Left is faster. Higher reads more hard parts right.
Soniox TTS RT v2 is better on stopping, accuracy and price. AssemblyAI TTS 1 is better on non-streaming latency, streaming latency and naturalness.
Last updated Methodology by Dileep Chagam
Measure by measure, the same sentences, the same place.
An estimate: speech-to-text transcribers scored it, not people.
The same sentence, read by each voice.
AssemblyAI TTS 1. Soniox TTS RT v2: 241 ms. AssemblyAI TTS 1: 171 ms.
AssemblyAI TTS 1. Soniox TTS RT v2: 241 ms. AssemblyAI TTS 1: 171 ms.
AssemblyAI TTS 1. Soniox TTS RT v2: 183 ms. AssemblyAI TTS 1: 159 ms.
Soniox TTS RT v2. Soniox TTS RT v2: 0 ms. AssemblyAI TTS 1: 40 ms.
Soniox TTS RT v2. Soniox TTS RT v2: 83.6%. AssemblyAI TTS 1: 81.5%.
Soniox TTS RT v2. Soniox TTS RT v2: $14.81. AssemblyAI TTS 1: $15.00.
Cekura runs these simulated calls on your agent and scores every one.