One aspect might be that each voice is unique and that the singer and the voice are inseparable. A drummer or violinist, however, can simply replace his drum set or violin by another.herodotus wrote:There is a thread in the instruments forum about a hypothetical virtual instrument that could convincingly mimic a human voice.
That thread has lots of different kinds of drama in it. Some of it personality driven, which is something I make an effort to avoid. But some of the contention is more interesting.
Specifically, many people seem apprehensive about the idea of convincingly mimicking a human voice as opposed to convincingly mimicking drummers or violinists or pianists.
I will admit that this surprised me. Not that people would feel that way, but that so many people in this community would feel this way.
Anyway, I thought it would be interesting to explore these issues in a separate thread that is unencumbered by off topic personality issues.
Discuss.
Is the human voice 'special'
-
fluffy_little_something fluffy_little_something https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=281847
- Banned
- 12880 posts since 5 Jun, 2012
-
- KVRist
- 168 posts since 18 Oct, 2017
The Terminator can imitate every voice 
-
- KVRAF
- Topic Starter
- 5703 posts since 8 Dec, 2004 from The Twin Cities
I guess I can understand that.vurt wrote: i think peoples initial horror was based on the idea of an ai listening to artists recording and allowing any one of us to sound like that particular singer. while again, doesn't bother me either way, i can see why this might have legal problems down the line, so would probably end up being somewhat more expensive than was offered
I myself did feel an initial revulsion to the idea of algorithmic composition. But that was more because I was afraid of film music starting to suck even more than it already does.
But, to me, there is nothing really special about the human voice as opposed to some other form of musical expression. To be honest, I go for weeks at a time without listening to any music with vocals. I have nothing against vocalists, mind. I just don't share in the fetishistic admiration they inspire, as opposed to, say, oboists.
-
- KVRAF
- 25053 posts since 20 Oct, 2007 from gonesville
I remember well being in a little boutique record store in Charlotte with my gf. I was... 23 or like that. Some guy, a bit older was trying to mack on her, trying to impress her with his superior experience, you know. But the topic was singers, this or another singer and we were interested more in instrumental music. And one of these guys said something on the order of 'When you get older, you'll see.' And I already had projected that as I matured I would gravitate more to singers.
-
- KVRAF
- 25053 posts since 20 Oct, 2007 from gonesville
"fetishistic" is strong language. And "fetishistic admiration" a bit of a straw man.herodotus wrote:I have nothing against vocalists, mind. I just don't share in the fetishistic admiration they inspire, as opposed to, say, oboists.
-
- KVRAF
- 4798 posts since 25 Jan, 2014 from The End of The World as We Knowit
The "vocalist fetishist" label is a bit drama-queen and sounds like it was pulled from the Other Thread but I admire its semantic uniqueness and am proud to say it certainly applies to me! (I saved my pennies to hear Lena Horne in the Fairmont Room)
s a v e
y o u r
f l o w
y o u r
f l o w
-
- KVRAF
- Topic Starter
- 5703 posts since 8 Dec, 2004 from The Twin Cities
To be clear, I wasn't talking about anyone in this thread. But the fetishistic adoration I referred to is rife in our culture, in an obvious and undeniable way. American Idol is only the most crass expression of this adoration.jancivil wrote:"fetishistic" is strong language. And "fetishistic admiration" a bit of a straw man.herodotus wrote:I have nothing against vocalists, mind. I just don't share in the fetishistic admiration they inspire, as opposed to, say, oboists.
I do not lack respect or affection for all manner of vocalists (especially freaks like Yma Sumac, Diamanda Galas, or Mike Patton), nor am I complaining about people's preferences.
But I have long since seen music more as an art of design than as a form of emotional expression.
I am not saying there is an either/or kind of choice to be made. And if there is, I have failed to make it. I just admire composers more than singers.
This isn't an attempt to put the world in order. Just an expression of personal preference.
