I adjust to the sound of Larry King's voice by changing the channel.
"clear" may be a poor word choice- he could mean "quickest" or "biggest jump in frequency"
Ogged, you may want to clarify that our fanclub's beloved scruffy cherub did not write the passage you quoted. Until I clicked through, I thought the criticism in your post was aimed at Yglesias, and not at (the then-unknown-to-me) Franz de Waal.
There are various formants that go together to make up the timbre of our voice. It's not just one pitch, so yeah, there will be harmonics at different frequencies and there is a low pitched formant in adult males that sits around 500Hz.
It's also possible, to a degree, to alter them by moving your larynx. I trained in phonology as an undergrad and one of the things you are taught is to acquire conscious control of this. If you lift your larynx you get a higher-pitched slightly strangulated tone to your voice.
John Major, former british pm, had this tone.
If you drop the larynx a LOT you get what sounds to my ears like this absurdly resonant and slightly posh sounding deep voice. This is a killer move to make in phone calls. :) If you drop it to the max and then 'sing' a long low tone you get that Tibetan monk type deep resonant thrumming noise.
There is also quite a bit of 'matching' going on in conversations. Sneaky salespeople, for example, will deliberately match the pitch and pace of people's speech to lull them into a sense of security.
So the subconscious matching of formant pitch would seem like a possibility -- however, I'm not sure how much it's really under our control. Formant frequency is as much to do with the resonance caused by the shape of the mouth, throat, and larynx. It's also affected by the particular vowel sound we are pronouncing.
The human voice contains a blend of pitches, starting with the lowest harmonic. Middle C is about 261 hz, and the A above it is 440 hz, so this under 500 hz thing is in the normal speaking range.
Think of it as a stare down with voices. During our conversation we are going to "sing a duet" and we will eventually agree on the key signature. The person who blinks first (switches to the other guy's key) loses.
Humans do a lot of this syncing up without knowing it. It is one way you can make yourself "likable," although for the status thing you want to be dominant, not likable.
This stuff is pretty subtle but you can spot it if you are looking.
quoted byyggy
You misspelled "syzygy".
Matt McG's (I'm considering going by my first name because there aren't enough Matts commenting here, by the by) mention of salesmen reminds me: the Salesman example in Tipping Point was always the one who got matched.
Scientists used to consider the frequency band of 500 hertz and below in the human voice as meaningless noise, because when a voice is filtered, removing all higher frequencies, ne hears nothing but a low-pitched hum. All words are lost.
I'm confused by this. A 440 is the tuning concert pitch, and well above my normal speaking range. Is the claim that the harmonics which differentiate words are much higher than this, even though the primary sounds are all below 500 Hz except for the really squeaky voiced?
But the phenomenon itself seems reasonable. Matching another person's tempo, rhythm and patterns often creates a sense of familiarity.
I used to work in a call centre (in between undergrad and grad school) and we had fairly elaborate, and slightly sinister, voice training to help us control conversations on the phone.
Pitch and pace matching are only some of them. Speaking slowly and more deeply generally implies authority - as you'd expect - but other tricks include omitting the pitch rise in interrogative sentences. If you keep the pitch completely flat instead of rising in a question you can -- they claim -- use this to give yourself an air of authority and it enables you to use questions rather than imperatives to persuade people to part with their money...
Pointing out the obvious, so much about our social status, position in the heirarchy, social class, level of education and so on are communicated in the voice.
Adept communicators can manipulate this either consciously or instinctively.
I wonder if an extra zero got in there by mistake. 50 Hz would make more sense, considering that human hearing poops out at 20 Hz.
http://www.umanitoba.ca/faculties/arts/linguistics/russell/138/sec4/formants.htm
On formants. It gives 500Hz as the frequency of the F1 formant. Then there are many harmonics above that.
Did one of us get married to SB and change our name?
Yeah, that might make sense. According to some googling the F0 formant, the one below F1, is somewhere between 80 and 200Hz in men.
So 50Hz would fit with that.
A low-pitched hum would make more sense around 50Hz, and would more plausibly be the sort of thing barely at the edge of conscious. I'm usually conscious of concert As.
Here is the paper about the debates . It is 500 Hz, De Waals must have made a mistake about the "meaningless noise" part.
It's also possible, to a degree, to alter them by moving your larynx.
