SeeVoice Real-Time: How It Works

SeeVoice Real-Time converts voice stream into synthetic avatar video in real-time.
The product is based on SeeStorm Phoneme Recognition / LipSync technology, and is delivered as a server software product.

SeeVoice, installed on servers of a provider of mobile synthetic video conferencing service, performs real-time procession of the incoming voice flow to extract phonemes from the speech and convert them into commands for animation of an indicated 3D character (avatar). The resulting video stream with the talking avatar is transferred to the recipient's mobile handset.

The input is a voice file in AMR/PCM/G.711 format, and the IDs of the chosen avatar, background picture and emotional state.
The output is a video stream in H.261/H.263/H.264/mpeg-4 format.
Other output formats can be provided on your request.