They did a "great" job... for me, it's exactly ambiguous between a high-pitched male voice and a deep-pitched female voice. It's like the old "is it a vase or two faces?" illusion, or "which way is the ballerina spinning?" It keeps flipping in my brain.
But for the same reason, I would never choose it because it sounds so out-of-the-ordinary, an extreme outlier for human voices. When a computerized voice is speaking, I want it to sound "natural" (e.g. common, unexceptional) so that I'm paying attention to the message, not the voice.
Unfortunately a single "gender-neutral" voice turns out to sound quite uncommon, very exceptional. I only know a couple of people who sound like that, out of 1000's. I applaud the intention, but unfortunately I don't think it works in the end.
Regularly or randomly alternating between "average" female and male voices feels like perhaps a better practical solution, matching the two "humps" in the bimodal distribution of actual voice frequencies. [1]
To be honest, it’s not really all that unusual once you start paying attention to it — plenty of women have deeper voices than mine.
I transitioned from male to female some years back and trained my voice from a deep, masculine voice to one solidly in the female range. Pitch is only one aspect of how we gender speech, and I think that’s why it feels off to you.
I would say this voice uses pretty male-coded speech patterns with a higher pitch than you’d normally hear a man speak in. The pattern of clear starts and stops between words and limited range of pitch used sounds “male” to my ears.
>> The pattern of clear starts and stops between words and limited range of pitch used sounds “male” to my ears.
This sounds like an exaggerated stereotype of how men speak. Interestingly, the stereotype is different for men from Southern European countries who are said to speak in a sing-songy voice with vowels flowing together between words like vocal ligatures.
For example, this stereotype is used for satirical effect in the following strip:
I'd guess you're noticing voices a lot more because of your efforts. I don't really notice at all, possibly because I'm not paying attention. I wonder if I heard the Q voice in normal usage (like on a phone system) I'd notice?
Piggybacking this comment to say, has no one noticed that you can drag the little bubble around to adjust the frequency of the voice? It defaults to 153Hz, and varies about 145-175Hz. Which indeed does seem to be right in the middle, but with the 153Hz being more on the "Male" end to be honest.
Hmmmm, I think there is more to gendered speech than pitch or frequency. It's about subject and delivery. Anyone a fan of The Venture Bros.? Dr Girlfriend is a female character voiced by a male actor (Doc Hammer) Although the given voice is that of a woman with an incredibly exaggerated smokers voice; think patty and selma from Simpsons multiplied by each other. The characters voice is obviously male but the delivery is effeminate which is why it works so well. At least I always heard it as a very exaggerated voice of a woman who smoked heavily.
My take on the voice? Sorry, to me it's an effeminate male voice. I hear a british woman overlaid with Fez from that seventies show. In fact I was on a flight yesterday with a very obvious gay male flight attendant who sounded very similar to this. Think man's voice with a little helium in the air.
I’m not a fan of these kind of characterizations. I can think of two things that make someone obviously gay: (1) they tell you, or (2) you see that person engaging in homosexual conduct.
Nearly anything else is probably a prejudicial stereotype.
This is indeed a complex theme, in part because how it is linked to negative stereotypes and in part because people might not realize they are sometimes reinforcing the negativity in them.
It is not controversial to say that an obvious metalhead is indeed a metalhead.
Also I would like to point out that "(2) you see that person engaging in homosexual conduct" is not that much of a silver bullet here, as bisexual and trans people exist too :)
Also sometime it is not hard to know more someone that they know themselves, I had a couple friends that were known to be gay before they knew themselves. And also "a very obvious gay male" does not mean homosexual to many people, in the last years society and the internet became much more mature in distinguishing masculine/effeminate stereotypes from actual sexual orientations. I have no idea of what went through GP minds, but I believe that being able to separate the "very obvious gay" from a description of a sexual orientation into a personality trait can be very beneficial for society in the long term.
This view of the future seems very bland and uninteresting.
I thought the idea is to make gender a flexible thing, for example you can experiment with being a female or whatever gender profile, based on whatever you feel comfortable with.
Not to reduce the worlds gender representations (in business, culture, etc) into a homogeneous middle ground? Expanding what's possible in terms of personal expression rather than (further) restricting it. So I'm very confused why this exists other than as an interesting technical project?
I had the same experience. I hope this isn't going to spin off more academic work about why our brains are hard-wired to be sexist by continually and automatically trying to perceive the gender of speaker based on their speech.
It sounds a little bit like the voice of a moppet-type character from League of Legends that just huffed helium.
Because it is so stylized you don't have the problem of a
"perfect english gentleman" who pauses unnaturally, sometimes makes vowels like T-Pain, etc. that you have with more realistic voices.
It definitely comes across as female to me, but it is cartoony enough that it not so culturally coded as female -- so there is less of a sense of "female servicing" as there
is with most voice assistants.
Although these are gimmicky the sort of route of approaching the problem from what would a machine or non-human AI voice sound like would actually get you the goal of a Genderless Voice.
As it stands Q isn't genderless it's just gender non-conforming, which is a great alternative but I'd like to see some work into an actual "AI" voice rather than these voice assistants that try to imitate a human on the gender spectrum.
I mean no offense, but how is this really solving any problems? How is inventing a new voice many people are uncomfortable with really making the world a better place? A user configurable male/female voice could just work.
The idea, I think, is that it matters that/if we gender our digital servants—that as we use them more it will affect our sense of gender rolls. That Siri and Alexa are both female by default puts them in the stereotypical female-secretary roll.
I have Waze speak as a British male, of course then he sounds like a stereotypical butler. It would be fun if I could set it to be Batman or HAL9000, but it would probably get old.
And it's the same wrong idea as violent video games making children more violent. Both (a) fail to appreciate substantive causes and relationships, and (b) discount the nuanced contextual distinctions that people (even children) naturally make.
Aside from the motivations of the engineers (which are unclear and maybe even irrelevant), I think it's interesting technical work, including taking a stab at gender perception in speech patterns with a bold application.
I'm not sure I see the problem with a voice assistant sounding male or female, and I certainly wouldn't buy this. It just sounds wrong, and I can't quite say why. I guess after all the work tech companies have put into trying to sound like a person, this deliberately doesn't.
Also, it sounds blatantly female but pitched lower. Even when I use the little drag thingy to pitch it as low as it will go, it still sounds female. I think most of us know pitch alone doesn't define voice, and that's clear from this.
>Also, it sounds blatantly female but pitched lower.
That’s exactly what I hear. I’m not annoyed there is a market for such frivolous “woke” things, but I am annoyed it seems to be a mediocre attempt at it, it’s entirely female to me.
> There is no problem with a voice assistant sounding male or female; what's the problem with some other voice assistant sounding genderless?
This is literally the whole problem. It's less that it's genderless in a robotic sense and more that it's a hybrid manipulation of actual voices. It just sounds wrong.
> Maybe you're used to all people sounding male or female, so sounding like neither seems unfamiliar.
Maybe so. We try to make our robots imitate humans, so of course they sound male or female. From what the website said, it might also be flipping between. But even if I am just used to it, so what? That's how I'd prefer my voice assistant to sound, if I ever got one. It's nobody's place to tell me what I should prefer, or say I need to get some creepy-sounding thing just to be woke.
Concentrate harder; it also sounds like a high pitched male.
Basically, for me, I flip flop between two genders. For me, there is no such thing as "genderless" voice, just "gender meta-stable" at best.
Another thing to consider is that adult female voices sound like boy's voices. In animation, boy characters are often acted by women (who aren't doing anything to change their voices).
So "gender neutral" easily has the interpretation of "between boy and man".
I think it sounds feminine in a couple senses, but male in others:
* The speaker speaks less from the chest and more from the throat
* In terms of enunciation, females tend to speak more clearly on sounds such as "t"s, which can be observed here.
* The voice sounds a little more male in that the ends of many words trend lower in pitch, but more female in that they "bottom out" very early in the descent and dip heavily into vocal fry.
* There is less range of tone, which is more characteristically male.
All this aside, I wouldn't be surprised if it's actually multiple people. And I suppose the fact that we all disagree is evidence that they're doing something right. But I think most of us can agree it still sounds really wrong in some way.
To be fair, most cartoon boy characters don't really sound like real boys. If you listen to their voice without looking at the character, it becomes obvious that it's a woman.
They are close enough that you don't have to deal with the nightmare that is child talent.
I was expecting something sounding like whispering, or Wookie, or maybe a child.
It sounds a lot like a woman to me. Perhaps in the lower frequencies it sounds like an older male teenager.
To be quite frank, I think the goal is quite dehumanizing and I don't see how this helps, really.
If fighting racist stereotypes and prejudice is not achieved by means of avoiding the issue, why would it be any different with gender? Aren't simple respect for the other and fostering a culture of dignity the true goals?
Our brain's lock-in heuristics will always find a way to reinforce contrasts so that we can navigate the social world. If not gender, something else. What is needed is conscientiousness.
