|
0:17:40 | for |
---|
0:17:41 | the D C U and was to for all the prosody |
---|
0:17:45 | oh |
---|
0:17:46 | a a representation |
---|
0:17:49 | mm |
---|
0:17:50 | so i think it's can |
---|
0:17:52 | four |
---|
0:17:54 | hmmm |
---|
0:17:55 | i |
---|
0:17:56 | yeah |
---|
0:17:57 | hmmm hmmm |
---|
0:18:00 | hmmm |
---|
0:18:03 | yeah |
---|
0:18:06 | i |
---|
0:18:07 | i |
---|