Fascinating. I have a much better phone/radio/public speaking voice than my standard speaking voice--something about needing to project to be understood makes me semi-consciously think about my voice, and the sensation in my throat is one of pushing something lower. I'd always assumed it was psychological rather than physiological.
You are probably dropping your larynx. If you look in the mirror it's sometimes possible to see it. If you look and then consciously drop your voice you sometimes see the Adam's Apple descend.
It's possible through this visual feedback to acquire conscious control and thus to 'set' your voice to be deeper and more resonant at will.
the sensation in my throat is one of pushing something lower
[insert obligatory ATM]
I have a very deep and gravelly voice that sounded great on the radio back in college when I did that sort of thing. I sang bass when I was 13. Unfortunately, when I'm not behind a microphone (or not drunk), I also tend to talk quietly and my slight Southern accent mostly manifests itself by slurring. Older people ask me to repeat myself a lot.
Larry King is still the worst interviewer on television, though.
Now it's interesting to think about the voices of the folks here. Anyone want to cop to having a thin, reedy voice?
Nope, sorry. Deep and somewhat resonant (I have a hard time projecting to a large lecture hall, though), good for radio or reading out loud.
Aha. Now I know why I have a higher voice in French and German than I do in my native English.
Ohhh. Wait.
A 440 = 440 k-hertz. Hz is an abbreviation for kilo-hertz (much like kcal is often just Calorie) that gained popular (if not scientific) usage.
I think. Now I'm trying to find a citation.
I'm pretty sure any frequency beyond 20,000 cycles per second is inaudible.
quoted by yggy
The preferred terminology is "ygls," pronounced \'ig-əlz\. Rhymes with giggles.
Anyone want to cop to having a thin, reedy voice
Well, if reedy means high, then yes. I'm always horrified to hear myself on tape -- listening to myself through my skullbones I sound all resonant and stuff, but from outside my head I'm disturbingly thin and fluty.
So I was reading my comment 8, and I think I left out some words, or possibly a sentence or two, that would have made it impossible for any Tipping Point non-reader to have a clue what I was trying to say. What I meant is that the person in Tipping Point who is used as the example of the paradigmatic Salesman (one of the three sociological archetypes which Gladwell uses) would always have the timbre of his voice matched by the person he was talking with, rather than his matching their pitch.
De Waals must have made a mistake about the "meaningless noise" part.
I'm not so sure. I think you might be conflating "meaningless" with "inaudible"; "meaningless noise" here just means that words are indistinguishable by listening only to that frequency range. I think.
Hz is an abbreviation for kilo-hertz (much like kcal is often just Calorie) that gained popular (if not scientific) usage.
I think this is incorrect. Unlike calories, we deal with signals and events of a variety of types that can have frequencies ranging from a few Hz to several kHz; I don't see any impetus for a calorie-style omission of the prefix. [On preview, SB seems to have it covered.]
Anyone want to cop to having a thin, reedy voice?
My voice has a slight nasal quality and I tend to speak quietly.
I also tend to avoid many standard social dominance games probably to my detriment (my standard approach to interactions where hierarchy is important is that I'm happy to admit someone else's importance as long as they admit my intelligence).
This is a blog after all, surely there are other socially awkward geeks around.
I sometimes notice myself consciously lowering my tone (I guess lowering my larynx, after reading this) when speaking in class or talking to someone "important". Huh.
27 -- "Yggy" makes me think of Ziggy Stardust, which I think is only appropriate considering half of the commenters would claim he was the special man and we are all just Yggy's band and half bitch about his fans and wonder if they should crush his sweet hands.
(I was going to make a "well hung and snow white tan" joke but just couldn't bring myself to do it.)
The question in my mind is whether having a dominant voice was helping those candidates win the election, or whether they acquired vocal dominance in the debates because the underlying dynamic of the race was favorable to them.
Let us hope MY doesn't take it all too far. He's already in danger of being sucked up into his mind.
Hz is an abbreviation for kilo-hertz (much like kcal is often just Calorie) that gained popular (if not scientific) usage.
I think this is incorrect. Unlike calories, we deal with signals and events of a variety of types that can have frequencies ranging from a few Hz to several kHz; I don't see any impetus for a calorie-style omission of the prefix. [On preview, SB seems to have it covered.]
Yeah, that's right. But still, it seems really weird that if you stripped off all the higher frequencies but left the 200-500Hz, you'd have no words, and if you stripped off the 200-500Hzs, you'd have words but no timbre. I guess I'd have to hear it.
