How to Choose a Voice for Brand Character Voice-over
Why a 'Pretty Voice' Is Not a Criterion
Casting directors often hear from clients: «I want a voice like a newsreader from a federal channel» or «something pleasant, warm». This is not a brief, but a taste preference, and it's impossible to select an actor based on this. A brand character's voice is a set of measurable characteristics, each of which works toward a specific communication task.
- Timbre — high, medium, low; determines whether the character is perceived as light/friendly or weighty/authoritative.
- Speech Tempo — fast speech tempo reads as energy and youthfulness, slow — as deliberation and expertise.
- Voice Age — not the actor's passport age, but how many years the voice sounds in the recording.
- Accent and Diction — regional dialect or, conversely, deliberately «neutral» pronunciation.
- Emotional Range — the actor's ability to switch from neutral intonation to irony, anxiety, excitement without strain.
Each of these parameters must be fixed before casting, not selected intuitively during demo listening.
Voice and character archetype: how they are related
Voice is the voiced character, so the selection is always made from the archetype, not from the pool of voice actors. The same text read by the 'Sage' and the 'Joker' forms a different attitude towards the brand in the listener's mind, even if the words are identical.
| Character archetype | Voice characteristics | Example of application |
|---|---|---|
| Mentor / Expert | Low pitch, measured tempo, calm intonation without accelerations | Financial services, legal products |
| Friend / Companion | Medium pitch, conversational tempo, lively pauses, light smile in the voice | Applications for daily tasks, delivery |
| Rebel / Provocateur | Sharp accents, uneven tempo, contrasting intonations | Youth brands, streaming, gaming |
| Joker / Trickster | High pitch, fast speech, playing with voice height | Entertainment applications, children's products |
| Guardian / Care | Warm middle pitch, soft dynamics, minimum of sharp fluctuations | Health-tech, insurance, home products |
Case: how they selected the voice for the delivery voice assistant
The team developed a character for the voice assistant of the food delivery service. According to the brief, the character was supposed to sound like an 'energetic courier-friend', not like a formal call center operator. Six actors were invited to the casting, each read the same text from three scripts: order confirmation, apology for delay, and a joking remark upon a bonus.
An interesting finding emerged: two voices that perfectly sounded in the positive script completely failed in the apology script — excessive cheerfulness was perceived as falseness and irritated the test group. The final choice was a voice with a middle pitch and slightly slower tempo than initially required in the brief: it maintained energy in positivity and did not lose trust in a difficult situation. The team's conclusion: voice testing should not be done on a 'lucky' text, but on the most inconvenient usage scenario — that's where its real range is revealed.
Technical criteria: live actor or synthetic voice
A separate question is whether to use actor recordings or AI-based synthesized voice. Synthesis is cheaper to scale: the same voice can be used in an app, on an answering machine, and in a commercial without repeated studio changes. A live actor provides emotional nuances that synthesis currently reproduces worse, especially in sarcasm, embarrassment, and genuine joy.
| Criterion | Live actor | Synthetic voice |
|---|---|---|
| Cost of scaling | Grows with the volume of text and languages | Almost does not grow after initial setup |
| Emotional Range | High, subtle shades | Medium, requires manual tuning of intonations |
| Speed of content updates | Requires studio re-recording | Generation in minutes |
| Legal protection of the voice | Contract with the actor for transferring rights | License for the voice model from the vendor |
Frequent errors in selecting a character's voice
Failures in this task repeat from brand to brand, and almost all of them are not related to the actors' talent but to the selection process.
- Selecting a voice based on one short demo without testing on complex emotional scenarios.
- To orient on the personal taste of the client, not on the archetype of the character and the target audience.
- Do not check the voice on long texts — vocal fatigue manifests only after several minutes of recording.
- Ignore age perception: the voice of a 45-year-old actor may be perceived by the audience as 'a parent's voice', which does not suit a product for teenagers.
- To forget to document the voice parameters, which leads to the next contractor selecting a new actor 'by eye'.
How to fix the voice in voice guidelines
The chosen voice needs to be described just as strictly as a logo is described in a color and spacing guideline. The document should include: the pitch range in hertz or a verbal description (e.g., 'medium-low, warm'), tempo in words per minute, the acceptable emotional range with example phrases, a list of prohibited intonations (e.g., 'without sarcasm', 'without office-style pauses'), and also reference audio recordings for verification when switching contractors. Without this document, each new voice studio will reinvent the brand voice — and sooner or later it will 'blur' across different content carriers.
