Perception of Smiles in the Voice – Pablo Arias, IRCAM – Voice Tech Podcast ep.007

Pablo Arias

Episode description

Pablo Arias is a final-year PhD student in perception and cognitive science at the audio research lab, IRCAM, in Paris. We discuss Pablo’s work on how people perceive smiling and non-smiling voices.

First Pablo explains what cognitive science, neuroscience and perception are, and why research into these areas is so important. He then takes us through the aims, methods, and results of his latest research paper into smiling in the voice, and we discuss the academic and technological implications of his work.

Pablo shares his advice for budding PhD researchers, and the importance of having a mentor. Then we ‘go deep’ on how to learn how to listen, and outline the big unanswered questions in the field of perception right now.

Links from the show

Episode transcript

Click to expand

Powered by Google Cloud Speech-to-Text

because it’s actually just sensors inside people’s brain

welcome back to the voice Tech podcast the show brings you the latest research in developments in the field of voice technology to a series of entertaining inspirational and informative conversations with voice technology experts by listening to this podcast you’ll be introduced to new products and Concepts get new ideas for your invoice project and learn about the tools and techniques that will turn your ideas into reality

today would be talking about the perception of Smiles in the voice which has implications for how we communicate how we listen and how we design conversation interfaces

one of the main aims of this podcast is to stimulate discussion around voice technology topics and something that I’ve been meaning to do for a while now is craft place where I listen is can gather and interact so to my end I set up voice Tech subreddit for him on Reddit why Reddit wow Festival it’s much more open to the public and much easier to join then say yes at group the website’s been recently redesigned and they have a fantastic mobile app as well but bike forums on for the chance that I invited in it but really the main reason for me is that the social norms on red at The Forum structure in the reputation system encourage reason debate as opposed to light the flame was he got on Twitter and it’ll save us a place where people engage that functions as opposed to just straightforward business networking like it shows on LinkedIn from Sims

and most importantly volitional Facebook so I invite you to head over to Read it And subscribe to the voice Tech subreddit that you can post links about voice Tech news and put it launches ask any questions that you have technical loan on tactical coming on the post already there and get involved in the discussion and somehow I hope will grow what kind of care of a time until available resource for listeners on the wall by the server that can be accessed on Reddit why does searching for voice Tech I’m over by the easy to remember to redirect voicetechpodcast.com Reddit so check it out

okay so now it’s my pleasure to share with you a wonderful conversation I had in person with Pablo Arias who is about to finish his PhD in perception and cognitive science at the audio research lab are come here interest paper published in a very influential Journal he has made some important discoveries around house smiling changes the voice and how people perceive smiling and now I’m smiling in the voice in this episode we cover what is cognitive science neuroscience and deception and why Research into these areas is so important

the many and varied activities of the research is in the cream team I have come up with Pablo’s a member and then probably takes us through the aims methods and results his latest research paper into smiling in the boys and we discuss the economic and technological implications of his what

Pablo says his advice to a budding PhD researches and the importance of having a mental and then we go deep on how to learn how to listen and outline some of the big on Elsa questions in the field of perception right now

this is my first face-to-face interview how did the professional recording studio ever come and I know appreciate the difference it makes to have a conversation face-to-face with someone make for a much more engaging experience for both parties and it really results in a productive and enjoyable Joy Villa interview experience so I’ll definitely more of these for sure so I bring you Pablo Arias

okay so I’m here with Pablo Arias Pablo is currently completing a PhD in conception on cognitive science SF time which is the institute for research and coordination in Acoustics in music hair in Paris France Euro you’ve been working hard on your page day it’s it’s almost done and you’re about to release so have just released a paper in a very influential journal and that’s causing a little buzz I’m so thanks so much for coming out to the show to tell us a bit about that and congratulations on the paper to talk to you about all the research that is going on here at your command on the team and the invoice in general in the

deception team that rhyme was there so I’m working out here come here in Paris and the bad parts of the perception and sound design team okay but I have a little project with a little project inside the chamber is called cream stands for cracking the emotional code of Music supervisors off of you who said that and working on that emotional aspects of voice and music to how does voice communicates and the use of emotions and how does music also do it how music that influences the emotions and how we can like how does it communicates emotion induce emotions throats music as well it’s a speech but we have been working a lot of the reasons your songs on voice speaking voice singing voice

okay and Seal Team focuses more on the effect of music and voice on the human body where interesting interested on the acoustic part of emotion communication how does the voice communicates emotion what are the acoustic features that the voice uses to communicate emotions but also how hard this specific pictures perceived and integrated by the listeners so how are they included in the brain how are they processed right and ultimately what are the physiological consequences of this acoustic features for instance right yeah it’s time to go team is is quite diverse you have a mix of neuroscientists and Signal processing experts

because voice is such a complex thing we want to approach it from different perspectives so we are approaching it’s from the physical description part so very sound analysis perspective Sound Processing but also because we are setting the perception we also are interested in. The cognitive science and Neuroscience behind the desert this Pacific Physical signals right now making the sounds of when Harry met Lloyd sounds yep yep so we have a team that is kind of composer of the neuro scientist and the and Signal processing engineers and we try to work together to understand this some of this mechanisms okay

