If dictating to the system via a microphone, the sound card is critical to success. Historically (1990's), this was a major issue in speech recognition. There has been steady improvement, and now (2011) most inexpensive PC's have acceptable sound cards built into their motherboards.
Discrete sound cards work well too, perhaps slightly better than the built-in sound systems typically found today. But for the most part they are becoming irrelevant because USB sound systems are the logical path for most users when a built-in sound system is not adequate.
An alternate to a sound card is to acquire a USB "pod," a device where you can plug in your favorite microphone. These are sometimes called USB Audio Adapters, USB sound systems, or other similar names. They are particularly suitable for laptops or other systems where you can not or your company does not allow you to open the system case.
Plug in a good microphone and you have yourself a nice sound system.
Another alternative is a USB microphone -- a microphone with the USB electronics included. Plantronics, Andrea, Philips, Insync, Sennheiser, VXi and others make such microphones. We most often recommend the VXi TalkPro line of these microphones when a headset is desired. Others work well for many persons.
Some dealers claim USB audio is better. Back in 1998 that was generally a fair statement. But with the better audio systems supplied with computers today, we view USB audio as a convenient, safe alternative.
A few older USB devices have problems on Windows Vista and Windows 7, but as of mid-2011 these problems are less and less common as vendors improve their Vista support. With Windows 7, USB audio has had few problems except with older hardware/microphones that do not meet today's standards.
Professional recording cards (typically costing several hundred dollars) work very nicely but are sometimes difficult to set up. Some do not even produce the relatively low-quality sound needed for speech recognition on releases prior to 7. After Release 7, NaturallySpeaking is able to convert high-quality sound to the format it needs for processing.
Several sound cards will produce good results for particular speakers, or with specific microphones, or have specific production runs that produce very good results. Many of the cards listed on the Nuance web site fall into this category.
IF YOU DON'T ENUNCIATE THE WORDS, THE SOUND CARD, MICROPHONE, AND PROGRAM CANNOT DO THEIR JOBS.
Users with laptop computer systems must accept whatever sound system is delivered with the laptop computer system or else use a USB microphone. A USB microphone bypasses the built-in sound card.
Users concentrating on processing dictation from recorded speech using digital recorders will not be affected by the sound system except while trying to learn the system. The recordings are transferred without using the sound card. The quality of the recorder microphone IS important! We suggest recorders supporting the .ds2 formats for best results. These include the Olympus DS-5000 and the Philips 9600 recorders.
A few people need two sound systems. For instance, they may want to use VoIP or record podcasts while using Dragon for computer control. Generally we suggest using one USB system and one PCI sound card or "built-in sound" in these situations.
Introduced in 2006, most PC's now offer "HD Audio." These systems conform to a new audio specification aimed at improving the overall audio experience on new PC's. Almost all implementations of HD Audio are satisfactory for speech recognition. But some, typically older, microphones don't work well with these audio systems. We suggest a USB Pod if you like your old microphone, then you can plug it into the USB Pod, and the USB Pod into your system, and you almost always have good audio once again.
top