How to make voices sound more energetic?

I want them to sound like they’re enjoying singing. I’ve got tuning down mostly, but my results seem kind of bored despite the source material not sounding it at all.

You haven’t given any information about the song, the voice database you’re using, etc., but in general if you have a reference track you can drop it into Melodyne (or any similar pitch correction/analysis tool), study the pitch patterns of the original, and try to recreate them in SynthV Studio.

The pitch patterns generated by the AI engine are based on how the original voice provider sings, so unless the original recordings were very energetic, the automatic pitch curves won’t be either. This means you either need to pick a difference voice database, or create the pitch curves yourself.

2 Likes

Hi, yes. I’m using Hayden and covering Not So Secretly by Dylan Conrique, with a template created with the audio processing function of the beta version. I guess I’m more wondering what kind of pitch curves, etc. make a vocal seem energetic? Or what does in general. Sorry for not providing enough information.

There are a lot of things, but they’re all very situational and the only real solution is learning by trial-end-error, or studying reference material.

A general list:

  • Analyze reference material (ie in Melodyne, Vocalshifter, etc.) and try to reproduce those patterns
    • Of course, you could just import the reference into SynthV Studio using Voice-to-MIDI, but copying the pitches directly won’t help develop your own skills
  • Adjust the various parameters at specific moments, not just as baseline values
    • Increase tension at faster segments, or during belted parts
    • Reduce the gender parameter slightly at more energetic moments
    • Experiment with tone shift in conjunction with tension and gender adjustments
  • Include breaths
  • Use pronunciations that match how people actually speak/sing, rather than relying on dictionary-perfect enunciation
    • This generally involves reproducing connected speech patterns, as well as using “lazy” phonemes (such as ax instead of certain vowels, or dx instead of t)
1 Like

Thanks for the pointers. I’ll go out and do a trial-by-error.