I love to quote Beauvoir's plato page [1], which is particularly well written I find (emph mine), and relates very well to what you've just brought up:
"Before The Second Sex, the sexed/gendered body was not an object of phenomenological investigation. Beauvoir changed that. Her argument for sexual equality takes two directions. First, it exposes the ways that masculine ideology exploits the sexual difference to create systems of inequality. Second, it identifies the ways that arguments for equality erase the sexual difference in order to establish the masculine subject as the absolute human type. Here Plato is her target. Plato, beginning with the premise that sex is an accidental quality, concludes that women and men are equally qualified to become members of the guardian class. The price of women’s admission to this privileged class, however, is that they must train and live like men. Thus the discriminatory sexual difference remains in play. Only men or those who emulate them may rule. Beauvoir’s argument for equality does not fall into this trap. She insists that women and men treat each other as equals and that such treatment requires that their sexual differences be validated. Equality is not a synonym for sameness."
Anyone citing platos republic do so with great disservice if they don't mention that it use satire as a mechanism for political commentary. Below is a few quotes and extracts from plato's republic, and just imagine this being spoken with a straight face.
The guardian class is a selected breeding program where the best women and best men exclusive breed with each other, and any deformetives or lower quality personage got put in an “unspeakable and unseen place”.
When a child is born it get immediately taken from the parent into the hands of breeding pens to be trained into protecting and valuing the collective community above all else. The breeding pens is located in a certain section of the city and apart from the guardians in order to preserve equality and avoid personal possession from the parents.
The children, being pre-destined to participate in wars, should accompany and observe the guardians in battle as “spectators of war”. The involvement of the children in war serves as an opportunity for the Republic to instill a sense of patriotism to the state and admiration of the mighty Guardian class. They should help out and serve in the whole business of war.
To sum up: We are talking about Eugenics, genocide, removal of personhood and individualism, extermination of the concept of family, and creating child armies from the elite. This in a time period where nobility, monarchs, family and arranged marries was the focus point of high society, and the bulk of armies came from lower classes. It is about as subtle political commentary as Monty Python, which is why drawing conclusions such as "the price of women’s admission to the guardian class is that they must live like men" is supported by very loose ground. At best it is just show the cultural assumption of connecting war with masculinity, and at worst asserts that personhood and individualism is feminine.
I tend to agree, it sounds more like a woman's voice to me. I guess some people will hear a man's voice. It will be like "the dress is blue/yellow".
Case and point, this is probably not perceived as "neutral" but either male or female, which kind of defeats the point.
As a side note, it doesn't feal natural at all, which is a big downside in term of acceptance.
In term of objectives, I consider the fact that the vast majority of vocal assistants have female voices more of a symptom rather than a cause, specially since they are not (yet) ubiquitously used in homes. It's more indicative of deeper issues in our societies regarding gender equality. On a very simple level, it's a basic reflection to the fact most human secretaries and assistants are women.
I would love to see as many voice assistants using male voices as ones using female voices. Jarvis in the "Iron Man" series kind of proves this can be done. Pushing for a neutral voice is more of a band-aid.
There are languages, ie cultures, where iphones default to male. Since it is not random, the assumption must be that market studies and tests has shown a male voice being favored in those countries vs a female voice in other countries. It more than proves that a male voice can be used as default.
The discussion of which gender the voice assistant defaults to can basically be distilled down that people rating the experience of using the female version higher or lower than the male version in specific cultures. The question is why people of some cultures are biased towards a female voice, while other countries are biased towards a male voice.
I think there is nothing wrong with dehumanising and degenderising computer assistants, they are neither human nor do they have gender.
We tend to think making computer interfaces appear human makes it easier for us, but I disagree. We have to remember we are not talking to a person, but a thing. It has no feelings, no real personality and no way to interpret what we're saying other than a series of orders.
What strikes me as odd is the objective of making the interface more human-like in some dimensions but less in others. Like you say, no one's talking to a human, but a thing. Users should be aware of that whichever the case.
And I should add that the onus of not perpetuating vicious stereotypes should fall on the shoulders of the users, not the designers. It seems to me that the idea of ever needing a genderless voice assumes that people are the passive victims of some evil ether that permeates their demeanour.
This is not to say that the project is not interesting - trying to capture the essences of human speech is quite a noble endeavour in itself, and it makes us more in touch with who we are.
Surely you'd refer to this as a "sexless" voice. Even if you believe that gender and sex are unrelated, voice is a product of your physical body and determined by sex. It's not affected by gender. In my opinion all this does is highlight that the "he or she" mindset is an unavoidable part of our reality, which is presumably the opposite of the intention.
Gender, ie cultural context, does in fact play a role on how we speak and both the pitch and intonation of how we speak.
You ever hear of upspeak/uptalking? Vocal fry? It's not so simple as just sex driven characteristics. (Although testosterone does impact the development of vocal chords.)
Slightly, but what we consider male and female voices are correlated with sex and not gender. This is apparent in almost any transexual person, unless they happen to have an androgynous voice or physical alterations to their vocal cords.
Interesting concept, but it sounds slightly female to me.
I wonder what difference is there between male and female voices besides pitch (excluding cultural differences in way of speaking)? If some are exclusively binary, using or not using it will obligatorily lead to one or the other, and if so, a totally "gender" neutral voice becomes impossible.
(I used gender enclosed in quotes because I'm only talking about the general biological differences between male and female voices, not elements such as way of speaking, choice of words, etc)
Maybe one approach could be mix aspects from both male and female voice.
I've heard it's not just the average pitch that varies; female voices also tend have a wider range of pitch. In other words most men speak in more of a monotone than most women. Whether this is biological or cultural, I don't know.
It sounds feminine to me. I predict that the truth is in the ears of the beholder, and subconscious bias will cause people to sense a more masculine or feminine voice.
I agree, perception probably plays a lot into this. It seemed mostly feminine to me but wavered a lot between what I perceive feminine and masculine voices are. This doesn't seem like the answer but I agree with whoever created this in that it would be nice to have non-gendered computer voices. My reasoning is just that I don't want to give machines a fake gender - it's something that's always bothered me. That thing is not a him nor a her.
If we could figure out how to make a genderless voice, then it would be intolerant. This is not a gender neutral voice. It sounds like it's cycling between female-type and male-type voices with a very strong bias towards female-type.
I applaud the effort, but it falls very short. The voice itself is intolerant because it's almost a parody.
That pitch control isn't really controlling musical pitch. It seems to mostly moving the peak of a band-pass filter. The actual change in pitch is very slight.
Think about it; if you change the pitch of a note from 100 to 150 hz, that's should sound like someone is singing a perfect fifth interval. Nothing like that is heard here.
To those asking: the problem with gendered voice assistants is that they are almost always female. The argument is that this trains people to think of women as subservient, or things to be ordered around.
I love the idea behind this, although I don't necessarily agree that they have achieved their goal, but perhaps that says something about me more than it does about the service itself. To me, this still sounds like a female voice, just pitch-shifted.
> The argument is that this trains people to think of women as subservient, or things to be ordered around.
But isn't that like people who assumed that violent games make people more violent (which was proven false in research)? To me this looks more like pushing some agenda trough from some activists in the gender debate.
There are technical reasons to use a female voice. A female voice better matches the frequency response of the human ear than a male voice, and the frequency response of small loudspeakers like those found in phones.
However, a pitched-up male voice could have the same frequency content, and it wouldn't sound any weirder than this "neutral" voice.
> The argument is that this trains people to think of women as subservient, or things to be ordered around.
Sounds like an extraordinary claim going unsubstantiated. Unlike that claim, the reason why voice assistants are overwhelmingly female is well known. It's because both users of both sexes overwhelmingly prefer a female voice in focus groups. That facet of Siri et al has been studied very thoroughly.
And around the other way, the reason people respond better to female assistant voices is because our culture has historically treated women as subservient. This important change (to our old gendered culture, and new nearly universally female automated assistants) is coming, so at least Q is pushing the conversation.
Great, but there's a whole bunch of languages out there differentiate between male and female spoken language. Bit disappointed that it doesn't mention anything regarding that.
Another vote for the voice sounding like a deep voiced Female.
Also, if their criteria for recording was what people "identify" as that's not going to help much.
They would need people who are biologically both, hermaphrodites, or neither, neuters (i.e. hormonally castrato). And given that humans are by default biologically female (i.e. without hormones a human becomes female) I'm not surprised the voice sounds female.
Also, their "future" of people not identifying as male or female (I assume they mean that most people would be like that, not a minority) is not a future I desire.
Having males and females makes life far more enjoyable (for most).
The minority where that doesn't work for them shouldn't have a problem with the fact that others are. Anymore than someone atheist should have an issue with someone religious.
People (or voices) don't have to be similar to you in order for you to interact with them.
It doesn't sound genderless to me, it sounds like something that just fails to be both male or female.
What I actually find the weirdest is when my ears hear female, it hears something British or South African. When my ears hear male, it hears a more American voice.
So is this just synthesised from two different speakers with different accents?
I like the idea and I'm glad there is effort being put into this.
I imagine though if one of the big players would implement this we would see an insane amount of public debate and media coverage about whether the voice sounds like a man or a woman.
It seems like the typical "tech tries to solve social issues"
Recently, United Airlines decided to allow for a non-binary gender option for booking flights. All the comments on social media were people screaming about how the liberal agenda was coming to get them, how there's only two genders, get over it, etc. "What, am I supposed to guess if that's a man or a woman sitting next to me now?!" But nobody is forcing these people to pick a new option. They can just select the one they've been selecting their whole lives. It literally does not affect them at all. But they still have an irrational hatred for an irrelevant option.
