Connect with us

How would you calibrate a sound level meter

Discussion in 'Electronic Design' started by Norm Dresner, Aug 25, 2004.

Scroll to continue with content
  1. Norm Dresner

    Norm Dresner Guest

    Any microphone with a suitable amplifier and an AC voltmeter makes a sound
    level meter. Suppose I wanted to make one to set up a stereo system where
    the input is single tone at a time and the goal is to adjust a multichannel
    equalizer to get approximately overall flat response from voltage/power in
    to sound level out. I don't need OSHA or any particular "weighting" scheme.
    Can someone give me a rough idea of the sequence necessary to calibrate
    something like this?

    TIA
    Norm
     
  2. James Meyer

    James Meyer Guest

    I don't think it is practical or necessary to calibrate such a device.

    If I wanted a warm fuzzy feeling about the measurements, I would connect
    the amplified audio from the microphone to a set of good headphones and make an
    A-B comparison using some live audio source(s). If I couldn't hear a difference
    between the source directly or through the 'phones, I would assume the sound
    level meter is "flat" or "calibrated".

    Jim
     
  3. BobGardner

    BobGardner Guest

    the input is single tone at a time and the goal is to adjust a multichannel
    ============================
    If you try to use tones to eq a room, there are nulls all over because of the
    standing waves. You move the mic a couple inches and the level is off by 10dB.
    Room analyzers put nice flat pink noise out in the speakers, and then all the
    leds on the analyzer should be about flat. If you use a tone, you could warble
    it back and forth half an octave, and then the spl meter would read the avg
    level in that octave.
     
  4. Pooh Bear

    Pooh Bear Guest

    If you're only making 'relative' adjustments at various frequencies then the
    meter doesn't need to be calibrated.

    3 problems.

    1. Only 'measurement grade' mics have a suitably flat frequency response.

    2. Room acoustics result in all manner of reflections and 'interference'
    that make the type of measurement you suggest impractical.

    3. EQing on a graphic equaliser to get a flat response rarely gives a great
    sounding result in any case. The result will only be valid for a single
    measurment point anyway.


    You might like to google for 'smaart live' a program that's used to eq live
    sound systems that'll run on a PC. There's a free eval version.

    In the end - use as much EQ as you like so it 'sounds ok' !


    Graham
     
  5. Norm Dresner

    Norm Dresner Guest

    I originally wrote:

    and several responders cuationed me that room acoustics would make the
    measurement process at best questionable.

    But perhaps I simply gave a bad application of a perfectly good instrument
    so let me try again. I have several speakers of differing
    acoustic-electrical efficiency by which I mean that in a free-field
    environment the acoustic energy output per unit of electrical energy input
    differs -- the usual measurements come from the factory in the form of dB
    SPL re 1 meter per 1 watt input but those details are unimportant. What is
    important is that I have several speakers, each with a restricted frequency
    range, which I want to use to create an array with flat response, i.e. the
    acoustic energy output per unit electrical energy input is approximately
    constant across the relevant bandwidth. This is usually done with
    cross-over networks which are in reality nothing but electical bandpass
    filters. But in terms of matching the acoustic energy from the various
    speakers I also need to attenuate the response of some of them relative to
    the others. In general, tweeters are much more energy efficient than
    woofers, sometimes as much as a decade in power. To do the matching I want
    to be able to measure the acoustic energy out of each speaker as a function
    of its electrical input. SO ...
    How would one go about calibrating a sound-level meter?

    Norm
     

  6. I still don't understand why you need absolute calibration. Seems to me all
    that matters is that you get them all equal; for that all you need is a flat
    mic and flat electronics (and a free-field testing environment).

    But if you insist: there are devices you can buy, for under $100, that
    produce reference sound levels (inside a sealed cylinder sized to fit the
    reference mic). Check audio gear suppliers, like Musician's Friend,
    Sweetwater, etc.; or, I'm sure I saw one in the latest MCM catalog. They're
    called "microphone calibrators".

    Frankly I'll be awfully surprised if this approach actually leads to a
    speaker system with flat frequency response in your listening environment.
    Seems to me it would be much more productive to use the spectrum analyzer
    approach already suggested, with pink noise. I don't see how you can
    meaningfully test the speaker response out of the enclosure and sans
    electronics, because those things interact. But I suppose it wouldn't be
    the first time I've been awfully surprised.
     
  7. mike

    mike Guest

    I've spent a LOT of time in the past messing with speaker systems.
    1) you don't need absolute calibration. You just need flat frequency
    response.
    2) for multiple speaker systems, you'll have phasing problems out the
    wazoo. Measuring one speaker at a time won't tell you the whole story.
    3) Room acoustics make tone measurements USELESS, USELESSS, USELESS.
    You'll have a much better chance getting good sounding results with pink
    noise.

    If I were gonna do this again, I'd use one of the sound programs to make
    a pink noise CD. I'd play that CD in my system as the source.

    I'd use that same sound program with a "flat" microphone into my sound
    card to view the result.

