Electronics > Projects, Designs, and Technical Stuff

How good of an audio meter can be made with basic parts ?

(1/2) > >>

lordvader88:
Assuming I could find a microphone or ?? in my junk parts, I have lots of transistors and op-amps, a few types of CD4000 logic chips, timers, what kind of an audio meter should I be able to make ?

For instance, something was humming outside, I can't work out the pitch off hand, maybe ~ 2kHz. Maybe around 40dB-50dB

So what is a good audio frequency and audio power meter ? Perhaps I'm talking about a basic audio spectrum analyzer, that shows power across say 10kHz. Thats too complicated.

I've made frequency to voltage circuits before. I could use that and calibrate it to known frequencies.

I also have things like decade counters, CD4026 7segment drivers, and several 7seg displays

What are some good circuits that will lock on to the loudest source, and give some reasonable freq. and volume level ?

lordvader88:
I have Arduino and even a broken soundcard (never even looked at that yet). So someday I can really make some cool audio projects...........but today I want something analog and basic, with basic silicon parts, and maybe 555/CD4000 era stuff.

DaJMasta:
Human voice is generally around 100-900Hz, so you don't need super fast stuff to get the fundamental pitch, but there are higher overtones in the sound as well as in articulations (start and stops of syllables or other percussive noises).  If you're familiar with musical pitches, the A in the center of the treble clef staff is 440Hz, so the one above the staff is 880Hz (and is somewhat high for a Soprano singer) and the A at the bottom of the bass clef staff is 110Hz (lowish for a Baritone/bass).

It sounds like what you're looking for is a tuner, but perhaps without the note letter display and just a frequency readout.  You could take the input from a mic, amplify it, and then use a comparator and a digital timing circuit to time the period of the amplified wave, but it takes a fair amount of finesse to get that to be consistently the sound you're expecting.  Generally you'd want some sort of automated gain control (or automatic attenuation, depending on your configuration), and you may have to play around with the comparator trigger point to get a reliable detection of the tone you want, but it could still be tricky to get in the presence of noise or with certain timbres of sounds measured.

One trick could be actually to dramatically distort the audio waveform by amplifying it to the point of clipping, then using your comparator at the center.  That makes the fundamental a little easier to pick out by pushing lower harmonic content into the clipped regions of the signal, but then if the signal you're interested in is slightly lower amplitude than higher frequency content, only the higher frequency reads out, even if the lower one sounds louder (human ear frequency response is far from linear).


Another interesting option could be a variation on a strobescopic tuner.  basically, you spin a physical disk with black and white markings on it at a fixed speed (controlled by an oscillator and a stepper motor driver), then flash a light against it using the amplified audio input as your light trigger.  The resulting pattern makes the lines on the disk appear to stop moving when the pitch going into the light is the same as the speed of the disk (well, speed of the disk and number of lines per revolution).  You'd have to manually tune it around to find your frequency of interest, but with extra bands on the disk, you can see several octaves at once.


If you went the DSP route, you could get a pretty good idea of the frequency and amplitude of the tone with an FFT.  You run an FFT on the band of interest and search for the peak, then read out the results.

KK6IL:

For simply seeing the frequency of audio, it can be as simple as using most any electret microphone, a battery, pullup resistor, feeding the mike output into the audio input of a laptop, and looking at the resulting signal with Audacity or other audio analyzer program which has built in spectrum analyzer.

A good analyzer can be made without a lot of expense.  Start with a good microphone.  the Panasonic WM-61A is out of production now, but plentiful on ebay. It's been analyzed for this application in several articles.
http://www.johncon.com/john/wm61a/
https://issuu.com/150176/docs/audioexpress_-_testing_panasonic_wm
The Linkwitz microphone modification is recommended if working with higher audio levels.
http://www.linkwitzlab.com/sys_test.htm#Mic

The audio output from the microphone could be measured directly with an audio voltmeter, but is at a low level. My test microphone, based on a TSB-60 electret from JLI Electronics produces about -35 dBm, 0.15V RMS with a 105 db signal after Linkwitz mod. 
https://www.jlielectronics.com/content/JLI-61A.pdf
A preamplifier such as in one of the above articles will get low level audio high enough to be measured by the voltmeter or fed into a laptop. To calibrate the system would require access to a sound meter calibrator, which typically output 94 or 104 db at 1 kHz, or access to a calibrated sound level meter. 





tooki:

--- Quote from: DaJMasta on June 11, 2019, 11:40:12 pm ---Human voice is generally around 100-900Hz, so you don't need super fast stuff to get the fundamental pitch, but there are higher overtones in the sound as well as in articulations (start and stops of syllables or other percussive noises).  If you're familiar with musical pitches, the A in the center of the treble clef staff is 440Hz, so the one above the staff is 880Hz (and is somewhat high for a Soprano singer) and the A at the bottom of the bass clef staff is 110Hz (lowish for a Baritone/bass).

--- End quote ---
Well, while you’re right about fundamental pitch, but even basic whistling is easily up in the KHz. (I suck at whistling, and a basic free spectrum analyzer app put the highest pitch I could whistle as being around 2KHz.) But given that the so-called “voiceband” in classic analog phone lines was 300–3300Hz and still sounded kinda muffled, I think one would want at least 100Hz to perhaps 5KHz in a spectrum analyzer for the described purpose.

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod