This is the first in a series of blog posts where I will be focussing on Dragon for PC (software formally known as NaturallySpeaking) and how it can help established businesses who are already transcribing audio. Those of you who work as transcription typists or perhaps manage a firm, likely legal or medical, where you have staff transcribing letters or notes this one is for you.
What I will cover in this post is:
- Using Dragon to your benefit, let it do the hard work and type, you just proof read
- You do not need the person dictating to train a Dragon profile, in fact they don’t even need to know, just leave them to dictate as they always do
- You can reduce your physical typing time substantially, giving you time to work on more productive and fulfilling work tasks
- Reduce your voice-to-text turnaround times from 3-4 times the audio length to only 1.25 times the audio length. i.e. a one hour audio file which used to take 3-4 hours to transcribe will now only need around 1 hour and 15 mins to proof read
In the course of my job I regularly speak with transcription typists. Mention the word Dragon and the defensive body language kicks in and comments like “that is the software that is going to replace me” or “I’m not worried it will never work” start flowing. Just recently I have switched my focus from predominantly Mac based voice recognition back to having a look at the Windows world and specifically what has been happening in the last couple of years, you will like this so read on.
Currently Dragon for PC Professional from Nuance is at version 14, the updates are fairly regular and with each update comes the promise of converting voice to text faster and more accurate. Olympus too have been busy working on their digital dictation and transcription offerings releasing version 6 of their dictation and transcription module software now called ODMS (Olympus Dictation Management System). Version 6 from the users point of view looks exactly like version 5 (called DSS Player Pro) but under the hood some major changes were applied specifically around linking more closely with Dragon. Olympus users who dictate and transcribe can get the full benefit of voice recognition software with the ODMS software and Dragon NaturallySpeaking.
Let’s start with a demo. The video below shows a voice recording made on an Olympus DS-7000. The audio in typical transcription workflow style arrives at the typists transcription module but this time instead of transcribing the audio we push it to Dragon to do it for us. All we will need to do is proof the transcribed audio.
You will see from the above video that the audio was transcribed very accurately. All that needs to be done from a transcription typists point of view is proof the transcript with the audio, fixing any misheard words and applying any formatting. As you can see the Olympus ODMS transcription module makes this easy, controlling the audio playback with your existing RS-28 or RS-31 USB foot pedal. Proofing not only gives the completed transcript the formatting and accuracy required but also serves to update the users voice profile in Dragon. Effectively Dragon is learning from its mistakes, the mistakes that you the transcription typist is fixing.
This will only work for single speaker, that is one person only, who speaks in a dictation audio file. So your typical medical/legal practice where many letters are churned out on a daily basis. Dragon is still not able to transcribe multi speaker audio such as interviews and meetings.
The person who sends you dictation does not have to change their workflow at all. They can simply continue recording into their digital dictaphone. You, the typist, can work smarter by creating a Dragon voice profile based on their recorded audio. Dragon needs only a minimum five minute audio sample to be able to create a profile, you can use existing audio that you have already transcribed for this training process.
Do you have questions about this post? Please leave a comment so that we can answer them for you.
Main image from Flickr, beautiful photo titled “Typing Pool” by jST.