This video is a demonstration of using voice synthesis tools to model a voice from scratch. I build up a masculine voice then I feminize it. This video is much different than usual. I just felt like making it and it was fun. As always, infinite thanks for being here.

► Support the Creation of Free Trans Voice Lessons:   / transvoicelessons  
► Want to take lessons? https://www.transvoicelessons.com/

►INSTAGRAM :   / digital.harmonic  
►PATREON :   / transvoicelessons  
►ORIGINAL MUSIC : https://zheaerosemusic.bandcamp.com/
►CONTACT: [email protected]
►WEBSITE: Transvoicelessons.com

VIDEOS YOU MAY HAVE MISSED:
► How To Master Voice :    • HOW TO MASTER ANYTHING: 7 Guidelines ...  
► Overview of Voice Feminization :    • The Big Picture of Voice Feminization...  
► Reduce Vocal Weight :    • FEMINIZE YOUR VOICE: 8 Exercises to R...  
► How to Hear Resonance :    • HEARING VOCAL RESONANCE  

ABOUT:
Basically voices are harmonic sources. The first step to making a voice is to make a harmonic series. Then, to avoid it sounding to hyperbuzzy or digital, you'll need to roll off the amplitude of the higher harmonics. This is called spectral tilt/spectral rolloff and it's the key variable behind "vocal weight".

Then, we will need to add some microvariance since no human being has perfect sound generation. We will also need a faint bit of white noise to create the sound of aspiration under the basic tone.

Next, we will need to add vocal resonance. For this video, I decided to start with the vowel /i/ 5% through the vowel scale range. This makes it quite masculine but still discernably an /i/ vowel.

Following that, I raise the pitch of it to set us up for feminization. The higher pitch makes the masculine resonance stand out as extra masculine and hollowish. Then, I shift the voice up it's /i/ vowel scale to about 90%. This point is actually more feminine than average. However, we then hear the weight we had was too high, so we lighten the voice a bit.

and... voila~! we now have a feminine voice singing the /i/ vowel. sounds like what we end up with form the foundation of the voice. articulation and coarticulatory behavior are far more complex to generate but are possible with bespoke effort.

note:
- resonance can and should scale with pitch. in these demonstrations resonance structures are static so if the pitch moves too far, the vowel identity will tend to skew in an uncanny way. That pitch against resonance tracking can be accommodated for in digital modeling but its beyond the scope of what I wanted to do here.

- there is also no reverb or attack envelope. both of these things would make it sound much realer.

warmly,
z

#VoiceScienceWorks #Voice #Feminization #Masculinization #Speech #Linguistics #TransVoiceLessons