Hi:

I received my 3-in-1 today.  Thanks Lunis.  That was fast shipping.  

I set it up according to the directions.  I am coming from the Olympus ME-31.  The volume on the 3-in-1 is much lower than my Olympus.  I started the volume check at 10 inches and repeated the paragraph over and over at the volume I am used to speaking.  I had to cancel because it wasn't picking up my voice.  I tried it closer and it wasn't much louder unless I was right next to the mic.  Finally I spoke closely and loudly and it picked it up.  I listened to my voice in playback and also in Windows recorder and it is very soft.  Is there a way to get the volume louder.  My Olympus is a lot louder but I think it depends on which USB sound card I am using with it.  I recall trying 2 different ones and one was also very soft.  Thanks for any help.  Kent

Kent,

Take a look at my post in the following KnowBrainer link:

SpeechWare 3-in-1 set up

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

Education is when you read the fine print. Experience is what you get if you don't. - Pete Seeger

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

Hi Chucker,

in your excellent instructions you say to"select the one for 1 channel, 16 bit, 44100 Hz (CD Quality). Do not select the 48000 Hz option."

 The documentation with the 3–in–1 specifically says to check the 48000Hz (DVD quality) option(the last sentence in "operating instructions" 2.1)

Is this a mistake? Does it matter very much?

thanks,

RG

RG,

The documentation isn't wrong per se. That is, I don't think that it matters that much. However, one of the things that I told SpeechWare some time ago was that 48000 Hz wasn't necessary and might not function as effectively as 44100 Hz. However, if you're not experiencing any kind of problems, then it probably doesn't matter. Regardless of the sampling rate, DNS extracts what it needs so that the sampling rate for Dragon is always 22.05 kHz 16-bit Mono. In short, it doesn't use either of the sampling rates shown for the SpeechWare 3-in-1 except except for extracting the 22.05 kHz 16-bit Mono sampling rate used by Dragon. So, if it ain't broke, don't fix it.

However, the other settings or the checkboxes does seem to provide a benefit by limiting potential conflicts.

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

"In theory, there is no difference between theory and practice. But in practice, there is." - Yogi Berra

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

Thanks for all the advice, Chuck.  I tried everything you suggested in the post to Philip and the volume is still very low.  Some words are not picked up while dictating.  I tried something interesting.  I plugged the Olympus into the back of the 3-in-1 and the volume was a lot louder, so I think this means it is the mic that is causing the low volume rather than the sound card portion.  I am not sure what else to try.  I don't think I should need to talk louder.  I need to practically yell and do it really close to the mic for the volume to be decent.  I appreciate your assistance.  

Kent,

Sorry to hear that my suggestions didn't fix the problem. I would be very surprised if the problem was with the SpeechWare 3-in-1 and/or the microphone unless your unit was defective. However, if plugging the Olympus into the microphone input on the back of the SpeechWare 3-in-1 base produces a better volume, then the base unit is probably fine. Because I haven't received my final release version of the SpeechWare 3-in-1, I mostly use the Sennheiser MD 431 II one into the SpeechWare 3-in-1 base, which gives me excellent results. However, I have no problems with the folder microphones that aren't adjusted for sensitivity in terms of volume. The problem that I have with those is picking up and attempting to resolve too much of the background sounds. Therefore, I use the Sennheiser MD 431 II because it, in an of itself, is the best noise canceling microphone available. So, what I get is the best of all worlds.

Do you have another system on which you can test the SpeechWare 3-in-1? The reason that I ask is that there is a remote possibility that the problem has to do with your USB ports. The AGC (Automatic Gain Control) is designed to up the amplitude of the microphone input to a predetermined level depending upon your distance from the microphone element. Therefore, at least theoretically, it shouldn't make any difference what microphone you're using.

