Choosing a voice

Female AI voices, grouped by sound

Female voices differ in three measurable ways, so we group the nine English picks by them. The three with the lowest spectral centroid sound warm and soft, led by Nova at 1,197 Hz. The three with the highest sound bright, led by Rosa at 3,917 Hz. The three with the narrowest intonation range sound calm, led by Isabella at 5.0 semitones.

Recommended voices

Nova

Female · American

STUDIO

Warm and soft: lowest spectral centroid

Centroid 1,197 Hz, the lowest of the nine. Median pitch 157.0 Hz, 8.9-semitone range, 222 words per minute.

Pitch
157.0 Hz
Range
8.9 st
Pace
222 wpm
Use voice
Aoede

Female · American

STUDIO

Warm and soft: second group member

Centroid 1,221 Hz. Median pitch 176.9 Hz, 7.9-semitone range, 197 words per minute.

Pitch
176.9 Hz
Range
7.9 st
Pace
197 wpm
Use voice
Michelle

Female · American

GLOBAL

Warm and soft: American narrator

Centroid 1,345 Hz. Median pitch 184.2 Hz and a 7.3-semitone range, at 167 words per minute.

Pitch
184.2 Hz
Range
7.3 st
Pace
167 wpm
Use voice
Rosa

Female · Philippine

GLOBAL

Bright: Philippine English

Centroid 3,917 Hz, the highest of the nine. Median pitch 201.8 Hz, 157 words per minute.

Pitch
201.8 Hz
Range
7.1 st
Pace
157 wpm
Use voice
Ezinne

Female · Nigerian

GLOBAL

Bright: Nigerian English

Centroid 2,953 Hz. Median pitch 191.7 Hz and a 6.7-semitone range, at 164 words per minute.

Pitch
191.7 Hz
Range
6.7 st
Pace
164 wpm
Use voice
Asilia

Female · Kenyan

GLOBAL

Bright: Kenyan English

Centroid 2,832 Hz. Median pitch 198.4 Hz and a 6.8-semitone range, at 158 words per minute.

Pitch
198.4 Hz
Range
6.8 st
Pace
158 wpm
Use voice
Isabella

Female · British

STUDIO

Calm: narrowest range

Intonation range 5.0 semitones, the narrowest of the nine. British Studio voice, median pitch 199.4 Hz, 203 words per minute.

Pitch
199.4 Hz
Range
5.0 st
Pace
203 wpm
Use voice
Emma

Female · British

STUDIO

Calm: British Studio voice

Intonation range 5.1 semitones. Median pitch 181.2 Hz, 200 words per minute.

Pitch
181.2 Hz
Range
5.1 st
Pace
200 wpm
Use voice
River

Female · American

STUDIO

Calm: quick American Studio pace

Intonation range 5.6 semitones. Median pitch 166.3 Hz. At 228 words per minute on our test passage, it suits short-form video.

Pitch
166.3 Hz
Range
5.6 st
Pace
228 wpm
Use voice

What the three groups mean for narration

Spectral centroid is the centre of a voice's energy, in Hz. Lower values sound warmer and softer, and higher values sound brighter. Intonation range, in semitones, shows how far the pitch moves within a sentence.

Warm and soft voices suit gentle narration such as personal stories. Bright voices carry more high-frequency energy and suit explainers and announcements where clarity matters most. Calm voices have ranges of about 5 to 6 semitones, so the delivery stays level across a long passage.

Pitch shows only part of the sound. Nova and Rosa differ by 45 Hz in median pitch, 157.0 Hz against 201.8 Hz, yet Rosa's centroid is more than three times Nova's.

Matching a female voice group to your script

Match the group to the text before you choose a voice. Personal stories suit a warm voice, product explainers suit a bright one, and long lessons or guided tours suit a calm one.

Punctuation controls pauses. Use a full stop between separate points so each one lands on its own.

  • Put each key point in its own short sentence.
  • Keep one idea per sentence in bright-voice scripts so the pace stays easy to follow.
  • Check the spelling of names and foreign words before generating.
  • Preview the same paragraph in two groups before you choose.

Male or female voice: choose by sound, not gender

The nine female picks run from 157.0 Hz (Nova) to 201.8 Hz (Rosa) in median pitch. Brightness varies far more: the centroid runs from 1,197 Hz to 3,917 Hz, which is why the groups above use centroid and range rather than pitch alone.

Choose the group that matches the tone you need, then compare two voices on your own script.

Same criteria in other languages

Hindi

Swara · Female · 255.5 Hz

Swara, a female Indian-accent voice at 255.5 Hz, the highest pitch of the three female Hindi voices measured.

Urdu

Uzma · Female · 206.4 Hz

Uzma, a female Pakistani voice at 206.4 Hz with a 6.6-semitone range, the higher of two female Urdu voices measured.

Bengali

Tanishaa · Female · 255.5 Hz

Tanishaa, a female Indian-accent voice at 255.5 Hz with a 12.8-semitone range, the wider of two female Bengali voices.

Arabic

Salma · Female · 214.5 Hz

Salma, a female Egyptian voice at 214.5 Hz with an 11.4-semitone range.

Spanish

Ximena · Female · 246.4 Hz

Ximena, a female European Spanish voice at 246.4 Hz with a 17.4-semitone range, the widest of the 24 female Spanish voices measured.

Tamil

Saranya · Female · 227.3 Hz

Saranya, a female Sri Lankan voice at 227.3 Hz with a 10.9-semitone range, the widest of four female Tamil voices.

Portuguese

Raquel · Female · 201.9 Hz

Raquel, a female European Portuguese voice at 201.9 Hz with a 12.8-semitone range, the widest of four female Portuguese voices.

French

Ariane · Female · 208.4 Hz

Ariane, a female Swiss voice at 208.4 Hz, higher than 83% of the seven female French voices measured.

Frequently asked questions

Nova (1,197 Hz) and Aoede (1,221 Hz) have the lowest spectral centroids of the nine English picks. Michelle (1,345 Hz) is next.
Rosa (3,917 Hz), Ezinne (2,953 Hz) and Asilia (2,832 Hz) have the highest spectral centroids of the nine picks.
Isabella (5.0 semitones), Emma (5.1) and River (5.6) have the narrowest ranges, so the delivery stays level. River is faster, at 228 words per minute.
Start with the warm and soft group: Aoede (1,221 Hz) at 197 words per minute, or Michelle (1,345 Hz) at a slower 167. Keep sentences short and preview a paragraph before generating.
Swara, an Indian-accent female voice at 255.5 Hz with an 11.9-semitone range, is the Hindi female pick on this page.
Tanishaa, an Indian-accent female voice at 255.5 Hz with a 12.8-semitone range, is the Bengali female pick on this page.

Try these voices on your own script

Paste your text in Voice Studio, pick a voice from this page and preview it before you generate.

Open Voice Studio