What is the vowel space?

The vowel space is a two-dimensional acoustic representation of vowel sounds, plotted using formant frequencies F1 (first formant) and F2 (second formant). Unlike the traditional IPA tongue diagram which estimates tongue position, the vowel space represents what we actually hear; resonances in the vocal tract. The three corner vowels (i, a, u) form a roughly triangular shape, with the colorless schwa vowel at the center.

What are formants in speech?

Formants are resonant frequencies produced by the vocal tract that characterize different vowel sounds. F1 (first formant) correlates with vowel height: high F1 means an open vowel like 'ah', low F1 means a close vowel like 'ee'. F2 correlates with vowel frontness: high F2 means a front vowel like 'ee', low F2 means a back vowel like 'oo'. These can be measured from any audio recording using speech analysis software.

How does the interactive vowel synthesizer work?

The p5.js sketch uses four bandpass filters (peaking filters) set to the F1, F2, F3, and F4 formant frequencies. A sawtooth oscillator provides the fundamental frequency (pitch), and the filters shape the harmonics to produce vowel-like sounds. Moving your mouse changes the F1 and F2 filter frequencies in real-time, allowing you to glide continuously between vowel qualities.

Why is the vowel space triangular?

The triangular shape reflects acoustic physics, not tongue position. The three corner vowels (i, a, u) represent maximally distinct acoustic positions: 'i' has high F2 and low F1; 'a' has mid-range F1 and F2; 'u' has low F2 and low F1. These three points define the perimeter of perceivable vowel qualities. This is similar to how the CIE color diagram is triangular because human eyes have three types of color receptors (cones).

How is this different from the traditional IPA vowel chart?

The traditional IPA vowel chart is based on articulatory assumptions about tongue position (high/low, front/back) that don't accurately represent what we hear. Dr. Geoff Lindsey's acoustic vowel space uses actual formant frequencies, which can be measured objectively. The IPA chart also conflates lip rounding with tongue position, while the acoustic space accounts for lips independently affecting the sound.

Explore the Vowel Space: Interactive Sound Synthesis with Dr. Geoff Lindsey's Model

⬅️ Back to Tutorials

📺 Interactive Tool

Open the Vowel Space Synthesizer →

Click and hold anywhere on the vowel space diagram to hear the corresponding vowel sound. Drag your mouse to glide continuously between vowel qualities. It’s just like moving between colors in a color picker.

🎯 HOOK

For over a century, phoneticians taught that vowels could be mapped on a simple chart based on tongue position. But what if the real story (the one your ears actually hear) is completely different? An interactive synthesizer lets you explore the vowel space firsthand.

💡 ONE-SENTENCE TAKEAWAY

This tutorial reveals the acoustic reality of vowel sounds through an interactive synthesizer that demonstrates how formant frequencies (not tongue position) create the vowels we hear, a shift from the 19th-century IPA tongue diagram to modern phonetic science.

📖 SUMMARY

The vowel space is one of the most fascinating concepts in acoustic phonetics, yet it remains largely unknown outside specialized circles. Developed and championed by Dr. Geoff Lindsey, the vowel space offers a scientifically accurate way to understand and visualize how vowel sounds actually work.

Traditional vowel charts, dating back to Alexander Melville Bell in the 19th century, describe vowels in terms of tongue position; high, low, front, back. This “tongue space” became the foundation of the International Phonetic Association’s vowel quadrilateral. But as Dr. Lindsey explains, this approach is fundamentally flawed: phoneticians were describing what they thought their tongues were doing, not what their ears were actually hearing.

The breakthrough came when researchers like George Oscar Russell published X-ray studies in 1928 showing that vowel quality is determined by acoustic resonances (formants), not tongue position. The first two formants (F1 and F2) are sufficient to differentiate vowels audibly, and these can be measured objectively from any recording using free software like Praat.

The interactive p5.js synthesizer at the heart of this tutorial implements Dr. Lindsey’s formant-based approach. By mapping F1 vertically (high = open vowels like “ah”, low = close vowels like “ee”) and F2 horizontally (high = front vowels like “ee”, low = back vowels like “oo”), you can explore the full vowel space by clicking and dragging. The synthesizer uses four bandpass filters to shape a sawtooth wave into vowel-like sounds, demonstrating in real-time how acoustic manipulation creates the illusion of different vowel qualities.

This tool transforms abstract phonetic concepts into tangible auditory experiences. Rather than memorizing that /i/ is a “high front vowel,” you can hear exactly what happens as you move from /i/ to /u/; the F2 frequency drops, creating that characteristic “back” quality. The continuous nature of the space also reveals something crucial: there are no sharp boundaries between vowels, only gradients.

🔍 INSIGHTS

Core Insights

Formants, not tongues: Vowel identity is determined by acoustic resonances (formants), not articulatory gestures. Two speakers with completely different tongue positions can produce identical vowel sounds if their vocal tract resonances match.
Continuous space: Vowels exist on a gradient, not discrete slots. Moving from /i/ to /e/ to /a/ produces smoothly changing qualities. You can stop anywhere and still produce a valid vowel sound.
Three-corner reality: The vowel space has a roughly triangular shape with three maximally distinct corner vowels: /i/, /a/, and /u/. This parallels the CIE color diagram’s three corner colors (red, green, blue).
Schwa at center: The most “neutral” vowel, schwa /ə/, sits at the center of the space. It’s the vowel you produce when not intentionally shaping your mouth.
Lip rounding matters: Unlike the traditional IPA chart, the acoustic space accounts for lip position independently. Rounding the lips while keeping the tongue still lowers both F1 and F2, creating a distinctly different sound.