To try to see what may be going on, try running the Audio Setup Wizard (Check Microphone - DNS 11) again with the SpeechWare 3-in-1 microphone. Then, immediately open your Dragon log (say open show Dragon log) open it in notepad by just simply double-clicking on it and go to the very bottom. It will show the audio setup SNR results. Copy those and post them here. You don't have to provide the entire Dragon log just the start and stop sections to the SNR section, which should be pretty close to the very bottom of your Dragon log if you open the Dragon log immediately after running the Audio Setup Wizard (Check Microphone - DNS 11). Also, attach the audio.wav file that is created during the Quality Check test. This file is located as follows:

C:\ProgramData\Nuance\NaturallySpeaking11\Users\<your user profile name>\current\voice_container\audio.wav (Windows 7/Windows Vista)

C:\Documents and Settings\My Documents\All Users\Nuance\NaturallySpeaking11\Users\<your user profile name>\current\voice_container\audio.wav (Windows XP)

In short, it's located in the same folder in your user profile regardless of what version of Windows you're using, it's only the path to such that changes between versions of Windows. The audio.wav file will give us a chance to listen to the results of your attempt to run the Quality Check test. This is the file that you listen to the after running the Quality Check and then hitting the Play button. It's what DNS uses.

The other possibility is that your speaker output is too low. In other words, there may not be any problem with the SpeechWare 3-in-1 microphone at all, just with the volume of output coming from the playback device. What happens if you plug a set of earphones into the SpeechWare 3-in-1 base and set the output (i.e., Playback device) to the SpeechWare 3-in-1 output? Sometimes the weaknesses there and not with unit at all.

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

"If the automobile had followed the same development cycle as the computer, a Rolls-Royce would today cost $100, get a million miles per gallon, and explode once a year, killing everyone inside." -- Robert X. Cringely

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

Thanks Chuck.

I tried it on another system with the same results.  I have also tried 3 different dedicated USB ports with the same results.

As you say it may be the playback that is too low.  However, when I listen to my dictation in Dragon Playback, sometimes words are cut off or not sounding at all because I assume the level is too low for the mic to pick up.  If I speak uncomfortably loud (for me) or very close to the mic then it picks it up.

 I have attached the audio file you requested.  This was recorded from about 10 inches away.

 Here are the results for the SNR.  I did it twice.

 10:02:59 Input Volume :  65535 ( 100 percent )

10:02:59 Level: Normal 2346, SNR: Average 24.439614

10:02:59 --- End volume calibration -------------

10:03:03 --- Begin SNR check -------------

10:03:19 Level: Normal 3093, SNR: Average 18.571797

10:03:19 --- End SNR check -------------

 

09:59:38 Input Volume :  65535 ( 100 percent )

09:59:38 Level: Normal 2298, SNR: Average 18.796523

09:59:38 --- End volume calibration -------------

09:59:51 --- Begin SNR check -------------

10:00:07 Level: Normal 3662, SNR: Average 22.181618

10:00:07 --- End SNR check -------------

That volume level does sound a bit low. We recommend forwarding this recording to bovvio@speechware.be at SpeechWare support.

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

Thanks Lunis.  I tried sending the file to the e-mail you mentioned and it was returned undeliverable.  Just tried boccio and that seems to work.

Kent,

The correct e-mail address is:

In addition, your voice is exceedingly soft. Is that your normal volume and pitch? To a certain extent it doesn't surprise me that you're having a problem after listening to your audio.wav file.

Take a listen to mine. This is my normal volume and pitch when I'm speaking to anyone or dictating. This is at 14 inches from the microphone element.

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

"In theory, there is no difference between theory and practice. But in practice, there is." - Yogi Berra

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

Chuck:

I hear the difference with yours.  I am going to try it with a louder voice.  In the meantime, here is the one recorded with my Olympus.  It is much louder.

And here it is with the 3-In-1 speaking what I would consider forced even if talking to someone right next to me.

Kent,

The audio file "audio3in1loud.wav is the best one of the bunch despite the fact that it is a slightly lower volume. The question that I have is, is the vocal pattern that I hear a direct result of your physical condition? If so, I can understand why your voice is softer and somewhat gravelly.

