Mister Opera vs Pseudoscience (featuring SINGWISE): Harmonics and Formants (re)explained
HTML-код
- Опубликовано: 5 окт 2024
- Some things are just are hard to understand. Other things are hard to understand AND surrounded by gatekeepers feeding you pseudoscience. Next up: Dr. Ken Bozeman, who owes us all an explanation.
I'll update this section later with better thoughts. Karyn, I hope you reconsider your position soon. We have work to do.
Update: it’s a week later now and this video was a great idea. I’ve met really cool people as a result. I did hear from Dr. Bozeman and he largely agreed with what I said here.
Do some respectable people believe in the separate container model of formant behavior? Yes. Is it pseudo-science? Yes I still believe it is, and that the basic reason for it is to avoid mentioning the constriction needed to make clear, ringing, and loud sounds. I think instrument makers have understood the acoustics of making loud and clear sounds with tubes for a very long time, and I think perturbation theory predicts how those things function, and that it’s only in voice that we wind up with these ideas that clearly don’t work.
videos referenced:
Perturbation Theory
• CSD3084 Perturbation T...
@Stefan Frisch
The Phenomenon of the Missing Fundamental
• The Phenomenon of the ...
@The Royal Institution
What are FORMANTS and HARMONICS? VOCAL FORMANTS AND HARMONICS Explained!
• What are FORMANTS and ...
@singwise
Music:
Bye, Bye Brain! by Alexander Nakarada | www.serpentsou...
Music promoted by www.free-stock...
Attribution 4.0 International (CC BY 4.0)
creativecommon...
Dandelions by Aposine | / aposine
Music promoted by www.free-stock...
Creative Commons Attribution-ShareAlike 3.0 Unported
creativecommon...
Sand Castles by Purrple Cat | purrplecat.com
Music promoted by www.free-stock...
Creative Commons Attribution-ShareAlike 3.0 Unported
creativecommon...
Discovery by Scott Buckley | / scottbuckley
Music promoted by www.free-stock...
Attribution 4.0 International (CC BY 4.0)
creativecommon...
Hey thank you for the video! It's always interesting to learn about formants and acoustics! Since I work as a vocal coach myself, i want to educate myself to not make mistakes and spread fake news. In my education this topic was not explained. None of my teachers shared information with me. Do you have some books or sources where I can find reliable information about formants and vocal tract acoustics? It's hard for a beginner in that topic to differentiate between real science and pseudoscience. :) Thank you very much!
I tend to look for primary sources and then a person who understands them. It’s far superior to finding a book on the subject, IMO.
This site has a clear explanation splab.net/apd/g400/
But before looking of vocal acoustics I recommend this www.phys.unsw.edu.au/jw/brassacoustics.html to learn about more simple closed pipes.
For a “simplified” model I would not use the theory wit multiple containers. I’d instead just talk about impedance matching and forget altogether about formant behavior.
unraveling a deep conspiracy, voice teachers teaching half truths, keeps students coming for more...
I had to end the video at some point but it’s only the beginning. The more I dig the more shit I find. It’s looking worse than what I imagined. Much worse.
@@MisterOpera I just watched the missing fundamental video, which was awesome. And I now wonder aloud if when we hear the cupo or cover sound, if this phenomenon of combinations of higher frequencies are creating the low, rich vibrational sound...
it would be a whole diferent approach on how to 'teach cover'
The low, rich sound is probably a nice big scoop of the second harmonic.
Higher frequencies help with cutting through, lower frequencies sound warm and rich.
Most of the time to make a cupo sound people are rounding the lips, which would be lowering all resonances. At the same time, the other end of the vocal tract is constricted to raise them, but more importantly to get a nonlinear acoustic effect that assists the larynx. The “cupo” sounds nice but also is easier than a more “open” configuration. You don’t have to be quite as precise, and you can do more vowels with the same mouth position.
Opening the throat also lowers resonances so sounds sort of similar, but it doesn’t give you that nonlinear effect so you can’t make a very strong sound and it uses a whole lot of air.
