Voice activated / speech recognition software
Friday 21 February 2014
Speech recognition, a tried and tested technology or is it reinventing the relationship between people and technology?
Speech recognition has been proven to unlock or increase productivity and creativity potential for individuals in a multitude of different working environments. With the rapid technological advances we have seen in recent years, speech recognition is now becoming commonplace, emerging all around us. It is now closer than ever to becoming the next dimension to computing and collaboration, by providing a much needed assistive technology link for employees and employers to become more efficient. The webinar will take you step-by-step through the practical use of voice activated software with a live demonstration.
Jonathan has spent over 15 years working in the field of digital dictation and speech recognition technology – he is the UK & Ireland Channel Manager for Dragon, and responsible for 65 specialist partner companies. We can safely describe Jonathan as an expert in the subject, and will gladly answer questions at the end of the webinar.
Q: In my business we have a large population of employees where English is not their first language, how does the system cope with accents?
A: Dragon will cope with most accents as long as the full user profile training procedure is followed and completed.
A: Costs will vary on the version type but typically single user box copies for Dragon Professional RRP £549 inc VAT, Dragon Premium RRP £149-£199 (without/with Bluetooth headset/portable digital dictation device), Dragon Legal RRP £699. As volume increases the unit cost reduces, because multi-user licensing is then prevalent.
Q: Who is the target end user? You mentioned carpel tunnel, RSI injuries and dyslexia etc. Any other target end users?
A: From an assistive technology perspective, anyone at an early age of 7 years old to mature student and employee, either with a learning difficulty or physical difficulty/disability. Anyone producing documentation and/or who wishes to become more productive is also a potential user of speech recognition technology.
Q: Is the Dragon software compatible with iPad/Apple programs?
A: Dragon is designed for the PC and the Macintosh. Other freeware will provide basic speech recognition via the Cloud onto the device (any one of the FOC Dragon apps via the iTunes store) are designed for smartphones and tablets.
Q: How does the system cope with ambient noise? Do you need to be in a quiet place? Do you need hardware, headset speaker etc?
A: No, as long as you have a good quality headset microphone or even a simple noise cancelling microphone, background noise can mostly be removed from the recorded audio. Dragon has a very simple process of checking the audio sound levels and quality of the microphone, and it is advisable to use this simple procedure when changing from one environment to the next. At the end of the day common sense should prevail that the better the quality of the audio in, the better the recognised transcription results will be.
Q: Can it be used with MS Excel and PowerPoint?
A: Yes, however, MS Excel is an application within the Microsoft Office suite primarily used for numerical, data calculations/analysis tools. Excel is not designed to be used as a document creation application, although transcribing into Excel in the local written language will provide similar transcription results as you would do, for example, with Microsoft Word.
Q: Can this be used with Apple products using office software?
A: Dragon Dictate uses the generic Dragon speech recognition engine and is fully compatible with the latest Mavericks MAC iOS.
Q: How disruptive is it to use with other colleagues all around? What has experience, in practice, shown?
A: Using speech recognition on a PC or Mac requires the user to more commonly use a headset and microphone and would not be seen as no more intrusive, or disruptive, as having a conversation on the telephone.
Q: Does Dragon learn regional accents well? I am in Northern Ireland.
A: As long as the Dragon user profile training is followed clearly then accents should not present too much of a problem. It is never going to produce 100% accuracy every single time, for everybody all of the time.
Q: When using the dictaphones and trying to transcribe the recording, do we need to do a profile for all speakers, as students have many lecturers?
A: If using a digital dictation recording device, the user will require Dragon to have a unique profile for that user and device. It is not possible for someone to gain the same level of accuracy with their own user profile, by recording a third party without the software having a pre-recorded profile for that individual.
Q: How does it work if you wanted to use it and were on a conference call at the same time?
A: If you used Dragon to transcribe the contents of a conference call, it would only accurately be able to transcribe the known users profile side of the conversation.
Q: How can staff use Dragon in an open area without irritating colleagues? We currently have staff in offices who use it but will need to move to open plan soon.
A: One way to do this within an open plan environment is to persuade the individual Dragon user(s) to perhaps speak a little more quietly which would require that specific user to adjust Dragon, or perhaps they could move to a different location when they’re using Dragon. Some organisations have created a dedicated room or space for employees wishing to use Dragon in an undisturbed area.