Try the following and see what the results are:

1. Run the Audio Setup Wizard (Check Microphone - DNS 11) at the distance from the SpeechWare 3-in-1 microphone that you want to dictate speaking at the volume in that wav file to which I refer above.

2. After establishing that level in the Audio Setup Wizard (Check Microphone - DNS 11), open DragonPad or notepad and dictate a couple of sentences at your normal speaking volume. Check the accuracy and then select those sentences and use the "play that back" command and listen to them (i.e., your voice). How do they sound? Is the volume satisfactory for you to hear what you said clearly? Does the SpeechWare 3-in-1 microphone transcribe your dictation accurately?

One of the tricks and tips that we teach court reporters, because they use a stat oh mask microphone and have to whisper in the court room, is to have them train their user profile at a normal volume and then try running the Audio Setup Wizard (Check Microphone - DNS 11) the same way. Then we have them check their accuracy when they whisper. This usually works for most court reporters. You may get a lower volume of playback, but the key is accuracy. In other words, if you use this approach as I've suggested in #1 and #2 above, what is the resulting accuracy regardless of what the volume of the playback is. If the accuracy is high, then you don't have to stress your vocal cords by speaking at an uncomfortable volume when dictating. Just speak at your normal volume and the SpeechWare 3-in-1 should accurately transcribe what you say. The only time that you want to speak louder is when you're running the Audio Setup Wizard (Check Microphone - DNS 11).

Let us know what the results are.

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

"In theory, there is no difference between theory and practice. But in practice, there is." - Yogi Berra

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

Thanks again Chuck for all the time you are taking to help me.  

No, I don't have a physical condition that affects my voice.  I have never been told that I have a soft voice.  I teach in a classroom situation and have never been told that it is hard to hear me.  My voice is a little gravelly though as you can hear.

I have done what you suggested with the training and I think it helps some.  I am definitely speaking louder now than I was with my Olympus microphone.  I was speaking softer with my Olympus because it is a louder microphone.  I have attached another audio file this time with the Olympus plugged into the  3-in-1.  The other Olympus file was not plugged into the 3-in-1.  You can hear that it is much louder. 

Even training as you suggested and speaking louder with the 3-in-1, I get some dropped words or punctuation is not heard. I have to be sure to keep the volume up at all times.  And also I cannot speak from very far away or else the accuracy suffers.  I am speaking closer than 10 inches for all my dictation now.  I certainly get no where near the range I have with the Olympus.  I heard back from Speechware and they are kindly sending me another microphone boom so I will see if that changes anything in case it is defective.  I may not get it for a couple weeks because they have a holiday there right now.

Also, I am not sure if the automatic gain control is working.  As I move back from the microphone it does not seem to amplify my voice so it is the same volume but I will have to experiment with this some more.  It sounds softer as I move back.

Thanks again and I will keep trying different things and also let you know what the new boom does for me once I have that in a couple of weeks.  

PS I am also trying typing with the mic on.  It is really nice that it doesn't pick up the clicks as words.  However, sometimes the cursor jumps to another place and so if I start dictating again, the words are put in the middle of a sentence instead of the end of the sentence.  I will also have to keep experimenting with this.

 

Kent,

Are you running the SpeechWare 3-in-1 in Green mode (green LED lit)?

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

"Tell me and I forget. Teach me and I remember. Involve me and I learn." - Benjamin Franklin

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

Yes green mode. Though I have tried blue just to see if it changed anything and it didn't.

It's possible that your microphone element is simply defective. On around December 16, we should be receiving some replacement microphone elements (microphone with gooseneck) and if you're still having trouble, contact us at (615) 884-4558 x2 or Support@KnowBrainer.com for a replacement.