A few years ago I started to read about the theory of formants, because one of my vocal teachers mentioned the singer's formant as the amplifier that enables the singer to be heard over an orchestra. And she stated that the sf was located in the area behind the nose, more precisely the sthenoid sinus. (And there are more teachers out there that associate the sf with nasal resonance, giving the ng exercise as example)
Personally I find the theory of harmonics and formants really difficult tot understand. So wrong explanation of this theory makes it even harder. So thanks for correcting this.
The “singer’s formant” has nothing to do with sinuses, and it’s only helping project the male voices. Sopranos and mezzos are already too high in pitch to need it.
The singers formant is in the 2500-3500 range. Orchestra peaks at 500-1000. The second harmonic of A4 is 880Hz. That’s only the second harmonic. The second harmonic of A5 is 1760 Hz.
It’s also important to remember that the singer’s formant doesn’t mean you sound good. It means you can be heard more easily. That’s not always a good thing.
It was a good observation, but even Johan Sundberg now admits maybe people expected too much from this one detail.
@@MisterOpera What do you mean by 'one detail'?
I mean that the singer’s formant is a detail of vocal acoustics you can probably just ignore.
Much more consequential is “source-filter nonlinearity”, which explains how your vocal tract shaping affects your vocal folds (I’m not kidding!).
This nonlinear “surfing” is how you can make really loud, clear sounds without really taxing the vocal folds all that much, and with only a little bit of air coming out. You don’t have muscles that can do it. It’s done using SOUND reflected back to your larynx.
I feel I have to comment a few things regarding the acoustics. Although I am not a vocal scientist per se, I do actual research in acoustics; my publication list can be found under my profile on Google Scholar, for instance.
(i) The word “formant” may actually refer to two related but different phenomena. With “formant”, some people just mean the observed frequency peaks in a spectrum, whereas both you and Singwise use the word to refer to the actual vocal tract resonances (VTRs). A formant, in the sense of an observed frequency peak, can only be found if the source tone contains partials falling in the region of a VTR. I personally prefer the term VTR, instead of formant, for what you seem to discuss (although I am likely not completely consistent in my terminology)
(ii) The nose cavity. When singing nasals “Mm” or “Ng”, the sound is radiated from the nose, and to me it seems reasonable to view the vocal tract effectively as a series of two resonators, divided by the entrance to the nasal cavity, so that the total length of the tract is actually from the vocal folds to the nose. Howver, when singing (nasally) with an open mouth and velum, I agree that the nasal cavity mostly serves as a Helmholtz resonator that will filter the sound, but not completely so, since some radiation, although small, will come from the nose.
(iii) The big one. You say that “formant behavior is predicted by perturbation theory”. Also, you say something like that the speed of sound is selectively slowed down or sped up by changes to the the tube diameter. These are *really* confusing statements! The speed of sound in air does *not* change with tube diameter. It is a thermodynamic quantity that scales like the square root of the absolute temperature. It’s about 353 m/s for air at body temperature, that’s it. No relation to vocal tract geometry.
Also, no perturbation theory is needed to predict formant (VTR) behavior. That’s a weird statement! These days, the most accurate way of predicting formant behavior is to make simulations using numerical solutions of the 3D wave or Helmholtz equation, like for instance my good friend Jarmo Malinen and coworkers in Helsinki have done.
I can sort-of guess how your statements came about, though. Numerical solutions to the full 3D governing equations are computationally costly and require accurate vocal tract geometry data (from MRIs eg.). An old-fashioned and very simplified modeling strategy is to use the 1D so-called Webster equation instead. There is a lot of tinkering going on to create such a model. Such a model is somewhat accurate for low frequencies, but not for higher frequencies like these in the singer’s formant region. Solving for the so-called eigenvectors and eigenvalues of the Webster equation, we can get some information about the VTRs. And the eigenvalues can likely be computed by some classical perturbation method. This is very old-fashioned stuff. And not very accurate. Also, it is possible to extract from the Webster equation effective propagation velocities of the waves, and these will indeed depend on the diameter of the tube. However, these are only artificial, “effective” velocities showing up in the simplified model, and should not be confused with the real speed of sound!
