KnowBrainer Speech Recognition
Decrease font size
Increase font size
Topic Title: Alternatives to the Sennheiser MD 431 II ?
Topic Summary: Searching for the ultimate accuracy and noise rejection for $300 or less
Created On: 12/10/2011 04:10 PM
Status: Post and Reply
Linear : Threading : Single : Branch
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/10/2011 04:10 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/10/2011 09:33 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/11/2011 10:00 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/11/2011 03:23 PM  
 Alternatives to the Sennheiser MD 431 II ?   - techlaw - 12/11/2011 11:05 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/11/2011 11:34 PM  
 Alternatives to the Sennheiser MD 431 II ?   - supee - 12/12/2011 03:41 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/12/2011 07:01 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 12/12/2011 10:35 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/12/2011 12:05 PM  
 Alternatives to the Sennheiser MD 431 II ?   - GDS - 12/12/2011 03:55 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/12/2011 09:33 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 12/13/2011 08:08 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 12/12/2011 05:22 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/12/2011 07:12 PM  
 Alternatives to the Sennheiser MD 431 II ?   - R. Wilke - 12/12/2011 07:34 PM  
 Alternatives to the Sennheiser MD 431 II ?   - gabor5454 - 12/12/2011 11:13 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 12/13/2011 07:57 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Lunis Orcutt - 12/12/2011 09:11 PM  
 Alternatives to the Sennheiser MD 431 II ?   - MDH - 12/13/2011 08:55 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 12/13/2011 09:56 AM  
 Alternatives to the Sennheiser MD 431 II ?   - R. Wilke - 12/12/2011 03:32 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Nelson - 12/11/2011 04:42 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/11/2011 06:22 PM  
 Alternatives to the Sennheiser MD 431 II ?   - R. Wilke - 12/11/2011 08:39 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/11/2011 08:55 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/13/2011 08:44 PM  
 Alternatives to the Sennheiser MD 431 II ?   - threevok - 12/14/2011 05:13 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Lunis Orcutt - 12/14/2011 11:12 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 12/15/2011 02:20 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/23/2011 07:33 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 12/23/2011 09:41 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/23/2011 10:38 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/23/2011 11:21 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 12/24/2011 10:32 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Amin Sabet - 12/24/2011 10:38 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Wolvenmoon - 03/22/2012 03:46 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Lunis Orcutt - 03/22/2012 09:37 PM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 10/24/2012 10:01 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Lunis Orcutt - 10/25/2012 12:51 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 10/25/2012 02:41 PM  
 Alternatives to the Sennheiser MD 431 II   - Nelson - 10/26/2012 12:03 PM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 10/25/2012 11:02 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 10/26/2012 12:16 PM  
 Alternatives to the Sennheiser MD 431 II ?   - MDH - 10/26/2012 09:13 PM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 10/26/2012 05:30 PM  
 Alternatives to the Sennheiser MD 431 II ?   - David.P - 11/23/2012 06:12 AM  
 Alternatives to the Sennheiser MD 431 II ?   - supee - 11/24/2012 02:15 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 11/24/2012 09:16 AM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 11/25/2012 08:38 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Lunis Orcutt - 11/25/2012 10:54 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Lunis Orcutt - 03/25/2021 01:04 PM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 11/26/2012 01:26 PM  
 Alternatives to the Sennheiser MD 431 II ?   - phils - 11/26/2012 04:21 PM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 11/26/2012 05:06 PM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 02/07/2013 07:49 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 02/07/2013 08:08 PM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 02/07/2013 09:31 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Chucker - 02/08/2013 04:09 AM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 02/08/2013 11:02 AM  
 Alternatives to the Sennheiser MD 431 II ?   - rnwolf - 09/03/2013 06:31 AM  
 Alternatives to the Sennheiser MD 431 II ?   - Nelson - 09/03/2013 10:44 AM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 09/03/2013 10:12 AM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 09/03/2013 01:21 PM  
 Alternatives to the Sennheiser MD 431 II ?   - elterry - 09/03/2013 02:58 PM  
 Alternatives to the Sennheiser MD 431 II ?   - phils - 09/03/2013 07:05 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Nelson - 09/03/2013 09:16 PM  
 Alternatives to the Sennheiser MD 431 II ?   - elterry - 09/04/2013 05:56 AM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 02/04/2014 11:27 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Lunis Orcutt - 02/05/2014 12:38 PM  
 Alternatives to the Sennheiser MD 431 II ?   - David.P - 03/25/2021 04:15 PM  
 Alternatives to the Sennheiser MD 431 II ?   - wheelstb - 02/05/2014 06:26 PM  
 Alternatives to the Sennheiser MD 431 II ?   - Lunis Orcutt - 03/25/2021 06:32 PM  
 Alternatives to the Sennheiser MD 431 II ?   - David.P - 03/26/2021 04:56 AM  
Keyword
 12/10/2011 04:10 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

I love the Buddy Desktop Mic 7G with Filtered Audio, and my Sennheiser ME-3, but I am intrigued by the possibility of using a desktop mic with even better noise rejection than the ME-3, and people say the Sennheiser MD 431 II offers just that.  

I've been practicing with my Buddy Desktop Mic, and I think I can live with keeping my mouth within an inch of the Sennheiser mic.  I can live with the size and weight of the 431 II.  I can also live with the price of the Sennheiser mic.  However, I keep thinking there have to be alternatives.  

I wonder whether the only reason the 431 II keeps coming up in speech recognition discussion is because of its history, ie that it was supplied with the IBM Personal Dictation system 18 years ago.  There are many other fine microphones of that class.  The vocalist forums don't mention the 431 II much anymore, and they are equally interested in performance and off axis rejection.  Can anyone suggest an alternative they love? 

 12/10/2011 09:33 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

When I research the idea of off axis rejection in the vocalist forums, the model which keeps coming up for me is the Audix OM7: http://www.audixusa.com/docs/products/OM7.shtml

This video is an awesome demonstration and really drives it home: http://youtu.be/MvUfXxalD7Q

I think I'm going to try out this mic.  Only question I have is whether I need a pre-amp or can go straight into the Andrea USB pod.  Supposedly the OM7 has about 7db lower output than most other dynamic mics in its class. 

 12/11/2011 10:00 AM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

The other microphone I am strongly considering is the Heil PR40.

Check out Bob Heil demonstrating the rear rejection of this microphone: http://www.youtube.com/watch?feature=player_detailpage&v=X7WSlq5dqWo#t=2034s

When I work at home, my kids usually make the loud noise directly to the rear of the microphone.  The supercardiod mic on my ME-3 picks up some sound in that area, whereas it looks like the Heil would not.

Here are the pros and cons as I see them based on reports:

Audix OM7: ~$230. Hypercardiod mic giving maximum side noise rejection and good rear rejection. Unforgiving in terms of need to stay "on the mic" (close distance and directly in front).

Heil PR40: ~$315. Cardiod mic giving maximum rear noise rejection but not as much side rejection. More forgiving in terms of distance to mic and angle of approach to mic.

 12/11/2011 03:23 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

I found a solution which so far seems to work extremely well. I purchased a Shure Beta 58a, which is a $159 supercardioid microphone. I have it running into an Andrea Half Duplex USB Pod (http://www.knowbrainer.com/NewStore/pc/viewPrd.asp?idproduct=234).  

After running the Audio Setup Wizard, I used the Windows controls to manually decrease the microphone input level from "24" (the result of running the wizard) to "4".  This makes it far less sensitive to ambient noise yet still sensitive enough to work maintain superb accuracy if I am right on top of the mic.  

As a system, it is extremely accurate and almost completely unaffected by the noise around me (keyboard clicks, loud TV, loud kids, etc).  When the kids go to sleep, I push the mic level back to "24" which gives me even better accuracy and also lets me back away considerably from the microphone. 

Check the attachment to see my setup in our basement and why noise rejection is so important for me in my "home office" .

For those of you lucky enough to have quiet workspaces, a regular desktop mic like the SpeechWare 3-in-1 or Buddy Desktop Mic 7G DSP is no doubt much more convenient than a traditional handheld mic like the Shure Beta 58A.  The Buddy has been excellent for me at work, but this Shure Beta 58A setup seems to be just the ticket for my shared basement space. It's my "poor man's Sennheiser 431 II".



DSC00643.jpg
DSC00643.jpg  (15 KB)

 12/11/2011 11:05 PM
User is offline View Users Profile Print this message


techlaw
Senior Member

Posts: 111
Joined: 10/19/2006

Very interesting postings, Amin.  I had exactly the same suspicion and reasoning, and reached the same conclusion. My conclusion was that although Sennheiser MD 431 is a top performer, it does not create a unique category by itself. I tend to be suspicious about establishments such as these, especially when there is lack of scientific proof.  The only scientific proof to me regarding microphone performance is side-by-side comparison under controlled conditions.  It doesn't matter how much anecdotal and even expert reviews there are if they didn't do side-by-side comparisons under a reasonably controlled condition.

I eventually chose Shure Beta 87A over the Sennheiser. (By the way, I use a professional grade preamplifier, instead of USB pod.)  Both had almost identical performance in my tests, but I went with Shure simply wanting to take a stance against a popular myth. Of course, had the Sennheiser demonstrated noticeable and verifiable performance advantage, I would not go with a different microphone just for the sake of being different. But the Shure Beta 87A is every bit as good as the Sennheiser, in terms of noise cancellation, noise floor and overall accuracy. Besides, it is lighter and smaller. The weight didn't matter to me because I don't use it handheld. But the size does matter somewhat, because having a huge microphone in front of you all the time makes the workspace crowded and less functional.

 

 

 12/11/2011 11:34 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

Hi techlaw,

Thank you for your reply! It's good to know that I'm not crazy, or at least not the only one who is crazy .

Kidding aside, it's great to hear from someone who not only was thinking along the same lines but actually tried the Sennheiser against another supercardioid microphone. I agree with you about side-by-side comparisons under controlled conditions, and when possible I prefer to see/hear the primary data.

I'm thinking about trying some alternatives to the Andrea and Buddy USB pods which I'm currently using. In particular, I had my eye on the M-Audio Fast Track Pro, which I know a lot of the podcasters are using. Which preamplifier are you using?
 12/12/2011 03:41 AM
User is offline View Users Profile Print this message


supee
Top-Tier Member

Posts: 277
Joined: 10/28/2006

how do you use the Shure Beta 58a, as it does look like a big microphone to me and I just wonder how you  Will be able to  use it  conveniently. Do you use it handheld, or do you put it in  a suitable standard?

-------------------------

DPI 15.3, XPS 9575,  Intel Core i7, 16gb ram, 1TB SSD. Sennhesser ME3,  Philips Speechmike Premium 3500, Windows 10, Office 2016.

