Don't clone someone else's voice! The 4 best tools to clone any voice using artificial intelligence
It’s safe to say that 2024 will be the big year for AI for the general public. Text-to-image generation apps are booming, and language models like ChatGPT are changing entire industries overnight.
Voice cloning using AI has also seen tremendous developments in recent months, and can now be considered a technology mature enough to have many practical uses while achieving the most reliable results.
If you are still unsure about the purpose of voice cloning, you will be interested to know that it is a deepfake technology capable of identifying, copying and replicating the tone of voice, inflections and nuances of a person's speech through artificial intelligence.
Once the AI is trained and learns how to make this voice, we can ask it to read any text using that person’s voice. This allows us to use AI to make that person appear in a video or audio clip speaking in another language, but also for many other things. In fact, today you can use voice cloning for audio videos, texts or audiobooks with your own voice without having to go through a tedious recording process or have good audio equipment.
Recently, many memes and humorous montages with cloned voices of celebrities have gone viral. Be careful, and remember that cloning someone else's voice without their permission is unethical, and no matter how "funny" it may be at a given moment, it may also have legal consequences. Use these types of tools only for legitimate purposes.
- VoiceLab de ElevenLabs
ElevenLabs has one of the best voice cloning software out there. You can clone a voice with just a 1-minute voice sample, and achieve very realistic and professional results using as little as 30 minutes of your voice.
An interesting detail is that the app has a verification system to verify that we are not cloning someone else's voice. From there, we can clone our voice in up to 28 languages and more than 50 different dialects.
In its free version, ElevenLabs lets you use up to 10,000 characters per month, with 3 custom voices and high-quality audio at 128 kbps.
- Deep Voice 3
It is an open source voice cloning software that you can install on your computer for local use. It is available through GitHub and is a neural text-to-speech (TTS) application that uses deep learning techniques to synthesize human voices.
The process of installing and configuring Deep Voice 3 is quite advanced and requires technical knowledge, so it is only recommended for those who have a lot of free time and want to "sauce" a little.
- They look like AI
With Resemble, you can clone your voice by recording just 25 phrases into the microphone, or in other words, providing about 3 minutes of data. From there, the AI will be able to train itself and synthesize your voice correctly, avoiding the typical robotic voices that sound so bad in some primitive voice clones.
Thanks to its highly realistic results, Resemble currently has over a million users. Of course, keep in mind that this is a premium tool aimed at the professional market, with a base price of $0.006 per second.
- Tortoise-TTS
It is another open source tool for advanced users. The application runs on Google Colab Notebook and requires the use of a recording program like Audacity to provide a voice sample for the AI.
The advantage of this type of tool is that it is free. If you want to try it out, you can take a look at the project’s GitHub site . You can also find an interesting tutorial on YouTube on how to do the whole cloning process step by step.
- LOVO
It is a web application that specializes in voice cloning. It requires a voice recording of at least one minute, which can be through a microphone or an audio file. The truth is that the process is really simple and straightforward.
Although this is another premium app, LOVO has a 14-day free trial and clones up to 5 different voices, although the generated audios cannot be downloaded.