There was a time when speech recognition programs were all the rage – simply record your messages and have your PC turn it into text. However, language is such a large moving target, consider how many accents you hear on a daily basis before you even start thinking about abbreviations and slang, and you’ll see why most applications can’t keep up.
In recent times this has left the market a little thin on the ground but Dragon Naturally speaking continues to plough, an increasingly single, furrow.
The latest edition is Version 9 and Nuance suggests that it’s the first version you don’t need to train to understand your voice. This is something of a big claim, considering previous versions have taken anything up to 2 weeks to get to grips with even the most basic of accents, let alone a strong Northern one.
To this end, we’d recommend running through the training program as only the likes of Joanna Lumley will be able to get away without it. One neat little trick is allowing the program access to your Word documents. This way it can pick up any written inflections you may use. You’ll end up with a list of queries at the end of the process but it should help speed things up in the long term.
Nuance also claims this version is 20% more accurate that Version 8 and that it is as much as 99% accurate. This is being a little over generous, as it didn’t quite like my natural voice. However, when speaking slower, it was far more accurate – almost impressively so.
The Standard version, the one we were sent, is as the name suggests the entry-level product and is ideal for those new to dictation. It will work within Word, Outlook Express and even Internet Explorer.
The more expensive versions, such as Version 9 Preferred (£160 inc. VAT) offers Bluetooth support so can be used with enabled headsets. The most logical reason to opt for the more expensive version is the inclusion of support for a wider variety of applications, including Firefox and Thunderbird, as well as Corel products.
Besides using it within Office applications, the most common use for the program is to transcribe audio files, perfect for the lazy journalist or student. When you’re using the program you’ll find the DragonBar, which is a small bar that sits at the bottom of the screen indicating the program is active.
Right-click on it and you can set your preferences and generally set sound levels and enter new words the trained vocabulary may not know. You can also switch users, so as not to confuse it.
Voice recognition is one of those tools that people naturally warm to, hoping it will help them speed up their performance.
Train it and it’ll work well for you but the claims of working straight from the box are a little ambitious.
While Version 9 is a vast improvement over earlier versions, it feels as though there is still some way to go before it’s an every man product.