I was a bit about your background while you’re from and how you ended up in France and yeah I’m in Columbia and I live there for a while like for 18 years and then move to France to do my studies and I always because I wanted to work and sound at work and The Sound of Music

and the swiping key from 10 years and I’ve been working on a lot of projects involved with Sound and Music I can dance to. He’s originally and yeah I know you’ve always been involved in music your your Music musician yourself you do music production sister how big is a passion as well as a as a career it’s all three and the two I I came to France about 10 years ago now it’s actually going to be 10 years on the 28th of August so it’s been ten long years but I’ve I’ve learned a lot here and the meaning cut the thing is that I’ve tried to be in contact with as much as sound as I could get in contact with so I have been working on developing a plug-ins for a processing sounds with the company’s processing the music

instruments now we’re doing some research really on the perception I’m psycho acoustic part of Sound Processing trying to do everything that you can do with sound that was very difficult to do it back in Columbia right now because it’s not that developed yet in terms of of research actually grind grind should mention that are, said research lab but they also offer Masters Masters courses so instead of going to a unit tradition unless you can do your master’s degree then you got the role of Tiana which means like acoustic signal processing and the computer science apply to music with somebody New York City

and the basic the learn a lot about the science of Music the science of sound but there’s also another composition of class the last one but in that in that class it’s it’s only composers that come in threes composers that begin to learn about electronic real-time electronic composition and the Samsung disease for applied to contemporary music creation okay great end and then we know that you’ve been doing okay stay what you’re going to tell us a little bit about in a moment you mentioned that you you’re so involved in music production and record

record label is that right yes that’s right I mean we are starting with my record label is called black Macau black Macau black Macau from Colombia Macau and we are doing a lot of how do visual projects or do visual Productions producing some techno mainly techno and electronic music but with a lot of visual content is coming on on September or October around the more we look out for that one black Macau records electronic music mainly techno music

like I said before we dive into the details of the paper can you disposable explain what is cognitive science neuroscience and and the study of perception for those who don’t know yeah absolutely in the last episode tomorrow towards the technology of the speech of how to process speech we are approaching and we are not the only ones but I mean what the old the domain is approaching speech is a more like

how does the brain process speech that’s kind of the very very broad very broad question is the study of commission and your science is the study of how technician is happening in the brain for commissioner how do you Define cognition cognition is starting to the thought process to its operation basic of mental operations happen in the brain condition is very very broad right you can perceive for instance a color but you can also like think of a memory write all this is happening in the brain and all these things are cognition is the study of all these things right so it’s very very broad very very Broad and it’s just the study of mental activity mental representations right and then you can go from

for instance or ability to do maths right to our ability to understand the speech to a several several things I knew resigns specifically is the study of

it’s kind of I would even say that Neuroscience is a subfield of cognitive science in a way in which which is very very tune to what is happening inside the brain so what is the encoding of things in the brain for instance

you know this if they fit the face area in which we know if a phase of its a model of the brain which which is known to be responsible for the perception of faces for a face identity modules in the brain that allows to access to mental representations to be two-faced identity bee-eater voice search identity recognition even beat colors or beads or be eaten

phonemes recognition right right down to the low level what’s cognitive science was asked what is consciousness princess and they will try to ask this with a very philosophical Concepts and then your science will try to find correlations or a brain areas that are responsible for this specific thoughts that’s a good distinction and Sciences is ready the study of what happens I’m near a science is the study of how I actually happens in the brain neurons firing and you know subdividing the brain to read

it’s so that’s again very low level and if you can be princess in the sound in The Sound Room and we can link it to Cyber Acoustics right which is really the study of let’s say physical intensities let me give you an example for instance the colors you perceive colors right you perceive a yellow door you perceive the black screen right. That’s physical intensities what I mean is that it’s for color specifically it’s wavelengths that are arriving to your eyes right and we’re going to perceive as being colors right that’s perception receiving physical physical