If some people want a genderless voice, yeah, I think tech companies should adopt it. If it makes someone more comfortable, and it doesn't harm me in any way, I'm not going to bag on it.
This is really good. I've been raising the voice from up and down to go from make to neutral to female and then let it drift in and out. I feel like that was the best. It was great paying attention to familiarity while also getting neutral.
That doesn't sound like a genderless voice. It sounds like a mix of gendered voices. You move the pitch up and down to make it sound more "male" or "female", and it's supposed to sound "genderless" when it has an equal mix of both. You can clearly hear the different voice strands superimposed on each other, like a sort of chorus.
In any case, a blend of male and female is still just a blend of male and female. Genderless should stand for something that is neither male nor female, and I can't really see how you can achieve that by mixing male and female in any proportions.
The number of people saying that this voice sounds absolutely female is just an extension of the tendency to imagine anything even slightly deviating from the most stereotypically masculine as feminine by default. The one comment that I saw comparing this to a masculine voice was comparing it to a "gay flight attendant" that they had on a flight recently - and there, again, is the tendency to assume someone's gay just because they deviate slightly from the stereotypically masculine. The voice sounds genderless to me.
It seems that most people don't really like the way this sounds, but honestly I had the opposite response. I despise the 'natural' sounding synthetic voices because there always seems to be something off about it. The fact that this sounds so weird is what I appreciate about it.
Perhaps one day we'll have natural sounding voices, but until then I'd always prefer a more artificial sounding voice. Something something uncanny valley, I guess.
It sounds like a poor attempt at voiceover editing. It's nasally and monotone and dull and dreary. There's this strange vibrato throughout it all and there are sudden rapid shifts between a really low and high tone of voice. I'd honestly much rather have something blatantly robotic than this almost successful voice, because all I feel is that I'm being played for a fool.
Gender issues aside, speaking as a daily text-to-speech user this voice doesn't sound particularly natural or pleasing. The accent seems to constantly shift between American and British, and there is some weird audio phasing going on. It sounds more like a Hollywood director's idea of what robots should sound like than something I'd actually want to listen to.
The Hacker News crowd is very concerned about people "inventing something that doesn't need to be invented" when that thing strays anywhere near social and cultural issues rather than a shiny new startup or web framework.
If you're not sure what problem it solves, just google 'alexa female' and you'll get plenty of articles to read, and, hopefully, consider.
just google 'alexa female' and you'll get plenty of articles to read
Generally from those who have decided that their job is to find ways to be offended by everything. If Alexa and Siri had male voices by default, I have no doubt that the same people would be complaining that tech bros are erasing the historical contributions that women made as secretaries and assistants.
> Generally from those who have decided that their job is to find ways to be offended by everything.
You know, I never understood why people trot this idea out so often. It's so irrational.
If people were rational, they would say "hmm, X doesn't offend me, but many people say that X offend them; perhaps they have a point."
Instead, only those who are incapable of understanding their own confirmation bias say "All these people say that X offends them; how silly."
I think this is a highly irrational take.
The better take would be: "Some people seem to be offended by everything; other people are not. What are the principal factors of a person that go into predicting 'will this person be offended by X'?"
There's a lot of people saying that alexa is a girl because it's perceived as more helpful. But have we thought about the fact that maybe it's perceived as more nice or as less threatening? There's a reason movie villains are stereotypically male voice actors - deeper male voices sound less friendly, forget helpful.
Is that natural instinct or societal training? If we start using non-female voices for voice assistants, maybe we can train people that a voice doesn't have to be female to be helpful. Otherwise if we use female voices because society is just used to it, then we're just reinforcing an unnecessary stereotype for convenience.
I like this, but it's probably too late for all tech companies, except maybe Google.
I was pretty disappointed that all the major voice assistants have female voices and most of them have explicit female personas and names. I find it a little creepy to be honest.
At least Google didn't go for the whole personality crap. They could switch Google Assistant to this type of voice and it'd fit in. You can't switch Alexa to this type of voice while keeping that name.
> when that thing strays anywhere near social and cultural issues rather than a shiny new startup or web framework.
I don't know what you're reading, but HN is generally highly skeptical of those too, perhaps overly so. This is especially true with your web framework example.
It's simply not an invention or even something to go on about.
You hear 'genderless' voices every day, which is to say, voices which one couldn't place as being male or female with any reasonable accuracy.
You'd be surprised at who in your immediate vicinity has more or less a non gendered voice if you didn't have all the other 'clues' as to their gender.
In the production of synthetic voices, surely we've come across many which are de-facto genderless already.
As a blind person I'm going to have to pretty strongly disagree with this. A vast majority of the voices I hear do in fact read as male or female even without visual indicators.
It sounds like a deep female voice, not just in terms of frequency but timbre; perhaps a transwoman who has undergone a lot of voice training. Even deepening it into the male range didn't eliminate the female timbre.
Colossus (from Colossus: the Forbin Project) arguably had a more genderless voice than this.
"Technology companies often choose to gender technology believing it will make people more comfortable adopting it.
Unfortunately this reinforces a binary perception of gender, and perpetuates stereotypes that many have fought hard to progress.
As society continues to break down the gender binary, recognising those who neither identify as male nor female, the technology we create should follow.
Q is an example of what we hope the future holds; a future of ideas, inclusion, positions and diverse representation in technology."
If technology is needed to create this kind of voice -- IE it's not really natural, and as others have suggested it creates an uncanny valley of voice that confuses people, wouldn't that actually perpetuate the idea that there is a gender binary?
(I don't have a dog in this fight, I just don't see that this accomplishes what you want it to)
> stereotypes that many have fought hard to progress
What stereotypes would this help to alleviate?
> the technology we create should follow.
Why it's a service that communications via artificial audio. Why is it important that you can't have a male or a female voice? People still exist, the voices in question don't have a physical appearance.
> As society continues to break down the gender binary, recognising those who neither identify as male nor female, the technology we create should follow.
People who neither identify as male nor female AND want to listen to a gender-neutral voice seems like a very niche market.
it is an interesting idea. technology of course has no gender, and any synthesized speech can only express an arbitrary simulation of gender, so making this explicit reveals the underlying reality that all technological simulations of gender are by definition constructed. a piece of art that demonstrates the construction of gender probably will not catch on well as a product though.
I instantly recognized the voice as M2F, before it even confirmed that by the description of how it was created. If I had a device with this voice, I would feel a strong urge to reenact the final printer scene from the movie Office Space.
Companies want their voice-enabled products to be purchased. Revulsion doesn't help.
And silly, too. Regardless of what is going on inside a person's head (which we can choose to respect, absolutely), biologically there is definitely just two genders. You can identify however you like, modify if you wish, but in the end your body will be one or the other, right along with your voice.
> A computer assistant doesn't have a gender so why should it represent itself as male or female
A human has to be in the recording booth to create the vocal library. Also you can change the Siri voice to not only be Male, but also have a different accent.
The page asserts that audio content to many listeners is firstly "defined" (though perhaps "subconsciously emotionally engaged with to some extent" would be more accurate) by perceived gender, before semantics. The notion is that by adding a genderless/neutral voice, this would be removed. Of course, you then need to discard genders in names as well, hence 'Q'. It's pretty forward thinking... I don't doubt assistant tech may go this way. However, wouldn't "dehumanization by association" for non gender identifying individuals be a risk if that popularizes? A nice thought-provoking art project, if nothing else.
It's the least gendered synthetic voice I've heard but there is still a slight wobble between being first a hint more male and then a hint more female.
Also seems to me like it's somewhat camp - a neutral voice in that regard would also be worth exploring
It sounds a bit like a vocal Necker cube illusion. (I.e. I find myself swapping between either male or female depending on how I am listening to it.) It's a bit like the Yanny/Laurel thing for me where I hear both.
this is really cool. my personal reaction was to feel it was a little odd and off, but as I moved the Hz meter up and down it went from "oh a womans voice" to "oh a mans voice" pretty clearly. that is just plain interesting to know, I'd never seen how clearly quantifiable and single-axis that distinction can be. I bet if i spent a few weeks or so using a device in those Hz ranges i'd get used to it.
While we're at this, I'd rather prefer that my virtual assistant would be based on Majel Barrett-Roddenberry voice she used in ST LCARS voice (?) interface. Pretty sure she did record enough lines to synthesize her voice.
It's not something most people support, including most LGBT activists, notwithstanding that they may believe otherwise.
If you pay close attention, most of the cultural debate over gender identity issues relates not to eliding distinctions and expectations, but to reshaping the boundaries and expectations of those identities.
It's easier to see what I mean by this with the debate over racism and affirmative action. In the U.S. most white people think equality means being "color blind"--so blind that some refuse to even admit that non-overt racism is possible. In truth civil rights in the U.S. has never been about eliding or ignoring [racial] distinctions, but changing expectations. As long as power differentials exist (and they will always exist), these expectations will matter and will constantly shift.
Another way to look at it is that any identity is by definition exclusive. And identities only matter because they imply a default set of expectations. If you say that people should be literally free to choose their identity, you're allowing people to freely reinterpret and redefine the implied expectations and shared _meaning_ of that identity, which can be injurious to those who previously chose that identity. There's an inherent conflict, there. It's why as a white male I can't go around identifying as a black female, certainly not without making substantial changes to my behavior and perhaps even physical appearance, changes that should be consonant with what being a "black female" means. To do otherwise would be another kind of violence.
