Deep voice

I wish to do a voiceover. The processed audio should sound deep, grainy. Unfortunately, I don’t have a deep voice. I tried several VST plugins. Eventually, they all sound fake or demonic. The pitch shift is better in Cubase with some success.

Next, I tried to model the human vocal tract. I tried splitting the audio into two channels. To the first channel, I added an audio chopper with a chopping triangle wave at 28 Hz. This is somewhat how the vocal cords distort in our vocal tract for deep-voiced people. This output was again made to go through a low pass filter of 1000Hz because the grain mostly affects the base voice and not the high-frequency sibilance sounds like T, F, P sounds.

The second channel was added to a high pass at 1000Hz allowing only sibilances to pass. The result was interesting, but there was something wrong. Some Uncanny valley in human listening.

Any suggestions?

Santanu -

my best suggestion is to hire a person with a deep voice :wink: there are plenty of options out there for hire.

Sulfur Hexafluoride?


Melodyne perhaps. Works better for me with the formant than Cubase

***I would move the pitch down in Cubase a 4th or 5th then mess with the formant shifting it down as well

I just did one here in Melodyne:

  1. Select all and move the pitch down
  2. Select formant tool
  3. Select all and move down to desired formant sound

Works great

Years ago, I did this quite a bit, changing voices(for games and edu stuff).

Convincing results depend quite a bit on a particular voice. Some sound great, others(with lots of noise components)not so much.

A female speaker with whom I worked quite a bit was great for this, I could make a convincing man from her, but she had to adapt her manner of speech otherwise she sounded too much like a girlish cartoon or comedy character.
She found this an interesting challenge, and became quite good at this.