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

Hi,

 

Just received the 3:1 Speechware today. I have been using the 2:1 Speechware table version before this. No matter what I do in the setup of the 3:1 microphone, the volume is so soft that the program (Powerscribe 4.7) will not accept the volume setting. If I get to about 1 inch of the mike I get adequate volume but pretty poor recognition. In contrast the 2:1 mike works extremely well up to about 12 inches away and gives adequate volume and good recognition. I was hoping the 3:1 would be incremental improvement but seems to be too low to be useable. I am using the using the press to talk setting in the "green" mode. The blue mode is even softer. 

 

I am on Windows XP. I have opened the sound control panel and can manually adjust the volume to maximum with no improvement. The AGC box is checked. This setup works perfectly with the 2:1 Speechware microphone that I have been using for the past 6 months. 

 

Any other ideas on setting or how to increase the volume? I just read the posts in this thread and have no voice volume issues and never had any volume issues on the 2:1 microphone.

 

Thanks, Mark.

 

I am having volume issues still.  I followed Chucker's advice (thank you) from the set up thread.  here is my Log and Wav file...Any thoughts?

 

Thanks again,

 

Phillip

 

09:26:28 Info: multimedia device ID 0 name <Microphone (USB TableMike)>

09:26:28 Info: default live mic sample rate is 22050

09:26:28 Info: live mic <Microphone (USB TableMike)> sample rate 22050

09:26:30 LOG (Perf): DgnSAPI paused -> BeginUtt = 140 msec

09:26:31 Info: turning off microphone immediate

09:26:31 Warning: skipping OUFA, missing wordspec for word <Show Dragon Log>

09:26:31 SigQual[ uD 2665 | Nd 0 | sN 43 | cD 0 | wF 0x00 | aF 0x00 | dA 0.0 [63158] | sA 43.0 [63158] | Rt C ]

09:26:31 Info: SDChannel_UpdateSigProc

09:26:32 Final Termination of temporary instance 

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

I tried  the microphone at home. It is still quiet. When I play back my dictation it is apparent that it is missing several words in the sentence.  The quality sounds ok...it is just quiet.  (Yes I am using it in "green light" mode).

~ Phillip 

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

It appears that some of the 3-in-1 microphones actually have the wrong microphone element. It's easy to tell if you have the wrong microphone element. Although you may have to choke up on the microphone to pass the Audio Check, you should be able to dictate from at least 10 inches away. If you have to choke up from 2 to 8 inches then you probably have the wrong microphone element. We can dictate effectively from at 18+ inches on our microphone. If your microphone isn't handling your dictation from 10 inches, contact us at Support@KnowBrainer.com for a replacement mic boom.

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

Getting it right the first time was the reason advanced for the long delay in shipping.

You're absolutely right and we suspect there was a bit of a rush at the last-minute. It probably would've been a good idea to have taken an extra day double checking

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

                        Welcome to the World’s Most Popular Speech Recognition Forum

We are not testing the microphones before shipping (for obvious reasons) but manufacturer mistakes do happen. Would you mind looking at the top back of your microphone just behind the microphone boom, for a small red triangle? If you have a red triangle, you have a manufacturer second-generation microphone. We probably don't have to ask but we assume that the USB TableMike is 1 of your choices when running the Audio Check and if you are still unable to pass the Audio Check, please contact us for an RMA so that we can test your microphone and report back. We should take care of this relatively soon because the 3-in-1 microphones have almost sold out.

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

Lunis - as usual, I'm confused.  I have replaced the circuit board with the new version and cut the resistor out of the microphone element, but I don't have any red triangle on the back near the microphone boom.  And, I do suffer from dropped words which I had attributed to my lousy enunciation and until I read these posts, I thought drop words was only happening to me.  Is my 3 in 1 still not up to date?