differences service I said signing on a meaning is that right now media to the the physical stimuli that where exactly is this like sound you we perceive sound intensity which people called Columbus right now with all these dimensions and why they are so interesting to study is that when you actually begin to look at how the actual physical intensities relates to the percents there is no linear relationship between the two for instance if you hear a very loud sound at very base frequencies know if you hear

if you hear a sound that that have exactly the same physical energy in them right but that are in different frequencies for every base 1 and a very high one the one that is going to be in the midst like around 3 to the Hertz is going to sound much more stronger to you okay so it’s not just about the loudness who sang yeah it’s not just about the physical intensity but it’s all a perceptive the physical intensity but we are our ears are tuned to be more precise are certain frequencies so we are going to hear sound that are around the princess pinky toe hurts 3 kilohertz much much more how much more stronger than sounds that are around the 200 hurts because they do ear is tuned to hear some specific sounds

Evolution I guess maybe one more chance to have a baby crying than that the waves crashing in the city or and it’s yes that’s our auditory system has equal to hear the system that it needs to hear to two interesting account so the study of perception is inside looking at what humans perceive and near science will help us understand why they why they could save these things released Howell these things of deceived in in the brain yep okay I’m starting to go to pictures I know it’s a complicated set of a set of subjects and let you say that there will fibroid in the runway as well say to underline at least a basic concepts Great Depression is so why is it important to study all these things and say why if we I guess for Neuroscience apps that identify the cause of disease

I’m stuck exception I mean what why is it important to understand how sound is perceived what does not allow us to do so why should we study perception we can see it from 2 from two perspectives

the first perspective would be a very fundamental research question. The second perspective would be of course more applique applicative application applied applied Technologies or suffer from the fundamental perspective is very important to understand perception because it will somehow or how we work actually do you will see when I talk a bit about a work that there are a lot of things going on that we don’t actually know that we don’t actually know that they are going on so I’m conscience and conscious aspects that we are not aware that that happened when we process sounds that we were supposed to say for instance speech right

I’m thinking of an example that on a recent breakthrough on

oh I think it was last year by Miss gahani actually found how the full names are included in the brain while I’m okay structures that are represented in the auditory cortex rice using neuroscience and using some kind of brain scanner they’ve actually look to the brain while having phone in holiday dinner juice scanners they use what do we call a cogwheel is actually just sensors inside people’s brain implant Sciences its people with epilepsy that have already the implants are these people hearing voices for a lot of time and then they they saw how this voice is how this phone is with map to the auditory cortex right and when you begin to understand those things right

you begin to understand what time the other line mechanisms of voice perception right and that can have a lot of medical applications in the future if we really begin to to to know how and what is the sound doing inside the brain so people that had a Prius of people that have strokes you can have a president permits afterwards of the apartments if they don’t they don’t process they don’t process some chores in the voice I mean they can’t interpret to understand all these mechanisms little by little step by step we will be able to create new new medical

procedures and the best understand the what is happening in the brain in which areas to know how how how we actually work and develop new sorts of medicine Sora interventions I would say more leads to also says that inventions installations that we can’t you can dream up

so that brings us nicely on to your way I’ll be a little bit about your page do you know what you what you been working over that so the last 3 years is yes of course as it was saying earlier we are setting how emotions they have a speech conveys emotions speech communicate emotions and how this emotions are perceived and the physiological it’s called the physiological reactions that Des that the speech can cause in listeners. That’s how what’s our team has been focusing about in for the last two they say 3

a lot of experience in which represents people with her on what they run voices or other’s voices and the Sea of different reactions to a lot of people I know don’t like hearing their invoice and we also work for instance with up with a manual phone so who was working on dominance list social attitudes on speech he developed a method to actually measure the social attitudes and speech he was working with dominance and trustworthiness so what does it mean to adopt what what is a dummy invoice right when I suppose worse than S Voice model eat right and the is it used in the same way in different cultures when I guess eventually can we synthesize in what can we meet today around voices to have to be more.

yes that’s right yes and there is this late a pioneering work work by noon for the team work by my supervisors don’t really know who was working on the emotional voice processing just took things very quickly he did so he created digital outdoor algorithms that comes from the voice in different emotional directions right direction such and such as happiness Sadness the secret of the tool that can produce happy sounds and voices except okay so what he did is that can transform people’s voices and people hurt their own voices sounding with different emotional colors but this transformation not very subtle right

how to read a text for 10 minutes and what she showed was that actually people would congruently move their emotional state to wear the effect that he was using so people hiring some people reading out loud but having their own voice played back into that rare in real-time modified slightly to be happy or sad but then the gum happy hour outside of the result yeah that’s just me yes that’s the main results yet okay well it sounds like a fascinating to watch and every one of these different area of this what is it that you can see what’s actually to study the Acoustics of Smiles in speech Smiles in space that is very interesting about smiles