You can't really understand any of these issues without understanding and appreciating the genesis--power. It's all about power over others. We can never have a society without power differentials because that would imply the power to _refuse_ to recognize the identities others have chosen for themselves. Power matters not because it's per se evil, but how it's used and how it's distributed and what are the concrete results in people's lives.
In adjudicating conflicts over what the expectations for an identity are, one must take into account how power is used and abused and by whom. For example, it's legitimately problematic for a man to argue that transgender women shouldn't be allowed in women's bathrooms because of the threat of violence. It's problematic because it's another instance of a man making decisions about what it means to be a women, and men have historically abused their social and economic positions when drawing these boundaries to disadvantage women more generally. The logic of the argument is secondary to the power politics, because fundamentally it's really about the power politics. It's more legitimate for women to make that argument, and in so much as they do make it it carries more weight (notwithstanding that women as a group are certainly capable of abusing their relative powers), but in actuality they've been less vociferous about it.
The line of thinking that says power differentials are inherently wrong and/or that we should elide the implications of identities are inconsistent. If only things were so simple....
Supporting that for people who want it is one thing, living in a world where we act like there is no gender is another. A lot of us enjoy being and/or identifying as male/female.
> The key word there is defined. Defined by gender means boxed in by the expectations that being a male or female thrusts on us.
Identity is by definition a set of expectations. To lose those expectations is to lose a facet of identity.
There's a group of people who want to remove that facet of their identity, cool, I support them. But I like that facet of my identity, and I like the fact that most people around me also express that facet of their identity. I don't want want to be genderless or to live in a genderless world. Do not marginalise me for being that way.
Typical meatbag voice. The future belongs to formant synthesis, I can easily get to 1000wmp without any loss of coherency and the sound is fundamentally inhuman.
It's intentionally made to be hard to listen to. OP spliced multiple copies together with a slight offset, varied the rate while recording, and intentionally selected a passage which repeats very similar words in proximity, which further increases the difficulty.
> Formant synthesis does not use human speech samples at runtime. Instead, the synthesized speech output is created using additive synthesis and an acoustic model (physical modelling synthesis. Parameters such as fundamental frequency, voicing, and noise levels are varied over time to create a waveform of artificial speech. This method is sometimes called rules-based synthesis; however, many concatenative systems also have rules-based components.
This sounds fine to me. I don't care if a clearly artificial voice doesn't sound human. Computers should have their own voice.
I think the stated purpose is stupid, though. Gendered roles are good. They give people a purpose and make them happier. I really pity the little girls who get forced into things like engineering just because their parents wanted to be hip and "genderless".
I've given input on political issues as well as gender issues under multiple identities and personas to gain consensus
Almost no progress on consensus regarding political issues but on some sex and gender issues the male persona was unqualified to "mansplain" or had to much privilege to contribute or complain, where the same information was agreed upon and deemed helpful coming from the non-male persona - assumed to be female
This is more about the decision path people make to discredit a someone. When confronted with information that deviates from your script, people typically look for faults in the messenger instead of the actual information. You look for words that a different party would use exclusively to pigeonhole the messenger as the other, you look for other biases.
I felt that even if an A.I. was created to try to aggregate perspectives and offer its observations, people would try to disagree with it based on who created it. A valid criticism, but what would be left with to offer potentially unifying perspectives.
I like this and I imagine how it can be trained to create a voice that you agree with, but that nobody else would hear.
..... based on your cookies.
Then everybody gets to hear a different voice and mostly the same information - perhaps slightly different words as to not undermine the illusion.
I would humbly suggest that the bevy of reactions to this innovation that amount to 'eww ugh this creeps me out' is evidence of precisely the cisheteronormative bias that this voice was designed to mitigate. I invite you to introspect about your (dis)comfort around folks who are gender-nonconforming.
I rather a female assistant voice, because mothers in human societies are women, and we're hard-wired to find motherly voices more soothing.
This is getting out of hand, it's unnatural, fits strictly in the uncanny valley, and is borderline disgusting; both in terms of sound and in terms of politico-cultural values and intentions.
I was very curious to see how this sounded.
They did a "great" job... for me, it's exactly ambiguous between a high-pitched male voice and a deep-pitched female voice. It's like the old "is it a vase or two faces?" illusion, or "which way is the ballerina spinning?" It keeps flipping in my brain.
But for the same reason, I would never choose it because it sounds so out-of-the-ordinary, an extreme outlier for human voices. When a computerized voice is speaking, I want it to sound "natural" (e.g. common, unexceptional) so that I'm paying attention to the message, not the voice.
Unfortunately a single "gender-neutral" voice turns out to sound quite uncommon, very exceptional. I only know a couple of people who sound like that, out of 1000's. I applaud the intention, but unfortunately I don't think it works in the end.
Regularly or randomly alternating between "average" female and male voices feels like perhaps a better practical solution, matching the two "humps" in the bimodal distribution of actual voice frequencies. [1]
[1] https://erikbern.com/2017/02/01/language-pitch.html
To be honest, it’s not really all that unusual once you start paying attention to it — plenty of women have deeper voices than mine.
I transitioned from male to female some years back and trained my voice from a deep, masculine voice to one solidly in the female range. Pitch is only one aspect of how we gender speech, and I think that’s why it feels off to you.
I would say this voice uses pretty male-coded speech patterns with a higher pitch than you’d normally hear a man speak in. The pattern of clear starts and stops between words and limited range of pitch used sounds “male” to my ears.
>> The pattern of clear starts and stops between words and limited range of pitch used sounds “male” to my ears.
This sounds like an exaggerated stereotype of how men speak. Interestingly, the stereotype is different for men from Southern European countries who are said to speak in a sing-songy voice with vowels flowing together between words like vocal ligatures.
For example, this stereotype is used for satirical effect in the following strip:
https://pbfcomics.com/comics/automatic-business/
6 replies →
I'd guess you're noticing voices a lot more because of your efforts. I don't really notice at all, possibly because I'm not paying attention. I wonder if I heard the Q voice in normal usage (like on a phone system) I'd notice?
1 reply →
Piggybacking this comment to say, has no one noticed that you can drag the little bubble around to adjust the frequency of the voice? It defaults to 153Hz, and varies about 145-175Hz. Which indeed does seem to be right in the middle, but with the 153Hz being more on the "Male" end to be honest.
I had noticed it, but the voice sounds female to me even when dragging the bubble way down into the "male" region.
Dragging it all the way down left sounds like a woman with a deep voice.
I agree and I don't really see the value in making this a battlefield. Pick a male or female voice, whatever you like best.
Having an AI-voice sound artificial could be an advantage in certain situation but I doubt that was the goal here.
Hmmmm, I think there is more to gendered speech than pitch or frequency. It's about subject and delivery. Anyone a fan of The Venture Bros.? Dr Girlfriend is a female character voiced by a male actor (Doc Hammer) Although the given voice is that of a woman with an incredibly exaggerated smokers voice; think patty and selma from Simpsons multiplied by each other. The characters voice is obviously male but the delivery is effeminate which is why it works so well. At least I always heard it as a very exaggerated voice of a woman who smoked heavily.
My take on the voice? Sorry, to me it's an effeminate male voice. I hear a british woman overlaid with Fez from that seventies show. In fact I was on a flight yesterday with a very obvious gay male flight attendant who sounded very similar to this. Think man's voice with a little helium in the air.
> … a very obvious gay male flight attendant
I’m not a fan of these kind of characterizations. I can think of two things that make someone obviously gay: (1) they tell you, or (2) you see that person engaging in homosexual conduct.
Nearly anything else is probably a prejudicial stereotype.
This is indeed a complex theme, in part because how it is linked to negative stereotypes and in part because people might not realize they are sometimes reinforcing the negativity in them.
It is not controversial to say that an obvious metalhead is indeed a metalhead.
Also I would like to point out that "(2) you see that person engaging in homosexual conduct" is not that much of a silver bullet here, as bisexual and trans people exist too :)
Also sometime it is not hard to know more someone that they know themselves, I had a couple friends that were known to be gay before they knew themselves. And also "a very obvious gay male" does not mean homosexual to many people, in the last years society and the internet became much more mature in distinguishing masculine/effeminate stereotypes from actual sexual orientations. I have no idea of what went through GP minds, but I believe that being able to separate the "very obvious gay" from a description of a sexual orientation into a personality trait can be very beneficial for society in the long term.
5 replies →
https://en.wikipedia.org/wiki/Gay_male_speech
Whether you are a fan of those characterizations or not. It's real.
6 replies →
Sexuality is not just about intercourse. It is also a form of culture, identity, and expression.
Sounds female, usually, but clearly wavers between male and female. Primarily though it sounds unpleasant to me. Not a fan.
This is [uncanny valley][1] applied to voice. They intentionally made a voice that listeners can't quite identify, and it ends up feeling unnatural.
[1]: https://en.wikipedia.org/wiki/Uncanny_valley
This view of the future seems very bland and uninteresting.
I thought the idea is to make gender a flexible thing, for example you can experiment with being a female or whatever gender profile, based on whatever you feel comfortable with.