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

Bill
DNS Pro v12.5, KB 2012, Mtech Desktop PC (i7 960 3.2 gHz with 12 GB RAM), Windows 7 Pro 64-bit, 240 GB SSD, Philips SpeechMike 3500, SpeechWare 3-in-1 TableMike, Philips SpeechMike 5274 Classic, MS Office 2010 Professional

The red triangle was added by the manufacturer to identify second-generation units coming from the factory. The difference between a 1st generation and 2nd generation unit is the elimination of the resistor and swapping out the circuit board; both of which you have accomplished. The only part of the 2nd generation unit you don't have is the red triangle but we can probably find 1 for you at Staples  

However, missing words and delays are a red flag to substandard accuracy which can be caused by speaking softer or louder than when you ran the Audio Check or some form of interference. It can also be caused by a bad sound system but as of yet, we have not identified a bad 3-in-1. We have received 2 returns but there was nothing wrong with either unit.

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

Here is another that I did about 3 inches away instead of 10 as I did with the others.

 

Edit:  The new audio files replaced the ones in the old posts because they had the same names so I had to rename them and reattach.  I couldn't figure out how to delete the original.  Sorry for the confusion.

Unfortunately, I have to report similar problems with the 3-in-1.  Volume is low, and I hear background static that I would never hear using a good microphone (e.g., the 2-in-1).

Here's my ASW log entries, which show SNR clearly below average:

17:29:02 ASW: Start

17:29:02 ASW: parameters:  /ns /l enx /recorderui=1 /mode=LiveInput /title="3-in-1" /help="C:\Program Files (x86)\Nuance\NaturallySpeaking11\help\enx\premium\dragon_enx.chm" /outputwave="C:\ProgramData\Nuance\NaturallySpeaking11\Users\3-in-1\current\voice_container\audio.wav" /source=USB /SampleRate=22050 /v5 /outputsettings="C:\ProgramData\Nuance\NaturallySpeaking11\Users\3-in-1\audioout.dat" /inputsettings="C:\ProgramData\Nuance\NaturallySpeaking11\Users\3-in-1\audioin.dat"

17:29:24 --- Begin volume calibration -------------

17:29:31 Input Volume :  65535 ( 100 percent )

17:29:31 Level: Normal 6651, SNR: Below Average 16.629694

17:29:31 --- End volume calibration -------------

17:29:35 --- Begin SNR check -------------

17:29:52 Level: Normal 6224, SNR: Below Average 16.017830

17:29:52 --- End SNR check -------------

17:29:52 

 

I've also attached my audio.wav files.  Compare the 2-in-1 (both loud and clear) against the 3-in[1 (soft and filled with background noise).  Same office, same distance from mic element, same laptop, etc.

And I followed all the directions suggested by Chucker.  I've been using speech recognition for many, many years, so no rookie mistakes here. 

Looks like there's a pattern here. 

You may have received 1 of the -58 dB units which should have been -47 dB. What we will need to do is cut the resistor out of your microphone element which are more than happy to talk you through but you would probably feel more comfortable having us do it for you. Just contact us at Support@KnowBrainer.com and we can fix your microphone. The entire procedure only takes 10 seconds.

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

More proof that life is not fair…I did a lot of complaining about the lateness of the microphone, and mine works flawlessly. The more I use it the more impressed I am with the quality and accuracy of the transcription. On the other hand, after creating a new user as suggested by the instructions, I reverted to my well-trained user and find it far more accurate. I suspect that the older user, which was also a USB microphone (in fact it was the 2-in-1) has nailed my idiosyncrasies and in my case, that's more important than whatever benefits a new user might provide. Another reason may be is that my system is admittedly bizarre – my voice goes from the microphone via a USB port to Linux to a virtual soundcard driver in a virtual machine and only then to NaturallySpeaking running in a virtual machine. The volume can be adjusted at several points in the information stream, starting with Linux, then Windows, finally by Dragon in the check microphone setup. The 3-in-1 works best with the Linux volume set to unamplified, whereas the CS 50 USB works best with the volume set to an increase of about 30%.

