 |
VO-BB - 20 YEARS OLD! Established November 10, 2004
|
View previous topic :: View next topic |
Author |
Message |
Mike Paul Contributore Level V

Joined: 05 Oct 2021 Posts: 159 Location: Sacramento
|
Posted: Fri Nov 18, 2022 9:12 pm Post subject: State of the art: AI voices in 2022 |
|
|
Back in the yesteryears of computing, I used to visit a website called MacHeist. They often created bundles of software for Mac that, the more people bought, the more software titles would be unlocked in the bundle for everyone. It was really cool, back when software ran on computers instead of crappy web browsers. (One step forward, two steps back?) For reasons unknown, I went there tonight and was surprised that they were still alive.
I was more surprised to discover a piece of software called Micmonster. With a tagline of "The future of voiceover" under the title, I had to click. Not saying they're good (they're not), just that they exist, for those interested in this side of the industry and how soon we'll be replaced by it. I don't think this will replace us anytime soon, though it may allow for the creator economy types to add voiceovers to their stuff if they don't do it themselves and wouldn't be paying pros (or amateurs, haha) for that type of work. My gut says we're 10 years or less away from not being able to tell the difference in general/neutral speech, and then it'll take a long, long time to convincingly portray all of the nuance a human can emote.
Thankfully you can still tell it's AI and not human. You'd love to have a law or some type of legal guardrails that say "you can't make this too real, now, ya hear?" but like that'll do anything.
If you can make it through the whole video, well.... I don't know why you would do that to yourself. You'll never get the time back.
That opening sentence in the video, though...
https://www.macheist.com/sales/micmonster-lifetime-subscription |
|
Back to top |
|
 |
Lee Gordon A Zillion

Joined: 25 Jul 2008 Posts: 6863 Location: West Hartford, CT
|
Posted: Sat Nov 19, 2022 1:12 am Post subject: Re: State of the art: AI voices in 2022 |
|
|
speakingofmike wrote: | we're 10 years or less away from not being able to tell the difference in general/neutral speech, |
To a certain extent we're there already. I listen to the VO on some YouTube videos and honestly can't tell whether they're AI like in your link, or some human VO they got from Fiverr who just sucks. Either way, it's a turn-off and that's generally exactly what I do. _________________ Lee Gordon, O.A.V.
Voice President of the United States
www.leegordonproductions.com
Twitter: @LeeGordonVoice
 |
|
Back to top |
|
 |
Philip Banks Je Ne Sais Quoi

Joined: 20 Jun 2005 Posts: 11067 Location: Portgordon, Scotland
|
Posted: Sat Nov 19, 2022 2:34 am Post subject: |
|
|
We need to add a category to the list people who announce to unsuspecting guests at a party "Well, as a voice actor..." It's that of Utility Voice. The days of people making a living by supplying their voices to utility projects/productions are numbered.
Yesterday I needed to flip/mirror a photo sent to my iphone by a friend. Couldn't find anything, lost the will to live. AI would've done the job - Utility. A few weeks ago I did the introduction to a charity event, the copy needed the mind, heart, body and soul of a Human Being. AI consists of 101010101001010101010101, we don't! |
|
Back to top |
|
 |
Bob Bergen CM
Joined: 22 Apr 2008 Posts: 976
|
Posted: Sat Nov 19, 2022 6:59 am Post subject: |
|
|
For years I was saying that we have nothing to worry about because computers can't act. Today, they can. Just listen to James Earl Jones in Obi-Wan-Kenobi, which was 100% AI created from previous recordings. AI is being used for both group ADR as well as sweetening lead actor's performances. This is here, now. It's used more often than most people know. Those who say they can hear the difference are just hearing lousy examples and are oblivious to the good ones cuz they are just that good.
VO is just the latest to be affected by this tech disruption. Set dressers and make-up artists found much of their work replaced by computers years ago.
I remember when I worked on A Bug's Life and marveled at the CGI grass, water, etc. I asked John Lassiter what his current challenge was. He said, "We are trying to figure out how to do the ocean." Then came Finding Nemo. Tech brilliance just improves.
Computers can act. They can take a single line read flat and churn out dozens of emotions and interpretations. Is it perfect? Of course not. Neither are most actors.
This is very real. I used to say this will not impact the vo industry in my lifetime. I now say I don't think this will make vo actors obsolete in my lifetime. But tech has historically demonstrated how fast it makes the previous generation obsolete. Silent films did away with vaudeville. Sound did away with silent films. TV impacted film. Cable impacted network TV. Streaming impacts everything. AI impacts vo. |
|
Back to top |
|
 |
ricevoice Cinquecento