Not to reduce the worlds gender representations (in business, culture, etc) into a homogeneous middle ground? Expanding what's possible in terms of personal expression rather than (further) restricting it. So I'm very confused why this exists other than as an interesting technical project?
I agree. But most of the time it just sounded like the typical deep female computer voice that scifi has used for decades.
I had the same experience. I hope this isn't going to spin off more academic work about why our brains are hard-wired to be sexist by continually and automatically trying to perceive the gender of speaker based on their speech.
If we really want a genderless assistant voice bring back the novelty voices from Mac OS X like Zarvox, Pipe Organ, and Cellos.
https://youtu.be/Z4hqUxb9MmY?t=186
I like the voice "Ivy" from Amazon's Polly.
It sounds a little bit like the voice of a moppet-type character from League of Legends that just huffed helium. Because it is so stylized you don't have the problem of a "perfect english gentleman" who pauses unnaturally, sometimes makes vowels like T-Pain, etc. that you have with more realistic voices.
It definitely comes across as female to me, but it is cartoony enough that it not so culturally coded as female -- so there is less of a sense of "female servicing" as there is with most voice assistants.
Decepticons, attack!
* Soundwave G1 voice compilation part I - YouTube || https://www.youtube.com/watch?v=qOu3hYXE_tM
Although these are gimmicky the sort of route of approaching the problem from what would a machine or non-human AI voice sound like would actually get you the goal of a Genderless Voice.
As it stands Q isn't genderless it's just gender non-conforming, which is a great alternative but I'd like to see some work into an actual "AI" voice rather than these voice assistants that try to imitate a human on the gender spectrum.
+1 for whisper.
Yeah. For some reason it reminds me of S.E. Lain.
1 reply →
My brain has no trouble wrapping around the genderlessness of the voice Colossus from The Forbin Project:
https://www.youtube.com/watch?v=iRq7Muf6CKg
A non-human voice is a valid goal, but a human non-gendered voice is a different thing.
Please recognise that agender people are both human and genderless.
I never said they aren't. Nice virtue signaling though!
I mean no offense, but how is this really solving any problems? How is inventing a new voice many people are uncomfortable with really making the world a better place? A user configurable male/female voice could just work.
The idea, I think, is that it matters that/if we gender our digital servants—that as we use them more it will affect our sense of gender rolls. That Siri and Alexa are both female by default puts them in the stereotypical female-secretary roll.
I have Waze speak as a British male, of course then he sounds like a stereotypical butler. It would be fun if I could set it to be Batman or HAL9000, but it would probably get old.
And it's the same wrong idea as violent video games making children more violent. Both (a) fail to appreciate substantive causes and relationships, and (b) discount the nuanced contextual distinctions that people (even children) naturally make.
Aside from the motivations of the engineers (which are unclear and maybe even irrelevant), I think it's interesting technical work, including taking a stab at gender perception in speech patterns with a bold application.
I mean, to really address this problem, don't we need genderless employees in the workplace?
2 replies →
I vote for Jeeves.
> That Siri and Alexa are both female by default puts them in the stereotypical female-secretary roll.
But isn't this happening only if some marxist is searching for this "problem"? I never considered it this way, until your comment.
It's like with the service OP is presenting... they are offering "solution" to made up problem.
How are you solving any problems?
Edit: my point being that it's a rude and combative way to pose that question
what
I'm not sure I see the problem with a voice assistant sounding male or female, and I certainly wouldn't buy this. It just sounds wrong, and I can't quite say why. I guess after all the work tech companies have put into trying to sound like a person, this deliberately doesn't.
Also, it sounds blatantly female but pitched lower. Even when I use the little drag thingy to pitch it as low as it will go, it still sounds female. I think most of us know pitch alone doesn't define voice, and that's clear from this.
>Also, it sounds blatantly female but pitched lower.
That’s exactly what I hear. I’m not annoyed there is a market for such frivolous “woke” things, but I am annoyed it seems to be a mediocre attempt at it, it’s entirely female to me.
There is no problem with a voice assistant sounding male or female; what's the problem with some other voice assistant sounding genderless?
> It just sounds wrong, and I can't quite say why.
Maybe you're used to all people sounding male or female, so sounding like neither seems unfamiliar.
> There is no problem with a voice assistant sounding male or female; what's the problem with some other voice assistant sounding genderless?
This is literally the whole problem. It's less that it's genderless in a robotic sense and more that it's a hybrid manipulation of actual voices. It just sounds wrong.
> Maybe you're used to all people sounding male or female, so sounding like neither seems unfamiliar.
Maybe so. We try to make our robots imitate humans, so of course they sound male or female. From what the website said, it might also be flipping between. But even if I am just used to it, so what? That's how I'd prefer my voice assistant to sound, if I ever got one. It's nobody's place to tell me what I should prefer, or say I need to get some creepy-sounding thing just to be woke.
> it sounds blatantly female but pitched lower.
Concentrate harder; it also sounds like a high pitched male.
Basically, for me, I flip flop between two genders. For me, there is no such thing as "genderless" voice, just "gender meta-stable" at best.
Another thing to consider is that adult female voices sound like boy's voices. In animation, boy characters are often acted by women (who aren't doing anything to change their voices).
So "gender neutral" easily has the interpretation of "between boy and man".
> it also sounds like a high pitched male
I think it sounds feminine in a couple senses, but male in others:
* The speaker speaks less from the chest and more from the throat
* In terms of enunciation, females tend to speak more clearly on sounds such as "t"s, which can be observed here.
* The voice sounds a little more male in that the ends of many words trend lower in pitch, but more female in that they "bottom out" very early in the descent and dip heavily into vocal fry.
* There is less range of tone, which is more characteristically male.
All this aside, I wouldn't be surprised if it's actually multiple people. And I suppose the fact that we all disagree is evidence that they're doing something right. But I think most of us can agree it still sounds really wrong in some way.
2 replies →
To be fair, most cartoon boy characters don't really sound like real boys. If you listen to their voice without looking at the character, it becomes obvious that it's a woman.
They are close enough that you don't have to deal with the nightmare that is child talent.
1 reply →
I was expecting something sounding like whispering, or Wookie, or maybe a child. It sounds a lot like a woman to me. Perhaps in the lower frequencies it sounds like an older male teenager.
To be quite frank, I think the goal is quite dehumanizing and I don't see how this helps, really.
If fighting racist stereotypes and prejudice is not achieved by means of avoiding the issue, why would it be any different with gender? Aren't simple respect for the other and fostering a culture of dignity the true goals?
Our brain's lock-in heuristics will always find a way to reinforce contrasts so that we can navigate the social world. If not gender, something else. What is needed is conscientiousness.
I love to quote Beauvoir's plato page [1], which is particularly well written I find (emph mine), and relates very well to what you've just brought up:
"Before The Second Sex, the sexed/gendered body was not an object of phenomenological investigation. Beauvoir changed that. Her argument for sexual equality takes two directions. First, it exposes the ways that masculine ideology exploits the sexual difference to create systems of inequality. Second, it identifies the ways that arguments for equality erase the sexual difference in order to establish the masculine subject as the absolute human type. Here Plato is her target. Plato, beginning with the premise that sex is an accidental quality, concludes that women and men are equally qualified to become members of the guardian class. The price of women’s admission to this privileged class, however, is that they must train and live like men. Thus the discriminatory sexual difference remains in play. Only men or those who emulate them may rule. Beauvoir’s argument for equality does not fall into this trap. She insists that women and men treat each other as equals and that such treatment requires that their sexual differences be validated. Equality is not a synonym for sameness."
[1] https://plato.stanford.edu/entries/beauvoir/
EDIT: from an artistic point, imagining what a sexless voice would sound like is an interesting project !
Anyone citing platos republic do so with great disservice if they don't mention that it use satire as a mechanism for political commentary. Below is a few quotes and extracts from plato's republic, and just imagine this being spoken with a straight face.
The guardian class is a selected breeding program where the best women and best men exclusive breed with each other, and any deformetives or lower quality personage got put in an “unspeakable and unseen place”.
When a child is born it get immediately taken from the parent into the hands of breeding pens to be trained into protecting and valuing the collective community above all else. The breeding pens is located in a certain section of the city and apart from the guardians in order to preserve equality and avoid personal possession from the parents.
The children, being pre-destined to participate in wars, should accompany and observe the guardians in battle as “spectators of war”. The involvement of the children in war serves as an opportunity for the Republic to instill a sense of patriotism to the state and admiration of the mighty Guardian class. They should help out and serve in the whole business of war.
To sum up: We are talking about Eugenics, genocide, removal of personhood and individualism, extermination of the concept of family, and creating child armies from the elite. This in a time period where nobility, monarchs, family and arranged marries was the focus point of high society, and the bulk of armies came from lower classes. It is about as subtle political commentary as Monty Python, which is why drawing conclusions such as "the price of women’s admission to the guardian class is that they must live like men" is supported by very loose ground. At best it is just show the cultural assumption of connecting war with masculinity, and at worst asserts that personhood and individualism is feminine.
That's a wonderful point; I need to get around to reading Beauvoir.
I tend to agree, it sounds more like a woman's voice to me. I guess some people will hear a man's voice. It will be like "the dress is blue/yellow".