Why a 'Pretty Voice' Is Not a Criterion
Casting directors often hear from clients: «I want a voice like a newsreader from a federal channel» or «something pleasant, warm». This is not a brief, but a taste preference, and it's impossible to select an actor based on this. A brand character's voice is a set of measurable characteristics, each of which works toward a specific communication task.
- Timbre — high, medium, low; determines whether the character is perceived as light/friendly or weighty/authoritative.
- Speech Tempo — fast speech tempo reads as energy and youthfulness, slow — as deliberation and expertise.
- Voice Age — not the actor's passport age, but how many years the voice sounds in the recording.
- Accent and Diction — regional dialect or, conversely, deliberately «neutral» pronunciation.
- Emotional Range — the actor's ability to switch from neutral intonation to irony, anxiety, excitement without strain.
Each of these parameters must be fixed before casting, not selected intuitively during demo listening.
Voice and character archetype: how they are related
Voice is the voiced character, so the selection is always made from the archetype, not from the pool of voice actors. The same text read by the 'Sage' and the 'Joker' forms a different attitude towards the brand in the listener's mind, even if the words are identical.
| Character archetype | Voice characteristics | Example of application |
|---|---|---|
| Mentor / Expert | Low pitch, measured tempo, calm intonation without accelerations | Financial services, legal products |
| Friend / Companion | Medium pitch, conversational tempo, lively pauses, light smile in the voice | Applications for daily tasks, delivery |
| Rebel / Provocateur | Sharp accents, uneven tempo, contrasting intonations | Youth brands, streaming, gaming |
| Joker / Trickster | High pitch, fast speech, playing with voice height | Entertainment applications, children's products |
| Guardian / Care | Warm middle pitch, soft dynamics, minimum of sharp fluctuations | Health-tech, insurance, home products |
Case: how they selected the voice for the delivery voice assistant
The team developed a character for the voice assistant of the food delivery service. According to the brief, the character was supposed to sound like an 'energetic courier-friend', not like a formal call center operator. Six actors were invited to the casting, each read the same text from three scripts: order confirmation, apology for delay, and a joking remark upon a bonus.
An interesting finding emerged: two voices that perfectly sounded in the positive script completely failed in the apology script — excessive cheerfulness was perceived as falseness and irritated the test group. The final choice was a voice with a middle pitch and slightly slower tempo than initially required in the brief: it maintained energy in positivity and did not lose trust in a difficult situation. The team's conclusion: voice testing should not be done on a 'lucky' text, but on the most inconvenient usage scenario — that's where its real range is revealed.
Technical criteria: live actor or synthetic voice
A separate question is whether to use actor recordings or AI-based synthesized voice. Synthesis is cheaper to scale: the same voice can be used in an app, on an answering machine, and in a commercial without repeated studio changes. A live actor provides emotional nuances that synthesis currently reproduces worse, especially in sarcasm, embarrassment, and genuine joy.
| Criterion | Live actor | Synthetic voice |
|---|---|---|
| Cost of scaling | Grows with the volume of text and languages | Almost does not grow after initial setup |
| Emotional Range | High, subtle shades | Medium, requires manual tuning of intonations |
| Speed of content updates | Requires studio re-recording | Generation in minutes |
| Legal protection of the voice | Contract with the actor for transferring rights | License for the voice model from the vendor |
Frequent errors in selecting a character's voice
Failures in this task repeat from brand to brand, and almost all of them are not related to the actors' talent but to the selection process.
- Selecting a voice based on one short demo without testing on complex emotional scenarios.
- To orient on the personal taste of the client, not on the archetype of the character and the target audience.
- Do not check the voice on long texts — vocal fatigue manifests only after several minutes of recording.
- Ignore age perception: the voice of a 45-year-old actor may be perceived by the audience as 'a parent's voice', which does not suit a product for teenagers.
- To forget to document the voice parameters, which leads to the next contractor selecting a new actor 'by eye'.
How to fix the voice in voice guidelines
The chosen voice needs to be described just as strictly as a logo is described in a color and spacing guideline. The document should include: the pitch range in hertz or a verbal description (e.g., 'medium-low, warm'), tempo in words per minute, the acceptable emotional range with example phrases, a list of prohibited intonations (e.g., 'without sarcasm', 'without office-style pauses'), and also reference audio recordings for verification when switching contractors. Without this document, each new voice studio will reinvent the brand voice — and sooner or later it will 'blur' across different content carriers.
Recent articles
MegaFon, BMW, VTB and more 240+ companies. See how it looks in the showcase reel:
Our branding agency is approached to order:
Branding agency services