I actually like individual do many they have been studied for centuries right since Darwin they’ve been started on living actually shown to be recognized end of everywhere in the world and the independently of of cultural trait Brian smart as universal ever been in contact with the Western cultures they found that they would also produces gesture and

categorize it to be appreciative and positive and positive gesture so what we asked was okay the thing is that smiling can be heard in the voice as well and we can communicate the same emotional facial expression which is very important throughout the voice. They always play outside you should smile on the phone trying to sell them something that subject

trying to find the

it’s a the weather weather there was an acoustic signature they say like some sort of acoustic fingerprint of that sort of voice right and weather. Acoustic fingerpicking hide consequences on the listeners right that’s kind of the main broad broad subject we were setting up the experiments about how to do I tackle the public storing this a lot of things for when we did was the production of specs so we fight we just ask people to come and record the different sorts of names with meaning of Widow smiling tigers

right on time right now can I come see you or you’re smiling in in what you producing an SSI hissing sound while you’re smiling and you can shoot the Resonance of speech going up from 3 to 3000 first to 7000 Hertz depending on how much I’m smiling and actually this is my name is produced by a muscle muscle muscle how much you contract your booty muscles and how much will change the sound of your voice and the summer

to which the cell production experiment just work by recording on starting the production aspects of smiles but also we did it starting the perception parts so really just asking people to a different sounds generates algorithmically which is called the rivers correlation actually measure the mental representations of what the what the participant use to

to do a certain task to let me be a little more specific what we did was to

how to generate thousands of sounds right some sounds you mean not even speaking a kind of filters and then what we do is ask participants to write whether the sound is smiling or not and this way when they do hundreds and hundreds and hundreds of Trials we can look at all the randomly randomly generated filters and the juice with inverse correlation what are the aquatic creatures that are changing the sound that are making participants crazy as they as if they are or not smiling you rather than just ask people to come in and say things for the unit with a smile not what you need right unscientific you actually went to the trouble of creating a tool Festival. Could inject various levels

smiling on not smiling into a into a speaker phone in part of a part of it part of speech and then use that to induce a response in a listener asking them whether they they could Harris model not and then because you notified the the phone names with a filter that you know you are controlling you know exactly what modifications to the sound caused people to Harris montellano not have his mama’s just a i n a conducting a survey on the street and actually performing real science is he had to the trouble I mean how long did not tell you to that paper we publish recently owners as that which is to journalist jacket cigar Society of America and we end up to

that’s a reverse correlation relay experiment took about 1 year what year the whole thing was working on this on this message during his cage so he got this he had really a good understanding of how to do it properly and quickly from his experience of course that’s okay so yeah I know the end you had a good understanding of parts of a speech on how to modify speech to allow people to have a smile in the boys yet so that’s when where I think we are a little bit closer to what the voice Tech podcast is right because we created actually an algorithm

to transform the voice in real time and to make it sound either not smiling or smile yeah you use the the findings from the study talk secret to a conduction modify in your time so I can speak and you can just have a smile to my voice so they can show you some examples if you want let’s do it then so what have you got for us so maybe I’ll play some French sounds but also some English sounds like to me that you are going to hear two versions of the sounds first the answer my my probation and then this might manipulation but the thing is that because we are changing only the smiles on this sounds you will hear that it’s exactly the same contents and it’s very important for us because because because we need to have control of all the emotional

dimensions of speech right do we need to really really control for all other aspects like pitch Contour International content the speech rates for everything with a fight with us is better to keep everything the same except for this morning but except that we found in specific to smile you have the norm smiling manipulation in this money manipulation does that mean does that mean both the files have been modified in some way and the thing is that we sort of linear transformation I would say I mean I can either go up or from take something that is up and take it down right as you have the ability to remove smiling from The Voice as much as you have a smile that hasn’t been smiling

saying that says something about not smiling thing you can make that voice even less Smiley right to not putting my lips pursed like we have just the model of this particular articulation and how it affects speech okay right now

love you my passcode Pro

does microscope

we push could flow

it’s a pretty chilly day it’s really chilly day

what would you like to do today what would you like to do today

justice justice

is contralto is twitching muscle

okay yeah I can really have the difference between the sound father’s have to smile at is in the sound father’s have the small removed interesting okay so that’s the effect of the tool that you’ve applied to the sound files yeah so that’s another reason that is actually I can be used on any voice and it’s an Adaptive algorithm we will adapt to the incoming voice content and the incoming voice identity to try to match the acoustic manipulation consequences Albert and develops in house 8 months now available is it is it going to be available at 2