I also noticed that the signal-to-noise of the 3-in-1 volume (or loudness or magnitude or whatever you'd like to call it) is really astonishing. I've seen two sources of evidence for this. First, in the check microphone settings the background noise is virtually zero until I began to speak. Second, while Dragon is waiting for me to speak, and the green microphone is on, the amplitude feedback bar is again virtually zero even while music is playing in the background – not a hint even of yellow – but as soon as I begin speaking the volume bar goes directly to the center of the green level. This is true whether I'm 4 inches from the microphone, 10 inches from the microphone, or even 2 feet away.

Go team! I'm sure that Lunis and SpeechWare will make good on the microphone elements that need replacement.

I still need to do a lot of testing but I just installed the new board and the sound quality is completely different.  It is still rather soft surprisingly.  However, it is completely clean and even when I speak softly it is picking up everything so far.  I did cut the resistor so I am a little surprised how soft it is but if I talk within just 4 or 5 inches instead of 10 it is noticeably louder.  I am impressed how well it is working though.  Thanks Lunis for getting the boards and making this work for us!

One other thing I noticed is that when I plug another microphone into the back of the 3 in 1, such as my Olympus microphone, it is very soft with the new board.  That, too, is completely different from the first board where the same microphone was much louder.  I wonder why they changed this.  I don't think I will need to use the plug-in but I was just wondering.  It is so soft as to be practically unusable.  I used the "check microphone" and received the low-volume message when dictating.

Same here.

 

I also noticed that when I play back dictation the words sound like they're surrounded by silencing envelopes--the loudness is attenuated at the start and end of each word.

We're currently in the middle of discussing this with SpeechWare but our understanding is that they lowered the volume input so that it would be ideally suited for their FlexyMike which is a very light small low impedance microphone capsule. Unfortunately, this is going to limit the use of other larger headset microphones such as those made by VXI, Sennheiser, Audio-Technica and the KnowBrainer hands-free. We would have preferred a more universal approach with a hotter input but as of yet, we are not certain what SpeechWare will do down the road. However, we have made the suggestion to implement this change in future generations.

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

Have you had any chance to check the FlexyMike?  Their website says it's been "specially 'tuned' for maximum performance with Dragon NaturallySpeaking." 

Thanks, 

Fred

The FlexyMike designed for use with the SpeechWare 3-in-1 is not yet available. Can't test what isn't yet available.

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

Be careful what you wish for.  You may find that what you get is not what you expect, or what you want. - Aesop (620 BC - 700 BC modern interpretation)

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

Has anyone tested the FlexyMike?  If so, how does it perform individually and when used with a 3-in-1?

-------------------------
DNS professional 10.1.  Windows XP. Dell latitude D620.  4GB RAM.  xTag Microphone.

The new FlexyMike's aren't here yet but the old FlexyMike utilizes the same condenser microphone element in a slightly different frame. FlexyMike accuracy is very good but the real benefit of the FlexyMike is its almost complete lack of weight and the 3-in-1 microphone input jack is tuned specifically for the FlexyMike.

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012 

 

 

 

Quote:

The FlexyMike designed for use with the SpeechWare 3-in-1 is not yet available. Tan't test what isn't yet available.

FBA,

Well Duh… We've had 10 FlexyMikes sitting on a shelf in our Office for close to 2 months because we were too retarded to notice them sitting 3 inches away from the dwindling pile of 3-in-1 microphones. We've been using the FlexyMike for the last 3 hours and actually stumbled across this question. For the time being, we can only quantify that the FlexyMike is reasonably noise canceling, very light/comfortable and accurate but we don't have any real numbers to share because we are also in the middle of testing the new Samson SE10 and SE50 microphones which also appear promising. Note that these wired headset microphones are primarily designed for end-users requiring an extremely light titanium frame and a low-profile (popular with physicians who dictate in front of patients). These microphones probably aren't for the masses but again, we don't have any conclusive information because we haven't finished testing. We will report back…

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

Click KB 2012 REV D to Download a 30 Day Evaluation of KnowBrainer 2012