The Helmholtz predictions don't work. Proponents try to overwhelm with math, but the models don't make convincing vowels, and the shapes are quite limited - it's not that we can't make those vowels that way, it's just that there is almost always more than one way to get the same or close formant values, because for just F2, we have 5 different ways to tune it (3 of them also tune F1).
We could get into fourrier transformations and all that if you want. I'll just go get someone math literate and ask them to come speak to you, if that's something you're interested in. My issue is quite simply that Helmholtz models don't work. In the sense that like, a car won't run. And proponents are a bit like used car salesmen, trying to overwhelm others into ignoring that the car just doesn't do what it claims (move you about).
The reason we can speed up and slow down is we have pressure changes. As a pressure antinode, the area of the laryngeal vestibule is experiencing almost no displacement but very high pressure.
A simple tube, narrowed or expanded in the places predicted by perturbation theory, using only two formants, makes much more convincing vowels than even the three-tube Helmholtz models.
Using inverse filtering, it's easy to confirm these four things:
1) constricting area of interest 2 lowers all resonances (yes this is how we use the word formant in this domain)
2) constricting AOI 3 raises all resonances
3) constricting AOI 4 lowers F2
4) constricting AOI 5 raises F2
AOI - area of interest
AOI 1 - overall length
AOI 2 - the pressure node of the 1st formant
AOI 3 - displacement node of the same
AOI 4 - the middl pressure node of F2
AOI 5- the middle displacement node of F2
These AOI locations are predicted by perturbation theory and work using simple tubes, as demonstrated by Sundberg.
We know the speed of sound is changing because that's the only thing that could be happening. If you have another explanation that's great, but I won't accept any OBSERVATION that claims Helmholtz models make convincing vowels orare shaped like vocal tracts. Literally they are interconnected tubes. The vocal tract isn't, and does not behave like them.
We can not go from bad OBSERVATION and jump into theories.
@@MisterOpera “The Helmholtz predictions don’t work” … “and the shapes are quite limited - it's not that we can't make those vowels that way, …”. I am not trying to be rude or disrespectful, but this is a response that does not make any sense. I don’t doubt your statement that vowels can be made in several ways, but that is a red herring that has nothing to do with the Helmholtz equation. The Helmholtz equations is the frequency-domain version of the wave equation, and these are extremely accurate mathematical models of the acoustic pressure in air (under some reasonable limitations), but costly to compute. It may be that you have not seen good results in the literature, and that may be; the calculations are difficult and require accurate geometries, but that does not invalidate the equations!
Is the second paragraph an attempt of intimidation? I am interested in giving an honest feedback from an acoustic background, since I am an academic active in acoustics research. I have watched a few of your videos and liked them a lot. I am just looking to help out, not to fight. I am happy to talk to you in a nonconfrontational way.
Then you talk about acoustic pressure maxima (antinode) vs. acoustic velocity minima (note that it is *velocity*, not displacement). It is correct that these are complementary, that’s what standing waves are all about! But the “speed up” and “slow down” is something you see in effective quantities in the 1D model, it is not that the speed of sound actually is changing.
Your information about constriction and so on seems fine, and those can be confirmed by tube models (that is the Webster equation I talked about), Note that the Webster equation is *derived* from the Helmholtz equation after making simplifications, the most basic one being that the acoustic pressure does not vary in the cross-sectional planes inside the vocal tract. That is fine for low frequencies, as I said, but not for, say, the singer’s formant region.
The last two paragraph I do not understand.
Actually, Helmholtz resonance requires physical Helmholtz resonators. It's one thing to theorize that the vocal tract is a series of Helmholtz resonators. Now you have to FIND THEM. Showing that your numbers add up only implies that IF you found your resonators somewhere, you might have a theory that passed the most basic of tests. But your theory FAILS this test! Because we can't find the separate resonators!