okay and we kind of put a link on the description down below to the taper also waited we explain all the algorithm and everything that we did to create it. And yeah why did you try this algorithm then what was it in the Pacific that I could see features and see what is the consequence on the listeners right so what we did was to try to understand what is the processing of this cues manipulate that has you heard as we are very subtle but are very controlled right and what we did was an election myography experiments I like from my ography a way to measure muscle activity sensors that you put on on people’s faces for instance if you’re doing facial lift my Agra fee to measure the official react

when they are hearing you snore when they are seeing a similar picture so you can tell her that smiling or Not by 6:35 participants to come in rice and we put some medical myography sensors on their face measuring zygomatic muscles which is the one used to smile and their corrugator muscles which is the one used to frown on the phone on the forehead and actually just asked participants to hear different sounds of the soundtrack to me that you hurt the French ones and just to rate the emotion of contents and the save-the-dates smiley face of the sounds okay okay and what we wanted to know is what it’s whether the facial activity that people will have during the listening of the sounds

will in some way reflect the acoustic manipulation that we were doing to this sounds okay I’m so you want just taking the the participants word for at weather like it hair or small or not you actually wanted to look at that the muscle activity in the face yep some evidence with visual faces so people have been studying this but but when exposing people two faces emotional faces smiling but different emotional facial expressions we tend to vary implicitly and very microscopic Kelly imitate the facial emotion that we are saying that’s that’s like a theoretical it. Very important but I won’t go into those details true but okay

we are proceeding also this articulation in speech

does this perception also triggered the same mechanism in participants in people sound like some sort of contagion or imitation imitation yeah okay so when we recorded the oldest muscle activity and then we looked at the muscle activity during the listening of the sounds right and we we can do is that we can look at the muscle activity for all the sound that with manipulated with the on Smite affection and all the sounds that were manipulated with a smile effect right and what we find that was that they were significantly different muscle activity between those conditions in the voice really affected the muscle activity of this month

does did smile more when they hurt to smile effect and they did not take me to decreased when they heard the announcement effect like you feel like you feel they were really hearing this articulation of processing it and then going up to reflecting it in their own face

but what is cool is that because we had all these control of the equity Futures we can be sure that the difference between the conditions are exactly what we want we measured right you know exactly how to modify the sound in your to make people physically smile back or not and we are it’s not for instance that people just were telling a joke you know yeah they were listen to something funny or between the conditions there is absolutely

there is only one that mentioned that changes in the conditions in this different dimension that we change

content speech rate everything is comes to everything you could be sure the aspects of Timber in speech is triggering this physiological reactions in participants this is brand-new what is to conduct original Research into discover things I’ve never been discovered the fall and you primarily of the best padded that so you must feel pretty good about it was very

bird encouraging and the you know what you try to win when you work in science will try to make it work and 90% of the time you fail and then there is the distemper sent when you actually find something and that that was one very nice 10%. We were all very happy in the gym because theoretical is very important you know we find evidence that we process the articulation in the voice with in a similar way that how we process articulation visually from Little from the mechanisms implicated new release from the from what is happening in the brain that say it is it’s like it gives us a clue on how how gestures are perceived right absolutely very interesting to the nearest haunted so now I’m sorry explore the areas of the brain responsible for both the divisional round the old oak tree Passage

maybe these are shared that which is something that wasn’t no because I’m sure because we have the same reactions to emotional faces on the motion of the voices right to articulatory cues in voices and in faces we we we could say okay that’s strange that it’s the same reactions between modalities maybe it’s the same brand area that is responsible for these reactions or at least similar mechanisms website yeah I know so technological ones that I can see I mean applications of this discovery news about the tool that can add smiling to a voice just as you know I’m in the process of buying explain this phenomenon in which I can see this on this island could be used to to modify people’s voices in day-to-day lives that could be used to modify the voice of these I voice assistance

devices that are becoming All the Rage now yes I do like voice synthesis companies and you know who is Google and Amazon and all these people beginning to do a voice synthesis at some point we will need some emotion into this voices and will need them to be highly intelligible highly accurate very precise but also very human and to give them we will need to give them that we will need to give them emotions we will give to give them too many aspects of voice for them to be here for them to be a real estate Camp Verde to convey messages in the best possible way right now. Could be used in several situations as you say yes of course it’s pretty exciting from a human augmentation