Case and point, this is probably not perceived as "neutral" but either male or female, which kind of defeats the point.
As a side note, it doesn't feal natural at all, which is a big downside in term of acceptance.
In term of objectives, I consider the fact that the vast majority of vocal assistants have female voices more of a symptom rather than a cause, specially since they are not (yet) ubiquitously used in homes. It's more indicative of deeper issues in our societies regarding gender equality. On a very simple level, it's a basic reflection to the fact most human secretaries and assistants are women.
I would love to see as many voice assistants using male voices as ones using female voices. Jarvis in the "Iron Man" series kind of proves this can be done. Pushing for a neutral voice is more of a band-aid.
There are languages, ie cultures, where iphones default to male. Since it is not random, the assumption must be that market studies and tests has shown a male voice being favored in those countries vs a female voice in other countries. It more than proves that a male voice can be used as default.
The discussion of which gender the voice assistant defaults to can basically be distilled down that people rating the experience of using the female version higher or lower than the male version in specific cultures. The question is why people of some cultures are biased towards a female voice, while other countries are biased towards a male voice.
I think there is nothing wrong with dehumanising and degenderising computer assistants, they are neither human nor do they have gender.
We tend to think making computer interfaces appear human makes it easier for us, but I disagree. We have to remember we are not talking to a person, but a thing. It has no feelings, no real personality and no way to interpret what we're saying other than a series of orders.
What strikes me as odd is the objective of making the interface more human-like in some dimensions but less in others. Like you say, no one's talking to a human, but a thing. Users should be aware of that whichever the case.
And I should add that the onus of not perpetuating vicious stereotypes should fall on the shoulders of the users, not the designers. It seems to me that the idea of ever needing a genderless voice assumes that people are the passive victims of some evil ether that permeates their demeanour.
This is not to say that the project is not interesting - trying to capture the essences of human speech is quite a noble endeavour in itself, and it makes us more in touch with who we are.
Surely you'd refer to this as a "sexless" voice. Even if you believe that gender and sex are unrelated, voice is a product of your physical body and determined by sex. It's not affected by gender. In my opinion all this does is highlight that the "he or she" mindset is an unavoidable part of our reality, which is presumably the opposite of the intention.
Gender, ie cultural context, does in fact play a role on how we speak and both the pitch and intonation of how we speak.
You ever hear of upspeak/uptalking? Vocal fry? It's not so simple as just sex driven characteristics. (Although testosterone does impact the development of vocal chords.)
Slightly, but what we consider male and female voices are correlated with sex and not gender. This is apparent in almost any transexual person, unless they happen to have an androgynous voice or physical alterations to their vocal cords.
2 replies →
Interesting concept, but it sounds slightly female to me.
I wonder what difference is there between male and female voices besides pitch (excluding cultural differences in way of speaking)? If some are exclusively binary, using or not using it will obligatorily lead to one or the other, and if so, a totally "gender" neutral voice becomes impossible.
(I used gender enclosed in quotes because I'm only talking about the general biological differences between male and female voices, not elements such as way of speaking, choice of words, etc)
Maybe one approach could be mix aspects from both male and female voice.
I've heard it's not just the average pitch that varies; female voices also tend have a wider range of pitch. In other words most men speak in more of a monotone than most women. Whether this is biological or cultural, I don't know.
It sounds feminine to me. I predict that the truth is in the ears of the beholder, and subconscious bias will cause people to sense a more masculine or feminine voice.
I agree, perception probably plays a lot into this. It seemed mostly feminine to me but wavered a lot between what I perceive feminine and masculine voices are. This doesn't seem like the answer but I agree with whoever created this in that it would be nice to have non-gendered computer voices. My reasoning is just that I don't want to give machines a fake gender - it's something that's always bothered me. That thing is not a him nor a her.
Hmm, I wonder if "genderless" will be more distracting than a clearly male or female voice. It seems unnatural.
I'm half serious when I say this, but calling it "unnatural" can border on intolerance, for somewhat obvious reasons.
If we could figure out how to make a genderless voice, then it would be intolerant. This is not a gender neutral voice. It sounds like it's cycling between female-type and male-type voices with a very strong bias towards female-type.
I applaud the effort, but it falls very short. The voice itself is intolerant because it's almost a parody.
1 reply →
That pitch control isn't really controlling musical pitch. It seems to mostly moving the peak of a band-pass filter. The actual change in pitch is very slight.
Think about it; if you change the pitch of a note from 100 to 150 hz, that's should sound like someone is singing a perfect fifth interval. Nothing like that is heard here.
It came across to me as slightly more feminine, I moved more towards the 140hz range to make it sound more gender neutral to my ears.
I don't know, I thought it was pretty spot on. I thought 140hz sounded consistently male.
To those asking: the problem with gendered voice assistants is that they are almost always female. The argument is that this trains people to think of women as subservient, or things to be ordered around.
I love the idea behind this, although I don't necessarily agree that they have achieved their goal, but perhaps that says something about me more than it does about the service itself. To me, this still sounds like a female voice, just pitch-shifted.
> The argument is that this trains people to think of women as subservient, or things to be ordered around.
But isn't that like people who assumed that violent games make people more violent (which was proven false in research)? To me this looks more like pushing some agenda trough from some activists in the gender debate.
There are technical reasons to use a female voice. A female voice better matches the frequency response of the human ear than a male voice, and the frequency response of small loudspeakers like those found in phones.
However, a pitched-up male voice could have the same frequency content, and it wouldn't sound any weirder than this "neutral" voice.
> The argument is that this trains people to think of women as subservient, or things to be ordered around.
Sounds like an extraordinary claim going unsubstantiated. Unlike that claim, the reason why voice assistants are overwhelmingly female is well known. It's because both users of both sexes overwhelmingly prefer a female voice in focus groups. That facet of Siri et al has been studied very thoroughly.
And around the other way, the reason people respond better to female assistant voices is because our culture has historically treated women as subservient. This important change (to our old gendered culture, and new nearly universally female automated assistants) is coming, so at least Q is pushing the conversation.
So why not just use a male voice then?
Trains is hard job
Sounds like an average 14-year-old squeker who's in the middle of puberty hurling out insults in a video games voice chat.
Great, but there's a whole bunch of languages out there differentiate between male and female spoken language. Bit disappointed that it doesn't mention anything regarding that.
Another vote for the voice sounding like a deep voiced Female.
Also, if their criteria for recording was what people "identify" as that's not going to help much.
They would need people who are biologically both, hermaphrodites, or neither, neuters (i.e. hormonally castrato). And given that humans are by default biologically female (i.e. without hormones a human becomes female) I'm not surprised the voice sounds female.
Also, their "future" of people not identifying as male or female (I assume they mean that most people would be like that, not a minority) is not a future I desire.
Having males and females makes life far more enjoyable (for most).
The minority where that doesn't work for them shouldn't have a problem with the fact that others are. Anymore than someone atheist should have an issue with someone religious.
People (or voices) don't have to be similar to you in order for you to interact with them.
It doesn't sound genderless to me, it sounds like something that just fails to be both male or female.
What I actually find the weirdest is when my ears hear female, it hears something British or South African. When my ears hear male, it hears a more American voice.
So is this just synthesised from two different speakers with different accents?
Huh. Cool, but I'm not really sure what problem it solves.
It doesn’t solve a problem, it’s just not “problematic.”
I like the idea and I'm glad there is effort being put into this.
I imagine though if one of the big players would implement this we would see an insane amount of public debate and media coverage about whether the voice sounds like a man or a woman.
It seems like the typical "tech tries to solve social issues"
Recently, United Airlines decided to allow for a non-binary gender option for booking flights. All the comments on social media were people screaming about how the liberal agenda was coming to get them, how there's only two genders, get over it, etc. "What, am I supposed to guess if that's a man or a woman sitting next to me now?!" But nobody is forcing these people to pick a new option. They can just select the one they've been selecting their whole lives. It literally does not affect them at all. But they still have an irrational hatred for an irrelevant option.
If some people want a genderless voice, yeah, I think tech companies should adopt it. If it makes someone more comfortable, and it doesn't harm me in any way, I'm not going to bag on it.
It sounds female to me, but I like it. It sounds like something out of a ships computer in a sci-fi film.
This is really good. I've been raising the voice from up and down to go from make to neutral to female and then let it drift in and out. I feel like that was the best. It was great paying attention to familiarity while also getting neutral.
Sounds like a South African accent and a recording (or several) at that. Then some filters.
That doesn't sound like a genderless voice. It sounds like a mix of gendered voices. You move the pitch up and down to make it sound more "male" or "female", and it's supposed to sound "genderless" when it has an equal mix of both. You can clearly hear the different voice strands superimposed on each other, like a sort of chorus.
In any case, a blend of male and female is still just a blend of male and female. Genderless should stand for something that is neither male nor female, and I can't really see how you can achieve that by mixing male and female in any proportions.
This is awesome. My computer shouldn't need a gender. It's a computer.
A computer is a masculine noun in two of the languages that I use.
I'm sure you can find plenty of synonyms where it's feminine.
1 reply →
The number of people saying that this voice sounds absolutely female is just an extension of the tendency to imagine anything even slightly deviating from the most stereotypically masculine as feminine by default. The one comment that I saw comparing this to a masculine voice was comparing it to a "gay flight attendant" that they had on a flight recently - and there, again, is the tendency to assume someone's gay just because they deviate slightly from the stereotypically masculine. The voice sounds genderless to me.