-
- KVRAF
- 4798 posts since 25 Jan, 2014 from The End of The World as We Knowit
American Idol are reality tv performers. True vocalists / song stylists / interpreters / scat singers / singer-songwriters / gospel-soul are also composers (within their chosen constraints.). Listen to different singers interpretations of standards.
s a v e
y o u r
f l o w
y o u r
f l o w
-
- KVRian
- 1428 posts since 27 Apr, 2012
Basically what el-bo said. I think to some degree the human voice/language is like the human face - everyone is mentally attuned to what it should be like and when it's "off" we can tell. Compare that to, say, a violin. I'm no violinist or even orchestral music aficionado and I probably couldn't tell the difference between a well-programmed sampled violin and a real violinist, even though the software probably can't fully replicate the nuances of a real performer. A "good enough" emulation of a violin can work just fine for a popular audience, but "good enough" for a lead vocal is not going to cut it. That's one reason realistic vocal synthesis is especially challenging - it has to be virtually perfect or else people can tell it's fake.
-
- KVRAF
- Topic Starter
- 5703 posts since 8 Dec, 2004 from The Twin Cities
I grew up listening to different singers interpretation of standards. My mom was a singer, in both churches and night clubs. She owned hundreds of L.P.s by vocalists from Sarah Vaugn and Ella Fitzgerald to Leontyne Price and Eileen Farrell. And she was a fanatical admirer of Judy Garland. I myself admit to getting tears in my eyes when I hear 'The Man that Got Away'.Michael L wrote:American Idol are reality tv performers. True vocalists / song stylists / interpreters / scat singers / singer-songwriters / gospel-soul are also composers (within their chosen constraints.). Listen to different singers interpretations of standards.
I am perfectly aware of the power of the human voice.
But I am also aware of the power in the sheer abstract design of the 3 part fugue from Bach's Musical Offering, or of Bartok's Music for Strings, Percussion and Celesta.
I happen to prefer the latter. And if you want to tell me that it gets as much respect as the former in our culture, I am going to have to tell you that I think you're mistaken.
-
- KVRAF
- 25053 posts since 20 Oct, 2007 from gonesville
OK, but you do have beef with 'our culture' which fetishizes singers. I can't do anything with 'our culture' as singular. I would have to interview individuals to know they have a fetishist's approach to a thing.herodotus wrote:To be clear, I wasn't talking about anyone in this thread. But the fetishistic adoration I referred to is rife in our culture, in an obvious and undeniable way. American Idol is only the most crass expression of this adoration.jancivil wrote:"fetishistic" is strong language. And "fetishistic admiration" a bit of a straw man.herodotus wrote:I have nothing against vocalists, mind. I just don't share in the fetishistic admiration they inspire, as opposed to, say, oboists.
I do not lack respect or affection for all manner of vocalists (especially freaks like Yma Sumac, Diamanda Galas, or Mike Patton), nor am I complaining about people's preferences.
But I have long since seen music more as an art of design than as a form of emotional expression.
I am not saying there is an either/or kind of choice to be made. And if there is, I have failed to make it. I just admire composers more than singers.
This isn't an attempt to put the world in order. Just an expression of personal preference.
I have observed particularly recently that commercial establishments play 99.9% songs with vocals to do whatever they think they're doing by subjecting customers to wall-to-wall music. And I have some juicy rants on FB regarding certain details of this.
But the way I think of this is, isn't Diamanda Galas a composer? It's an unnecessary dichotomy, I can't use this kind of thinking. I can easily say however that at one time I will have likely made a similar argument as 'abstract design vs the emotive aspect' or really smart music versus the operas a lot of people think is the highest expression, or whatever.
I will say that, because the voice is purely the body and needs no other paraphernalia, it is more personal. Now, if the aesthetic thrust is to elevate the transperonal over the personal, this is a whole can o' worms...
-
- KVRAF
- 4798 posts since 25 Jan, 2014 from The End of The World as We Knowit
Now you are comparing the public's preference for musical styles.herodotus wrote:I myself admit to getting tears in my eyes when I hear 'The Man that Got Away'. I am perfectly aware of the power of the human voice.