 12/12/2011 07:01 AM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

I use a basic desktop microphone stand, which works very well for me: http://farm8.staticflickr.com/7170/6494792299_ba73b2f7d0_b.jpg

Eventually I may adopt a swivel mount boom arm if I get tired of having to sit in the same position.

 12/12/2011 10:35 AM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Quote:
Don't you think it is interesting that the acapella singers, podcasters, and others who demand quality microphones with noise/feedback rejection almost never mention the Sennheiser 431 II, yet the speech recognition community brings it up again and again? I strongly suspect that one of the primary reasons for this is that the speech recognition community is relatively small and doesn't experiment a great deal with a wide variety of microphones.  I realize that I may be wrong about this, but this new Shure microphone is working out incredibly well for me so far.
 
I don't mind spending $450+ on the microphone, but I don't like the idea of doing so because of dogma which has been passed down over the years. I would love to see a speech recognition site do rigorous testing of all of these dynamic and condenser microphones, but I suppose that the target market is too narrow to justify this. 

Amin,

I think it would be just as interesting to find out how many a cappella singers, pod casters, and others who "…demand high quality microphones with noise/feedback rejection…" use speech recognition?

I have no question but that the microphones being discussed in this post would work equally as well as or better than the Sennheiser MD 431 II with speech recognition. I also think you're correct in that there has not been a significant amount of side-by-side comparison or testing of these microphones in a controlled speech recognition environment precisely because of what you point out. That is, speech recognition is a narrow focus application.

Side-by-side comparison testing Re: speech recognition of all of these microphones in the proper and controlled environment would be expensive. Proper microphone testing requires, as was done by L&H (Lernout & Hauspie) using a soundproof environment and a piece of equipment that was called the "Mouth", and specifically designed to control all the variables relative to speech recognition, between 2000 and 2002 would be ideal relative to this discussion. However, that testing environment cost almost $100,000. I don't know whether or not ScanSoft, when they acquired the Dragon assets in December of 2001, ever kept that equipment because it was set up in the old Dragon Systems offices in Newton Massachusetts and they divested themselves of that building shortly thereafter. I seriously doubt whether Nuance's been doing the kind of testing with microphones that was done at L&H when that equipment was available. The only reason that I point this out is because if you're going to do side-by-side testing of microphones with Dragon, you can't do it haphazardly because you simply can't control all the variables as we were able to do in that test environment. That being said, and even though you can't place any scientific validity on real-world testing, the results would still be comparable as long as you use a digital sound meter and test that the appropriate axis points relative to the polar pattern.

Regardless, I doubt very seriously that the average Dragon user dictates in an environment that really requires any of these microphones. That is, nice to have but accuracy in speech recognition is dependent upon more than simply the quality of your microphone. Basically, I concur with Rüdiger. You can go out and spend the money on one of these microphones and still not get good accuracy. They only lay the groundwork. The rest is dependent upon the user.

I get comparable results with both the Sennheiser MD 431 II and the Audix OM2. I haven't tested the Audix OM7, or any of the other ones listed, so I can't talk to any of these. However, my reason for using the Sennheiser MD 431 II is that I became aware of it and purchased it for the purpose of giving demonstrations in tradeshow environments or before large audiences where the ambient background noise level is high (anywhere from 88 dB to 93-103 dB). If I were doing so today, I might choose one of the other listed microphones. Nonetheless, the only reason that I put any kind of emphasis on the Sennheiser MD 431 II is because it gives me 99+ percent accuracy in the environments in which I use it. For normal dictation in my work environment, I use the SpeechWare 3-in-1 as I'm using right now at a gives me the same accuracy as the Sennheiser MD 431 II coupled with the Andrea PureAudio USB-SA USB SoundPod.

My main point is that I would never argue that there are microphones out there that are less expensive and equally as comparable to the Sennheiser MD 431 II, and, in some cases, may even be better. The main point is: (a) were talking about speech recognition under reasonable background noise level environments and not as the lead singer in a rock group, and (b) once you get a microphone that gives you 99% or better accuracy, is one of these microphones going to improve on that? The important point has to do with the latter. That is, accuracy is an asymptotic function. Once you get to 99% + accuracy, there isn't much that's going to increase that accuracy levels significantly, including and especially going with one of these microphones. Consider that if you're getting 99% accuracy on average with every 100 words dictated, improving the accuracy to 99.9% only means that you're getting approximately one less misrecognition. Is that worth paying an additional $150-$450 for one of these microphones? I don't think so.

PS, I use the Sennheiser MD 431 II with the desktop spring bloom just to keep it out of my face and away from interfering with the screen.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, we make it better!

"Many of the things you can count, don't count. Many of the things you can't count, really count." Albert Einstein



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 12/12/2011 12:05 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

Chuck,

I take your point about the expense and complexity of microphone testing, and I agree with you that most people don't need to dictate with loud environmental noise.  However, lots of people do.  Whether it's a noisy basement like mine, an active emergency department, or a trade show, conditions do exist where a desktop mic like the SpeechWare 3-in-1 will basically be useless.

For those conditions, and I face them every single evening in my own home (loud TV, loud video games, loud kids, and I need to get work done in the same room), the "conventional wisdom" is that a person needs to shell out $450 for the 431 II plus stand/cable/hub.  I really doubt this is the case.

With respect, I think you sort of missed the point of this thread.  Fact is, you sometimes need your 431 II.  Other people sometimes have a specific need for such a tool.  The question is whether the 431 II is "the one" or whether there are other microphones that can also deliver under the specific circumstances for which the 431 gets recommended time and again. 

 12/12/2011 03:55 PM
User is offline View Users Profile Print this message

Author Icon
GDS
Top-Tier Member

Posts: 844
Joined: 01/16/2009

Quote:
Whether it's a noisy basement like mine, an active emergency department, or a trade show, conditions do exist where a desktop mic like the SpeechWare 3-in-1 will basically be useless.

Amin,

This has been a great thread, and I agree with your point that we can all benefit by trying different microphones under different circumstances. To Chuck's point that accuracy is an asymptotic function, and to Rudiger's wonderful point that it's the singer and not the song, I humbly offer my own recent experience:

In Praise of the KnowBrainer Hands Free Headset

I recently dictated for hours on a noisy trade show floor using the $50 KnowBrainer hands free headset and didn't make a single mistake. Obviously circumstances are different for a demo than for a free form dictation, but no matter how well-planned the demo there is plenty of unplanned speech recognition goes on during the course of a whole day.



-------------------------

Eric Wright


DPI 15.3. 

 12/12/2011 09:33 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

Amin,

This has been a great thread, and I agree with your point that we can all benefit by trying different microphones under different circumstances. To Chuck's point that accuracy is an asymptotic function, and to Rudiger's wonderful point that it's the singer and not the song, I humbly offer my own recent experience:

In Praise of the KnowBrainer Hands Free Headset

I recently dictated for hours on a noisy trade show floor using the $50 KnowBrainer hands free headset and didn't make a single mistake. Obviously circumstances are different for a demo than for a free form dictation, but no matter how well-planned the demo there is plenty of unplanned speech recognition goes on during the course of a whole day.

Eric, I missed your post but appreciate it very much.  The KnowBrainer headset sounds like a great value.

Here's the bottom line for me:

If you want a headset and ambient noise is low, there are many affordable options, including even the OEM microphones and some very inexpensive Plantronics and Logitech headsets.  I have an absolutely cheap Dynex USB headset from Best Buy that works well in a quiet room.

If you want a headset and ambient noise is sometimes high, there are several well proven and affordable options, from theBoom models to the KnowBrainer headset and the Sennheiser ME-3.  

If you like working with a desktop mic with some working distance and ambient noise isn't too bad, there are several well-proven and affordable options, including the SpeechWare and Buddy options.

If you don't want a headset and ambient noise varies from good to bad to awful, then there is a class of performance microphones including the Sennheiser MD 431 II and probably a bunch of other models which will do the trick.  I think there are plenty of us who would like a microphone for these particular circumstances and would be interested in the options other then the 431 II; hence the thread.

 12/13/2011 08:08 AM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Quote:
Here's the bottom line for me:

If you want a headset and ambient noise is low, there are many affordable options, including even the OEM microphones and some very inexpensive Plantronics and Logitech headsets.  I have an absolutely cheap Dynex USB headset from Best Buy that works well in a quiet room.

If you want a headset and ambient noise is sometimes high, there are several well proven and affordable options, from theBoom models to the KnowBrainer headset and the Sennheiser ME-3. 

If you like working with a desktop mic with some working distance and ambient noise isn't too bad, there are several well-proven and affordable options, including the SpeechWare and Buddy options.

If you don't want a headset and ambient noise varies from good to bad to awful, then there is a class of performance microphones including the Sennheiser MD 431 II and probably a bunch of other models which will do the trick.  I think there are plenty of us who would like a microphone for these particular circumstances and would be interested in the options other then the 431 II; hence the thread.

Amin,

Nicely put. Now were cooking on all four burners.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, We make it better!

"We are all faced with a series of great opportunities brilliantly disguised as insoluble problems." - John W. Gardner



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 12/12/2011 05:22 PM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Quote:
I take your point about the expense and complexity of microphone testing, and I agree with you that most people don't need to dictate with loud environmental noise.  However, lots of people do.  Whether it's a noisy basement like mine, an active emergency department, or a trade show, conditions do exist where a desktop mic like the SpeechWare 3-in-1 will basically be useless.

Amin,


How many is "lots of people"? I agree that "some people do". However, in the same breath you also say "most people don't". The only reason that the Sennheiser MD 431 II has been brought up both currently and in the past is as a comparison to other types of microphones and as part of a discussion about "what is good noise canceling". With the advent of this collection of threads, I'm more than willing to step off that high horse and acquiesce to the fact that the Sennheiser MD 431 II isn't the only comparable microphone of its type. In fact, I've had the opportunity in the last year to test some of the other vocal microphones, particularly the Audix OM2 and the OM5, the latter of which is pretty close in performance to the OM7. As has been already pointed out, there are many comparable and some better, as well as less expensive, alternatives to the Sennheiser MD 431 II that, in all likelihood, perform as well under high noise conditions.


It doesn't make a whole lot of sense to throw the baby out with the bathwater. I purchased my Sennheiser MD 431 II years before any of these microphones were available or even known to exist. Further, advances in microphone technology have been coming fast and furious in recent years. Nonetheless, taking what I've said out of context doesn't really validator or justify some of the current comments.