Or maybe because it kinda does sound like a gay flight attendant.
Sometimes a pipe is just a pipe.
It’s disappointing to see you pull out such an easy, cheap, and tired stereotype.
3 replies →
It seems that most people don't really like the way this sounds, but honestly I had the opposite response. I despise the 'natural' sounding synthetic voices because there always seems to be something off about it. The fact that this sounds so weird is what I appreciate about it.
Perhaps one day we'll have natural sounding voices, but until then I'd always prefer a more artificial sounding voice. Something something uncanny valley, I guess.
Even when I move the frequency near the perceived male area, it still sounds female to me. There's more to voice than frequency alone.
It sounds like a poor attempt at voiceover editing. It's nasally and monotone and dull and dreary. There's this strange vibrato throughout it all and there are sudden rapid shifts between a really low and high tone of voice. I'd honestly much rather have something blatantly robotic than this almost successful voice, because all I feel is that I'm being played for a fool.
This is a solution in search of a problem. The overwhelming majority of people do not care what gender their digital assistant's voice is.
Gender issues aside, speaking as a daily text-to-speech user this voice doesn't sound particularly natural or pleasing. The accent seems to constantly shift between American and British, and there is some weird audio phasing going on. It sounds more like a Hollywood director's idea of what robots should sound like than something I'd actually want to listen to.
This reminds me of the audio clip where you can flip between whether you heard Yanny or Laurel: https://www.nytimes.com/interactive/2018/05/16/upshot/audio-...
Sounds like female. Plus, even when I drag the bubble to the bottom, it also sounds like female but with an additional filter on it.
When I think about it, I've never heard anyone who is genderless, I've heard only male and female voices, so I wouldn't know how to compare!
The voice itself is well done. I'd use it, but in more "bubble to the top" form.
The Hacker News crowd is very concerned about people "inventing something that doesn't need to be invented" when that thing strays anywhere near social and cultural issues rather than a shiny new startup or web framework.
If you're not sure what problem it solves, just google 'alexa female' and you'll get plenty of articles to read, and, hopefully, consider.
just google 'alexa female' and you'll get plenty of articles to read
Generally from those who have decided that their job is to find ways to be offended by everything. If Alexa and Siri had male voices by default, I have no doubt that the same people would be complaining that tech bros are erasing the historical contributions that women made as secretaries and assistants.
> Generally from those who have decided that their job is to find ways to be offended by everything.
You know, I never understood why people trot this idea out so often. It's so irrational.
If people were rational, they would say "hmm, X doesn't offend me, but many people say that X offend them; perhaps they have a point."
Instead, only those who are incapable of understanding their own confirmation bias say "All these people say that X offends them; how silly."
I think this is a highly irrational take.
The better take would be: "Some people seem to be offended by everything; other people are not. What are the principal factors of a person that go into predicting 'will this person be offended by X'?"
2 replies →
Siri does default to male in some countries, the UK one is male.
The top results are from lead tech editors at the guardian and the Wall Street Journal.
1 reply →
There's a lot of people saying that alexa is a girl because it's perceived as more helpful. But have we thought about the fact that maybe it's perceived as more nice or as less threatening? There's a reason movie villains are stereotypically male voice actors - deeper male voices sound less friendly, forget helpful.
Is that natural instinct or societal training? If we start using non-female voices for voice assistants, maybe we can train people that a voice doesn't have to be female to be helpful. Otherwise if we use female voices because society is just used to it, then we're just reinforcing an unnecessary stereotype for convenience.
1 reply →
These gender stereotypes are the reason why this voice exists.
i'm pretty sure it's because hollywood has been conditioning us to hear flat affect male robot voices as sinister for 50 years
I like this, but it's probably too late for all tech companies, except maybe Google.
I was pretty disappointed that all the major voice assistants have female voices and most of them have explicit female personas and names. I find it a little creepy to be honest.
At least Google didn't go for the whole personality crap. They could switch Google Assistant to this type of voice and it'd fit in. You can't switch Alexa to this type of voice while keeping that name.
It varies by country. Siri has a male voice in the UK.
5 replies →
> when that thing strays anywhere near social and cultural issues rather than a shiny new startup or web framework.
I don't know what you're reading, but HN is generally highly skeptical of those too, perhaps overly so. This is especially true with your web framework example.
It's simply not an invention or even something to go on about.
You hear 'genderless' voices every day, which is to say, voices which one couldn't place as being male or female with any reasonable accuracy.
You'd be surprised at who in your immediate vicinity has more or less a non gendered voice if you didn't have all the other 'clues' as to their gender.
In the production of synthetic voices, surely we've come across many which are de-facto genderless already.
As a blind person I'm going to have to pretty strongly disagree with this. A vast majority of the voices I hear do in fact read as male or female even without visual indicators.
It sounds like a deep female voice, not just in terms of frequency but timbre; perhaps a transwoman who has undergone a lot of voice training. Even deepening it into the male range didn't eliminate the female timbre.
Colossus (from Colossus: the Forbin Project) arguably had a more genderless voice than this.
What problem does this solve?
"Technology companies often choose to gender technology believing it will make people more comfortable adopting it.
Unfortunately this reinforces a binary perception of gender, and perpetuates stereotypes that many have fought hard to progress.
As society continues to break down the gender binary, recognising those who neither identify as male nor female, the technology we create should follow.
Q is an example of what we hope the future holds; a future of ideas, inclusion, positions and diverse representation in technology."
If technology is needed to create this kind of voice -- IE it's not really natural, and as others have suggested it creates an uncanny valley of voice that confuses people, wouldn't that actually perpetuate the idea that there is a gender binary?
(I don't have a dog in this fight, I just don't see that this accomplishes what you want it to)
3 replies →
> stereotypes that many have fought hard to progress
What stereotypes would this help to alleviate?
> the technology we create should follow.
Why it's a service that communications via artificial audio. Why is it important that you can't have a male or a female voice? People still exist, the voices in question don't have a physical appearance.
7 replies →
> As society continues to break down the gender binary, recognising those who neither identify as male nor female, the technology we create should follow.
People who neither identify as male nor female AND want to listen to a gender-neutral voice seems like a very niche market.
Society still views gender as binary.
it is an interesting idea. technology of course has no gender, and any synthesized speech can only express an arbitrary simulation of gender, so making this explicit reveals the underlying reality that all technological simulations of gender are by definition constructed. a piece of art that demonstrates the construction of gender probably will not catch on well as a product though.
I instantly recognized the voice as M2F, before it even confirmed that by the description of how it was created. If I had a device with this voice, I would feel a strong urge to reenact the final printer scene from the movie Office Space.
Companies want their voice-enabled products to be purchased. Revulsion doesn't help.
Am I the only person who thinks that the text above is some elaborate kind of trolling?
This sounds incredibly sad and dystopian.
And silly, too. Regardless of what is going on inside a person's head (which we can choose to respect, absolutely), biologically there is definitely just two genders. You can identify however you like, modify if you wish, but in the end your body will be one or the other, right along with your voice.
7 replies →
Ahh, so delusion
I find it artistically interesting. A computer assistant doesn't have a gender so why should it represent itself as male or female?
I find something offputting that so many agents pick a female voice (e.g. Alexa, Siri)
> A computer assistant doesn't have a gender so why should it represent itself as male or female
A human has to be in the recording booth to create the vocal library. Also you can change the Siri voice to not only be Male, but also have a different accent.
2 replies →
I find something offputting that so many agents pick a female voice (e.g. Alexa, Siri)
Its research based, not merely "sexism":
https://www.cnn.com/2011/10/21/tech/innovation/female-comput...
2 replies →
> A computer assistant doesn't have a gender so why should it represent itself as male or female?
One could argue that a computer assistant should have a human-like voice and an overwhelming majority of humans have distinctly male or female voices.
Best attempt at a judgement-free guess: some people don't like the bimodal manifestation of gender and would like to not be reminded of it.
The number of people in this category would be small right? Why should I spend time to worry about it?
4 replies →
Is there any market signal that this is the case?
Maybe we should try to figure out why our society is producing people with this rejection of basic biology rather than ignore the problem
The page asserts that audio content to many listeners is firstly "defined" (though perhaps "subconsciously emotionally engaged with to some extent" would be more accurate) by perceived gender, before semantics. The notion is that by adding a genderless/neutral voice, this would be removed. Of course, you then need to discard genders in names as well, hence 'Q'. It's pretty forward thinking... I don't doubt assistant tech may go this way. However, wouldn't "dehumanization by association" for non gender identifying individuals be a risk if that popularizes? A nice thought-provoking art project, if nothing else.
It sounds like the creepy female changeling from Deep Space Nine but with a slight British twang.
It's the least gendered synthetic voice I've heard but there is still a slight wobble between being first a hint more male and then a hint more female.
Also seems to me like it's somewhat camp - a neutral voice in that regard would also be worth exploring
It sounds a bit like a vocal Necker cube illusion. (I.e. I find myself swapping between either male or female depending on how I am listening to it.) It's a bit like the Yanny/Laurel thing for me where I hear both.
