top of page

comevis presents:

Voice Synthese [TTS]: 

Synthesizing your corporate voice

comevis Top 100 T100_21_Member.jpg
German Design Award_GDA22.png
Slide 1 - Top
Anker 1

Because behind every synthetic voice there is a human being

Synthetic voices in voice assistants and apps are not - as many believe - completely artificially generated voices. Behind them there is always a human being and a complex recording process. This is exactly what we have specialized in. At comevis, we are experts in the field of voice synthesis and offer you extensive know-how, a broad network of speakers, and technology tailored specifically to the application area.

On Air Anmelden
Mikrofon Nahaufnahme

Reach your goal faster with comevis experts

What normally takes up to 10,000 sentences and about 4 weeks of work, we can do in just 300 sentences and 3 to 7 days, for example, thanks to neural language models.

This not only saves time and money, but also gives you the opportunity to make your corporate voice as individual as possible and tailored to the brand. Special tonalities, whispering, shouting, speaking with understanding or encouragement - it's all possible. Stand out from the competition with branded voice marketing and personalize your corporate voice so extensively that it no longer sounds like a synthetic voice, but like real conversation partners.

We bring you together with top speakers

Thanks to our broad network in the voiceover scene, we have exactly the right professional voiceover artists for your corporate voice. Together, we develop fair and transparent contracts and ensure that ethical information on this far-reaching use of the voice is clear and understandable at all times.


The right technique for great voices

Since a recording studio and normal microphones are by far not enough to produce a good synthetic voice, we at comevis work with special technology that guarantees the highest quality. We record your speakers in our specially designed TTS cube and use it to ensure that a voice recorded over several days sounds as if it has been cast in one piece. The cube always ensures the same acoustic conditions with reduced reverberation time and can bring the SSR (signal-to-noise ratio) to -80 dB. This means for you: a clear voice without any noise. For comparison: a normal studio just manages an SSR of 40 to 60 dB.

comevis quality for your voices

Standard quality tests such as an MOS test, in which a subjective quality perception is checked with users, or a WER test (Word Error Rate) are not enough for us. At comevis, we have developed a special test procedure which, among other things, tests the naturalness, tonality and listening effort for users. This is how we ensure that your synthetic corporate voice is at an excellent speech level.

Custom TTS application areas







Branded Voice Messages

Personalised content

Swisscom Hintergrund.png

Brand Custom Voice: Best Practice Swisscom 

Since 2017, we have been supporting Swisscom with our corporate voice/brand voice services.

We continuously record Swisscom Voices in German, English, Italian and French for various application areas. Custom voices were synthesized by comevis for the Swisscom product TV Box. In particular, our special know-how for recording & generation of synthetic custom voices is used. This makes new digital applications possible.

comevis_2018_mit_Claim_white (1).png

Microsoft Reference Statement

We are proud that Microsoft shows the excellent quality work results, especially of the German and French voice on their website.

Microsoft also proudly presents the language models for which we have developed the training data.

Grünes Muster

Swisscom Voice Assistant

Voice German, Human

Voice German, TTS

Voice French, Human

Voice French, TTS

Aufnahmestudio 2-Vodafone

The Voice of TOBi - Vodafone Voice/XR-Bot

vodafone logo

Together with Vodafone, we have created a human voice for the chat/voice/XR bot TOBi.

TOBi is Vodafone's digital assistant. He has been supporting Vodafone customer service since April 2019. Its aim is to categorise the various concerns of customers in the initial contact and, if possible, solve many of them directly itself. Until now, customers could only chat with TOBi. This is set to change, as TOBi has now learnt to speak and is developing into a digital service employee who will support his colleagues on the hotlines to help customers more quickly in future.

Roter Schweif

TOBi Voice Assistant


N-TTS Sprachsythese

Neural synthesisation

This process describes the artificial generation of a human voice - spoken texts are converted into speech. Highly complex neuronal processes and TTS (text-to-speech) technologies are at the heart of the process: this not only analyses and learns the spoken word, but also its phonetics, i.e. the special features of the voice in terms of timbre and pronunciation.


The Vodafone speaker had to record just 2,000 sentences during the entire process in the recording studio. With the help of its artificial intelligence, Voicebot TOBi can create completely new sentences from these - without Thomas having to speak them again in the recording studio.

Bild vom comevis Team mit o2 Telefónica
Image by Rich Smith
Logo von Telefonica

From silence to voice:
How we made the O2 Bot "Aura" speak

By developing a voice for Aura, we have managed to make the bot more human and accessible. We made sure that Aura can talk to our customers in a natural and friendly way to answer their queries and direct them to the right place when needed.

A.I. Training data in perfection

Quality In - Quality Out. 

Only if the training data for creating a synthetic voice is perfect in all nuances is it possible to create a high-quality voice model.

For this purpose, we at comevis have developed a special production environment in our Sonic Lab which, in contrast to conventional recording techniques, can bring out significantly more details and quality.

Bild von Bjoern Wegner beim Arbeiten

Target group-specific custom voice for blue

Comevis und Blau Team
Logo von blau

We created a particularly charismatic synthetic voice for the target group of the mobile phone brand Blau in collaboration with Nuance. Here too, our specialised know-how and the possibilities in our Sonic Lab led us to a first-class result.


bottom of page