7 Osage from a conversation in Facebook Nights in White Satin so I did not connection elements of voice is an update so boring to me in 10 years to those voices will be amazing you know what the extended then I mean it sounds like there’s so many directions to this good guy to have ideas of you know how you could build on this work do you have plans to build on that or the other members of your team that we built on that okay starting the processing of audio visuals files so we really work on the processing of auditory smarts why they call auditory smiles and now we are working on the audio-visual Smiles so how does how do Visual and auditory Smiles interact to create a shared mental representation of shared Visalia

yeah a shipment of representation okay so how do you steal dettori part influence the visual part of the divisional party plus the auditory part which one is most important maybe are the weights how do they add together or several questions never done it like you really enjoy it

okay listen to this listen up while you’re enjoying the show out of it to write it and subscribe to the new voice text subreddit they will find other listings of the show not posting links about voice Tech news and product launches asking questions and commenting on the post December that can be easily reached at voicetechpodcast.com Reddit it go check it out

yeah I wanted to ask you I know there’s going to be a lot of people listening you think while I’m really inspired by that so I would love to get involved in a project like that maybe one day and maybe I’ll even talking discover something you know I’m breaking is that what you have to get can you describe the life of a PhD student Harry Potter and so I know just in general and we know what your plans must be the most fun part and you know how has changed you ever heard of the three heads

yeah it’s been really an amazing journey an amazing journey and the tell if there’s one thing that I really enjoyed learning that I I said I didn’t have to I was going to learn was learn how to fail a grade for that actually I find that to be the most gratifying thing now

and because when you are doing research you will you will you will it is going to be hard because you will come eat a lot of mistakes will do a lot of mistakes a lot of Errors day after day because because you are a new territories territories difficult to explore when you say commit mistakes I’ve made mistakes do you mean the youth run experiments and they failed to prove what you were hoping to prove button in a sense that they’ve they’ve prove something or do you mean this takes is just made the experiment completing all them boys out there are so many types of failure all of them all of these failures you learn

and that’s it that’s that’s what I learned and that’s what I like about it is that you can do an experiment for instance and you just did it wrong because you can have a lot of technical issues right now for instance you didn’t measure the good muscle on the face like your other child was a beats up two up front so your signals are too noisy right now or you can have noise coming from the electricity that is blowing all your signals to the Samsung or the recording for the most from the muscles you have to throw throw them away but each time you fail you when you find why you failed you are improving a little bit you’re improving a little bit so you’re doing a little step forward and the next time you do it now you’re controlling for I don’t know for the placement of the electrodes for the technical issues in the sound for the

etcetera and after doing this a lot of times you arrive to something that actually is is is working right like really working but of course there’s the old you figure out you just said which is also very interesting one

Indy’s the unpredicted outcomes but sometimes you do your experiment which is very well controls and you actually don’t find what you were expecting to find right so you have also to two

requesting yourself on your cervical knowledge okay Riggs I like this is there a bias somewhere right okay really a lot of types I like to call it favors but it’s just a way to go forward and just going to go for a good advice run for all of us ready in whatever Endeavor you are involved in yeah that’s a good attitude to have towards the tree every Friday I no matter how it comes about as a learning experience for me. That’s what I learned on the phc what’s what’s that what’s that

do just fail day after day but after awhile you will stop fading just and you will learn so much by Saving it’s it’s it’s it’s amazing how much we can learn by doing things even if you fail you learn a lot of things a lot of things that’s really interesting thinking about doing a PhD then what would your advice to them day and then from your experience if you have to do it again like what would you do differently

so I think my advice for a PC would be start a PG if you are passionate about it and while doing it try very hard and be very persistent

and the it will be hard to be persistent but you keep on trying to keep going to keep fighting keep learning from your face even if your face that’s okay if you are learning from your failures if you are failing and not learning that’s really hard 3 hard and that’s because you may not be doing things right for the support I can come inside you talk to people in those cases marvelously okay interesting feeling because you are not doing great job great job that’s from your experience that if you have to do it again you have to do another page to see what would you do differently this time I think I would take more holidays and then I mean is really seriously because you tend to

because you put your pressure on yourself yeah and you tend to because you like what you’re doing you tend to put a lot of pressure to the extent that you don’t to rest

and would you don’t rest you your mind isn’t as clear as you don’t have as much ideas just weren’t sure what we need is to have ideas that what we are working on that’s our work is to have ideas developed ideas and if you are too tired to develop ideas on to have new ideas you are not good so I if I had to do something different I would maybe take a little bit more holidays at some point to overload you with things I’m just not thinking straight and just not having enough to hook you to return the distance from your work also you can keep driving forward the expensive expensive take a moment and when you when you come back

everything will be so much clear the ideas will be closed and we will be into it and probably just working on that okay alright seven needs to sleep to process information define memories that’s been proven yet you have to respect the physical in the makeup of yourself with yourself to do good work you’re not the first person to say that in the park on sexually surprisingly the same thing if I wait until she’s been out and everything will fall apart