But I am also aware of the power in the sheer abstract design of the 3 part fugue from Bach's Musical Offering, or of Bartok's Music for Strings, Percussion and Celesta.
I happen to prefer the latter. And if you want to tell me that it gets as much respect as the former in our culture, I am going to have to tell you that I think you're mistaken.
Pop music IS more popular than Classical. No argument about that!
But if we normalise, and compare pop vocal vs pop instrumental, I would say that many people prefer amazing instrumental solos to amazing vocal melodies (but prefer harmonies between voices).
Many people also prefer classical instrumental to classical vocal (solo or choral).
Why one over the other? That is an interesting question.
I think it depends on how easily someone can identify the different-but-simple patterns (hooks) in a musical idea. If a solo vocal hook in a single-line pop melody is simple enough to identify, and the difference makes it memorable, it can become a hit.
I think we learn the most simple patterns through exposure to everyday speech (hence the reason pop melody intervals are small) while more complex patterns are learned through repeated exposure to them (e.g. atonal oboe quintets).
s a v e
y o u r
f l o w
y o u r
f l o w
-
- KVRAF
- 2008 posts since 11 Aug, 2012 from omfr morf form romf frmo
I don't think it's the voice itself that is so special. The plugin described in that post specified a lot more technology that just doesn't exist yet.
There is extant technology where sampled and/or synthesized voice is reasonably intelligible and understandable. The physiology that produces formants is well understood and modelled. It's not going to take long until it sounds indistinguishable from a Real Live Person. One thing that is a little odd to me, if you've heard Cortana, Alexa, et. al., is that random mouth noises/clicks don't exist. Sure, they're edited out in many recordings so we're kind of used to it, but in the context of conversing with someone, it's a bit more natural if they're still in there.
Anyway, the other half of the equation is understanding and executing intonation automatically and that's going to require AI. Whether that's a lot of weak AI that can run on the chips in phones and countertop devices, or a scientific breakthrough that gives us strong AI is another question. But it's a difficult problem to solve. Everything from emotion to physiology to regional accents and cultural background contribute to it in various ways. So we'll see it first with an emotionless "newscaster" accent (as we already do with American interactive speech devices which adopt the "mainstream" Midwest accent).
The other half to that half of the problem is learning various intonations based on all the possibilities and creating a model the first half can use. For now, it has to be manually programmed. But it's a very difficult problem to program every single possibility, e.g., a slightly flustered 26 year-old amateur actress with a Texan accent trying to perform an English Received Pronunciation accent in a local play. If you've tried programming Vocaloid, you know how difficult, if not impossible, this would be. Since a strong AI is equivalent to a human's mental ability, I'm almost confident we'd need a superintelligent AI to do this task.
However, if we can do that, then modelling and executing singing can't be too far behind. On the other hand, some people think the advent of superintelligent AI would bring on the technological singularity, which leads to a rapid technological explosion (think superintelligent AI with an IQ of 3000 creating even more intelligent AI). So it's hard to tell just how society would react to perfect AI singers given the way society is transforming around them.
If you've heard a lyrical song for the first time and it struck you very deeply, how would you feel if you learned it was produced by entirely by AI? If the humanity is lost, does the piece lose meaning? Then there's the question of whether this superintelligent AI recognizes itself. If so, we would be compelled to give it personhood. If AI is a person with super-intelligence and the sum total knowledge of humanity, doesn't that make it more equipped to understand and express the human condition than humans? Would the fact that it's easy for the AI to do take away from the work's perceived value? I don't think it would matter. There are many works that deeply touch people where the creator is simply irrelevant to them.
