Speech at CES; Test Driving Dragon's Naturally Mobile
The 1999 Consumer Electronics Show (CES) was held in Las Vegas in January. For the first time in its long history CES offered a session on speech recognition (which I moderated) and had a speech pavilion in its exhibition (sponsored by Speech Technology Magazine).
Consumer electronics covers a broad spectrum of its products that we use in our daily lives, from toasters to PDAs. Needless to say, the speech processing industry has long viewed consumer electronics as a major pathway to universal acceptance of speech recognition and text to speech synthesis. Consequently, the inclusion of speech recognition in the CES show was a positive indicator that the consumer-electronics industry is beginning to reciprocate that interest.
The title of the speech-recognition session "Voice Recognition: Selling Speech to the Techno-Phobe" reveals the nature of that interest: speech recognition is still an unknown quantity. Hopefully, the success of the pavilion and the session will encourage CES to include speech recognition in future shows.
I recently evaluated Dragon NaturallyMobile, a new entrant in the emerging consumer-electronics category of palm-size digital recorders that can be linked to speech-recognition for transcription. Recorder/transcription systems are among those consumer products that have the potential for enhancing widespread acceptance of speech recognition. The NaturallyMobile CD contains Voice It linking software (for communicating with hand-held devices) and NaturallySpeaking modified for the NaturallyMobile environment. It ships with a headset, a handheld recorder (jointly designed by Voice It and Dragon), a cable for linking the recorder to a PC serial port, three manuals, and two laminated foldouts.
Since busy professionals are likely to be attracted to these devices for their time and laborsaving features, my evaluation focused on usability. The contoured recorder fits nicely into either hand with easy access by the thumb to the function buttons. It is sensitive to noise from hand movements so left-handed people must be careful not to drag their thumbs across the microphone. The recorder stores up to four dictated files (called "folders"), offers all functions (e.g.,add, delete, insert ) needed for a standard palm-size dictation recorder, and holds from 20 to 80 minutes of dictation in "voice-transcription" mode, depending upon the memory card that is installed. The manual "Dragon NaturallyMobile Recorder" offers clear instructions on setup and use of the recorder.
The first contact users have with the NaturallyMobile system is the audio setup. Therefore, usability for this task is critical for user acceptance. Unfortunately, the NaturallySpeaking audio setup wizard makes no reference to the NaturallyMobile device, so I used the "Getting Started" and "Dragon NaturallyMobile Recorder" manuals to guide me. After a frustrating hour of attempting to set the audio input level for the recorder, I contacted Dragon's technical support who backed me through the wizard to the point where I made my original incorrect selection regarding "line in" versus "memory" input. The technician also directed me to one of the laminated foldouts, which describes how to do the setup and enrollment processes.
One could say that the problem was due to my not reading all the manuals and foldouts. However, I was behaving like a typical user for a product of this type. I strongly suggest Dragon reduce the number of manuals from three to one and include only one laminated foldout. I also recommend that Dragon suggest that users establish separate models for dictation using the recorder and editing using the headset. Finally, I recommend that the audio setup use non-technical language such as "Please indicate how you will dictate to NaturallySpeaking: a headset plugged into the sound card; Dragon NaturallyMobile; a handheld recorder plugged into the sound card...had I had such choices I might have flown through the audio setup.
Enrollment, linking, and transcription are straightforward. In fact, this column was created using the system. After dictating a folder into the recorder, the user connects the recorder to the PC cable and starts the Voice It linking software. There are simple icons to link the recorder, upload files, and transcribe. The transcription icon automatically loads the Dragon NaturallySpeaking and initiates transcription. I needed to load my headset speaker-model to edit the file. The Dragon NaturallyMobile is a nice addition to the recorder/transcription product category but correcting the usability of the audio setup will be critical for user acceptance.
Judith Markowitz is the technology editor of Speech Technology Magazineand is a leading independent analyst in the speech technology and voicebiometric fields. She can be reached at (773) 769-9243 or jmarkowitz@pobox.com.