The fact that you can tune a Helmholtz resonator, like you can ANY resonator, doesn't make the vocal tract function anything like Helmholtz resonance based physical models. They are OBVIOUSLY not similar, and it's OBVIOUS that the researchers who developed perturbation theory were struggling with the exact same problem: Helmholtz resonance doesn't explain vocal tract function, nor does the modeling resemble it. (I'll give you a small victory on the few sounds simple Helmholtz make, versions of uh and u that use vocal tract shapes resembling uniform tubes.)
IF you had actual separate tubes you could pinpoint as the Helmholtz resonators, ok maybe. But look at the working physical models. They do NOT have similar structure, nor CAN they have similar structure to a vocal tract, because it isn't made of three tubes. It's ONE RESONATOR. (excluding the nose, and any nitpicking about basically inaudible radiated sound).
On the other hand, a simple rubber tube and a duck call, manipulated according to predictions made by PT, makes very convincing vowels.
You're deluded if you think these two things function in remotely similar ways. Or, I am the deluded one. Either way, we don't agree on what is even occuring in the first place. I don't see any convincing physical models that support your view, and even a rudimentary tube and oscillator gets closer to human speech than the most sophisticated Helmholtz resonator PHYSICAL models. There's footage of those at Arai labs if you're interested.
You can easily test if you have one resonator or three (actually if you listen to the deluded, there's an infinite number of separate tubes). Change the LENGTH of the vocal tract, and ALL resonances will move on a spec or inverse filtering. If it's separate tubes, why is it not possible to lengthen just one of them, and change the corresponding formant without altering all the others?
If it's separate tubes, why does narrowing at AOI 3 raise all resonances? This should not happen in your model, but it very obviously does!
I CAN think of one group of people particularly unskilled at vocalizing, who might fail to notice you can narrow AOI 3: voice experts! Perturbation theory is maybe less obvious to this group. They are happy to say the first formant must be in the mouth, because that's the only way they know to detune it from the fundamental.
Perturbation theory holds up to direct observation and testing. Helmholtz models don't. When people ask what's up with that, you try to intimidate them with technical words. But none of it makes the modeling ACTUALLY WORK, nor is it directly applicable to vocalization techniques. Perturbation theory is directly applicable, with no modifications, hand waving, or cognitive dissonance required. It just works exactly like it says on the box.
Yea, if you had am infinite number of wine glasses you could individually tune and play at the same exact time, you couldake something resemblimg a vowel. But the closest you'd ever get to speech is somewhere in the uncanny valley. Not so with a tube and an oscillator. AND you don't need to guess where to constrict or expand the tube. All you need is the length of the resonator and knowledge of where the displacement nodes and antinodes are. And that is very easy to do: to get the next formant up, add an extra node and antinode, and distribute the total evenly along the tube. This will predict nodes and antinodes for EVERY formant your resonator makes.
Also btw this same theory predicts how the flare of a trumpet bell-end or narrowing in th mouthpiece will raise all resonances of a trumpet. Helmholtz modeling, afaik, doesn't even try to explain trumpet design.
Btw it gets even better, because PT helps predict and manage source changes arising from nonlinear interaction with the filter. Never mind jist making all the vowels Helmholtz models can't produce - we still need to account for source-filter interactions. In some cases that means formant-harmonic couplings (that's why we use formant in the sense of VT resonances btw - we ALSO have harmonics and they are not affected by perturbations).
Practically speaking, Helmholtz modeling doesn't allow for this, because each value comes from a discreet resonator. It doesn't allow for using F1 to raise F2, and definitely doesn't allow for it without moving the fundamental.
Or tell me this: describe a model for me, physically based on helmholtz resonance, in which F1 is below value given for a uniform tube? Where's the resonator doing that, when your tubes aren't long enough?
splab.net/apd/g300/
Here is a demo of a physical model.
