Using Uberduck AI Voice Generator to Create New Music
Ever wish you could hear a famous vocal artist featured on your track? Thanks to Uberduck and other popular AI voice generators, it's becoming easier every day. Just type in a short prompt and in less than a minute, you've got a video of Drake rapping and singing about your topic of choice.
Uberduck is a free text-to-speech (tts) web app that synthesizes celebrity voices instantly and gets them to say anything you want.
In the video above, YouTuber Ed Talenti uses a free AI lyric generator to come up with some content. He feeds that into Uberduck, exporting the audio to his DAW and chopping it up to match one of his beats.
Talenti's imitation of Drake, Kendrick Lamar, and Playboi Carti are entertaining to watch. They probably would never be mistaken for the actual artist, but that doesn't matter. Youtubers are proving that with some pitch shifting and vocal effects, the right engineer can make these ai voices sing.
1. Creating Vocal Tracks with UberDuck AI
2. Using GPTChat for Lyrics and AudioCipher for Melodies
3. The Uberduck Discord Channel
4. AI Alternatives to Uberduck
Creating Vocal Tracks with Uberduck AI
The interface says it all. You just enter a phrase and let the text-to-speech feature do its magic. Within moments, you’ll have an audio file that’s free to download.
To try the app for free, visit uberduck.ai and explore their various tools. The site offers over two thousands voices to choose from. Pick from rappers like Kanye West and 21 Savage. Explore fictional characters like Family Guy's Peter Griffin or Breaking Bad's Walter White. The world is your oyster.
Fair warning — expect some trial and error to get the vocal takes you want. Most of the time, the initial audio output sounds like a garbled version of the voice it’s imitating. I found that the speech synthesis works better with some words than others.
Using AudioCipher + ChatGPT + UberDuck
In December 2022, music producers on IG and TikTok started using OpenAI's free tool, ChatGPT, to create funny meme videos of famous rappers. Influencers are coming up with prompts like "Write a drake verse about how he doesn't like beans in his chili" and used Uberduck to impersonate their voices (poorly).
AudioCipher is a text-to-MIDI generator that lets you type in words and spit out melodies in your key signature of choice. You can use the randomize rhythm feature to audition different patterns until you find something that fits.
So if you have a fun idea for a song, you can ask ChatGPT to produce the lyrics, UberDuck to produce the vocals, and experiment with AudioCipher to build melodies and chord progressions that fit the vibe. It's a fun way to break outside your everyday songwriting habits and lighten things up a bit.
Once you have a melody you like, you'll just need an autotune plugin to make the spoken text sing. If you're looking for a free option, try GVST's GSnap VST. If you come up with a banger, you can always strip out the UberDuck vocals and bring in a human vocalist to improve the song quality. Think of this is as a quick way to prototype new ideas.
Uberduck Discord
Uberduck launched in 2020 and within a year, YouTube videos about the app started going viral. Video editors had rappers performing fake verses on unreleased beats. People also started using Uberduck to create voiceovers for their TikTok videos, when the platform made it difficult for users to do it otherwise.
Exciting as it may be to revive Juice WRLD and XXXTentacion for one final collaboration, TTS voice generators are capable of a lot more. Founder Zach Wener publishes articles on the Uberduck AI blog regularly, exploring different applications for their technology.
Earlier this year, Wener wrote an article for developers about a Discord chat bot you can configure to turn text to speech within a voice channel.
Gamers and streamers who don’t want to use their voice but still want to participate in voice channel now have that option. The demand for TTS voice overs has already been proven with the popularity of Roblox stories on social media.
With Uberduck's “private voice clone” plan, users can actually use the API to train on samples of your own voice. When that process is complete, the Discord chat bot could technically be set up to turn text into an AI version of you!
If that sounds like science fiction, check out this video from a viral youtuber who personally became one of the celebrity voices that users could imitate.
Free alternatives to Uberduck AI
With a growing ai community of over 14,000 users on their Discord server, the Uberduck text-to-speech framework could evolve into something powerful in the future. The pricing model is generous for free users and scales up only for those who have a serious need for it.
We've typed up a full list of the best AI voice generators for musicians here.
The one knock against Uberduck is that they don’t currently offer an option in the app store. If you’re looking for something that loads on your phone, with or without the internet, try Apple’s classic Text to Speech! app. Otherwise, here are the two best options available currently.
SynthV by Dreamtronics
SynthV is a vocal synthesizer designed specifically for creating music. It's more powerful than Uberduck in many ways. Synthesize V leverages a deep neural network-based synthesis engine to generate its realistic singing voices. You can control vocal dynamics like breathiness, nasal tone, and more.
The main downside is that it has not been trained on celebrity voices, so you won't be able to imitate famous rappers. It was developed in Asia, so there's a strong focus on anime characters and non-english voices.
Text-to-Song by Voicemod
Voicemod is a company responsible for a popular popular voice changer tool on Discord. Gamers have historically used it to mask their voice and sound like someone else while in a group chat.
Text-to-Song was released by Voicemod in December 2022, allowing users to type in words and hear them sung back with pre-created melodies. You don't have much control over the sound but if you're just trying to make some funny memes for social media, this is a good option.
Check out their site here: https://tuna.voicemod.net/text-to-song/
Google's Text to Speech API
It probably comes as no surprise that Google has a text-to-speech API available as well. So while Uberduck may be the one attracting all the fun memes, tech giants are building their own transformers on siloed machine learning frameworks under their control.
This may be one of the best current demonstrations of how AI text-to-speech could join with generative AI music software to create mechanized super-musicians in the future.
For now, while we fend off the robot apocalypse, any music producer can use Uberduck to generate audio samples quickly. If you're someone who doesn't like to sing or rap, but has ideas in their head that they want to get out, this might be a good starting point.
It's also just kind of fun to mess around with. Enjoy yourself and if you haven't already, be sure to visit our homepage to learn more about AudioCipher.