Independent historical guide to SpeechFX, Inc. embedded speech products. Not affiliated with the former company; does not offer SDK downloads or support.

Voice-In — Embedded Speech Recognition SDK

Voice-In (marketed as the VoiceIn Standard Edition SDK) was SpeechFX, Inc.’s embedded automatic speech recognition stack for adding hands-free control to appliances, industrial equipment, and mobile products. Historical pages positioned it as the input half of a complete embedded voice stack, paired with engines such as FonixTalk TTS or DECtalk.

SDK Features (Historical)

• Platform range: From small embedded RTOS targets to server-class multi-channel deployments
• Speaker independence: No per-user training required in marketing materials
• Noise robustness: Neural-network recognition promoted for background speech and music
• Continuous speech: Command-and-control without forced pauses between words
• Resource efficiency: Tuned for constrained RAM and MIPS budgets on embedded silicon
• Languages: U.S./U.K. English, Canadian and European French, German, Japanese, Korean, Castilian and Latin American Spanish, Swedish, and Italian listed in period spec sheets

Phonetic / Animation Component

Historical documentation also described a phonetic alignment feature for animators — mapping recognized phonemes to character mouth movements so game and avatar lips could track synthesized or recognized speech.

Product Context

SpeechFX marketed Voice-In alongside Fonix VoiceGaming and broader “embedded voice solutions” for video games and smart devices — part of the same Utah/Lindon-era OEM licensing business as DECtalk and FonixTalk.

Heritage Guide Scope

This page does not distribute SDKs, demos, or specification PDFs. It preserves URL and product context for researchers tracing inbound links to /voicein/ (legacy voicein.html).

Related: text-to-speech guide · DECtalk on Linux