Joined: 28 Dec 2007 Posts: 532 Location: Sacramento, CA
|
Posted: Sat Nov 19, 2022 8:01 am Post subject: |
|
|
One thing computers can't do: be directed by a client. Clients who really enjoy directing a v-o session and having the talent "this time maybe try it this way" and making it a collaborative process... I don't see those clients switching over to AI. Thankfully, in my experience they also tend to be the clients with the bigger budgets! _________________ Chris Rice - Noisemaker
www.ricevoice.com |
|
Back to top |
|
 |
Bob Bergen CM
Joined: 22 Apr 2008 Posts: 976
|
Posted: Sat Nov 19, 2022 8:28 am Post subject: |
|
|
ricevoice wrote: | One thing computers can't do: be directed by a client. Clients who really enjoy directing a v-o session and having the talent "this time maybe try it this way" and making it a collaborative process... I don't see those clients switching over to AI. Thankfully, in my experience they also tend to be the clients with the bigger budgets! |
You are correct, the computer cannot direct. But the programmer can!
This is primitive compared to the advances made since it launched.
https://www.youtube.com/watch?v=-7x3CbbR-ns
This of course will not affect the entire vo industry. You are correct there will be a level of professional actors and buyers who will want to work human to human. But this absolutely will make a dent in the day to day that is vo. This is history repeating itself, as the union vo industry never thought vo actors would ever take non-union work, nor did they think top buyers would turn to non-union talent for network commercials, as non-union talent just were not as good.....until they were. |
|
Back to top |
|
 |
George Been Here Awhile

Joined: 09 Dec 2019 Posts: 283 Location: Seattle, WA
|
Posted: Sat Nov 19, 2022 8:43 am Post subject: |
|
|
To piggy-back on what Bob said (i gotta stop doing that because it needs the word piggy and it makes me sound deliberately puntastic) and in response to Chris:
Bob is right: a programmer with the more expensive AI voices available, a programmer CAN direct a voice, and provide samples to a client, and the client can audition them and pick and choose which nuances they like. Line by line.
When clients find out they don't have to direct professional actors anymore and it ends up reducing cost and time and the end product is exactly what they're looking for, there we have it.
We're not fully there yet. But the rate at which it is improving is telling.
Remember how people used to say that MP3 as a music format could never get traction because the quality wasn't as good as a CD?
Remember when streaming couldn't ever possibly take the place of DVD because of the quality differences?
Remember when X couldn't possibly take the place of Y because it wasn't good enough? |
|
Back to top |
|
 |
Lee Gordon A Zillion

Joined: 25 Jul 2008 Posts: 6863 Location: West Hartford, CT
|
Posted: Sat Nov 19, 2022 12:25 pm Post subject: |
|
|
Bob Bergen wrote: | Just listen to James Earl Jones in Obi-Wan-Kenobi, which was 100% AI created from previous recordings. |
Technology in the hands of Lucasfilm is not quite the same as what the people most of us do business with are currently capable of. That's not to say it's not coming or even coming soon, but for the kind of work most of us do, it isn't significantly encroaching on our turf just yet.
But here's the thing AI will almost certainly never be capable of: personal relationships. The relationships we develop with our clients are the one thing AI can't exploit. At least not until our clients are replace by AI. _________________ Lee Gordon, O.A.V.
Voice President of the United States
www.leegordonproductions.com
Twitter: @LeeGordonVoice
 |
|
Back to top |
|
 |
Mandy Nelson MMD

