Robust Affordable Speech Recognition
LumenVox Comparison
The Vestec Speech Engine is a standards-based, Asterisk-compatible, Digium-certified product for enabling robust speech recognition with Asterisk at truly affordable prices. It has been specifically designed to help popularize and demystify speech recognition by lowering solution costs, increasing recognition accuracy, and enhancing ease-of-use.
Below is a comparative assessment of Vestec Speech Engine with LumenVox Speech Engine for speech recognition with Asterisk.
Retail Price Comparison
| Price per Port (Channel) | Vestec | LumenVox |
|---|---|---|
| Speech Starter Kit | $25 | $50 |
| Standard Engine | $99 | $199 |
| Standard Engine 5 Pack | $399 | $895 |
| Annual Maintenance | $15 | $40 |
NOTE: 1. Retail prices include reseller and distributor mark-ups.
2. For LumenVox: (a) Standard Engine denotes “Lumenvox Speech Recognition Lite for Asterisk, with 1 Lite Engine Port”; and (b) Standard Engine 5-Pack represents “Lumenvox Speech Recognition Lite for Asterisk, with 5 Lite Engine Ports.”
3. LumenVox Speech Engine is the property of LumenVox LLC.
2. For LumenVox: (a) Standard Engine denotes “Lumenvox Speech Recognition Lite for Asterisk, with 1 Lite Engine Port”; and (b) Standard Engine 5-Pack represents “Lumenvox Speech Recognition Lite for Asterisk, with 5 Lite Engine Ports.”
3. LumenVox Speech Engine is the property of LumenVox LLC.
Core Features Comparison
| Feature | Vestec | LumenVox |
|---|---|---|
| SRGS Grammar Syntax | ✓ | ✓ |
| SISR Grammar Tags | ✓ | ✓ |
| 500 Keywords Vocabulary | ✓ | ✓ |
| Advanced Noise Reduction | ✓ | ✓ |
| American English Model | ✓ | ✓ |
| Asterisk Connector | ✓ | ✓ |
| Distributed Recognition | ✓ | ✓ |
| Installation Support | ✓ | ✓ |
NOTE: LumenVox Speech Engine is the property of LumenVox LLC.
Vestec is working on MRCP (Media Resource Control Protocol) support, non-English acoustic models, and large vocabulary engines. In our opinion, the overwhelming majority of speech recognition opportunities in North America do not require MRCP for platform integration, non-English acoustic models, or more than 500 vocabulary items per recognition.