Here's a test for you. Turn on CNN Headline News and turn up the volume as loud as you can stand it. Make sure that it is behind you ( 0° ). Then, test any of these microphones. Do the same at 105° and 180°. Let us know what the results are. Also, with any of these microphones, dictate several paragraphs and report on recognition accuracy. This is one way to test how well any of these microphones work because CNN Headline News is what is termed "clear speech". This is how I initially tested the Sennheiser MD 431 II eight years ago. I was playing CNN Headline News through my surroundsound home theater system using a digital sound meter and measuring the background noise level at approximately 95 dB. The volume was so loud that my neighbors complained until I told him what I was doing. Nevertheless, I did this because I knew that this is the type of environment that I was going to be using this microphone in. Using this approach is the best measure of SPL. If you get 97% accuracy or better under this condition with any of the microphones discussed, then you have a microphone that compares to the performance of the Sennheiser MD 431 II. However, far be it from me, today, to argue that the Sennheiser MD 431 II is the end-all and be-all in noise canceling microphones.


On the other hand, I get the best of both worlds by coupling the Sennheiser MD 431 II up with the SpeechWare 3-in-1 base unit as my USB SoundPod. Consider the following quality check using this combination. Anyone will get the same results using any of the microphones discussed. I'm not touting the Sennheiser MD 431 II here, just using it as an example. I get the same comparable results using the Audix OM2 and the OM5, which I haven't had a chance to test in a tradeshow environment yet.



I wholeheartedly agree that under conditions that you specify, the SpeechWare 3-in-1 with the shipping microphone (desktop microphone element unit) probably won't perform as well under the same conditions, if at all., But you're not restricted to that combination you can use any external microphone and get excellent results using the SpeechWare 3-in-1 base unit as your USB SoundPod. I've done this with the Sennheiser MD 431 II, with the Samson Airline 77 wireless microphone, and the Sennheiser ME-3, all with comparable results. Nevertheless, microphones that are designed with the same basic configuration is the Sennheiser MD 431 II work best. So, it's the microphone design and  specifications that count, not the model number or manufacturer.

Quote:
For those conditions, and I face them every single evening in my own home (loud TV, loud video games, loud kids, and I need to get work done in the same room), the "conventional wisdom" is that a person needs to shell out $450 for the 431 II plus stand/cable/hub.  I really doubt this is the case.

All things considered, you're preaching to the choir. I've never claimed that one "needs" the Sennheiser MD 431 II. I've only touted its capabilities. So, please don't quote me chapter and verse because you'll get no argument from me.

Quote:
With respect, I think you sort of missed the point of this thread.  Fact is, you sometimes need your 431 II.  Other people sometimes have a specific need for such a tool.  The question is whether the 431 II is "the one" or whether there are other microphones that can also deliver under the specific circumstances for which the 431 gets recommended time and again.

Ancient history. The Sennheiser MD 431 II isn't the "one", or "only one". It's just one of a number of similar microphones with similar specifications. It's simply the one that I've done the most real-world testing with. Give me comparable results with any of the other microphones discussed, and I'll be the first one to get on your bus. I didn't miss the point.

Chuck "Aiming for the moon and missing it is better than aiming for the ditch and hitting it."   - Author Unknown



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support



 12/12/2011 07:12 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

How many is "lots of people"?

 
I don't know. I haven't done a study. However, I think it's reasonable to assume that my own situation (no quiet home office, have to work around noisy family members) is a common one. There are also many physicians working in noisy environments at the hospital. 
 

I agree that "some people do". However, in the same breath you also say "most people don't".

 
That's another assumption on my part, and it doesn't contradict the first part of what I said. There are many cases in life where "lots of people" do something, but "most people don't" (but "some people do" ). 

I am not preaching to anyone here. In fact, I am not making any claims. I am simply putting forth a hypothesis that other microphones match the speech recognition and noise rejection qualities of the Sennheiser, and it seems that you agree with this hypothesis.

My purpose in starting this thread was to look for others' experiences with some of those other microphones. Techlaw gave us his impressions of one such microphone, and I thought that was interesting and useful to read. I wish I had time to do rigorous testing of my microphones, but I do not. I can only say that so far it seems to be working well for my situation as described.
 12/12/2011 07:34 PM
User is offline View Users Profile Print this message

Author Icon
R. Wilke
Top-Tier Member

Posts: 7702
Joined: 03/04/2007

Amin,

if truth be told, there is no such thing as a valid scientific base to evaluate accuracy gained from different microphones when used for speech recognition. The problem with it, you just can't reproduce it consistently enough in order to come to terms with it. Even with the most adequate speakers performing such, which would involve having people do it who are particularly trained to it, such as broadcasters for instance, the input will vary any time the testing is conducted to such a degree that there are too many factors changing along the course.

Therefore, you would have to accept things the way they are and try to live with it - or get yourself a better environment to dictate in. Some kinds of environments just don't cut it, or will have an extremely detrimental effect on it that you just cannot avoid kicking in, such as when dictating under a waterfall.

Basically it all boils down to what it says in the general training text, and everything else that can be said is just perpetually repeating it.

Rüdiger

 



-------------------------



No need to buy if all you want to do is try ...

DragonCapture KB Download (Latest)
DragonCapture Homepage

 12/12/2011 11:13 PM
User is offline View Users Profile Print this message


gabor5454
Power Member

Posts: 46
Joined: 03/20/2011

Quote:
On the other hand, I get the best of both worlds by coupling the Sennheiser MD 431 II up with the SpeechWare 3-in-1 base unit as my USB SoundPod

In the high noise environment, are you using it in the regular ("blue"mode or the speech rec ("green" mode?

Tx

RG

 12/13/2011 07:57 AM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

RG,

To answer your question, in the "high noise" environments I use the SpeechWare 3-in-1 base unit in green mode (speech recognition mode).

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, We make it better!

"You can have brilliant ideas, but if you cannot get them across, your ideas will not get you anywhere."  Lee Iacocca



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 12/12/2011 09:11 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 39353
Joined: 10/01/2006

Nuance does little if any microphone testing which for the most part, leaves the job up to speech recognition resellers and end-users. The key reason why we have specifically avoided testing microphones like the Sennheiser MD 431 II and Samson Q7 (OK we actually tested those 2) is because these types of microphones are not ideally suited to speech recognition even though they're extremely accurate and in the case of the MD 431, extremely noise canceling.

Quick Personal History: In a previous life (late 60s through early 90s), I owned a commercial 24 track recording studio, wrote, arranged, produced several albums and utilized a wide array of microphones. In 1992, I started using speech recognition to automate mixing tracks. Since I already owned a vast assortment of microphones, I used what I had for speech recognition. However, I soon learned that speech recognition and vocals are 2 very different animals. Some microphones need to be able to handle a capsule jarring scream while others need to pick up the nuance of the split-second attack of a bow as it strikes a cello string. Studio and performing microphones are highly specialized. Many of these microphones are not well-suited to speech recognition because they're designed for higher volume levels, specialized frequencies (like percussion) or are simply unwieldy to handle. The Sennheiser MD431 II would be one of my personal favorites if I was doing a demonstration at Comdex because of the high accuracy and noise rejection. However, it's not my favorite microphone for dictation because I don't want to hold a long heavy performing microphone within 2 inches of my mouth. If I position the MD431 II in a standard weighted stand, I find that the extended length and width partially blocks my monitor and if I use a boom arm, I'm constantly having to move it. We concur with Chuck in that you could probably find equivalent or possibly even better microphones for speech recognition but how much better would you be able to appreciate if you're already scoring over 99% accuracy. You also have to factor in the convenience of a microphone. To some degree, even a headset microphone can be a hassle because you have to take it off and on and as soon as you get the microphone capsule lined up just right, the phone rings. Of course headset microphones also offer the advantage of allowing you to lean back in your chair.

We don't offer a lot of wireless microphone choices because we focus most of our attention on the
Samson Airline 77. As much as we love this microphone, like the Sennheiser MD431 II, it was never designed for speech recognition. The headset is little heavy and it goes around the back of your head. This gives you the option to wear a cowboy hat while dictating but most of our customers really don't require this option

Bottom line: Stage or recording studio microphones often make accurate speech recognition microphones but they're frequently expensive and not necessarily ideally suited to the task. We're not saying that a performing microphone isn't the right choice for you but for the average speech recognition end-user, they would probably be better off going with something that is specifically designed for speech recognition. Our current favorite desktop microphone is the
SpeechWare 3-in-1. It's not more or less accurate than the Airline 77 or Sennheiser MD431 II but you don't have to be within 2 inches of the microphone element, it costs less, is considerably smaller and includes a number of unique features such as a dedicated circuit board with NaturallySpeaking 11 algorithms and a specialized noise rejecting USB soundcard.


-------------------------

Change "No" to "Know" w/KnowBrainer 2020
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ex 1

 12/13/2011 08:55 AM
User is offline View Users Profile Print this message

Author Icon
MDH
Top-Tier Member

Posts: 2243
Joined: 04/02/2008

Quote:
To some degree, even a headset microphone can be a hassle because you have to take it off and on and as soon as you get the microphone capsule lined up just right, the phone rings.

Lunis,

And that is an advantage of a BOOM O as opposed to a BOOM C or any other headset microphone that covers both ears. I can leave the microphone on, call a patient, phone to my uncovered ear, hangup and dictate a note summarizing the conversation, then repeat without taking the headset microphone on and off. It is a convenience that goes unmentioned, but makes the BOOM O headset microphone a very attractive option for those that use it in this fashion.

MDH

MDH



-------------------------
 12/13/2011 09:56 AM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Mark,

You can do the same with the Sennheiser ME-3, the Samson Airline 77 wireless microphone, and the KnowBrainer hands-free microphones. I do it all the time with the Samson Airline 77 wireless microphone. I often even where the headset microphone for my wireless VoIP phone along with the Samson Airline 77 wireless microphone. I just have the Samson Airline 77 wireless microphone microphone on the left side and the wireless VoIP phone on the right. Or, I can pick up the handset and do exactly the same thing.

So, there are numerous options Re: microphones.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, We make it better!

"We are all faced with a series of great opportunities brilliantly disguised as insoluble problems." - John W. Gardner



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 12/12/2011 03:32 PM
User is offline View Users Profile Print this message

Author Icon
R. Wilke
Top-Tier Member

Posts: 7702
Joined: 03/04/2007

Quote:
Consider that if you're getting 99% accuracy on average with every 100 words dictated, improving the accuracy to 99.9% only means that you're getting approximately one less misrecognition. Is that worth paying an additional $150-$450 for one of these microphones? I don't think so.

I think it would even be worth of, if there was such a thing as built-in accuracy guarantee, but then there isn't. To give you another example, in a former life, one of them, I spent quite a few years working on building sites to make my living. I used to buy just the best tools avaible, and most of the time they were also the most expensive, but that payed in them lasting much longer. There are amazing things you can do with high quality tools indeed, but they hold no guarantee in themselves that you always do.

Rüdiger

 



-------------------------



No need to buy if all you want to do is try ...