I'm a big fan of this, as a non-binary trans woman. Also, it sounds like me ^^
no offense meant - no one sounds like this. record your voice and play it back side by side. there's some kinda uncanny valley effect at play
you're literally telling me what I sound like. that's one hell of a splain
2 replies →
Next up: vocal synthesis of males and females replying in chorus.
this is really cool. my personal reaction was to feel it was a little odd and off, but as I moved the Hz meter up and down it went from "oh a womans voice" to "oh a mans voice" pretty clearly. that is just plain interesting to know, I'd never seen how clearly quantifiable and single-axis that distinction can be. I bet if i spent a few weeks or so using a device in those Hz ranges i'd get used to it.
Meh, as soon as assistant got it i switched to a masculine voice. it's still pretty reedy, i'd like something like terry cruz's voice baritone
That's not Pat?
https://www.google.com/search?q=pat+snl
Yeah, I don't quite understand what the goal of this voice is.
Confident and helpful -- shouldn't that be the aim? Doesn't that sound different for men and women? And across different nationalities?
Q sounds unsure of itself to me... and just weird. I wouldn't prefer this voice to Siri or just about any of the old MacOS voices.
* Mac OS X TTS (Text-To-Speech) Voices - YouTube || https://www.youtube.com/watch?v=Z4hqUxb9MmY
Sounds like a transgender or queer male person.
I would have thought a deep learning approach trying to fool a classifier would have been a more politically neutral approach.
This is neat, but I don't care if my AI's voice is genderless. I'm fine with whatever. I usually pick female.
This actually worked for me. Sounded like the gender could go either way and the voice was pleasant enough.
I'd install it.
Sounds awesome.
But calling something "Q" is a weird choice, considering a recent conspiracy theory
Can you elaborate to which recent conspiracy theory are you referring to?
https://en.wikipedia.org/wiki/QAnon
I'm on Safari and I hear nothing. I guess silence is ungendered?
Sounds more feminine than Elizabeth Holmes, in any case.
That doesn't sound like John de Lancie at all!
While we're at this, I'd rather prefer that my virtual assistant would be based on Majel Barrett-Roddenberry voice she used in ST LCARS voice (?) interface. Pretty sure she did record enough lines to synthesize her voice.
Sounds like a strong, independent lesbian
Nope, sorry. US politics have made me develop a reflex disdain for anything new named like that.
> I'm created for a future where we are no longer defined by gender...
That's not a world I want to be part of.
The key word there is defined. Defined by gender means boxed in by the expectations that being a male or female thrusts on us.
Ergo, that sentence sounds like someone is asking not to be boxed in by their gender.
Why is that not something you can support?
It's not something most people support, including most LGBT activists, notwithstanding that they may believe otherwise.
If you pay close attention, most of the cultural debate over gender identity issues relates not to eliding distinctions and expectations, but to reshaping the boundaries and expectations of those identities.
It's easier to see what I mean by this with the debate over racism and affirmative action. In the U.S. most white people think equality means being "color blind"--so blind that some refuse to even admit that non-overt racism is possible. In truth civil rights in the U.S. has never been about eliding or ignoring [racial] distinctions, but changing expectations. As long as power differentials exist (and they will always exist), these expectations will matter and will constantly shift.
Another way to look at it is that any identity is by definition exclusive. And identities only matter because they imply a default set of expectations. If you say that people should be literally free to choose their identity, you're allowing people to freely reinterpret and redefine the implied expectations and shared _meaning_ of that identity, which can be injurious to those who previously chose that identity. There's an inherent conflict, there. It's why as a white male I can't go around identifying as a black female, certainly not without making substantial changes to my behavior and perhaps even physical appearance, changes that should be consonant with what being a "black female" means. To do otherwise would be another kind of violence.
You can't really understand any of these issues without understanding and appreciating the genesis--power. It's all about power over others. We can never have a society without power differentials because that would imply the power to _refuse_ to recognize the identities others have chosen for themselves. Power matters not because it's per se evil, but how it's used and how it's distributed and what are the concrete results in people's lives.
In adjudicating conflicts over what the expectations for an identity are, one must take into account how power is used and abused and by whom. For example, it's legitimately problematic for a man to argue that transgender women shouldn't be allowed in women's bathrooms because of the threat of violence. It's problematic because it's another instance of a man making decisions about what it means to be a women, and men have historically abused their social and economic positions when drawing these boundaries to disadvantage women more generally. The logic of the argument is secondary to the power politics, because fundamentally it's really about the power politics. It's more legitimate for women to make that argument, and in so much as they do make it it carries more weight (notwithstanding that women as a group are certainly capable of abusing their relative powers), but in actuality they've been less vociferous about it.
The line of thinking that says power differentials are inherently wrong and/or that we should elide the implications of identities are inconsistent. If only things were so simple....
Supporting that for people who want it is one thing, living in a world where we act like there is no gender is another. A lot of us enjoy being and/or identifying as male/female.
> The key word there is defined. Defined by gender means boxed in by the expectations that being a male or female thrusts on us.
Identity is by definition a set of expectations. To lose those expectations is to lose a facet of identity.
There's a group of people who want to remove that facet of their identity, cool, I support them. But I like that facet of my identity, and I like the fact that most people around me also express that facet of their identity. I don't want want to be genderless or to live in a genderless world. Do not marginalise me for being that way.
1 reply →
Why?
Typical meatbag voice. The future belongs to formant synthesis, I can easily get to 1000wmp without any loss of coherency and the sound is fundamentally inhuman.
https://ufile.io/9905a
I didn't quite catch that...
It's intentionally made to be hard to listen to. OP spliced multiple copies together with a slight offset, varied the rate while recording, and intentionally selected a passage which repeats very similar words in proximity, which further increases the difficulty.
> Formant synthesis does not use human speech samples at runtime. Instead, the synthesized speech output is created using additive synthesis and an acoustic model (physical modelling synthesis. Parameters such as fundamental frequency, voicing, and noise levels are varied over time to create a waveform of artificial speech. This method is sometimes called rules-based synthesis; however, many concatenative systems also have rules-based components.
3 replies →
Espeak's sibilance really does get painful at speeds like this. You could probably further increase your processing rate if you vary inflection a bit.
I've played around with the internals a bit, but I've not been able to substantially improve understandability past 1200 wpm no matter what I do.
The settings for that text snippet were
$ espeak -p 100 -s 1000 -v male7
That's really the optimum for me for speeds over 650 pwm.
This sounds fine to me. I don't care if a clearly artificial voice doesn't sound human. Computers should have their own voice.
I think the stated purpose is stupid, though. Gendered roles are good. They give people a purpose and make them happier. I really pity the little girls who get forced into things like engineering just because their parents wanted to be hip and "genderless".
I want my virtual assistants more realistic, not less. This sounds nothing like a real person.
This is the biggest nothing burger in the entire world. Why is this on hacker news?
This reminds me of Harrison Bergeron, a story about a dystopian future where any inequality is brutally suppressed.
Is this the state of start-ups again, inventing something no one wants or needs?
new career option for elizabeth holmes.
How often will you be wrong if you assume someone is gay because they speak in the stereotypical gay manner? Almost never.
I put my Siri into British. Servants should be British. Is that racist?
I can't help to hear it as the voice of the person interviewed in that infamous clip titled "Why are you gay (you are gay)".
This is a good step
I've given input on political issues as well as gender issues under multiple identities and personas to gain consensus
Almost no progress on consensus regarding political issues but on some sex and gender issues the male persona was unqualified to "mansplain" or had to much privilege to contribute or complain, where the same information was agreed upon and deemed helpful coming from the non-male persona - assumed to be female
This is more about the decision path people make to discredit a someone. When confronted with information that deviates from your script, people typically look for faults in the messenger instead of the actual information. You look for words that a different party would use exclusively to pigeonhole the messenger as the other, you look for other biases.
I felt that even if an A.I. was created to try to aggregate perspectives and offer its observations, people would try to disagree with it based on who created it. A valid criticism, but what would be left with to offer potentially unifying perspectives.
I like this and I imagine how it can be trained to create a voice that you agree with, but that nobody else would hear.
..... based on your cookies.
Then everybody gets to hear a different voice and mostly the same information - perhaps slightly different words as to not undermine the illusion.
I would humbly suggest that the bevy of reactions to this innovation that amount to 'eww ugh this creeps me out' is evidence of precisely the cisheteronormative bias that this voice was designed to mitigate. I invite you to introspect about your (dis)comfort around folks who are gender-nonconforming.
gosh I love the downvotes boys.
hypothesis confirmed
Imagine being triggered by your GPS devices voice
Grow up
This voice sounds like a third gender to me: trans.
Can we not pretend that a computer generated voice in a specific frequency range is some sort of political statement?
I rather a female assistant voice, because mothers in human societies are women, and we're hard-wired to find motherly voices more soothing.
This is getting out of hand, it's unnatural, fits strictly in the uncanny valley, and is borderline disgusting; both in terms of sound and in terms of politico-cultural values and intentions.
It's just a voice. You can dislike it without having to get worked up about it.
But it isn't "just a voice", isn't it?
It's part of a socio-cultural Zeitgeist which actively combats values which I think are worth defending.
If I could bring somebody here from anywhere between 100 to 10,000 years ago, to show them the absolute state of society...