Joined: 07 Aug 2008 Posts: 2911 Location: Wicked Mainah
|
Posted: Mon Nov 21, 2022 1:09 pm Post subject: |
|
|
I work closely with two private developers that know the right voice direction to give, that I mimic, that they can then put anywhere and you can barely tell the difference between my reading a full sentence and them putting it together. It is mind-blowing and, as said, certainly here. We've tested their stuff on people that know me well, know what I do and they can't get it right.
To be clear - they use me and my voice, pay me for my time, but ARE NOT allowed to use my voice for their programs. What they are developing will rely on programming the machine voices to sound flawless and by using a live human they are figuring it out. _________________ 006 member of the Sisterhood of the Traveling Mic. Bonded by sound.
Manfillappsoc: The Mandy and Philip mutual appreciation Society. Who's in your network?
Have you seen my mic closet? ~ me to my future husband |
|
Back to top |
|
 |
Bob Bergen CM
Joined: 22 Apr 2008 Posts: 976
|
Posted: Mon Nov 21, 2022 2:20 pm Post subject: |
|
|
Mandy Nelson wrote: | I work closely with two private developers that know the right voice direction to give, that I mimic, that they can then put anywhere and you can barely tell the difference between my reading a full sentence and them putting it together. It is mind-blowing and, as said, certainly here. We've tested their stuff on people that know me well, know what I do and they can't get it right.
To be clear - they use me and my voice, pay me for my time, but ARE NOT allowed to use my voice for their programs. What they are developing will rely on programming the machine voices to sound flawless and by using a live human they are figuring it out. |
And, this is a huge catalyst for the technology getting better and better. Actors are being hired to assist in this AI tech. Also, P2Ps have used booked work for the same purpose. (this I've been told by the major studios for years) Yes, some P2P bookings come with protections to prevent this without compensation. But this usage was implemented way before such protections. The irony is how today's actors are contributing to AI taking work from tomorrow's actors. I'm not saying this is not inevitable. It is what it is. |
|
Back to top |
|
 |
GunslingerWriting Contributor II
Joined: 23 Feb 2010 Posts: 62
|
Posted: Sun Nov 27, 2022 10:57 am Post subject: |
|
|
I am astonished at how little understanding some voices on here--especially older ones--have of the capacity of AI technology to do all of the things that are being dismissed as impossible.
People don't know what they don't know...and that's at their planning/occupational peril. |
|
Back to top |
|
 |
Mike Paul Contributore Level V

Joined: 05 Oct 2021 Posts: 159 Location: Sacramento
|
Posted: Sun Nov 27, 2022 8:24 pm Post subject: |
|
|
It's true, Gunslinger, we don't know what we don't know and I think we're having a great discussion about the state of the art because I clearly had no idea just how good it was, but now I'm getting one.
What are your thoughts on the topic? Have some experience worth sharing? |
|
Back to top |
|
 |
GunslingerWriting Contributor II
Joined: 23 Feb 2010 Posts: 62
|
Posted: Mon Nov 28, 2022 4:28 pm Post subject: |
|
|
I've worked with people who've been doing this longer than anyone else I know and I've seen/heard the outcome of their work. It can be breathtakingly real, down to inflection, passion and emotional evocativeness.
The scary thing is, what they showed me many months ago that blew my socks off is five times better now.
Technologically, the sky's the limit; and so is the gutter for those who choose not to at least have a passing awareness of what this stuff is, what it can do, how it can sound and whom it will, inevitably, appeal to. Because that will be lots and lots of people who want a voice on a budget. |
|
Back to top |
|
 |
Mike Paul Contributore Level V

Joined: 05 Oct 2021 Posts: 159 Location: Sacramento
|
Posted: Mon Nov 28, 2022 5:54 pm Post subject: |
|
|
Thanks for your thoughts Gunslinger. Very interesting. What, then, would be a good hedge against this as a voice actor? For example, the hedge against the disappearance of the "announcer" read was to learn "conversational".
Now I'm wondering what I should do when this type of work comes up. My gut says to just audition for it because whether I do or don't, this is happening whether we like it or not. Can't put the genie back in the bottle .
It also kinda feels similar to the sustainability problem we're having with the "race to the bottom" of rates. In both cases--AI and rates--we need to figure out how to keep this business sustainable as a primary income for those of us striving for that. I choose to align with the GVAA and sometimes have to turn down low-budget projects that I know someone else will snag. Not because I'm financially secure, but on principle.
Suddenly between the low-ball rates and AI, the future's looking frickin dim, man.
Wondering what that hedge is. |
|
Back to top |
|
 |
BruceG Been Here Awhile

Joined: 01 Jun 2012 Posts: 258 Location: just south of Boston, MA
|
Posted: Mon Nov 28, 2022 8:28 pm Post subject: |
|
|
No doubt that AI is more than catching up in terms of replicating the human voice. Recently there were times when I got a "robo-call" or listened to a narration on a YouTube video and had to really stop and ask: "Is this voice human?"
Here's my question: can AI truly "act"? Acting/Voice Acting is so integral to being human; we can improvise, take risks, try different directions, all on a whim, based on who we are as individuals. I just can't see AI touching that. Hopefully that means there will always be a demand for real voices.  _________________ "What was that? An exhibition? We need emotional content." - Bruce Lee, Enter The Dragon |
|
Back to top |
|
 |
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|