DragonCapture KB Download (Latest)
DragonCapture Homepage

 12/11/2011 04:42 PM
User is offline View Users Profile Print this message


Nelson
Top-Tier Member

Posts: 371
Joined: 09/02/2011

I have been using the Sennheiser MD 431 II for close to 12 years.  This microphone has been the most accurate and the most noise cancelling of the many, many microphones purchased or tested by me in an attempt to improve accuracy.  In my opinion, the only microphones that have surpassed the MD 431 II performance has been the Sennheiser MD 441U (800+) which is a dynamic microphone or the MKH 50 (1200+) which is a condenser microphone requiring phantom power.  If you are not willing to spend the additional money to get the better performance from the MD 441U or the MKH 50, it would be my recommendation that you stick with the MD 431 II as it has been well tested and recommended by many users and you will save money and much frustration in trying to get that better microphone.

-------------------------

DPI 15.61, Knowbrainer, VoiceComputer, Lenovo Thinkpad P50, Lenovo Thinkpad X1 Carbon, Sennheiser MD 431 II, Microtech Gefell MD 300, Sound Devices USBpre 2 and Sound Devices MixPre 3 - II.

 12/11/2011 06:22 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

Don't you think it is interesting that the acapella singers, podcasters, and others who demand quality microphones with noise/feedback rejection almost never mention the Sennheiser 431 II, yet the speech recognition community brings it up again and again? I strongly suspect that one of the primary reasons for this is that the speech recognition community is relatively small and doesn't experiment a great deal with a wide variety of microphones.  I realize that I may be wrong about this, but this new Shure microphone is working out incredibly well for me so far.
 
I don't mind spending $450+ on the microphone, but I don't like the idea of doing so because of dogma which has been passed down over the years. I would love to see a speech recognition site do rigorous testing of all of these dynamic and condenser microphones, but I suppose that the target market is too narrow to justify this. 
 12/11/2011 08:39 PM
User is offline View Users Profile Print this message

Author Icon
R. Wilke
Top-Tier Member

Posts: 7702
Joined: 03/04/2007

Quote:
but I don't like the idea of doing so because of dogma which has been passed down over the years

"Dogma" isn't in our vocabulary! - No, just kidding, of course it is, but I guess you are putting too much emphasis on particular devices when it comes to accuracy (just another dogma incidentally). Save yourself the money, or even more important, the time.

You know, it is the singer, not the song.

Rüdiger

 



-------------------------



No need to buy if all you want to do is try ...

DragonCapture KB Download (Latest)
DragonCapture Homepage

 12/11/2011 08:55 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

You know, it is the singer, not the song.

Rüdiger

I take your point, but I think we can all agree that there are environments that call for special tools.  The New York Stock Exchange is one, and the emergency department is another.  My basement is similar to those .

In the world of dynamic microphones developed for vocals, the Sennheiser MD 431 II is just one of many excellent microphones.  There are a great host of microphones that get discussed for their various qualities.  Even in terms of noise rejection, there are quite a few microphones which are more often discussed than the Sennheiser.  Yet a search of the voice recognition forums turns up very little outside of the 431, 835, OM2s, Samson Q7, and ???

A supercardiod like the 431 II isn't the best tool if the goal is maximum rear rejection.  A cardiod design would be better suited.  If the noise is coming from the sides, a hypercardiod like the OM7 would probably be the best for the job.  Yet the Sennheiser gets often billed as the unqualified "best" at noise rejection.

I'd love to see an analysis like this one but including the Sennheiser MD 431 II, OM7, and a few others: http://recordinghacks.com/2011/06/02/ultimate-podcast-mic-shootout/#rj

 12/13/2011 08:44 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

Some followup on my new Shure Beta 58A, which I'm using with an Andrea USB Hub (total cost ~$200 including mic stand):

 

This mic setup has two distinct personalities:

-When ambient noise is low, I can comfortably dictate from a working distance of 1-2 feet with excellent results

-When ambient noise is high, I can set the input level much lower and still get excellent results while dictating from a working distance of 1-2 inches.

 

Overall, I'm very pleased with this versatile setup.  Highly recommended.

 12/14/2011 05:13 PM
User is offline View Users Profile Print this message


threevok
Member

Posts: 20
Joined: 01/14/2010

I thought I had read on here that the MD 431 II has some noise canceling feature that makes it unique among dynamic microphones regardless of polar pattern. Chucker?
 12/14/2011 11:12 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 39353
Joined: 10/01/2006

Chuck can give you more specific information but as long as you speak within 2 inches of the microphone element, the noise rejection is nothing short of stunning. We're not particularly fond of this microphone because we find it a bit expensive, unwieldy, a little heavy and somewhat bulky but if we were presenting a demonstration of NaturallySpeaking in front of a large loud audience, such as Comdex, it would be our 1st microphone of choice.

-------------------------

Change "No" to "Know" w/KnowBrainer 2020
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ex 1

 12/15/2011 02:20 PM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Quote:
I thought I had read on here that the MD 431 II has some noise canceling feature that makes it unique among dynamic microphones regardless of polar pattern. Chucker?

threevok,

No, there is nothing unique about the Sennheiser MD 431 II as you can see from the collection of threads in this post.

Technically speaking "noise canceling" is a misnomer, particularly when it comes to speech recognition. The only true form of noise canceling is Active noise canceling. True noise canceling is what is used by noise canceling headphones, such as the Bose or the Sennheiser brands of noise canceling headphones for listening to music etc. Considering jet engines create 75 to 80 dB of noise inside the aircraft cabin, passive noise canceling headphone models have some serious limitations. That's where Active noise-canceling headphones come in. Active noise canceling adds an extra level of noise reduction by actively erasing lower-frequency sound waves. How do noise-canceling headphones accomplish this? They actually create their own sound waves that mimic the incoming noise in every respect except one: the headphone's sound waves are 180 degrees out of phase with the intruding waves.

When it comes to microphones used in speech recognition, Active noise canceling is employed by the Andrea ANC series microphones that use two microphone elements 180° opposed to one another, which results in the Active noise canceling process. For a period of time, the Andrea ANC microphones ruled the roost. However, they did have a tendency to be somewhat problematic because the manufacturing process never guaranteed perfect 180° out of phase Active noise canceling, and they tended to reduce the amplitude of the output (volume) as well as over canceling background noise as the volume of the background ambient noise level increased beyond approximately 80 dB as noted above. Nevertheless, it was a good idea at the time.

Active noise canceling is how your car muffler works. However, this approach uses baffling to mimic the primary engine sounds and then feed it back on itself 180° out of phase. If you have ever seen the most recent Audi car commercials, they use Active noise canceling by sampling the ambient background noise and feeding it back on itself, resulting in effect in "silence". Consider the following as a sample of how Active noise canceling works:

The form of noise canceling used in microphones is predominantly referred to as what is called Passive noise canceling. Passive noise canceling doesn't cancel noise, it reduces the impact of background noise on the quality and clarity of your speech output by "suppressing" the volume and amplitude of any signal input entering the microphone from specific access points. Microphones like the Sennheiser MD 431 II and the others referenced in this collection of threads used to additional passive techniques. On the one side they include built in wind masks and pop filters. Coupled with high quality, unidirectional microphone elements that emphasize input from the front of the microphone while desensitizing input from off access points. There are various polar patterns used in these types of microphones from cardioid, which looks like an inverted heart and the reason why there referred to as cardioid, to hypercardioid and super cardioid microphone elements which are similar except that they have an additional polar pattern (very narrow) at the back of the microphone element. What is often misunderstood with regard to hypercardioid and super cardioid microphone elements is that these types of microphones, of which the Sennheiser MD 431 II is one and uses the super cardioid microphone element, adding extra desensitizing resister to the portion of the polar pattern represented by the pattern on the back side of the microphone. This is actually an Active noise canceling technique. What it does is reduce this sensitivity of the backside polar pattern of these microphones to anywhere from -47 dB to -55 dB. This is because it not only reduces the signal amplitude coming from off access points, as well as the polar pattern at the back of the microphone element, it effectively inverts them 180° out of phase so as to effectively cancel unwanted background noise without impacting on your speech coming from the front of the microphone. This also reduces the overall size of the polar pattern such that you have to be within 1 or 2 inches of the microphone to get the best results. The claims that the cardioid polar pattern is best are not viable in terms of the way that hypercardioid and super cardioid microphones work in terms of the category of microphones that compose those that are listed in this collection of threads coupled with the Sennheiser MD 431 II. In actuality, the super cardioid polar pattern is the best because of the way that it is configured, with the hypercardioid being the second best. Nevertheless, properly configured, cardioid microphone elements can be as good while being slightly less expensive.

The last category of "noise canceling" is what is used in the SpeechWare 3-in-1. This is called Adaptive Noise Canceling. Adaptive Noise Canceling is actually a form of Active noise canceling. However, the way that it works is to reduce the noise floor making the users speech cleaner and clearer by using a number of algorithms (AGC, signal equalization, and ANC) to suppress the impact of external ambient noise to a level or volume to a level that the signal output or ambient noise is significantly reduced while the signal input or your speech coming from the front of the microphone is significantly enhanced in terms of both clarity and quality. This is what allows you to dictate with high accuracy at a distance between 10 and 14 inches from the microphone itself. It also works well with external microphones like the Sennheiser MD 431 II connected to the SpeechWare 3-in-1 base unit running in Green mode (speech recognition mode), although the distance that you can be from these microphones plugged into the SpeechWare 3-in-1 base station from the external mic input is not as significant. I'm currently using the Sennheiser MD 431 II as an external mic with the SpeechWare 3-in-1 base unit in speech recognition mode (Green mode) and I'm able to dictate with very high accuracy at a distance of about 6 inches. In fact, if I ran the Audio Setup Wizard (Check Microphone - DNS 11) under this condition keeping my mouth about 1 inch away from the microphone element, I get a poorer result (lower signal-to-noise ratio) that if I back off to about 6 inches. Backing off about 6 inches with the Sennheiser MD 431 II under this set up also reduces the propensity for picking up key clicks and other types of background noise that would interfere with both performance and accuracy. The advantage of Adaptive Noise Canceling is that the signal output can be increased to accommodate for distance from the microphone element without amplifying or over amplifying ambient background noise.

However, even with all of this being said using an external microphone such as the Sennheiser MD 431 II with the SpeechWare 3-in-1 base unit in speech recognition mode (Green mode) there is a limit to the background noise suppression as the volume increases (ambient background noise). It's best to use the speech recognition mode in low to moderate noise environments. If you need to use microphones like the Sennheiser MD 431 II or any of the other high-quality vocal mics discussed in this collection of threads, then you're better off switching to Blue mode and let the microphone itself handle the background noise (i.e., dictating at the normal distance of approximately 1 inch from the microphone element). While the playback volume may be lower under this condition, the ability of these microphone(s) to handle high-volume ambient background noise is better. Keep in mind that volume is the big accuracy killer. If the by him is too low, Dragon can't "hear" what you're saying clearly. If the ambient background noise is too loud, it's significantly impacts on Dragon stability to understand what you said, as well as tending to produce longer delays and poorer accuracy.