Anyone skilled at making vowels can tell you this is WRONG. It is not a given that the lips must be rounded for u or the back of the vocal tract constricted for ah. We also do not have anywhere near this range of motion with the larynx, the neck can protract and retract, AND you can even change length using only the lips.
Start mapping out the actual motions that produce ALL vowels, and it adds up to... Perturbation theory. That ONE theory covers any move you can make, and the resulting change in resonance, confirmed via inverse filtering.
Where is ö? How about ih? Why can't this model give us German long e and short e? The list just goes on and on!
What do you even make of a ventriloquist when your modeling is like this??? (It's perfectly unremarkable using perturbation theory).
These formant values and the way we get them are fungible for the vocalist, until we add requirements like matching impedance and nonlinear interactions. If you don't know at least instinctively you're probably learning to sing from voice researchers.
I used Voce Vista for a short time and was quickly frustrated with it because it took far longer to interpret the readings and make sense of them and then strategize how to improve my singing technique. However, it is a good tool for figuring out the differences of how each voice used the higher frequencies to have their own sound. Example: Corelli and Bjorling. I can’t sound like them even using a tool to guide my frequencies with in my voice to sound like them. It would be foolish to try and a detriment to my voice.So, yes, I would call it self guided gaslighting to do so. It is just a tool that does one thing.
Yeah I found the same thing. By the way if someone is considering buying the program they should check if they have an iOS device first. There is a free real-time spec in the App Store. Do a search for “wevosys”.
@@MisterOpera Thanks, that's great advice!
@@MisterOpera and on the matter: I find that kind of software very helpful to determine what you are actually hearing in a certain voice or even in your own voice, in particular if you are dealing with voices of a wider range of vocal genres. So in my opinion it works far better as a tool for analysing voices than as a teaching or learning tool.
Well if someone finds it helpful then I guess that’s good enough for me. Personally I prefer to use my ears and a good set of monitor speakers.
I will mention the free app in a video if that helps people. It works really well and you can also record the screen and your voice as you use it.
I’m not sure about the android app but I did find one.
@@MisterOpera I also prefer to use my ears as my primary tool of assessment. But whenever I need "proof" for other people or need a computer to understand my thoughts I have to rely on software.
Great video!
Thanks! Please share it if you can. Pseudoscience has a pretty big marketing campaign. The truth has to rely on volunteers.
27:27 Try it at home: take a cylinder with the same cross-sectional area but a longer longitudinal axis (and therefore a higher volume) and attach a drumhead. Is it a different pitch? If so, how? And why? If not, why not?
Or just gedanken it.
I must admit I have zero drums
@@MisterOpera ditto. Just my vocal perturbations.
Brilliant
The acoustical “ mistakes “ in my judgement gives each singer their unique voice. Sometimes I think we can suffer from the paralysis of analysis especially bad analysis.
Can you elaborate a little? I’m not quite sure what you mean by acoustical mistakes.
I am Curious as well. And are these "mistakes" not "harmful"
Upon reflection, I believe that “ mistakes”is the wrong tact to take. Each unique voice has its own built in acoustics. Voce Vista and like programs can adequately analysis those acoustics. They can for lack of a better word “aid” the singer in producing a better sound. I will say that when science is properly questioned and applied the results can be measured. In the end it is up to each individual singers curiosity and determination to make the most out of proper science .
I don’t think you can actually do very much at all with Voce Vista, as far as training a singer. The voice has nonlinear effects. That means for instance that the 3 ways of raising f1 are not fungible. Assuming they are fungible leads to all sorts of problems.
The main way I see people using voce vista is to gaslight themselves into believing they “improved” based on some objective they had regarding squiggles on a graph that don’t correspond necessarily to their real intentions. It’s pre or post-performance lawyering.
Individual characteristics of voices are mostly higher formants. They don’t really affect function much.
