Being a productivity tool, one thing that Dragon can’t do is work on your behalf. As relaxing as it may seem to kick your feet up when you work, the software is a conduit, there to improve your speed of work and facilitate work flow. You still have to drive the work, which is a good thing, otherwise you might be out of a job. The robot apocalypse will have to wait until AI is smart enough to get your reports done. While this may not offer all that much comfort when you see just how integrated the digital life is fast becoming, the good news is that we can leverage these technologies to our advantage.
Dragon is one such product, there to customise itself to the nuances of your voice and vocabulary. The more you use it, the more it gets to learn about the words you use more frequently, how you say things and when you typically correct misrecognised text. Another thing that makes it a breeze is that you’re able to dictate at up to 140 words per minute. Beyond this point, it becomes much less accurate because unless you’re a horseracing commentator or an auctioneer, chances are you won’t be able to keep the words from blurring into one another. Diction is still important, even if you can speak quickly and it’s knowing just how quickly you can speak without running into poor accuracy.
We’ve had clients who want to almost talk so quickly they do a verbal shorthand style. This is already difficult for the human ear to decipher, so it’s not the best for a programme that doesn’t even understand English. Humans can use context and familiarity to better transcribe fast-talking. Dragon just has the soundwaves, which if indistinguishable become almost impossible to interpret. We’ve also had clients who want the software to deploy transcribed text instantly. While this is close to what’s happening already, it’s only able to dump the text after having a split-second or two to relay the sound recording, interpret the soundwave and respond with an appropriate transcription.
So, now that you know you have to aim for clear dictation and the software needs a split-second to decipher your sounds – it’s easy to see its limits if you can call them that. While we’d love to attach electrodes to people’s heads and have them use telepathy to interpret thoughts rather than strain vocal chords at all, the technology isn’t there yet. What we can do is suggest you improve your system resources. If processor speed is how quickly your computer can read a page, memory enables your PC to read multiple pages simultaneously.
Working with Dragon, you may find you have several applications open at the same time. This can cause a drag on resources having to keep them active and within system memory. Being a resource-intensive software application itself, it does help to go in with as much memory as possible. When you start using Dragon it’s a good idea to slide the speed versus accuracy bar to favour the side of accuracy. Your PC system resources will actually decide what works best based on overall performance, but why not give Dragon an extra split-second or two when you begin using the software and it gets to build your profile from scratch. Having top-end hardware is a great way to get the best performance out of your Dragon speech recognition software!