photoman,

I think it's important to keep in mind that there is a real skill involved in dictating to a Digital Voice Recorder.  It takes a lot of practice even to get the Olympus DS 5000 to produce highly accurate results.  Although, the DS 5000 is more forgiving and generally tends to produce reasonably high accuracy even for those who are not skilled at using DVR's, my experience with doctors and lawyers over the years has shown that they tend to be, although not always, more skilled at what I call "blind dictation".  That is, being able to dictate into a recorder without actually seeing the text displayed.  That is generally not true with the average user of speech recognition and Digital Voice Recorders.  The average user is used to dictating directly into DNS where they can see the results of their dictation.

That being said, there is one issue that I feel the need to clear up.  The Olympus and Philips recorders use the proprietary DS2 compression algorithm, which was originally developed by Olympus.  This compression format yields very high compression ratios with virtually no loss in audio quality.  The problem that most DVR's have is that they use proprietary audio formats that must be converted to those that are supported by DNS.  This generally results in a double whammy.  The first problem stems from the fact that transcribing from a DVR means that you are not dictating directly into DNS with a good, high-quality, noise canceling microphone.  Being that this is a secondary audio source, there's a certain amount of accuracy that is lost is to the degree that the audio output from the DVR is less optimal than straight dictation into DNS.  In most cases the differences not significant to more than 1% or 2% average loss of accuracy.  Nevertheless, the quality of transcribing from a DVR is generally less optimal and subject to more variation than dictating directly to DNS with a good high quality, noise canceling microphone.  Less expensive DVR’s tend to incorporate more of the types of background interference than the DS 5000 or the Philips 9600 series.  In addition, there is virtually no loss of audio quality when transcribing directly from the Olympus DS 5000 into Dragon NaturallySpeaking by virtue of the lossless compression utilized in the DS2 format.  Where I generally disagree with Lunis is that the DS2 algorithm doesn't increase accuracy.  The reason is that it is simply a compression algorithm and does nothing to the quality of the dictation output.  The reason that it's effective is that it maintains that quality of audio signal during both the compression and decompression process without any loss of audio quality in either direction.  However, in an of itself, the algorithm doesn't improve accuracy.  In fact, data compression, whether it's data or audio, does only one thing.  It compresses the data or the audio period.  So, accuracy in this case is based on the degree to which the DS2 compression algorithm maintains the original quality of the audio input.  That's it, no more, no less.

What is all this leading up to.  Basically, users who are very heavily dependent upon the use of DVR's for accuracy in order to produce the types of highly professional documents necessary in their positions shouldn't cut corners and should use nothing less than the best, highest quality Digital Voice Recorders.  On the other hand, if the average user is simply taking notes and doesn't require the absolute recognition accuracy that many professionals do, then it isn't necessary to spend an exorbitant amount of money for a DVR.  In such cases the DS 2400 is probably more than adequate because, as in your case, losing between 1% and 3% overall accuracy is not critical.

Still, although the quality of your DVR will have an impact on the overall accuracy, most users will find that becoming skilled at "blind dictation" will often more than compensate for the difference.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC

Education is what survives when what has been learned has been forgotten. - B.F. Skinner (1904-1990)

I would agree and it is not only to a DVR but any type of dictation machine.

I think that you are probably right from my observations of persons to whom dictation into a voice recorder is a way of life.

I would think however that most of these persons were dictating long before the advent of voice recognition.

It is a question of what one is used to in the work environment and the earlier one is required to carry out one's work by dictation, the more profficient one becomes.

There are many occupations within the professions and business where dictation is not a required or essential tool to use. 

As one who has not had or needed to do any real dictation until later on in life, it is not so easy to master dictation in any form, DVR or DNS.

My fascination with voice recognition over the years is its potential to do much more than mere dictation into a document.

Although it is nice to see the words appearing on the screen in DNS, I cannot help wondering whether "blind dictation" would not also be advantageous in DNS.

I just got a call from Nuance today and they have a special where they are "encouraging" the package deal of DNS preferred plus a recorder. Since I already have the software, they are offering the Philips 660 Digital Voice Tracer for $129 plus shipping and tax; and 30 days to try it out to see if it works for me.

Does anyone have any experience with this particular model and where does it rank relative to the 860 that Leo mentioned?

Rüdiger,

That is great information to know. The good thing about DNS contacting me is not that the price was that good (I found it cheaper just by doing a couple Google searches). The benefit to me was it reminded me that I had wanted to get a good recorder that I could use with DNS plus they are offering a 30 day return, if I pay return shipping of course.

In any case I will look for a used version of the older recorder you mentioned on eBay and maybe get lucky. Thanks again for your input.

Photoman,

since I have mentioned the Olympus DS-330, one thing that should also be mentioned, it just came to my mind. The software (DSS Player) that comes along with it doesn't run in Vista, so one would also have to buy a newer version of it in order to download and transfer recorded files to Dragon. Just one thing to keep in mind, too.

Rüdiger

Lumis,

 Thank you for the education. Suffice it to say that the stratospheric $500 and $600 recorders are out of my price range. Based on a lot of "ifs," which of the following two recorders would you say would produce the best result with DNS 10 preferred: the Olympus DS-2400 or the Olympus DS-330 + the updated software?

Or, are both of these below the level of what you would consider providing the minimum accuracy levels?

Thanks for your help.

Has anyone had any experience with the Sony MX 20? 

 according to the page below it is the top ranked recorder:

http://support.nuance.com/compatibility/default.asp

I'm also a bit confused; is the Phillips LFH 660 the same as the Digital Voice Tracer 660?  according to the Dragon site, it is well ranked, ranking only slightly below the Sony.    I understand what you meant about the encoding algorithm that preserves the sound quality.  it does make sense.  I just can't justify the cost of those higher end models at this point.  I'm just wondering if you guys are perhaps more selective than the Dragon people themselves?

sorry about my capitalization, I haven't explored the forms enough to find a solution to fix it.