I was sitting in a “Voce Vista” presentation in 2008 presented by Donald Gray Miller, PhD. The man who made possible a way to look at voice analysis using your computer. His book “Resonance in Singing “ referenced Johann Sundberg and Ingo Titze and their foundation for his research in forment singing. Dr Millar would give everyone who attended a basic course in vocal anatomy and acoustics. On page 29 and 30 were the basics model and explanation of forments. His explanation based on the science of the time was another link in understanding of the vocal tract. It wasn’t until your series of video’s explaining constriction and diligently practicing overtone scales that the lights came on for this singer. I remember this presentation because Dr. Bozeman dropped in to listen to some of the presentation. While he was there he dropped into my brief case, which was open, a copy of his book on Vocal Forments. Needless to say, l never ordered his book.
I’m pretty sure voce vista also wasn’t the first real time spectrograph for users. I have had a free one on my iPad for a very long time, by a company called wevosys
I wouldn’t mind getting my hands on one of their EGG units tho.
Remover when watching those overtone videos: there are mistakes in my acoustics in there. Maybe if I bring real acoustics to the audience they can make a list of my mistakes to make it easier for me to fix 🤣
Philippe, you are a genius!👏🔥
I don't think you could conclude that from this video haha. I'm literally just reporting on some (deliberately?) overlooked science. Perturbation theory makes life EASIER than pseudo-scientific versions of vocal acoustics like the one published by Ken Bozeman and his vocology friends at voicescienceworks.org. sure, it yields 5 areas of interest to manage the first two formants instead of just 2 (3 for F1 and those plus 2 more for F2), but you don't need to think of all 5 at once. As a bonus, everything works exactly as printed on the label.
I found out because someone else corrected me, I had to admit it was true, and then I asked for better explanations. I didn't figure out any physics myself, I just figured out the social and political games that were keeping the science from me. I did interact with a genius, IMO. I'm hoping to have him on the show at some point.
I suggest dropping in on their Facebook group (acoustic singing pedagogy). You can watch their leadership (for instance Ian Howell) doing their best to make all of this look way more complicated and scary than it really is. Unnecessarily big technical words are a favorite strategy. Look right past it at the actual practical application and then you can see it isn't predicting results accurately. They are constantly explaining away why they can't produce any robust singing.
The reason is they have sabotaged their own efforts with poor modeling, and are stubborn about admitting the mistake and starting over.
Philippe, I was just referring to your entire contribution to the modern vocal pedagogy not only the topic discussed in this video. There is so much pseudoscience out there and watching all your videos really helps us understand how things actually work. Thank you 🍀
Well that's very encouraging, thanks! I feel like I rely more on organization than intellectual prowess, but I'll gladly take the compliment 🥰
Could you possibly send me resources to learn how to use formants and could u send me recourses that explain perturbation theory I don’t really trust many videos online for this kinda stuff LMAO but I think you’re pretty accurate or atleast a shot I’m willing to take
This site does a beautiful job and if you root around there you can also find other models to compare to. splab.net/apd/g400/
I think it would be great to form a discussion group as well. I'm very convinced that anyone who can talk already uses perturbation theory, and there's indirect evidence: when you force some position that takes a formant away from a target, then one of the other structures relating to that formant seems to adjust "involuntarily", resulting in a mixed feeling of control and chaos if perturbation theory isn't a conscious concept.
@@MisterOpera okay thank you! I’ll look into it when I have time
Enjoy! I don't know how I ever sang without it
@@MisterOpera ik about a discussion group held by studio west and ill likely join that eventually so that u can get better with formants also that’s an interesting theory (I’m not entirely sure on how formants work) but I’m assuming that what you were saying was that the body will automatically adjust itself when there’s a mismatch in structure if the person who sings isn’t consciously going against their bodies readjustment.. is that what you meant?
Dear Neil Patrick Harris, you mention changing the "speed of sound" by altering the diameter of the tube at the site of a node or antinode. This makes no sense to me. The speed of sound is determined by the properties of the medium it's travelling through, changing the diameter won't. Also, what do u mean when you say formants allow us to make certain waves "fit" that otherwise wouldn't? Your video may be "science-based" but it is also riddled with errors, missteps, and colloquialisms, and is pretty difficult to follow. Like it or not this IS an abstract and non-tangible concept that makes it difficult to explain to those without a physics background. Especially when the goal is application-based.