And 34 is what I meant to post before I got sidetracked by the data. All we have is a correlation; the vocal dominance itself could be affecting the voters, or it could be some other quality (such as confidence), that both causes one to be vocally dominant and win voters.
Maybe all the flappy smacky consonants live in the higher frequencies, so that a voice filtered above 500 Hz sounds like warbling vowels.
Being the audio guru that I am, I took 24 dB out of all frequencies greater than 500 Hz in a somewhat deep female voice sample (not Cala), and it was still quite comprehensible. However, 24 dB is not just a whole lot, so I applied the EQ a couple more times, and while the pitch was still apparent, the words quickly became incomprehensible, so I can verify the effect. The vowel information is all contained in the relative amplitude of the higher harmonics, and most frictives/siblants/plosives (but not P or B, which are deeper) produce noise whose primary frequencies are up in the thousands, so the remark seems to be accurate.
There are technical terms for the flappy smacky consonants, but I'm feeling whimsical.
In the actual journal article, they analyze the range "beneath 0.5 kHz". kilo is 10x less than mega, right? So this would really be the range below 0.05 megahertz?
The citation for the article is:
Gregory, Stanford W. and Stephen Webster, 1996. A Nonverbal Signal in Voices of Interview Partners Effectively Predicts Communication Accommodation and Social Status Perceptions. Journal of Personality and Social Psychology. 70(6): 1231-1240.
I got to it through EbscoHost; those of you without university library connections might be SOL in trying to get it.
I'd have to hear it also. Harmonics and resonance and everything is cool, but I couldn't for the life of me pay attention to acoustics in my physics classes and I can't seem to work this out in my head.
So, I just recorded my voice and then used a filter to take out everything about 500Hz.
It is more or less undertandable , but very muffled. Like a tv heard through a thick wall.
Oh, we're talking about actual hertz, not megahertz. So yeah, the cutoff is 500 Hz.
2 errors in my first unfogged comment! I'm so proud!
Re 43: are you sure that the dropoff was complete? Many low-pass filters and EQ effects actually only do a 20-30 dB drop, not a complete elimination.
Heh. For being just about the simplest unit this side of the meter, the Hertz is giving us a lot of shit today.
Ce n'est pas de ma faut!
Thanks, pdf23etc.
Actually, on a more extensive sample, I can confirm that even after five applications of a low-pass filter, unfamiliar recordings can still be understood about 15-30% of the time. So there's no complete information loss.
Ah, just saw that pdf23ds had already done this.
I can post a .wav but I suspect my filtering may not be as competent as pdf's. If anyone really wants to hear it I can whack it on some webspace.
If anyone really wants to hear it I can whack it
...at the Mi-- oh, never mind.
the simplest unit this side of the meter
The meater may be a dim unit, but it sure is a long one.
I just did a low-pass filter (i.e. it filters out everything but the low frequencies) at 100 Hz, and still heard a vauge, very soft mumbling that could actually be above 100 Hz, and my filter (Audacity) is crap. But it was directly correlated to the pitch of the voice, so the idea that it's actual subvocal sounds that are involved here doesn't make sense to me.
If anyone's interested, I might be able to get some short .ogg files (what else?) uploaded this evening for demonstration.
Just about 29". Only Labs knows precisely.
Too bad nobody ever convinced Barry White to run for president.
Do any of the articles say how this works cross-gender? I know nothing about audio, but the average pitch differences are pretty large; whatever this matching process is, do men and women match each other in conversation?
I sound like a Southern duck.
I have a voice made for the internet.
Just about 29".
This an in-joke I'm not up on?
I think SB is following up 53 (and it's 39", ahem).
I got my figure from ogged.
OK, teh audio uploaded.
Heavily low-pass filtered at about 300Hz:
http://www.mcgrattan.f2s.com/unfogged/low-filtered.mp3
Which is me reading the little bit of text that begins "Scientists used to consider the frequency band of 500 hertz..."
Also, I think that this study doesn't mean "deep voices win"; rather it's "the higher status person (probably determined already) sets the tone, literally, of the deep parts of the voices in the conversation." I don't know if this would affect the male/female voices question--could it be that the overall frequencies settle on the same frequency, even if women don't have as much of those frequencies?
I think that my voice is pretty nasal--which is part of the reason I don't pronounce my name "whiner." I can project to the back of the classroom pretty good, though. Sometimes even in ordinary conversation.
60: I'm not so sure...
61: ...because this kind of explanation is usually out there.
61: Ah, I was thinking the actual meter. Very well--it's an injoke you weren't up on (me either).
My voice gets hoarse if I have to project to a large lecture hall.
Going to be screwed on the job market, I fear.
re: 63
, yeah, I think presumption is that lower status males whose formant 0 frequency is lower than the person they are talking with will raise their formant 0 frequency to match the high-status individual.
66: There's no reason for the hiring committee to know that, if it doesn't show up in your evals. Unless they have you teach a large lecture class on your campus visit, but I think you're more likely to get a smallish class. (And IRL, often large lecture halls have microphones, which usually cut off all frequencies above 500 Hz.)
I wonder what the relative status is between you and a person you feel socioeconomically superior to but physically threatened by.
68: I'm sure most of the distortion is the speakers and the hall acoustics. But I relate. I skipped out on most of the lectures on one particular freshman course for that very reason.
re: 69
That's a good point. They are a bit vague about what status is supposed to represent.
So there would be no words anyway, without those high frequencies.
The solution is to modulate my voice to control the weak-minded, à la the Bene Gesserit.
Oh, right, from that book with Digby Arizona and Gulping Haddock.
I'm disturbingly thin and fluty
Once again, LB, you surprise me. You sound so authoritative here.
I was thinking about this in relation to my students. I was around meeting with student groups today, and I found myself matching my speaking to my students, not pitch so much as strength and loudness. Some students are smart and confident but pretty soft spoken, and I tend to follow their lead on this.
I don't lack for authority at the front of the classroom, but I tend to match students communicating one on one. This is a pretty high status group I'm teaching, I'm just wondering if I come off as 'low status' in small group interactions, or if this is a different sort of matching than these subvocal pitches. Also, I wonder if pitch dictates participation in classroom discussion. If people feel they can't match this pitch then does it discourage participation?
My impression is that what the researchers are describing is different from volume, speed, or forcefulness. So you might be speaking in a relaxed and quiet manner, but still maintaining your "status." It's really hard to know without voice samples.
Surely one could be trained to actually hear the difference, with numerous audio examples and some practice?
Absolutely. In one of the articles about it, one of the researcher's grad student catches the prof matching the dean at a party.
OT:
I heard that only 65% of porn viewers are male. I would expect the number to be higher.
I'm a baritone, which is a bore a tone - can't hit the high tenor stuff anymore and can't quite get the lowest low. Still, playing Lockstock I got a nice crash course of voice lessons on lowering the pitch and creating resonance. And then hitting the consonants, of course.
The voice is quite remarkable and often ignored. Subtle changes can have a big affect.
On the male/female thing I wonder if we harmonize an octave apart? It would make sense, but I dunno for sure.
I don't think the normal difference is anywhere near an octave. The pitch difference is more like a fourth or so. But the tone difference (maybe something to do with formants?) is the bigger one, apart from pitch, and it's how we can usually tell women with deep voices apart from men with high voices.
Also I think maybe women have more access to the higher parts of their singing range, because the changes men's voices go through during puberty restrict the range besides lowering it. But I'm really no expert in this area.
Once again, LB, you surprise me. You sound so authoritative here.
See, I don't type with my larynx. (Not that I haven't tried, but there's that pesky gag reflex.)
But the tone difference (maybe something to do with formants?) is the bigger one, apart from pitch, and it's how we can usually tell women with deep voices apart from men with high voices.
In college, my sister dated a guy with some vocal abnormality that left him, despite all other visible indicia of masculinity (muscle mass, facial hair), with a speaking voice in a normal women's range. Every time I answered the phone when he called, I unhesitatingly identified him as a butch friend of hers from her high school basketball team. I have the impression from that that the non-pitch differences in men's and women's voices are voluntary behaviors, rather than physical differences.
McGrattan has already covered a lot of the acoustics of this, but there are a couple of things that seem to be confusing people. Presumably what people are matching is fundamental frequency (F0) or something like it - this is usually around 500 Hz, so that's probably where that figure comes from. Men and women can have the same F0 - as pdf points out, the frequency difference isn't very great between men and women. What is different is formant structure, which is due to the shape of the vocal tract. Women usually have smaller vocal tracts, so their formants are at higher frequencies. It sounds like the guy LB's talking about probably had an unusually narrow vocal tract or something.
My memory, and it's been 10 years since I studied this properly, is that men have more formants than women.
