In addition to chatting with their animals, many frustrated engineers talk to their oscilloscopes. The activity may have some therapeutic value, but regardless of what is said, the scope does not respond.
Hewlett-Packard has changed the situation by introducing an oscilloscope that you can command verbally to take a particular action—and it will. Why? Because engineers increasingly are using both hands to position probes accurately on dense, fine-pitch SMT printed circuit boards. To avoid the possibility of shorting together two very closely spaced tracks, some users have resorted to pressing scope buttons with their nose or a pencil held between their teeth. Needless to say, user safety becomes an issue when probing multiple high-voltage signals.
With the Voice Control option for the Infiniium series of high-performance oscilloscopes, you can run and stop acquisition, autoscale waveforms, and choose the default setup. In addition, waveforms can be manipulated by changing the time base, the vertical scale, or position, and you can measure and print waveforms. All this can be done via English commands issued through a collar-mounted microphone.
The enabling technology for this $495 option is provided by the ASR1600, the Lernout & Hauspie (L&H) automated speech recognition engine. Proprietary signal-processing techniques and advanced speech algorithms deliver accurate recognition across a wide range of conditions. A sound card and a Windows-based component object module (COM) interface complete the option. The COM interface is the mechanism through which the voice application sends commands to the scope.
The interest in a verbally controlled scope was uncovered during a series of focus groups held worldwide. For HP, this was not a new activity. The basic design of the HP Infiniium product family is, in large part, based on an extensive customer- listening campaign to discover what today’s engineers needed most in an oscilloscope. Many of the features, such as analog-like front panels with simple controls for basic functions and a Windows®-based graphical user interface (GUI) for advanced features, stemmed from comments made in the original focus groups.
Further sessions identified the growing frustration caused by ever-smaller signal track or IC pin spacing. And with a little lateral inspiration from a voice-controlled fingerprint scanner in the “Blade Runner” movie, scope ease of use has been improved.
The engineer who developed the voice-control interface, Mike Karin, said, “Years ago, some customers wanted a foot pedal added to each HP scope so they could use both hands for probing. They soon will discover that voice control is much handier than a foot pedal.”
How easy is it to use, and how well does it work? The Infiniium’s voice-control option incorporates a natural-language English command set. Users don’t have to train the scope to recognize their voices. When a verbal command has been accepted by the scope, it provides visual confirmation.
Scott Sampl, manager of HP’s general-purpose instruments business for the Electronic Measurements Division, said, “We looked at several voice-processing engines, but we discovered that most of them were dictation engines with complex vocabularies requiring a high level of training. Based on the surveys we’ve conducted, we knew that scope users wouldn’t take the time to learn how to use these vocabularies. L&H’s engine, with its speaker- and gender-independent features and a more limited vocabulary of up to 1,000 words, closely fits our requirements.”
HP conducted usability tests in which users probed a demo IC board—two points close together and two points far apart. The tests revealed that a speech-enabled scope works best when the user speaks using correct phraseology, not pausing too long during the command, and enunciating clearly. The more proficient users were in operating an oscilloscope, the easier it was for them to adapt to voice control.
Engineers can take comfort in knowing that the ASR1600 engine must compare what they say with a stored grammar. Only if there is a close match will semantic parsing take place and the voice application receive a meaningful input. So, keep talking to your scope. Depending upon its software revision level, it may now understand more than you think. Hewlett-Packard, (800) 452-4844, ext. 6564.
Copyright 1999 Nelson Publishing Inc.
September 1999