I use the line-in input for speech recognition applications. I do this for two reasons: First, because I like electronic toys, and, second, because the line-in input is, at least presumptively, generally electronically cleaner than the microphone input channel on the sound card. The line-in input channel is "cleaner" than the microphone input channel, because the microphone input channel has its own built-in preamplifier which is not as clean and introduces more noise and distortion than my higher-quality, external preamplifier. I say "pPresumptively" because I cannot really prove that it's that way, although it's supposed to be according to the experts and according to the specifications for the sound card. I also say "pPresumptively" because I cannot prove that it yields better recognition accuracy. I think so.... It should... in theory... but I have not yet devised a method of proving it.

I use various professional quality performance style dynamic microphones, coupled to a small inexpensive preamplifier device called an M-Audio AudioBuddy, which boosts the relatively weak microphone signal up to line level. I switch between the different microphones, depending upon my whim -- not because I have any particularly sound reason for doing so, but because I can't leave well enough alone and like to play with electronic toys.

So, I switch between the Senneheiser e 835 hand-held microphone, the Shure Model 512 headset with a dynamic boom microphone (same element as model SM-10), a Behringer B-1 large diaphragm electret condenser studio microphone (requires 48 volt phantom power supplied by the preamplifier), and the Heil Pro 20 hand-held microphone. You can see the Sennheiser and the Heil microphones on Martin's WebSite. Both are excellent and I highly recommend either or, if you like toys, like me, get both both of them!

You should know, however, that using the line-in input allows me to also connect pay headset to my microphone input channel, at the same time. This allows me to switch quickly between channels, and between various headset microphones which plug into the microphone input, allowing me to alternate different microphones on that channel, such as my old favorite,the Shure VR250BT, and my newer favorite, the Sennheiser headset, like the one Marty sells, and othe microphones that work well with SR software, such as the VXi TalkPro. I have a little switch box that I made in my workshop which allows me to switch microphones in and out of either channel at will.

Again, I cannot prove whether not one channel is, really, cleaner and more electronically pure and with lower distortion than the other. The textbook analysis says that's true. And I believe it. But I also know that my sound card is a electronically much cleaner and as much lower distortion than any of the USB sound pods out there -- and I also know I cannot prove that my favorite sound card is really more accurate for speech recognition use them one of these relatively noisy sound pods.

Yet, I do think it proves that speech recognition software can be fairly forgiving of a certain amount of electronic noise and distortion in the background, but does seem to respond favorably to a high quality microphone. In other words, back from noise is one thing, and the clarity and purity of the actual microphone signal is another. Put another way, you need a high-quality microphone which produces a high quality recording signal, even know there is some background noise and background distortion in the sound input device your recording with. This makes perfect sense, of course, when we think about how the audio setup wizard is designed to compare the strength of the recording signal to the strength of background noise. As long as the recording signal is sufficiently strong relative to the background noise, the software seems to ignore the background noise just fine. The wider the signal-to-noise ratio, the better; assuming that the background noise is not altogether too high itself.

So, I frequently alternate between using a microphone on the line-in input, and on the microphone channel, and there are usually two or more microphones hooked up to my sound card at all times. As long as I use a very high quality microphone on either channel, I don't care which channel I'm using, because I think I get equally good results on either channel. You cited by previous piece on the use of preamplifiers, and I think my comments and then and now are consonant, without contradiction. You can use most microphones plugged directly into the microphone input without requiring even the 20 decibels preamplifier stage on the sound card. I believe, however, strongly, that the combination of an external preamplifier and line-in channel is acoustically and electrically superior than using the 20 decibels preamplifier stage on the sound card -- the external preamplifier is almost always of higher quality than the preamplifier built into the microphone input channel. Nevertheless, I do not think anyone has ever proved that difference materially affects recognition accuracy.

The bottom line seems to be that if you want to use the line-in input for speech recognition applications, you need an external preamplifier. Nevertheless, you can almost always use the microphone plugged directly into the sound card microphone input and get equal results. I do both, primarily because I can. There are many times in life when practical reality transcends the theoretical explanation, and we just go with what works. So, it is entirely up you how you want to proceed. I am pleased of my investment into an external preamplifier, but I hope I never have to go before Congress, or worse yet, my wife, to fully justify the financial outlay of the than to say I like electrical toys.

That's just my take.

////// Rest Regards. //// JH Richards ///
===================================================

Dubleagl2,

>>>>>>>>>>>>>>>>>>>>>>>>>>>>
You mentioned internal electronic noise. I recently purchased a new notebook, and love it, but sometimes hear thru my mic a buzzing electronic noise, which seems to be louder if I'm using AC or particularly if I'm using DNS with the mic and am charging the battery at the same time.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>

What you are describing is very common with notebook computers. Dell computers in particular are known to introduce noise when using the power charging unit rather than when on battery power. One thing you can try to do that helps in some cases is reverse polarity of the connection. In other words, turn the connector in the wall outlet 180°.

Using the Andrea USB sound pod you say you already have is the other solutions most people use to overcome unwanted electronic noise from a notebook computer.

--
Martin Markoe, eMicrophones, Inc.
The best microphones for Speech Recognition
See us at: http://www.eMicrophones.com/index.asp
Read, "Key Steps to High Speech Recognition Accuracy" at:
http://www.emicrophones.com/docDetails.asp?Documen...