That is, if you do a fourier transform of the spectrograms created by voices when they produce vowel sounds: partly because men's voices are lower there are normally more `harmonics' above the bottom `note' within the normal range of human hearing.
It's like the difference between the pure sine-wave like tone of a violin when compared with the less pure but more complex note of a cello.
[I hope I have this right]
"I have the impression from that that the non-pitch differences in men's and women's voices are voluntary behaviors, rather than physical differences."
There certainly are a lot of conventions and affectations that can identify one's gender or other characteristics, but I think there's a physical component too. Sure, men can practice imitating women's voices until they sound female. I think even formant content can be affected somewhat with practice, but the practice is of the same quality of stretching your muscles--actual physical changes are necessary.
"partly because men's voices are lower there are normally more `harmonics' above the bottom `note' within the normal range of human hearing."
Or better yet, compare the sound from a very low key on a concert grand piano (9' long) with the same key on a baby grand (4'-5.5') or upright (of otherwise similar tone). Because of the shorter and thicker string necessary to get the right pitch on the shorter piano frame, the string is much stiffer, and there is a much higher harmonic content and lower fundamental content. On poorly-made spinnets, the lower harmonics are so absent or off-pitch that one can't even identify the fundamental pitch of the bottom octave and a half; they're aural garbage. On middle-of-the-road pianos, it's about half an octave. On the very, very best pianos, the bottom two or three notes are kind of iffy still, and included more for effect in loud passages than for their melodic use.
Yeah, 84 is right. I had forgotten that.
A good analogy for the vocal tract is a saxophone (or other woodwind): you have a reed that provides the initial vibration, but the size (fixed) and shape (which the player can alter) change the sound that comes out.
So, what happens if you use a lower timbre than a person who is objectively higher status than you. Let's say an associate at a large firm speaks with a lower timbre than the managing partner. Does that make you attractive or does it make you look like someone trying to be dominant, and therefore a prick who didn't know his place in the hierarchy?
I don't think this research shows that lower voices are dominant. It shows that conversational partners will match tones (in some acoustic sense I don't fully follow) and the dominant person holds their ground, leaving the other person to match them. So the dominant person could be higher pitched, and could pull the other person up the scale to match pitch.
Does that make you attractive or does it make you look like someone trying to be dominant, and therefore a prick who didn't know his place in the hierarchy?
If the subservient person does the dominant voice thing, I'd guess that depending on other personal qualities, they're either perceived as unbearably arrogant, or they receive an inappropriate (by objective status standards) amount of deference. It would be interesting to teach researchers to control their matching on purpose, and see how people reacted to them when they behaved inappropriately with it.
Yeah, lazyweb. How does the adjustment, when it's not performed, actually affect the conversation differently? Do those tend to be more confrontational, perhaps? Then again, a situation where both people feel they're higher-status than the other isn't often very great one to be in. Perhaps there's a negotiation that goes on unconsciously, and all conversations of sufficient length soon settle on one person or the other as the dominant one?
BG, I don't think the lower timbre itself is associated with dominance, but instead whatever timbre the dominant individual is using will be adopted by the other, whether higher or lower.
Perhaps there's a negotiation that goes on unconsciously, and all conversations of sufficient length soon settle on one person or the other as the dominant one?
This is my impression. It seems like this is supposed to be a largely unconscious process.
pdf and LB--I get that now that I've actually read the articles.
It seems like this is supposed to be a largely unconscious process.
Which means it would be fun to mess with.
I don't think this research shows that lower voices are dominant.
Not this reasearch, no. My own evidence is purely anecdotal and doesn't have any of the controls in place to make it generalizable, but I'm pretty sure that being tall, broad-shouldered, and deep-voiced has many times made people take me much more seriously than my qualifications or experience would dictate. Much more. Not to say that I complain about it, but I'm aware of it.
HEY, OGGED, THEIR IS A SPELING MISTEAK IN YOU'RE UPDATE TOO THE POST. WEE DON'T TAKE KINDLE TOO SPELING MISTEAKS AROWND HEAR!
I can't help but wonder if 97 was also posted by Weiner. Anyway, good point in 96.
And why, I ask you, won't this administration admit it's error, correct it and apologize to the public, so that we may all move on? This stonewalling will not do! In light of this administrations irresponsible behavior, I hope you will join me in calling for a special prosecutor to investigate Kindgate.
Shouldn't you be opening a bakery instead of being a bane to my existence?