Some of what I've said is based on the concepts behind "noise canceling", but much of it is based on actual testing in specific environments. The Sennheiser MD 431 II represents a type of microphone with it specific set of characteristics and specifications. Many of the microphones referred to in this post have similar characteristics and would perform with equal levels of accuracy and "noise suppression". The performance that you get out of any of these microphones is dependent upon the individual microphone specifications, not on any one particular model.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, We make it better!

Absence of evidence is not evidence of absence.  Dr. Carl Sagan



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 12/23/2011 07:33 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

I wanted to give some follow up on my experiences with the Shure Beta58A.  The bottom line is that this mic delivers.  In a quiet situation, I can turn up the levels and dictate comfortably from a foot away with great accuracy.  In a loud situation, I turn down the levels and dictate from close distance.  Results are outstanding with both of my USB hubs (Andrea and Buddy) and a hair better with the preamps on my new mixer (Allen & Heath ZED-10FX).

If anyone wants a great desktop mic experience for a loud environment like an ER and doesn't want to spend 3X the money on the Sennheiser, I highly recommend the Shure Beta58A.  The Sennheiser may be even better at noise rejection, but then again, maybe not.

As an aside, I don't know why anyone would spend $450 on a dynamic microphone and mate it with a $50 USB hub when a $100 interface like an M-Audio Fast Track will give a much cleaner signal.  Just my 2 cents.

 12/23/2011 09:41 PM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Quote:
As an aside, I don't know why anyone would spend $450 on a dynamic microphone and mate it with a $50 USB hub when a $100 interface like an M-Audio Fast Track will give a much cleaner signal. Just my 2 cents.

Amin,

The only reason why someone would spend $450 on a Sennheiser MD431 II is if they need it. The Sennheiser MD431 II has a proven track record over the last several years for high performance and high accuracy over the broadest range of ambient background noise levels. I bought it because I use it at trade shows and for giving demonstrations before large audiences. However, at the time that I bought it, it was the only microphone on the radar screen with that kind of performance and accuracy that fit my particular needs.

Would I recommend it to the average user? NO. Given the recent discussions on some of the available vocal mics of comparable performance and accuracy (quality), many of those would be far more suitable for the average user at less cost. While the Audix OM2 has less noise suppression than the Sennheiser MD431 II, it's more than adequate for the average user and significantly less costly. The issue is not the cost, it's the need. If you don't need it, why by it. On that point we agree completely.

I would love to get a copy of your audio.wav (Audio Setup Wizard) file using the Shure Beta58A with the M-Audio Fast Track. I would be really surprised if it performs any better in terms of a cleaner, clearer signal output than the Sennheiser MD431 II using the Andrea PureAudio USB-SA SoundPod.

I'm attaching a copy of the audio.wav file using the Samson Airline 77 wireless microphone and the Andrea PureAudio USB-SA SoundPod, which gives me an SNR of 30. This was recorded using DNS 10.1 Medical under Windows XP. The Samson Airline 77 wireless microphone does not have a pop filter like the Sennheiser MD431 II does. Nevertheless, I ran the Audio Setup Wizard with the TV running through my home theater surround sound audio system playing at a reasonable level in the background, but you won't hear it in the audio.wav file at all. You may hear some background noise from my computer, but the speech is clear and the background, if audible is very minimal. My accuracy is also 99.9%. What you hear is all the Dragon needs for good accuracy, aside from the obvious dictation style and enunciation requirements.

PS, in update, I'm attaching a copy of the audio.wav file for the Sennheiser MD 431 II using the Andrea PureAudio USB-SA SoundPod. This was done in DNS 11.5 Professional.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make Dragon NaturallySpeaking, We make it better!

"The least questioned assumptions are often the most questionable." -- Paul Broca



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 12/23/2011 10:38 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

I would love to get a copy of your audio.wav (Audio Setup Wizard) file using the Shure Beta58A with the M-Audio Fast Track. I would be really surprised if it performs any better in terms of a cleaner, clearer signal output than the Sennheiser MD431 II using the Andrea PureAudio USB-SA SoundPod.

Chuck, I don't have an M-Audio Fast Track, I have an Allen & Heath ZED-10FX.  I also don't have the Andrea PureAudio pod, I have the Andrea Half-Duplex pod.  What I can say for sure is that my Beta58A going through the ZED preamp is noticeably cleaner than my Beta58A going through my Andrea pod.  I'm making the assumption that the less expensive Fast Track would be similar to my ZED in terms of performance, but that may not be true.  Where do the audio.wav files get stored?  I'll upload some when I get a chance. 

  Edit: I found the audio.wav files. I'll post some shortly.

 12/23/2011 11:21 PM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

Okay, I'm attaching two files. One is the audio.wav file using the Andrea half duplex USB pod, and the other is the audio.wav file using the Allen & Heath ZED preamp.  If memory serves, Dragon Medical 10.1 gave the former a SNR of 29 and the latter a score of 35.  

These were recorded under identical conditions with the same Shure Beta58A mic, software, and computer.  The only difference is the preamp.

PS, in update, I'm attaching a copy of the audio.wav file for the Sennheiser MD 431 II using the Andrea PureAudio USB-SA SoundPod. This was done in DNS 11.5 Professional.

I hear a fair amount of noise on that file during the parts where you are talking.  The noise is gone in between your phrases but clearly audible during them.




 12/24/2011 10:32 AM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Amin,

I see your point. The Allen & Heath ZED preamp definitely has extremely clean and clear output. Not particularly portable though from what I can see online, unless I'm looking at the wrong one. Nevertheless, given the clarity of the audio.wav file with the Shure Beta58A I can come close with the SpeechWare 3-in-1 (almost identical – I'll send you a sample after Christmas for comparison), but you certainly have a winner there.

Have a Merry Christmas or happy holidays, whichever is appropriate, and a safe and prosperous new year to you and your family.

Chuck Runquist
Technical Project Manager
VoiceTeach LLC
Home of VoicePower®: We don't make speech recognition, We make it better!

"Kindness is the language which the deaf can hear and the blind can see." -- Mark Twain



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 12/24/2011 10:38 AM
User is offline View Users Profile Print this message


Amin Sabet
Senior Member

Posts: 87
Joined: 11/04/2011

Chuck,

You sure have that right.  The A&H mixer isn't very portable at all.  It also wasn't cheap.

The incremental benefit over the Andrea pod is probably minimal for speech to text, but I want to try a bit of podcasting, and it definitely improves the sound of my voice .

Merry Christmas to you and your family, and best wishes for 2012! 

Amin 

 03/22/2012 03:46 PM
User is offline View Users Profile Print this message


Wolvenmoon
Junior Member

Posts: 12
Joined: 05/27/2010

Did you ever do the benchmark with the Speechware mic?
 03/22/2012 09:37 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 39353
Joined: 10/01/2006

We wouldn't necessarily call this a benchmark but we did write a Battle of the TableMikes review. Unfortunately the next shipment of SpeechWare 3-in-1 microphones isn't due until the end of April.



-------------------------

Change "No" to "Know" w/KnowBrainer 2020
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ex 1



 10/24/2012 10:01 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

I am looking at purchasing one of the alternatives to the M.D. 431. I have a couple of questions.

1. Which model will do the best with a soft voice? I always liked the idea of privacy when I am dictating and I do not want to have to speak very loudly all of the time.

2. Which microphone has the smallest profile and would work best on a desktop stand?

3. What are the major differences in the accuracy and noise reduction abilities among the alternative models if there are any?

This has been a great thread and I thank you all in advance for the advice.
 10/25/2012 12:51 AM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 39353
Joined: 10/01/2006

Our recommendation would be the SpeechWare 3-in-1 TableMike which is also our personal microphone of preference. The 3-in-1, has a much smaller profile than the Sennheiser MD431 II and has a significantly longer range which would also allow you to move a little closer the microphone and speak softly. However, this thread is very old so you may additionally want to check out 6-in-1 and 9-in-1 models.



-------------------------

Change "No" to "Know" w/KnowBrainer 2020
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ex 1

 10/25/2012 02:41 PM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Originally posted by: wheelstb I am looking at purchasing one of the alternatives to the M.D. 431. I have a couple of questions.


1. Which model will do the best with a soft voice? I always liked the idea of privacy when I am dictating and I do not want to have to speak very loudly all of the time.




wheels,


First, I assume that you're talking about the handheld equivalents to the Sennheiser MD 431 II. I don't know whether any of these or even the Sennheiser MD 431 II will adequately handle a soft voice. They just may do so and all of the microphones in this particular category (i.e., dynamic vocal handheld microphones) demonstrate good clear signal output. Off axis noise rejection is supposed to be better on the higher end versions of the Audix mics. I've only tested the Samson Q7 and the Audix OM2. Neither of these two microphones is very good at noise filtering/noise suppression/noise reduction. They're fine in quiet and relatively low noise environments, but they don't handle moderate to high noise levels very well. Amit tested the Shure microphones, the most current of which is the Shure SM58. The specs for that are very good and the price is right ($99).

Second, the above are all handheld microphones that would fit very nicely in a microphone stand, but they are fairly hefty in terms of bulk and weight. With all of these microphones your mouth has to be within 1 1/2 to 2 in. maximum from the microphone. You might find that they work very well for a soft voice if your mouth is very close to the microphone element (i.e., 1/2 in.). Unfortunately, one of the things I like about the Sennheiser MD 431 II is that it has an on off switch on the microphone that is silent. That is, it doesn't produce any noise when turning it on or off so that Dragon doesn't pick up anything when you use the switch. However, in your situation that's not any advantage.

The bottom line is that one of these microphones might work very well for you. On the other hand, you may find that you have the same experience that you did with the SpeechWare 3-in-1. I can't say for sure. The only place that I can test these is relatively quiet with no talking in the background. I can speak softly and the microphone picks me up very well. One of the things that I've been training some of the court reporters to do is to run the Audio Setup Wizard (Check Microphone) using the volume of their normal voice. They have to whisper when doing court reporting or taking depositions, but running the Audio Setup Wizard (Check Microphone) using the normal volume of their speaking voice seems to help improve recognition when they're whispering. If you choose to get one of these microphones, I would try that and see if it works for you in your situation.

Otherwise, there will be a new microphone introduced in the not-too-distant future that has a design similar to the Sennheiser ME-3 but has a very strong output gain that should be ideal for those who need to speak softly like yourself. The only drawback is that it's a wired mic.

Regardless of any of the above, choosing a microphone that works well with a soft speaking voice is kind of a crapshoot. This is simply because there's no way of knowing for sure whether the microphone that you choose will be up to the task. There is, unfortunately, no way to adequately test all microphone under all conditions. So, it's a gamble. The only thing that I can suggest is to choose the best microphone at the best price from among the dynamic vocal microphones that are available.

