How-to
How to Make a Talking Bass (Yoi, Wow and Growl Basses)
Updated 2 min read
To make a talking bass, run a rich, distorted synth through a formant (vowel) filter and move that filter between vowel shapes in time with the beat. The vowel movement is what you hear as words: "ee" to "oh" to "ee" sounds like yoi, "oo" to "ah" to "oo" sounds like wow.
Why formants make a synth talk
When you say a vowel, your mouth and throat act like a set of filters with two or three main resonant peaks, called formants. Their positions are what make "ah" sound different from "ee", whatever pitch you sing. A formant filter recreates those peaks on any sound. Move it from one vowel to another and a synth starts to pronounce syllables.
U
- Sounds like
- "oo" in boot
- Character
- Dark, closed, hollow
O
- Sounds like
- "oh" in go
- Character
- Round, warm
A
- Sounds like
- "ah" in father
- Character
- Open, loud, aggressive
E
- Sounds like
- "eh" in bed
- Character
- Bright, nasal
I
- Sounds like
- "ee" in see
- Character
- Thin, piercing, very bright
How to make a talking bass in any synth
- Start with a rich source. A saw, a growly wavetable or an FM patch: the filter can only shape harmonics that are there.
- Distort it. Put a distortion or wavefolder before the filter so there are plenty of harmonics for the vowels to grab.
- Add a formant filter. Most wavetable synths have one (often called Formant, Vowel or Vocal), or use a separate formant filter plugin.
- Modulate the vowel. Assign an LFO or envelope to the vowel position, sync it to tempo (1/8 or 1/16 for riddim, 1/4 for slower wows) and draw the shape of the word.
- Distort again, then EQ. A second distortion after the filter adds grit to the vowels; cut the low end out of this layer and put a clean sub underneath.
- Resample. Bounce a few bars, chop the best syllables and rearrange them for the drop.
The tricky part is step 4: drawing an LFO shape that says a specific word takes trial and error, and changing the word means redrawing it.
The fast way: type the word into Hydra
Hydra has a formant mouth in every voice and a sequencer for it called Speak. Type a word and Hydra turns the letters into vowel steps on its U · O · A · E · I scale: YOI becomes I, O, I and WOW becomes U, A, U. Set Speak to Sync and pick a step length, or to Per note so each note says the word from the start. Slur decides whether the vowels snap or glide, Size makes the mouth small and nasal or huge and throaty, and Rage adds the distortion before or after the mouth.
Speak
YOIBite
Mouth
Size +25%Rage
Motion
The real plugin engine, running in your browser: same DSP code and presets as the VST3/AU/AAX.
Want to steer the vowels with your actual voice instead? Throat reads the vowels you sing and maps them to a synth filter live; our guide on turning your voice into a synth covers that approach.
Tips for clearer words
- Use contrasting vowels. I to U is the biggest jump and the clearest word; A to E is subtle.
- Keep the filter resonant. Higher focus or resonance makes vowels more pronounced, at the cost of a thinner sound.
- Distort before the vowels for clarity, after them for aggression.
- Leave space for the sub. Talking layers live in the mids; a clean sine underneath carries the weight.
FAQ
What makes a bass sound like it's talking?
Formants. Every vowel has its own set of resonant peaks: "oo" has low ones, "ee" has a low and a very high one. Sweeping a filter that recreates those peaks from one vowel to another, in rhythm, makes a synth sound like it's saying words such as "yoi" or "wow".
What is a yoi bass?
A yoi bass is a dubstep and riddim sound that seems to say "yoi": the formant filter moves from an "ee" shape through "oh" and back to "ee" on each note, usually on a distorted wavetable. In Hydra you can type YOI into Speak to get that movement directly.
Can I make a talking bass with a vocoder?
Yes, by feeding a recorded word into a vocoder with a bass synth as the carrier, but it's slower to tweak and the words come out smeared. Most producers use a formant filter on the synth instead, because the vowel movement can be synced, automated and resampled.