ai group vocals

· By Will Harken

2 Ways To Recreate Group Vocals With AI

So, we all know AI can mimic any singer, right? But it's not just about one voice. Try merging multiple singers, and it gets complicated. This guide dives into how you can recreate group vocals from your favorite tracks.

Strategies for Handling Group Vocals

Using AI for group vocals? Two main strategies come to mind.

Option 1: Create a mono group vocal model and re-widen it. It's quick, though lower quality.

Option 2: Create each vocal layer with a solo vocal model. Takes longer, but sounds better.

Currently, public AI models don't support stereo group vocals. But it's doable - companies like Udio can generate realistic stereo group vocals. Just not for converting existing vocals.

Creating a Mono Group Vocal Model

If singers are in unison or octaves, a group model might work. Use all instances of group vocals from the song as training data. However, don't expect radio-quality results.

This method is good for demos or music made for the feels.

First, isolate the group vocals and train a model on them. The AI will attempt to convert it into a group vocal. But beware—pitch issues or warbling can appear, hard to fix even with Melodyne.

The upside? You save time by not needing to recreate every vocal detail from the original mix.

Creating Each Vocal Layer with a Solo Vocal Model

This involves re-engineering each vocal layer one by one. It's time-consuming and requires a good ear for harmonies.

Picture this: Freddie Mercury on lead, Taylor Swift with a counter melody, and James Brown on backing harmonies. You'd need separate conversions for each and then mix them.

The solo vocal method yields the most professional sound.

Combining Both Strategies

I often use a mix of both strategies for a lyrics swap (check out that service here). Layer a group AI vocal as a texture under high-quality solo vocal harmonies for the best of both worlds.

Hopefully, some AI genius will develop an automatic tool for this someday.

Mixing Tips for Group Vocals

When mixing, use a compressor on your vocal bus to "glue" them together. Saturation adds energy and masks AI flaws. I usually go with Nectar.

Consider plugging into tools like Nectar and Neutron. Personalized music with my help can elevate your projects.

Need a song? Let's make one:

A useful tool for mixing is iZotope's Unmask. It only ducks frequencies clashing with your vocals when they come in—better than an EQ affecting the whole track.

Check out iZotope plugins here. Keep harmonies simple since AI struggles with complex ones. Training on an octave? It works okay. Training on a group harmony? Not so much.

Interested in more about AI and music?

Blog Links

  1. Anyone Can "Sing" Now: AI Voice Cloning
  2. The AI Vocal Mixing Technique No One's Talking About
  3. How AI is Revolutionizing Music

Helpful Tools

  1. iZotope Plugins for AI Mixing and Mastering
  2. Synthesizer V - Digital AI Vocals That Sound Really Good
  3. AI Vocals:


Leave a comment