by Clones are not coming, they’re already here — because, the Clone, is you! Yes, the call is coming from inside your head! I now have an AI Voice clone of my voice trained on over 30 hours of my Human Meme podcast. I trained my AI voice on my Human Meme podcast so the source material was clean, and well-edited – and because of that early care in podcast recording, and production, my new AI voice is shockingly, and incredibly real and, frankly, Uncannily Fantastic! Plus, there is probably a little tinge of terror tossed in there too — because now my voice can live on without me. Let’s hope it behaves! From this moment on, I have touched true immortality. Would you ever want it any other way?
Here’s an example of the result of my AI Clone voice. AI upon AI. Here’s the first — “Mr Watson, come here. I want to see you.” — moment of my just-birthed AI voice reading AI generated text about the rise of Fascism in America:
That was a thrill! Now, here’s my AI voice hosting my Human Meme podcast! Or should that be the INhuman Meme podcast?
This voice cloning is amazing technology. It was a challenge to find 3 hours of me speaking where each audio file had to be at least 30MB but not larger then 500MB. I’m lucky that it all worked out. There are even some “sighs” and other guttural instances that are now part of my AI Voice profile, and as I understand it, those utterances are rare in the finished AI voice profile.
Now there are problems with vocal immortality. The AI voice exists. People with bad intentions could use that voice to make one say things that are terrible and irrevocable. Is there some sort of embedded ID in the AI voice to mark it as AI and not a natural voice recording? We hope so!
AI voice is also excellent for people who may lose their voice in the future due to an accident or medical condition. With an AI voice, you can continue to “speak for yourself” just by typing what you want to say into a text box. Yes, it’s just that simple.
Voice once. Live forever!
Now let’s turn to our AI research friends share these insights and history about Voice Cloning:
Voice cloning, which is the process of replicating a person’s voice using artificial intelligence, originated from early speech synthesis techniques. Initial speech synthesis programs utilized rule-based methods, where phonetic and prosodic rules were used to generate artificial speech. However, these had very robotic-sounding outputs.
The advent of machine learning and AI in the last few decades triggered significant improvements. In the mid-2010s, advancements were made in textual synthesizing voices, particularly with systems like DeepMind’s WaveNet. This marked a significant step toward more realistic, natural sounding cloned voices.
AI voice cloning today often utilizes deep learning techniques such as sequence-to-sequence models or other neural networks. One popular approach is a type of generative model called a Variational AutoEncoder (VAE) that is capable of outputting new data after being trained on a dataset.
Deep learning-based voice cloning methods also frequently use an architecture known as Transformer, which is often used in natural language processing (voices are converted to text, processed, and then converted back to voice). This approach has found excellent results in both voice cloning and other synthetic voice tasks.
AI voice cloning has a host of potential applications and benefits:
1. Accessibility: Voice cloning can assist individuals who’ve lost their ability to speak due to illness or an accident, by recreating their original voice.
2. Entertainment: In the film and gaming industries, voice cloning can be used to generate dialogues, removing time and resource-consuming traditional methods.
3. Personalization: In customer service, personalized AI voices can offer more natural interactions, enhancing the experience.
Despite the above benefits, the technology also comes with potential dangers and ethical concerns:
1. Deception: The realistic nature of AI voice clones can lead to deception and fraud, such as deepfake audios or scam calls.
2. Consent and Privacy: Using someone’s voice without their permission breaches their privacy and ethical guidelines.
3. Misrepresentation: In the wrong hands, fake voice clips can be produced to misrepresent someone’s views, potentially causing harm to personal or professional reputations.
With an AI Voice clone, there is no doubt your surrogate lives! You are no longer bound by space and time. You could record a podcast, or a message for the future, that can then be triggered and played for the world long after you’ve sprung from this mortal coil. Now that isn’t an afterlife, it’s real life!
If you want to clone your voice, beware that there are a lot of services out there, but they are not all the same quality. Some services are better, and easier to use, than others. Try them out. Ask around. Listen. Make voice waves. Live forever on your own terms. You won’t regret the mind-blowing experience!
@boles make it read the navy seal copypasta?
That’s a funny idea — but a little too rough language for here, I fear!
@boles Excellent article! You're your own Immortal Man!
You are not only a man of great taste, and learning, you’re figuring out this Fediverse propagation thing and commenting where and how! You need another YOU, not just another voice. Hmph!