I’m more people on projects of inspired you and how to you on your journey and in some way over see the people in your team do famous celebrities in annual field that you admire all you have a do you have mentals

yeah you might feel. I mean they’re not people doing an amazing job on this Fields be them from an evolutionary perspective from really the encoding neuroscience perspective right there really a lot of things I would say for a if you want to there’s a very good book for people that are going to be there with neuroscience and interesting music which is called this is your brain on music this is your brain this is an amazing book to read when you are not familiar with your science that’s right that’s very very nice this morning to science and biology and evolution and a little bit more into Data that are doing research it’s it’s more for people doing research you can read the language music and the Brain by Anna Patel

which is an absolutely amazing the guy doing absolutely amazing research account under share the processing shirt aspects between language and music and how it is to actually relate and Sons one of these helps you and guide you through the the pet stain for treated me a lot and help me to do big braids with work and help me to learn a lot of things but I’ve also learned a lot from guy who runs afflicts company is called Flex Plus Florence and they do I’ll do processing plugins price on the he really taught me how to

how do you mean how to listen he has a very precise years very precise set of yours like because he’s doing good I’ll do processing tools very high-end audio processing plug-ins for music production and when I was working with him I really do see him listen and listen to details in Music and Sound transients Dynamics Timber frequencies that’s that was something amazing yet to two. That sounds really fast it any shows it is a skill that can be learned that we really appreciate that visual visual medium much greater detail than we do we do the acoustic I mean if other people want to learn how to listen

what would you recommend is that some you have to be shown or I mean there are so many ways of listening to me there is Princess the guy in the way of listening it’s from a production is from a production perspective music music production perspective so you would really hear the bass you will hear the defies the frequencies the transience the Dynamics and how they all translate together and how they create the whole mix and you will learn to hear space right so that the music production aspects of it okay but now there’s other aspects which don’t need that position in the ears which is hearing how you would see your friend but yeah how to cure all the q’s that the voice conveys every time listening skills

listening skills that somebody is not in a good mood when somebody is not having a good time or whether somebody is doubting us or something right you can infer so many things from people’s voices and that’s another critical listening that is very different from the music production listening another school that’s only lacking in the time and money it’s really different skills but they both are listening skills right now and then there’s so many things like that right you can also learn to hear an instrument piece right at a player right to hear all the details on the playing and older they got to sand off the vibrate to sound of the little inflections on the plane to listen to classical music

I mean if you want to learn to produce music by mean if you if you want to produce music do you have to have a good set of ears and to have a good set of ears you have to listen listen and listen and listen to a lot of details right have a good sound setup and the just

work Harden on

I’m getting to the best sound quality possible– out of your setup stuff right okay and the other the other the other part the more interactive first I would say just be conscious about it constantly say so many things that we don’t hear right people get so many clues that we are not catching every consciously try to hear those things you hear them directly we’re very good at it and actually what new science and computer science has shown you that the processes underlying the cognition and the processing of speech is extremely develops extremely developed this capacity of language that we haven’t communication and social interaction I really mind-blowing and are compared to

animals are really one-of-a-kind interesting

it helps that it helps to know that we we already have all the all the equipment in our heads that can help us get in the house to have these things I guess what you’ll mentors has shown you is that the specific things in the sound to listen to what the pay attention to since we know what to listen to whether it’s the q’s in the voice of your friend whether it’s the the Timbre or the the Dynamics in music production then you chew in your ad you tune your attention to those things and it becomes apparent so yeah this is borrowing yeah okay

well the the big questions that in the field of deception you know what are people working on right now while the big the big guy topics that still still keeping the researcher is busy so for instance in cognitive science there’s a very hot topic today is and is trying to understand Consciousness show and that’s I think that’s very very interesting and like people are really beginning to grasp and Beautiful Beginning to have Clues on what it is that makes us think that you are conscious means a lot of research in the area area and the it’s really moving very fast so I think that’s very hot right now

cognitive incoming Sciences the Consciousness Consciousness Consciousness research yeah okay and also there’s a quite Hot Topic right now in cognitive science is social interactions the study of social interactions is the study of how we communicate how we Bond as a group at right all the aspects of group dynamics notice at one21 social interaction base in primates you can study social interactions in Inhumans rice and veggies that we we as humans have the capacity we are we are we are social specie we need the contact with others

