2
\$\begingroup\$

What are some of the factors that makes a microphone preamp good or bad? noise, cancelling reverb and other factors.

I'm trying to build a electrec Mic that will handle voice recognition, from a distance in a room. It will connect into a Computer Audio input.

I need to talk all factors into account. mic needed range around 10 feet. I have looked over several microphones pre amp schematics.

If there are good examples I would love it.

Thanks

\$\endgroup\$
  • 1
    \$\begingroup\$ I can't stress enough that if this is what you want to do for a senior capstone project, you sincerely reconsider! Signal conditioning, voice recognition, and now making a microphone? Please see that you might be in way over your head. \$\endgroup\$ – kevlar1818 Jun 13 '12 at 18:13
  • \$\begingroup\$ Kevlar1818 I appreciate the advice. The good news this isn't a current part of my senior project, but an add on if I get ahead. I know voice recognition from a distance is a bit of a stretch. I'm still curios to try it out, so it is a side project that could supplement my senior project. \$\endgroup\$ – Ashitakalax Jun 15 '12 at 4:37
5
\$\begingroup\$

There are three main characteristics of microphone amps relevant to this discussion:

  1. Gain. The output signal of different microphones varies a lot. Electrets (with appropriate power supply and pullup resistor) usually produce a stronger signal than a bare dynamic or piezo mic. Look at the datasheet, but experimentation may be needed. From mic to line usually requires a voltage gain of somewhere around 1000, but could be a few 100 to a few 1000 depending on the mic.

  2. Noise level. This is probably your toughest criterion. With audio, signal to noise ratio generally needs to be high. Figure 60 dB bare minimum, and that's pretty crappy. 100 dB or more would be much better.

  3. Frequency response. "HiFi" audio is considered 20 Hz to 20 kHz, because that more than covers the range of hearing for most humans. If you are only interested in voice signals, then you don't need such a wide range. In fact, narrowing the frequency range then helps because it attenuates noise that can't be voice due to frequency alone. Passable voice is around 100 Hz to 3 KHz, although that will sound noticeably squashed, like thru a tunnel. 50 Hz to 8 KHz is much better, and sounds pretty good. Beyond that you add more frequencies that aren't really included in the voice, but are ambient sounds that make it sound more "live". If you are trying to do voice recognition, you want to actively limit the frequency range, but probably not in the preamp. This is better done later digitally where it can be done more flexibly and accurately.

\$\endgroup\$
  • \$\begingroup\$ Olin I understand why you would want to filter the signal for voice, is there a reason why you would want to do this digitally vs with hardware. I would think it would be just as good to have a filter on the signal line after it is amplified. Would this be a good amplifier to use? sound.westhost.com/project66.htm \$\endgroup\$ – Ashitakalax Jun 13 '12 at 16:03
  • 1
    \$\begingroup\$ @ashit: As I said, digital filters are more flexible and accurate. You can do a tight digital filter as part of decimation to reduce the bandwidth, but it would be hard to do that in analog with a narrow frequency range. You can do some analog filtering of signals you know you don't want up front, within the limits of the analog parts accuracy. \$\endgroup\$ – Olin Lathrop Jun 13 '12 at 16:22
  • 2
    \$\begingroup\$ 4. Distortion. The little single-transistor amps usually used in mic pres are not very linear. Could be good or bad, depending on the application. \$\endgroup\$ – endolith Jun 13 '12 at 17:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.