Q: How can you move the mouse on-screen most efficiently?
A: You can use the mouse on-screen by verbal mouse up/mouse down commands, or simply just by using the mouse in the normal way if this is possible.
Q: Which brands of digital dictaphone are compatible with the software?
A: Most devices that produce an audio file in one of the following formats should be sufficient - .wav, .dss or .mp3
Q: How does it handle background noise?
A: The majority of reasonable quality headset microphones these days will have a microphone which is capable of background noise suppression. It is always best to check first.
Q: Do you have to train the software on your accent? How can that be done?
A: Any user wishing to use Dragon, we’ll have to create a profile so that the software understands the unique pronunciation of that individual. When Dragon is first used there is a very simple profile creation process and it can take as little as 10 minutes, up to a maximum of 20 minutes, to complete the entire process.
Q: What is the cost of the software?
A: Costs will vary on the version type but typically single user box copies for Dragon Professional RRP £549 inc VAT, Dragon Premium RRP £149-£199 (without/with Bluetooth headset/portable digital dictation device), Dragon Legal RRP £699. As the volume increases the unit cost reduces, because multi-user licensing is then prevalent.
Q: Do you know of a preferred or 'best way' of planning for speech e.g. mind-maps, bullets anything else? Speaking is not always natural.
A: Certain individuals may have difficulty with pronunciation. As was mentioned on the webinar, the software is not going to suit everybody all of the time, and it won’t be 100% for everybody all of the time. The general idea is to speak naturally and normally (as if they were talking to someone in person).
Q: Is UK English or American the default spelling?
A: In the UK it is UK English default spelling. If US English is required then this needs to be selected when the profile is created.
A: Not sure I quite understand the question.
Q: Shared/open plan offices?
A: Dragon should be fine for either environment. The software has the capability to re-tune the headset/microphone to take into account any changes in ambient surrounding noise.
Q: One of my staff uses wireless at home on his personal computer using a standard headset. What do you think about whether the quality is better with using wireless or cable?
A: There are so many different headsets available in the marketplace, both wired and wireless. Generally speaking, a hard wired headset will be less likely to lose contact/pair and is therefore likely to be more reliable.
Q: Do you know what the levels of success have been with the Dragon mobile app? There have been some unfavourable reviews. I have a member of staff who would like to load it onto their Kindle. Thanks. Sharon (Nugent) Dorset County Council
A: We have had some great levels of success with using portable digital dictation devices as well as the Dragon mobile app. The Dragon mobile apps are 'remote microphone' or as simple 'digital recorder. I’ll need to check if the Kindle is supported.
Q: We have just changed to Office 365 – is Dragon compatable?
A: Yes it should be.
Q: What about using this software in an open plan office?
A: Dragon can be used in an open plan office. If Dragon’s acoustic settings are quickly checked (30 seconds) then any differences in ambient noise level in an open plan environment will be quickly taken into account, and should not have effect on the resulting transcription.
Q: What happens to 'umms' and 'errs' - are they automatically eliminated?
A: If they are spoken in a normal way then the majority of 'umms' and 'errs' are automatically removed. If there’re pauses between then it we’ll try to recognise the sound of the word. More often than not it is far easier to try and concentrate on not using 'umms' and 'errs'.
Q: Will Dragon recognise different accents e.g. Scottish accents are difficult to pick up as the systems don't seem to 'hear' the ends of words?
A: Yes, as long as the individual user follows the correct procedure in creating the profile, and follows the initial profile creation process, then accents should not be a problem for Dragon. It will never be perfect for everybody all of the time. If there are individual words which are causing a problem, again there are slightly more advanced ways of overcoming this. This is something that a fully trained and accredited Dragon technology partner will be capable of helping with.
Q: How easy it is to correct mistakes within the document?
A: Very simply, there are a set of very simple commands that will become natural to use. Selecting a word/sentence can automatically provoke Dragon to reveal alternative suggestions, or the user just simply dictate over the top of the highlighted text with the replaced text.
Q: Will Dragon work in a noisy open plan office?
Q: Would Dragon be suitable for people with Dyspraxia?