Dear George Takei, you could easily Google perturbation theory to learn about how this works. I guess you were watching fiction instead, looking for jokes about my appearance.
Assuming you’ve done that and now understand I was telling the truth, is there any specific mistake in my video? I’m happy to address any specific mistakes, but my patience with you will wear thin if you promise to show me it’s riddled with mistakes, and then don’t explain a single one.
Do not make people hope unless you can make good on your promises. Even implied promises. Even disingenuous promises. If you claim to have info, but don’t have info, you are a bad person. It’s not much better if you know of specific mistakes, mention them mistakes, but don’t describe them with specificity. Not good.
As for your claim that a physics background is needed, I disagree. Perturbation theory is very intuitive, and you can simply learn the predicted positions of nodes and antinodes with a given vocal tract length, and use unstructured play (moving things around) to see that yes indeed, the vocal tract does work this way.
What’s needed isn’t a degree. What’s needed is less bias against ideas that imply something inconvenient.
Using perturbation theory isn’t more complicated than the currently popular theories. Despite opponents working really hard to make it seem hard to understand, it just isn’t that hard to understand, observe, and apply using experience and intuition.
It also helps describe important rules and guidelines about exchanging functions (for instance why, with a low larynx, widening in the pharynx detunes the 2nd formant in a similar way to constricting under the uvula).
Perturbation theory also explains that each formant is derived from the next lowest formant. Which helps predict that you move the 3rd formant when you move the 2nd, which helps predict that when you couple the second formant to the 3rd harmonic, for instance, you can still use 3rd formant detuning to “fake” vowels because the listener is already listening to the 3rd formant in speech (it’s a weaker change but still can be used to “fake” vowels that wouldn’t otherwise work).
Go ahead and try a simple experiment. PT predicts a 3rd formant node and antinode between the middle antinode of the 2nd formant and its antinode at the terminal opening. Lift the tip of your tongue and you’ll “tweak” the 3rd formant. It’s a weaker change than moving the 2nd formant but you can hear it. And you can use your fingertip in the same place to detune the 3rd formant in the same way.
These locations are predicted by PT and are right where they should be, and behave just as they should. You can also confirm your actual formant tuning using the free software called “Sopran” to do some inverse filtering. You will confirm that PT is in fact really good at describing how things work, and it does intuitively, in a way a child can understand.
Multi-tube models, in addition to making no sense at first glance (it’s like saying “suppose things are not as they are…”), keep giving wrong answers. They break down really easy and start giving wrong answers. They poison intuition as well, and lock people in to ideas that never even seemed to work all that well.
It’s not more complicated to base intuition on perturbation theory. Just less wrong.
@@MisterOpera First of all, looking like NPH is not an insult, so don't fret. Secondly, after googling this Georgi Takei, I'd say you're drawing dangerously close to a racial innuendo, which is reaaaaally uncool. I've dealt with enough racial discrimination in my life it would be disappointing to see an "educated" person like you perpetuate that kind of nonsense on here - never mind all this physics banter. In terms of your patience....how about you answer the specific questions I posed to you? Not specific enough? ok 4:20-5:25 there, when u mention one's ability to change the speed of sound, then go on to describe the function of formants as making sound waves fit into places they otherwise "shouldn't". I'd like you to explain the accuracy of these statements. You have no problem ripping into Karen so naturally, you should be open to the critique. Finally, I am not super invested in the topic so feel free to move along after reading. Either way, I know I will. :)
Ok, guy who doesn’t resemble George Takei. This should help you figure out how to get answers. bfy.tw/Sc8h
@@MisterOpera that Takei potshot was terrible. Yikes! In any case, what he is asking is valid. Speed of sound is defined by density. What you have claimed there and the "making sound waves fit" part needs some clarification or referencing imo.