If you wanted to try out and Audix OM2, I have one that I'm not using and I don't intend to use, which was given to me for testing purposes. I do not have the XLR cable to go along with it, but depending upon where you live, I would be happy to give it to you and send it to you at no cost except for the cost of shipping. Up to you. I won't be using it again so it's just sitting around here collecting dust.


2. Which microphone has the smallest profile and would work best on a desktop stand?


as I said above, all of these dynamic vocal mics are handheld and they're fairly hefty. However, I have my Sennheiser MD 431 II, which is one of the heaviest among them, in a spring boom that I can move up and out of the way, as well as letting me position it very easily anywhere that I want it. On the other hand, you could use some of the very good desktop stands as well.


3. What are the major differences in the accuracy and noise reduction abilities among the alternative models if there are any?


Without testing these under strict and controlled testing environments, it's difficult to say how significant any differences are in noise reduction capabilities. I know what the Sennheiser MD 431 II can do in that it handles very high noise levels very well. I would assume that the upper end of the other dynamic vocal mics are likely very close. I don't have any idea what the Shure SM58 can or can't do or how well it does or doesn't do it. Only Amit could answer that question, but I think he found it quite good in that respect. All of the dynamic vocal mics have good sound quality, so accuracy should be an issue. The only issue would be how well any of them handled background noise, keeping in mind that your working in an environment with there's a lot of talking and if people are talking in the direction of the front part of any microphone wherein such as directed at the front part of the microphone element, all the noise reduction/noise suppression/noise filtering in the world isn't going to cancel out foreground speech coming from a source other than yourself and particularly if the volume of such is high enough.

The bottom line is that I can't really make a good recommendation for you because I can't really test even the Audix OM2 without being able to exactly mimic your what you call speaking softly. All I can do, as well as anyone else can do, is speculate.

Chuck

"Many of the things you can count, don't count. Many of the things you can't count, really count." Albert Einstein



-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 10/26/2012 12:03 PM
User is offline View Users Profile Print this message


Nelson
Top-Tier Member

Posts: 371
Joined: 09/02/2011

Wheelstb

I have never found that when dictating with the Sennheiser MD 431 II there has been any issue with the type of voice (softness) employed. Very frequently, particularly when in hotels or public places, I find there is a need to dictate very quietly in order that others not hear what is being dictated.  It has been my experience that no matter how low my voice is, this microphone is more than able to deal with the variance in voice tone, loudness, or projection. 

Regardless of which microphone you eventually choose, an option that you may want to consider is the use a microphone amplifier to overcome your soft voice problems. But unless your microphone has good lateral sound rejection capabilities the amplifier will highlight any unwanted lateral sounds. This option of course means having to deal with extra gear.

An alternative to this, and since you cannot get the 3 in 1 to work for you or you are uncertain about the other microphones recommended, you might want to consider the Sennheiser ME 65/K6, which is a combination of battery operated or phantom power microphone that will allow you to dictate up to 24 inches away from the microphone and it has the same lateral sound rejection capabilities as the Sennheiser MD 431 II. But again it is expensive.

Altogether, it seems to me that the generous offer made by Chuck is one that you should grab, particularly since you are a recent graduate without a lot of disposable income.  At the very worst you may have to purchase an amplifier which generally are under $100.  This microphone may not be perfect, but until you can find one that will meet your needs or you can afford a more expensive microphone it ought to work satisfactorily.



-------------------------

DPI 15.61, Knowbrainer, VoiceComputer, Lenovo Thinkpad P50, Lenovo Thinkpad X1 Carbon, Sennheiser MD 431 II, Microtech Gefell MD 300, Sound Devices USBpre 2 and Sound Devices MixPre 3 - II.

 10/25/2012 11:02 AM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Lunis

Thank you very much for the quick reply. I have tried both the 3-1 and the buddy seven g. For some reason need a microphone was able to work correctly for me. I sent the 3-1 back to Knowbrainer to be tested and it was reported to be working fine. Literally both of those microphones would simply produce gibberish, a few of the words would wind up the correct however.

I have good luck with the boomo and the audiotachnica

I tried creating new profiles with both devices and I am also sure that I had the proper input device selected. I have no idea what the problem was.

I was kind of looking for a microphone that would work in at least the typical office environment. It's my understanding that the noise canceling ability on the 3-1 and the buddy 7G is not well suited to working in the typical office.

Possibly instead of noise cancellation ascertain noise rejection or suppression technology.
 10/26/2012 12:16 PM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Nelson & wheels,

The officer of the Audix OM2 stands. I will never use it and it was given to me, like I said, totally for testing purposes. It will give you the same quality of output and signal clarity as the Sennheiser MD 431 II. If the Sennheiser MD 431 II works with quiet voices and you can tolerate being between 1 and 1 1/2" away from the microphone itself, it should work with soft voices because it does perform similar to the Sennheiser MD 431 II. However, it does not possess the same degree of noise suppression/noise rejection as the Sennheiser MD 431 II. Therefore, the only question is how well it will work in your particular work environment.

Chuck

"Kindness is the language which the deaf can hear and the blind can see." -- Mark Twain

-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 10/26/2012 09:13 PM
User is offline View Users Profile Print this message

Author Icon
MDH
Top-Tier Member

Posts: 2243
Joined: 04/02/2008

Chuck,

Kudos to you for your generosity.

MDH



-------------------------
 10/26/2012 05:30 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Nelson,

Thank you for the information, it makes me feel much more comfortable that I will be able to find a microphone that will work under my desired specifications.

Chuck,

Thank you for the extremely generous offer, if the offer still stands I am definitely going to take you up on it. I have sent you a private message so that we can figure out where to proceed from here. Thank you again for your extreme generosity. I cannot tell you how much I appreciate it.

As far as the noise rejection specifications of the microphone, I'm just looking for something that will be capable of functioning in the typical office environment. So I cannot imagine that it would be overly taxing for the microphone.
 11/23/2012 06:12 AM
User is offline View Users Profile Print this message

Author Icon
David.P
Top-Tier Member

Posts: 617
Joined: 10/05/2006

Originally posted by: Chucker

"Consider that if you're getting 99% accuracy on average with every 100 words dictated, improving the accuracy to 99.9% only means that you're getting approximately one less misrecognition. Is that worth paying an additional $150-$450 for one of these microphones? I don't think so."



OMG Chuck! This statement (discovered just by accident) obviously still stands to be corrected!

First, 99% accuracy means (for example) that you get 100 errors in every 10.000 dictated words.

Second, 99.9% accuracy means that you get only 10 errors in every 10.000 dictated words (not 90 as your above statement implies). This is ONE TENTH the number of errors, not "one less misrecognition", as you put it.

Thus, 99.9% accuracy means TEN TIMES better accuracy than 99%.

Similarly, 99.99% accuracy means a HUNDRED TIMES better accuracy than 99%.

 

And so forth.

 

Therefore, if anyone manages to find a microphone that improves the accuracy from 99% to 99.9% (or to 99.99%, for that matter), this microphone simply would be ten times (a hundred times, respectively) better -- not "1 %" (or 10 %, respectively) better.

 

Originally posted by: Chucker

Is that worth paying an additional $150-$450 for one of these microphones?"



Hmmm...

Anyway, since everyone knows that about 147% of all people don't know how to calculate percentages, that's not TOO bad, is it

See ya around

David.P

--
Bonus:
Real Time Deep Neural Network Lecture Speech Recognition, with amazing accuracy (and Real Time translation into spoken Chinese)



-------------------------

Sennheiser MKH Mic
Visual & Acoustic Feedback + Automatic Mic Control



 11/24/2012 02:15 AM
User is offline View Users Profile Print this message


supee
Top-Tier Member

Posts: 277
Joined: 10/28/2006

if I was a marketing executive for the manufacturer of Dragon NaturallySpeaking, I will put my sales pitch the following ways:

After quoting the percentages as stated above, I would them make the following statement:

"for example, when you dictate a 10,000 word document, dragon NaturallySpeaking version 11 might make 100 errors(99% accuracy), with Dragon NaturallySpeaking version 12, we have improved this so much that version 12 only make 10 errors!(99.9% accuracy). Correcting recognition errors is one of the problems of speech recognition, and the one that consumes precious time. This new version means that you spend one 10th of the time that you used to spend correcting mistakes.This fact is more than worth the upgrade price alone."

in fact, when we use 100 word document to make the calculation, 99% accuracy is equal to 1% error rate while 99.9% accuracy is equal 0.1% error rate.

Therefore with 100 word document, with 99% accuracy, you make one error in an hundred word document (hundred errors in a 10,000 word document, or 10 errors in a 1000 word document); while with 99.9% accuracy, you make one 10th of an error (1/10 of a word, if there is anything like that! This is equivalent to 10 errors in a 10,000 word document or one error in the 1000 word document).

 

In statistics, a larger sample tend to make things clearer.


One would expect that people who were actually trained in marketing and advertisement will be able to say these things much better than I have said it above, but alas it is not always like this.



-------------------------

DPI 15.3, XPS 9575,  Intel Core i7, 16gb ram, 1TB SSD. Sennhesser ME3,  Philips Speechmike Premium 3500, Windows 10, Office 2016.



 11/24/2012 09:16 AM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

supee,

Well put. That's precisely the problem. Percentages and statistics are relative and generally imprecise. They are difficult to express in marketing terms. Long ago I suggested to the Nuance marketing team that they stay away from these types of accuracy explanations.

David's basically correct. The first problem with percentages and statistics based on accuracy is something that you already pointed out. That is, they would be valid only with an extremely large sample where the number of misrecognitions are precisely calculated based on the types of misrecognitions and what should be counted as a misrecognition and what should not. In addition, for the average user, what they are concerned with is not the percentage of accuracy that can be achieved with one version or another; they are concerned with how easy or how frustrating it is to use Dragon, particularly if they have to make an excessive number of corrections which kills productivity. No users achieve specific percentages of accuracy 100% of the time. Accuracy varies because: (a) we never say the same things in the same way twice, and (b) the only way to calculate an individual users accuracy would be to count the misrecognitions over many sessions and then average them. Tedious and unnecessary. Calculating the percentage of accuracy over time for one individual user might be useful for that user. Regardless, it takes a great deal of care and properly measuring accuracy in terms of counting the number of misrecognitions, correctly identifying those errors that are "true" misrecognitions (e.g., words that aren't in the vocabulary don't count as misrecognitions because Dragon wouldn't get them right anyway). There are also misrecognitions that are identified as insertions, substitutions, etc. For example, if you dictate something like "Dragon NaturallySpeaking" and the end result is "Dragon NaturallySpeaking at" or some variation, that is an insertion and it only counts as one misrecognition. In addition, there are substitutions that can be multiple words or phrase. Each one of the words in a multiple word/phrase doesn't count as a misrecognition, it's the entirety of that particular multiple word combination or phrase. I know I'll take some heat on this but that's the way Nuance development and QA calculate misrecognitions. The point is that each misrecognition has to be assessed on its own and that is a very tedious process.