What will be interesting is when superintelligent AI starts critiquing each other! At first, they would be catering to humans, but as superintelligent AI become consumers of music, who knows what kind of music will satisfy them? Especially since they don't have the same limits as human hearing. Might even be a curse, if they can pick out every partial in a signal. Maybe we'll see AI limiting itself with something like a filter modelled on the human ear/brain. Or maybe we'll see hypermicrotonal music. Or music that isn't necessarily heard (I mean, they're already "listening" to digitally encoded waveforms). Like a sequence of interesting numbers, and they'll leave us alone to art made for our physiology. Maybe we'll see new Grammy categories. No way to tell how it'll go.
There is extant technology where sampled and/or synthesized voice is reasonably intelligible and understandable. The physiology that produces formants is well understood and modelled. It's not going to take long until it sounds indistinguishable from a Real Live Person. One thing that is a little odd to me, if you've heard Cortana, Alexa, et. al., is that random mouth noises/clicks don't exist. Sure, they're edited out in many recordings so we're kind of used to it, but in the context of conversing with someone, it's a bit more natural if they're still in there.
Anyway, the other half of the equation is understanding and executing intonation automatically and that's going to require AI. Whether that's a lot of weak AI that can run on the chips in phones and countertop devices, or a scientific breakthrough that gives us strong AI is another question. But it's a difficult problem to solve. Everything from emotion to physiology to regional accents and cultural background contribute to it in various ways. So we'll see it first with an emotionless "newscaster" accent (as we already do with American interactive speech devices which adopt the "mainstream" Midwest accent).
The other half to that half of the problem is learning various intonations based on all the possibilities and creating a model the first half can use. For now, it has to be manually programmed. But it's a very difficult problem to program every single possibility, e.g., a slightly flustered 26 year-old amateur actress with a Texan accent trying to perform an English Received Pronunciation accent in a local play. If you've tried programming Vocaloid, you know how difficult, if not impossible, this would be. Since a strong AI is equivalent to a human's mental ability, I'm almost confident we'd need a superintelligent AI to do this task.
However, if we can do that, then modelling and executing singing can't be too far behind. On the other hand, some people think the advent of superintelligent AI would bring on the technological singularity, which leads to a rapid technological explosion (think superintelligent AI with an IQ of 3000 creating even more intelligent AI). So it's hard to tell just how society would react to perfect AI singers given the way society is transforming around them.
If you've heard a lyrical song for the first time and it struck you very deeply, how would you feel if you learned it was produced by entirely by AI? If the humanity is lost, does the piece lose meaning? Then there's the question of whether this superintelligent AI recognizes itself. If so, we would be compelled to give it personhood. If AI is a person with super-intelligence and the sum total knowledge of humanity, doesn't that make it more equipped to understand and express the human condition than humans? Would the fact that it's easy for the AI to do take away from the work's perceived value? I don't think it would matter. There are many works that deeply touch people where the creator is simply irrelevant to them.
What will be interesting is when superintelligent AI starts critiquing each other! At first, they would be catering to humans, but as superintelligent AI become consumers of music, who knows what kind of music will satisfy them? Especially since they don't have the same limits as human hearing. Might even be a curse, if they can pick out every partial in a signal. Maybe we'll see AI limiting itself with something like a filter modelled on the human ear/brain. Or maybe we'll see hypermicrotonal music. Or music that isn't necessarily heard (I mean, they're already "listening" to digitally encoded waveforms). Like a sequence of interesting numbers, and they'll leave us alone to art made for our physiology. Maybe we'll see new Grammy categories. No way to tell how it'll go.
-
- KVRAF
- 4798 posts since 25 Jan, 2014 from The End of The World as We Knowit
^^^ I think it will be easy for an AI singer of an average song to fool an average person. However, AI programmed by cubicle technicians will ever be able to exceed what a creative artist can do.
s a v e
y o u r
f l o w
y o u r
f l o w
-
- KVRian
- 806 posts since 7 Aug, 2015 from H2O
Then we will have beaucoup upon beaucoup of "uncovered" home demos of, for example, Kurt Cobain. None will be the wiser...Michael L wrote:^^^ I think it will be easy for an AI singer of an average song to fool an average person. However, AI programmed by cubicle technicians will ever be able to exceed what a creative artist can do.