we need to do this emotional bonding with neither to speak to others we need to contact with her mother’s during development with families we almost always have gathered in in in groups right and that’s that’s something that is beginning to beginning but that is being studied a lot right now and theoretical. It’s very very interesting and how does that relate to perception

everything relates to everything when you speak about mental abilities right side right yeah yeah it was social interaction Builds on perception so let me take an example for instance a music people people who have a museum which is the inability to recognize pitchers pitch pitch pitch in the voice in general in general height of pitch perception here the pictures the thing is that in social interactions we use a lot pitch count or security questions several things are pickled Peach Contours write this in two Nations so would you don’t have this perception you will have actually impaired social interaction spice

so actually I would say that is related to perception because we built on perception to communicate because if we can if we can’t communicate in the same channels as other communicate it’s very difficult right people who suffer from these apartments are being studied to two hours and for myself more January 4th at the committed prices of yes you can know whether people with them you can recognize emotions in the voice so yeah yeah I’m so we use these yeah test cases and you can know whether okay they have less accuracy at recognizing voices how does it affect

different different contexts in their lives or find people who don’t have that one aspect of perception we would we are doing in with the going back to the smile example that gave to the people imitated the spinosaurus sound right yes what we were doing we are doing is an experiment to get blind participants congenital blind participant that have never

seen a smile okay we are trying to see whether their inventory unconscious mechanisms of their face is actually preserved even if they have never seen a natural smile yeah that’s right because you would because you might seem that people are smiling because I copying other people who have smiled at that would be absolutely also developed just threw out the Sonic Express experience yasmani could just be something that’s a night to my physiology that’s that’s that would that would be kind of a computer but with some of discussions but that’s okay excellent let’s move on then I’m so what’s on the horizon for Pablo Arias what you mean working on for the next 6 to 12 months

to finish the PSG I have no defense at the end of the year and the papers related to the project with a smile Smiles some eye tracking stuff and blind experiment Sandra and I will try to write that but I think I’ll be ready after that I will have some rest yeah I’m just take to take a take a few months of break I actually haven’t started a lot of what I’m going to do after work that is better take your advice from before that that advice site I don’t often make it to work a lot of on the record label also

I put some ear piece out there and do some shows and that right there will be a lot of things to do

next to you I’m sure there is plenty around the corner Alright good stuff I can people find more about you and I know what you can go to the cream project website which is cream that are Camp. FR okay otherwise if you are interested in knowing more about the institution your cam you can go to aircam. FR at music show The Following on Twitter if you want to hear more on the record label go to Black Macau on Facebook or just send me an email to areas at your Camp. Fr

social yeah I see the cream team posting stuff all right. It’s been wonderful so it should have been such an interesting discussion in studio recording thanks to Pablo’s activities in Sound Engineering is help me set it all up in the in the studio I’m so yeah it’s nothing like face-to-face definitely have to do more of that but he was a struggling with Scott kozin and all that is just not worth it sometimes today

you just had from Pablo Arias a final-year PhD student at first smiling in the voice is really important for a number of reasons best it resulted in the production of software tools that can add or subtract a smile from The Voice which will later be made open source for the whole Community to enjoy

then it use these tools to produce clear evidence for a relationship between Herring Smiles in the voice and producing a physical smiling response even want to listen to self reports that no smart as hard

Publix advice for budding PhD researchers Western show that you’re passionate about the subject before you start so you’ll stick with it and what however the three is it was a stress the importance of having a mentor and taking care of yourself by taking a step back from what once in awhile that’s something else we had on the podcast before

and I’m surprised in the conversation was when he described how you been told to listen properly by other soundex but not here, it’s a skill that we’re any of knowledge exists that I’m considering potent and yeah and world of conversation interfaces would like to become much more important both the design off on the use of these devices as well as fell in the past nor interactions of course

that’s over today I hope you enjoyed last night as always you can find the show nights with links to the results is mentioned in the episode voicetechpodcast.com and also be sure to check out the cream Team website as that is loads of fun. Amazon map from other team projects you can follow me on Twitter @voicetechcarl and sign up for the monthly newsletter voicetechpodcast.com so I can use Alexa

to support the show just tell one friend or colleague about this episode and also don’t forget to check out a new subreddit voicetechpodcast.com Reddit

I’ll be back soon with another episode but until then I’ll give you a heist call Robinson thank you for listening to The Voice Tech podcast

Subscribe to get future episodes

How to support the Voice Tech Podcast

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.