I'm sorry, Dave. I'm afraid I can do that: Microsoft unveils Custom Neural Voice – synthetic, but human-sounding speech
- Reference: 1612459929
- News link: https://www.theregister.co.uk/2021/02/04/custom_neural_voice/
- Source link:
Unsettling, because unlike the usual text to speech we've come to know and love over the years, which require a substantial amount of data (10,000 lines or more, according to Microsoft) to sound fluent, Custom Neural Voice requires far less in terms of training audio. The result is disturbingly human-like.
[2]
"This new technology allows companies to spend a tenth of the effort traditionally needed to prepare training data," explained Microsoft, which will come as a delight to out-of-work actors looking to do some voiceover jobs on the side (it probably won't).
There is also a real risk of abuse, hence the GA gates now being entirely thrown open.
[3]
Remember OpenAI's GPT model that was too dangerous for mere mortals? Well, it's now for sale on Azure [4]READ MORE
Microsoft's own [5]code of conduct for the technology warns against using "photo realistic avatars with synthetic voices to represent real people" nor "using a synthetic voice with contents without editorial control." Sensible guidelines when choosing a use case, but unlikely to put off a determined miscreant.
As for the technology itself, three components are at play: Text Analyzer, Neural Acoustic Model, and Neural Vocoder. The trio take inputted text, convert it to a phoneme (a basic unit of sound) sequence, pass that through the model to predict acoustic features before finally spitting out audible speech.
The Neural Model itself is trained using neural networks and actual voice recordings. Those recordings are where things get sticky, and "Microsoft requires every customer to obtain explicit written permission from the voice talent before creating a voice model." Verification is also performed.
After all, once that model is up to snuff, the voice could say all manner of things. Microsoft also insists the use of a synthetic voice be disclosed to users, which could make some of the relentlessly perky chatbot-style use cases presented potentially awkward.
[6]
Adopters [7]have included AT&T , which had a voiceover artist churn out 2,000 phrases and lines in order to voice cartoon character Bugs Bunny with Custom Neural Voice. At least in that instance one knows that Bugs is a fictional character. ®
Get our [8]Tech Resources
[1] https://techcommunity.microsoft.com/t5/azure-ai/build-a-natural-custom-voice-for-your-brand/ba-p/2112777
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_emergent_tech/artificial_intelligence&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YBx8iQVc89i9CV@amaNmcgAAAJI&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_emergent_tech/artificial_intelligence&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YBx8iQVc89i9CV@amaNmcgAAAJI&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[4] https://www.theregister.com/2020/09/06/in_brief_ai/
[5] https://docs.microsoft.com/en-us/legal/cognitive-services/speech-service/custom-neural-voice/transparency-note-custom-neural-voice
[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_emergent_tech/artificial_intelligence&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YBx8iQVc89i9CV@amaNmcgAAAJI&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[7] https://blogs.microsoft.com/ai-for-business/custom-neural-voice-ga/
[8] https://whitepapers.theregister.com/
Voicemail
It’s only a matter of time before I get a voice mail from the CEO requesting a wire transfer for a new vendor in Nigeria
The magician (and technology first adopter, friend of Silicon Valley types like Gates and Jobs) Penn Juliette was presented with a magic trick where the contestant had trained an artificial voice on hundreds of hours of Penn Juliette speaking (from TV shows and podcasts). After performing the trick, the contestant told Penn that for erhical reasons he would delete the artificial voice - unless Penn would like a copy for himself.
Penn pointed out that he was the only person on Earth who had zero conceivable use for an artificial voice that sounded like Penn Juliette.
computer voices
Love them in movies, but dislike them in real life.
We know.
Of course we all know what it'll be used for....
That age old subject - pornography
Well, what other use has technology got?
synthetic, but human-sounding speech
Well, you wouldn't be able to pass it off as Stephen Hawking, but "human-sounding" is still a bit of a stretch if you ask me.
I see a use case
There is an enormous number of books out there in electronic text format and often quite expensive, if available in audio.
Personally, I use them on long drives, but there are a lot of visually impaired people out there who could benefit from improved "auto voicing" of text material. Some of the existing stuff is pretty good, but can not only get tedious when listening to longer texts such as novels, but can be quite disconcerting when they come across strange words, names or places that don't exist in dictionaries, eg SF and Fantasy.
I did once convert a series of 8 novels for a blind friend, manually finding all the highlighted "I don't know this word" markers and changing them by spelling and phonetics to pronounce them more realistically, adding them to the custom dictionary so at least I only had to do each new word once. It was ok, but a long chore and not ideal for long listening sessions. Things have moved on since then, but this sounds like a at least a small, possibly a large leap forward.
only a matter of time before I get a voice mail from the CEO asking for a wire transfer to a new vendor in Nigeria