Allison Smith is a well-known voice actress who is famous for her work in the entertainment industry. She has provided her voice for several automated systems, including hotel wake-up calls and workout apps.
A great story that Allison shares is about how she once experienced the oddity of hearing her voice wake her up in a hotel room. Allison's husband even downloaded a fitness app with her voice to motivate him at the gym, but eventually switched to a different voice due to her constant encouragement.
Despite these amusing experiences, Allison Smith continues to be a highly sought-after voice artist in the industry.
On today’s episode, we discuss the various advancements in text to speech technology.
We touch on the possibility of developing a system that can detect someone's truthfulness and the challenges of doing so.
We also talk about ChatGPT, which is capable of mimicking a human voice flawlessly. Allison suggests that a hybrid approach combining both human and AI voices will be the future of the industry.
Key Takeaways
- Detecting lies and synthetic voices: Allison discusses the challenges of developing a system that can detect when someone is lying or telling the truth, as people have different ways of interpreting what sounds truthful. They also talk about the rise of new synthetic voices that mimic human speech and are used in entertainment, language localization, and other areas.
- Good IVR and vocal inflections: Allison explains the importance of having good interactive voice response (IVR) prompts that flow naturally and sound like a human conversation. They also discuss the importance of using different vocal inflections for numbers, dates, and other information.
- Voice-over experience and AI voices: Allison’s experience as a voice-over artist for various clients, including medical ads that require them to sound cheerful while listing side effects. They also talk about the evolution of text-to-speech technology and the emergence of Chat GPT, which can mimic human voices flawlessly. The speaker predicts that a hybrid approach combining human and AI voices will be the future of the industry, and mentions having a voice clone built based on their own voice.
- Emotional metrics and speech synthesis: Allison discusses the use of bots to analyze emotional metrics in callers' voices to measure urgency in crisis situations, and the possibility of measuring fertility by analyzing the sound of women's voices.
Timestamps
[00:00:00] TV job led to tanning salon recording.
[00:06:22] Text-to-speech technology advances the alarm voiceover industry.
[00:10:14] Recording Cepstral speech from script fragments.
[00:14:10] Technology measures emotion and fertility in voice.
[00:17:48] Tech-created better voices, from parametric to concatenative.
[00:22:17] Synthetic voices used in the entertainment localization industry.
[00:24:25] Detecting truth and lies is difficult.
Quotes
- The Trend in AI Voices: "They [AI developers] want to be very conversational... they want it to sound that casual and that conversational."
- The Future of Emotion Detection: "I was really, absolutely blown away by some of the things that they can do. For example, there was one speaker that was talking about if somebody were to call into a crisis line, and they would have Bots that would actually gauge exactly how urgent their request is just by measuring the metrics of their emotion in their voice, which is astounding."
Connect with Allison
LinkedIn - https://www.linkedin.com/in/allisonsmith3/
Website - https://www.theivrvoice.com/
Twitter - https://twitter.com/voicegal