You bring up an interesting problem. How do you count 1/10th of word or misrecognition. That's why I use larger numbers (i.e., 1000 words, 10,000 words). Even so it's a relative measure and doesn't necessarily apply across all users or all of the time. Keep in mind that software is part of the equation, hardware is part of the equation, and the user is part of the equation, usually the largest part (70%). Even the human one-on-one communication is not 100% accurate. We've gotten so jacked up on this idea that we can measure things like accuracy objectively and empirically, that we forget that the basic aspect of getting good accuracy from the end-user standpoint is a practical issue. You can benchmarking, measuring, statistical testing until the cows come home. But, to the average user, as DavidP points out, is basically moot.

Nevertheless, your explanation is applauded.

Chuck

"Good enough never is." - Debbi Fields: founder of Mrs. Fields Bakeries

-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 11/25/2012 08:38 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

I did a little research on the sure beta 58 and the audix om 2. Not surprisingly, the only experience I was able to find recorded on the Internet were the experiences of local/performance users. In all cases that I found where the two microphones were compared the Audix om2 seemed to have the edge in noise rejection. I think what every review I found was referencing was the ability of the microphone to reject the background or musical instrument noise and promote the clarity of the vocalist.

I'm not sure how well these reviews will translate into the world of speech recognition. It seems like although they are different situations, the goal and end result would ultimately be the same. You want to reject background noise while promoting the clarity of the speaker.
 11/25/2012 10:54 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 39353
Joined: 10/01/2006

While you can use any microphone for speech recognition, it doesn't mean it's necessarily well-suited to the task. We come from the music industry and we've gone through a lot of Shure Beta 58s and a few Audix mics. We really can't recommend any of these types of microphones, including the Sennheiser MD431 II, for speech recognition because above all, they are unwieldy to the max. You either have to put them into a large weighted stand that's going to sit right in front of your monitor or you're going to need to put them into a spring arm and then constantly move them around in order to keep them in front of your mouth. In our opinion, this is a NOT very user-friendly and that's why you don't see many of these types of microphones in the microphone section of our website. If you want TableMike, I would strongly recommend a TableMike that is designed for speech recognition and doesn't constantly need repositioning because it has a healthy 20 inch range. Auto gain/EQ is another plus. If you're concerned about size, check out the new Philips SpeechMike Premium (LFH-3500) ~ (USB Handheld/TableMike) which is a handheld mouse/microphone featuring programmable buttons, built-in USB soundcard, small/light/portable, auto gain control with a 24 inch range, excellent noise rejection and more accurate. It even doubles as a TableMike with our TableMike it. Sometimes a microphone that is designed for other uses works extremely well for speech recognition but that's more of an exception to the rule. The Samson Airline 77 wireless microphone would be one of those exceptions but even the Airline 77 receives complaints about the headset being too tight on the head and the battery not being rechargeable within the unit.



-------------------------

Change "No" to "Know" w/KnowBrainer 2020
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ex 1



 03/25/2021 01:04 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 39353
Joined: 10/01/2006

The Shure Beta 58a should be excellent for speech recognition but it's a handheld music vocalist microphone that is not really designed for computer use. We stopped caring these types of microphones because even on a boom stand, they tend to be rather large and typically block part of the monitor. In order to obtain decent noise filtering you have to speak fairly close to the microphone and once you add a USB soundcard and boom, you are probably going over $200.

 

If you don't need noise filtering, check out the SoundTech GN-USB-2 TableMike. If you require noise filtering, check out the SpeechWare 3-in-1. Both of these microphones are designed for speech recognition and just as accurate as the Shure Beta 58a. The 3-in-1 includes additional perks such as auto EQ, auto-gain and a longer-range. Also note that the Shure Beta 58a frequency range is 50 to 16,000 Hz. The extended range is for musicians. 1000 to 9000 Hz are more applicable to speech recognition because they clip off high frequencies which can pick up more environmental noise. The Beta 58a supercardioid microphone element is directional but no more directional than the 3-in-1 cardioid element.



-------------------------

Change "No" to "Know" w/KnowBrainer 2020
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ex 1



 11/26/2012 01:26 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

I understand that there are a bit unwieldy. Ultimately, one am looking for is an easy way to use dictation and the telephone simultaneously. My first attempt at a setup was to use the 3-1 or the buddy 7g in conjunction with a telephone handset. I tried both of those microphones a couple of times, I even sent the 3-1 `Knowbrainer to be tested. I was never able to get either of those microphones to understand my dictation. They would literally just put gibberish on the screen maybe 40 to 60% of what was there was correct.

Typically I use the new audio technical binaural headset for the boom o. But the idea of wearing to headsets seems cumbersome and would make me feel like a cyborg. That is why I was thinking that using a handheld vocal microphone might be the best solution although I don't think there is an ideal solution.
 11/26/2012 04:21 PM
User is offline View Users Profile Print this message

Author Icon
phils
Top-Tier Member

Posts: 2624
Joined: 10/02/2006

Perhaps it is from using army surplus headsets with my ham radio when I was a kid and then using headphones for radio and recording but I don't mind wearing a headset. Plus one of the best features of the Airline77 is that by wearing it it doesn't move much relative to your mouth.


I use theBoom Quiet for my client Webex sessions and the airline 77 for dictation. I used to use only theBoom for dictation but switched to the Samson for DNS. I can wear both but since I don't dictate when on the phone I hang the two next to my chair. I also have the Sennheiser MD431 II on an articulated arm but the mic is too big even for use in bed.


Phil Schaadt



 11/26/2012 05:06 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Wearing a headset does not bother me either. I thought wearing a headset with an over the ear microphone like the Plantronics CS 70 might be a little bit awkward. I am certain he would name the look and feel like a cyborg.

There are a couple of things I'm trying to work out, perhaps someone here has been in the situation and can offer advice. I originally purchased the first audio technica headset to be carried by Knowbrainer because I thought it would be very comfortable to wear while still wearing an over the ear telephone Set up. But then I realized I would have no comfortable way of taking advantage of the text reading feature of Dragon. Which is something I would need to use quite frequently if I were working professional capacity.

Then I thought I would try the audio technica to speaker headset that recently came onto the market. Thinking I could put the Plantronics phone headset in one ear and is still have the other available to listen to the text as it was read back to me.

Theproblem with this is I intended to use the Plantronics headset with the telephone lifter. In order to activate the lifter you must press the button on the Plantronics headset. Every time I tried to access the button the audio technica headset would be sliding off of my head.

I could always use my boom o with the with the Plantronics telephone headset but for some reason having both ears covered up with a microphone makes me feel like more of a cyborg.

If anybody has any other advice I would appreciate it. Sorry I do not mean to hijack this thread. It is probably useful information to a lot of people however
 02/07/2013 07:49 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Chuck, does your offer for the microphone still stand? I am still interested.. I have sent a few private messages but, I'm assuming they did not go through. I would be more than happy to cover the cost of shipping and provide monetary compensation for the microphone. Thanks


Tom
 02/07/2013 08:08 PM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Tom,

I assume that you're talking about the Audix 0M2 handheld? If so, I realize that I don't have an XLR cable for it. If I'm correct, let me know.

Chuck

"Kindness is the language which the deaf can hear and the blind can see." -- Mark Twain

-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 02/07/2013 09:31 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Chuck,

That is what I was referring to. If you are still comfortable making the offer of the cable won't be a problem. I was expecting to have to get one anyway.

If the offer no longer stands I completely I completely understand.

Thanks

 02/08/2013 04:09 AM
User is offline View Users Profile Print this message

Author Icon
Chucker
Top-Tier Member

Posts: 14123
Joined: 10/10/2006

Tom,

The offer still stands. Send me an email or a private message with your shipping address and your preferred method of shipping and I will calculate the costs and you can pay me on PayPal (I can send you an invoice).

Chuck

Honor isn't about making the right choices. It's about dealing with the consequences. - Midori

-------------------------

VoiceComputer: the only global speech interface.

The views, thoughts and opinions expressed in this post are my own and do not reflect those of VoiceTeach LLC.

Chuck Runquist
VoiceComputer technical support

 02/08/2013 11:02 AM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Thanks for the extremely generous offer. I really do appreciate it.I will send you a private message with my shipping address.And my PayPal account information. Thanks again for your extreme generosity.
Tom
 09/03/2013 06:31 AM
User is offline View Users Profile Print this message

Author Icon
rnwolf
New Member

Posts: 1
Joined: 09/03/2013

I was inspired by Travis Rudd's "Coding by Voice" talk at PyCon 2013 earlier this year. http://pyvideo.org/video/1735/using-python-to-code-by-voice

When I bought my DNS many years ago I also bought a Plantronics mic/headset/soundcard. Unfortunatly I did not get the results I had hoped for and the kit got put in the loft.
I want to have another crack at this and want to buy better hardware and Travis said he used an "Audix OM7" microphone. http://www.audixusa.com/docs_12/specs_pdf/om7.pdf The spec seems good to a beginner like me. He even said that he had used it in a noisy coffee shop.
I share an office with a noisy computer and the fans make a continual noise and thus I am now looking to buy a mic that can work well in a noisy environment.

Travis Rudd's testimonial seems pretty strong evidence that it works. The "Audix OM7" costs £177 vs "Sennheiser MD431 II" £360 in the UK.
Is there anyone who can provide some advice on
1 ) The Audix OM7
2) Connecting the XLR connector to laptop (There seem to be a number of options, Direct cable to laptop mic all the way to XLR to USB acovertors that can also provide phantom power to mic.)
 09/03/2013 10:44 AM
User is offline View Users Profile Print this message


Nelson
Top-Tier Member

Posts: 371
Joined: 09/02/2011

rnwolfe:

 

I have never used an Audix 07 microphone.  However, I can clarify a couple of points for you.  The Audix microphone is a dynamic microphone, which does not require phantom power.  In other words, you do not need an independent power source to run your microphone.  Accordingly, all you will need is a cable that is designed for use with speech recognition that has a XLR to 3.5 mm fittings.  It is important that you buy your cable from a speech recognition dealer like Lunis, or you have the cable built for you that is intended for studio use.  Also, you should forget about plugging your cable directly into your computer as the sound cards are usually too deficient for speech recognition.  The majority of speech recognition users have either a Andrea or Buddy sound pod, which the 3.5 mm plugs into.  A new product that you might want to check out is the SpeechMatic MultiAdapter that seems to have raised the technology bar for sound cards.