How This Connects to Broader Topics

Language learning: Understanding the vowel space helps learners of foreign languages identify and produce vowel contrasts their native language doesn’t have
Speech technology: Text-to-speech systems use formant synthesis or neural vocoders that manipulate these same frequencies
Forensic phonetics: Formant analysis can help identify speakers from voice recordings
Historical linguistics: Vowel shifts can be tracked by measuring formant changes across time in recorded speech

🛠️ FRAMEWORKS & MODELS

The Acoustic Vowel Space Model

A two-dimensional coordinate system for vowel sounds:

Axis	Frequency Range	Correlates With	Low Value	High Value
F1 (vertical)	250–1000 Hz	Vowel height/openness	/u/, /i/ (close)	/a/ (open)
F2 (horizontal)	540–2600 Hz	Vowel frontness	/u/, /o/ (back)	/i/, /e/ (front)

Cardinal Vowels in the Interactive Tool

The synthesizer includes eight primary cardinal vowels (shown in the main diagram):

/i/ : 275 Hz F1, 2400 Hz F2 (close front)
/e/ : 412 Hz F1, 2150 Hz F2
/ɛ/ : 620 Hz F1, 1800 Hz F2 (open mid front)
/a/ : 900 Hz F1, 1350 Hz F2 (open central)
/ɑ/ : 710 Hz F1, 1050 Hz F2 (open back)
/ɔ/ : 530 Hz F1, 830 Hz F2 (open mid back)
/o/ : 380 Hz F1, 690 Hz F2 (close mid back)
/u/ : 275 Hz F1, 600 Hz F2 (close back)

Synthesizer Technical Implementation

The p5.js implementation uses:

// Four peaking filters create the formant resonances
F1F.set(yCoordtoF1(mouseY), 0.2); // First formant from vertical position
F2F.set(xCoordtoF2(mouseX), 0.4); // Second formant from horizontal position
F3F.set(f3, 0.8); // Third formant (adds "quality")
F4F.set(f4, 1); // Fourth formant (naturalness)

The sawtooth oscillator provides source harmonics, and the filters amplify specific frequency bands to create the resonant peaks that characterize vowels.

💬 QUOTES

“These early phoneticians were much like astronomers before Galileo… They thought they were describing the highest point of the tongue, but they were not. They were actually describing formant frequencies.” Peter Ladefoged, Vowels and Consonants
“Phoneticians are thinking in terms of acoustic fact, and using physiological fantasy to express the idea.” George Oscar Russell, 1928
“If you take any two vowels you can move continuously from one to the other and even stop halfway if you like.” Dr. Geoff Lindsey
“The parallels between color and vowels go further: languages can differ both in how many vowels they have and how many basic color tones they have.” Dr. Geoff Lindsey

⚡ APPLICATIONS & HABITS

Practical Uses of the Vowel Space Synthesizer

Pronunciation training: Move to a target vowel and try to match it with your own voice
Contrast identification: Practice hearing the difference between similar vowels in a foreign language
Dialect exploration: See where the vowels of different accents sit on the space
Teaching tool: Demonstrate to students that vowels are continuous, not discrete
Curiosity satisfaction: Just play around and discover how vowels “work”

How to Use the Interactive Tool Effectively

Start at the corners: Click on /i/, /a/, and /u/ to hear the three maximally distinct vowels
Trace the edges: Drag from /i/ to /a/ along the front vowel series, then /a/ to /u/ along the back series
Find schwa: Click the center of the space, as this neutral vowel appears in unstressed syllables
Explore your language: If you speak a Romance language, find where /e/ and /ɛ/ sit; for English, locate your TRAP, LOT, and STRUT vowels
Glide continuously: Don’t just click, hold and drag to hear the gradient between any two vowels

Recording Your Own Vowels

To compare your pronunciation with the synthesizer:

Record yourself saying vowels in isolation (e.g., “hee haw haa hoe who”)
Use Praat (free, cross-platform) to extract formant values
Plot your F1/F2 values on a chart like the one in the interactive tool
Compare positions with the cardinal vowels to identify any systematic differences

📚 REFERENCES

Dr. Geoff Lindsey’s blog post: The Vowel Space
YouTube video explanation: The Vowel Space - YouTube
Interactive p5.js synthesizer: Vowel Space Synthesizer
Praat speech analysis software: praat.org
Peter Ladefoged, Vowels and Consonants (recommended reading)
CIE 1931 chromaticity diagram (the color space analogy)

🔗 Related Concepts to Explore

Formant synthesis: The technique used in this tool, building speech from resonant filters
Source-filter theory: The acoustic model of speech production (source = vocal cords, filter = vocal tract)
Spectrograms: Visual representations of formant frequencies in speech
Vowel formants in different accents: How F1/F2 values vary across dialects

Credit for the vowel space diagram and model: Dr. Geoff Lindsey. His blog post and video provide extensive detail on this fascinating topic.

Crepi il lupo! 🐺