-------------------------

DPI 15.61, Knowbrainer, VoiceComputer, Lenovo Thinkpad P50, Lenovo Thinkpad X1 Carbon, Sennheiser MD 431 II, Microtech Gefell MD 300, Sound Devices USBpre 2 and Sound Devices MixPre 3 - II.



 09/03/2013 10:12 AM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Obviously I have never been in your office so I don't know exactly how loud it is. But I think something like the om7 might be a bit of overkill. Unless you are looking for a desktop microphone and the noise level is fairly high. There are several less expensive headset microphones that should work very well in an environment where the major source of noise is computer fans. There might also be other desktop microphones that have been specifically designed for speech recognition that can perform well in an environment with that type of background noise. I am I completely familiar with the so other people might have to give you their opinion.

If I were you I would definitely check out the microphone comparison matrix

http://knowbrainer.com/core/pages/miccompare.cfm  



Sorry but, I don't know specific information about how to connect that particular microphone. I know that there are different types of XLR cables for available depending upon the microphone, I don't know what is required for the one.



 09/03/2013 01:21 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Just a quick note to add to Nelson's great advice. The MultiAdapter will allow you to dictate with the microphone further away from your mouth. Usually when we talk about these type of microphones, you have to keep your mouth within a couple inches of the microphone itself. Depending on the noise level in your environment you may still have to do this. It still might be worth checking out that new soundcard.
 09/03/2013 02:58 PM
User is offline View Users Profile Print this message

Author Icon
elterry
Senior Member

Posts: 156
Joined: 07/10/2013

Hi
I'm trying to follow this thread, and would request clarification. Some of the contributors suggest they use preamps to some advantage over USB sounds cards, like my own Buddy 7g.
If that is the case, are these preamps plugged direct into the onboard sound card of the PC? Or have I got the wrong end of the stick?
It sounds very interesting, and some clarification would be much appreciated.
Thanks.
Steven

-------------------------

Dragon Naturally Premium 12.5, Samson Airline 77, Buddy 7G USB, Windows 7 64 bit. 6gb.


Amazon authorpage http://www.amazon.com/Eric-Meyer/e/B005EL5028/ref=sr_tc_2_0?qid=1415359382&sr=8-2-ent


 


 


 


 


 


 


 


Eric Meyer - Author (UK)

 09/03/2013 07:05 PM
User is offline View Users Profile Print this message

Author Icon
phils
Top-Tier Member

Posts: 2624
Joined: 10/02/2006

You don't NEED a preamp even though we own a full studio set up of audio gear and expensive microphones for tele-presence, telephone integration, recording training and online sessions. I use a Buddy 7G right now as my standard USB pod and get 98% to 99% recognition with an Audio Technica microphone. MAYBE the Speechware will be better but since I have yet to receive mine I can't tell you. The simplest clarification is that if you don't already own, use and understand the characteristics of audio pre-amps for other purposes, it is not necessary to buy them for speech recognition.

Philip Schaadt

 09/03/2013 09:16 PM
User is offline View Users Profile Print this message


Nelson
Top-Tier Member

Posts: 371
Joined: 09/02/2011

Elterry:

 

I agree there is no need to have a preamplifier to get effective results from speech recognition. 

 

The devices referenced in this post and others have sound cards built into them, which typically are of a higher quality than what you get from a sound pod.  As a result, these types of devices allow the user to obtain an improvement in the accuracy that is normally achieved.  It has been my personal experience using the Sound Devices USBPre 2, the type of accuracy achieved is always around the 99% range, whereas in the past using an Andrea sound pod in every day dictation the accuracy was more around 94%.  This in practical terms is a significant timesaver, which is important to me and worth the price of the device, which costs around $650. 

 

In terms of your original question not all preamplifiers or external devices are the same.  With the Sound Devices USBPre 2, it has the capacity to run from a computer or as a standalone.  It has the ability to run dynamic and condenser microphones on different channels.  When using it with the computer it connects through a USB port and no battery is required.  On the device there are two XLR microphone connections, which requires a cable that has XLR male and XLR female ends.



-------------------------

DPI 15.61, Knowbrainer, VoiceComputer, Lenovo Thinkpad P50, Lenovo Thinkpad X1 Carbon, Sennheiser MD 431 II, Microtech Gefell MD 300, Sound Devices USBpre 2 and Sound Devices MixPre 3 - II.

 09/04/2013 05:56 AM
User is offline View Users Profile Print this message

Author Icon
elterry
Senior Member

Posts: 156
Joined: 07/10/2013

Thanks for the replies. In fact, I have a Yamaha EMX 660 mixer amp, which could input my Airport 77 and line out to my laptop soundcard direct, or to my Buddy 7g and connect via USB.
It's a leftover from my singing days (amateur opera and musical theatre )
I was wondering if it would be worth the effort, having fine control over the mic settings via the preamp.
The Sound Devices kit looks pretty pricey. Then again, so does all good kit.
Thanks again
Steven

-------------------------

Dragon Naturally Premium 12.5, Samson Airline 77, Buddy 7G USB, Windows 7 64 bit. 6gb.


Amazon authorpage http://www.amazon.com/Eric-Meyer/e/B005EL5028/ref=sr_tc_2_0?qid=1415359382&sr=8-2-ent


 


 


 


 


 


 


 


Eric Meyer - Author (UK)

 02/04/2014 11:27 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Any idea how well one of these microphones might work for dictation? The first two microphone seem like they might be slight variations of one another but I'm not sure. Basically my goal is the same as the spirit of the thread. A good noise canceling desktop microphone that does not break the bank.


AE6100 - Hypercardioid Dynamic Handheld Microphone
http://www.audio-technica.com/cms/wired_mics/b8f2c39b7aa5fc9c/index.html  


ATM610a/S - Hypercardioid Dynamic Handheld Microphone with Switch
http://www.audio-technica.com/cms/wired_mics/b5fdfc48965498ee/index.html   


AE4100 - Cardioid Dynamic Handheld Microphone
http://www.audio-technica.com/cms/wired_mics/41ec6a9b054f882d/index.html



 02/05/2014 12:38 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 39353
Joined: 10/01/2006

From a speech recognition perspective (noting that we haven't tested these microphones) they should all work well from an accuracy point of view but we also want to point out that these microphones are both expensive and clearly not designed for speech recognition. For that matter, as impressive as the Sennheiser MD 431 II is, it is also not designed for speech recognition. It does a great job but do you really want a gigantic microphone blocking your monitor? You can put it on a spring stand but moving it from one position to another is going to get old fast and you shouldn't lock your body into one position for more than a few minutes. You will also likely need to be within 2 to 3 inches of the microphone capsule and you will still need to add the price of a USB soundcard plus additional amenities like cables. You could easily be looking at an excess of $350. These types of microphones are designed for vocalists.

By comparison, the SpeechWare 3-in-1 TableMike is less expensive, designed specifically for speech recognition and includes everything you need (built-in USB soundcard, appropriate cabling, some noise filtering, auto gain control, auto EQ and even a circuit board that includes Dragon algorithms). Now add to that a 20 inch pickup range (as opposed to a few inches) and a much smaller microphone capsule that doesn't obscure your monitor. The only reason we carry the Sennheiser MD431 II is because it's extremely accurate but more importantly, extremely noise canceling. Simply put, vocalist microphones may work very well for speech recognition but there are a number of caveats. Having said that, if we were performing a dog and pony show at a large convention like CES, the MD 431 II would be our 1st choice but when when it comes to typical work (on the road or in an office), we prefer a more traditional microphone approach like a Headset or TableMike.



-------------------------

Change "No" to "Know" w/KnowBrainer 2020
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ex 1

 03/25/2021 04:15 PM
User is offline View Users Profile Print this message

Author Icon
David.P
Top-Tier Member

Posts: 617
Joined: 10/05/2006

If any, I would go for a Hypercardioid. The more "hyper" the cardioid pattern is the better.

 

However, just adding to Lunis perfectly sage advice from my experience. I went about as "hyper" as it gets and have been using a Sennheiser MKH shotgun mic for 15+ years now (got a used one and paid about 1/10th of the price that these things cost when new)

 

While this microphone is great for picking up your speech even from across the room, let alone from a foot away sitting on your desktop, it will pick up any room's background noise just equally great. Therefore, don't expect any cardioid or hypercardioid (far field) studio microphone to be usable for Dragon in noisy conditions, unless it is placed very close to your mouth.

 

In fact, this microphone was about 1/30th of the price of a Sennheiser MKH and performs just as good as a desktop SR mic as my Senn.



-------------------------

Sennheiser MKH Mic
Visual & Acoustic Feedback + Automatic Mic Control



 02/05/2014 06:26 PM
User is offline View Users Profile Print this message


wheelstb
Top-Tier Member

Posts: 1036
Joined: 11/03/2008

Lunis, thanks for the info. I have tried to 3-1, it did not work well for me I think it had to do with the reverberation off of the wall behind my desk. I'm also looking for something with more noise cancellation. I have an Audix om2 that I use on a desktop stand similar to the one that Knowbrainer carries. It really doesn't block my view. And I don't mind using it. I'm thinking about purchasing a multi-adapter, to give myself more pick up distance. I know that the multi-adapter may only add an inch or two, from the experimenting I have done I think it will be a comfortable distance.

Thankfully at Amazon new models are selling for substantially less than the suggested retail price on the audio technica website
 03/25/2021 06:32 PM
User is offline View Users Profile Print this message

Author Icon
Lunis Orcutt
Top-Tier Member

Posts: 39353
Joined: 10/01/2006

3 speech recognition microphone resellers, including ourselves, discontinued offering the Telex M-560 microphone because it was simply too sensitive.

-------------------------

Change "No" to "Know" w/KnowBrainer 2020
Trial Downloads
Dragon/Sales@KnowBrainer.com 
(615) 884-4558 ex 1

 03/26/2021 04:56 AM
User is offline View Users Profile Print this message

Author Icon
David.P
Top-Tier Member

Posts: 617
Joined: 10/05/2006

Still my second best microphone of all times, see also the reviews over there at the link. And I also own a Senn MD 431 II (which I never use because it is too un-sensitive).

This also shows how noise-supression works best: use a very un-sensitive mic (or turn its volume down very low) and place the mic directly at your mouth's corner. This way, even whisper dictation works great.

Anyway, I never do that and have always been using shotgun mics on the desktop, in quiet offices that is.



-------------------------

Sennheiser MKH Mic
Visual & Acoustic Feedback + Automatic Mic Control



Statistics
32173 users are registered to the KnowBrainer Speech Recognition forum.
There are currently 1 users logged in.
The most users ever online was 12124 on 09/09/2020 at 04:59 AM.
There are currently 414 guests browsing this forum, which makes a total of 415 users using this forum.

FuseTalk Standard Edition v4.0 - © 1999-2021 FuseTalk™ Inc. All rights reserved.