    It was 20 years ago and all that was done with dedicated hardware, but
    the results were the same...I pitched the equalizer and all the random
    speakers. Went out and bought some good speakers.

    mike

    --
    Return address is VALID.
    Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
    Bunch of stuff For Sale and Wanted at the link below.
    Compaq Aero floppy,ram,battery.
    MINT HP-41CV, 2-METER AMPS, 200CH SCANNER
    http://www.geocities.com/SiliconValley/Monitor/4710/
     
  8. Pooh Bear

    Pooh Bear Guest

    Indeed. You would need to measure ina 'free field' environment to make the
    results meaningful. Out of doors would be reasonably ok.
    This is not a trivial task. Many ppl, have been here before you. Kinda like
    looking for the Holy Grail.


    Graham
     
  9. Mook Johnson

    Mook Johnson Guest

    If you want to do it right and not too Xpensive, put that Radio Shack SPL
    meter aside and use this.

    http://www.speakerworkshop.com/

    then go here to learn how to build the jigs and how to take proper
    measurements.

    http://www.speakerworkshop.com/Links.htm

    the RS meter has a response curve that is not linear. There are pages that
    show you ho to make it linear but you are likely better off just building a
    preamp for a $3 panasonic electro mic that IS linear (enough).

    good luck
     
  10. Leon Heller

    Leon Heller Guest

    Knowles makes very good electret mics, also, that are suitable for accurate
    measurements. Where I used to work our acoustic division often used them in
    the lab as they were much cheaper than the ones supplied with their
    measuring equipment.

    Leon
     
  11. mike

    mike Guest

    You're still focusing on the wrong problem.
    How do you intend to attenuate the drive to the speakers? Unless you
    plan to have an amplifier for each one, your attenuator is gonna really
    mess things up. That's why purists don't like passive crossovers.
    Attenuators are much WORSE.

    There are phase shifts. In the crossover, in the attenuator, due to the
    relative placement of speakers. You need to measure amplitude and phase
    from the listening point to all speakers. The more speakers you cobble
    together, the smaller the sweet spot will be...well, I guess there's a
    limiting case with an infinite number of speakers in an infinite space
    that might work well...

    Try this experiment. Put pink noise into your speakers. Use the sound
    card spectrum analzyer on the microphone. Experiment with the high
    frequency response as you translate/rotate the microphone in space.
    Also see what happens if you put a small flat plate behind the
    microphone. Just the shape of the microphone and any box that it might
    be in can have major effects on the measurement.

    Many have gone where you're going. Some have never come back ;-)

    Having said all that, it appears that you won't be satisfied until you
    calibrate your microphone.
    If you're in Portland, Oregon, I have a General Radio Type 1562
    Sound-Level Calibrator that we could try out. It put out sound, but
    I've never felt the need to use it.
    mike


    --
    Return address is VALID.
    Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
    Bunch of stuff For Sale and Wanted at the link below.
    Compaq Aero floppy,ram,battery.
    MINT HP-41CV, 2-METER AMPS, 200CH SCANNER
    http://www.geocities.com/SiliconValley/Monitor/4710/
     
  12. Norm Dresner

    Norm Dresner Guest

    Yes, ultimtely I plan to have a bi- or tri-amplified system but wanted a a
    way to determine the gain/attenuation in each channel.

    ...
    Sadly, I'm on the other coast in MD. But I'll see if I can find something
    like that.

    Thanks for your comments

    Norm
     
  13. legg

    legg Guest

    You are attempting to calibrate the amplifier and loudspeaker in the
    described exercise.

    To do this, the measuring equipment, signal source and test setup must
    be pre-calibrated and standardized to your application, even to the
    point-source or physical shape of reception (say - location of ears on
    a standard head).

    If the measurement is made with simple VU or SPL measurement, then the
    source must vary over the frequency range of interest - say a slow
    sweep or spot signals.

    Formatted signals for this are available from

    http://www.marchandelec.com/sweeps.htm

    http://62.242.11.12/midiutil/index.htm

    Whether what you get will be better than just dumping the equalizer
    altogether is moot.

    To calibrate metering - the standard technique is to compare it to
    reference calibrated standards. The calibration standards are produced
    by gnomes in secret laboratories in the Black Forest.

    RL
     
  14. Norm Dresner

    Norm Dresner Guest

    Do you have a URL for the Black Forest ;-))

    Norm
     
  15. (in <5nkZc.540775$>) about
    'How would you calibrate a sound level meter RESTATING THE PROBLEM', on
    Wed, 1 Sep 2004:
    http://www.blackforest-airport.com/E/
     
  16. Jim Thompson

    Jim Thompson Guest

    See that smog. You're looking west into France.

    ...Jim Thompson
     
  17. mike

    mike Guest

    There was one on ebay yesterday.
    mike


    --
    Return address is VALID.
    Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
    Bunch of stuff For Sale and Wanted at the link below.
    Compaq Aero floppy,ram,battery.
    MINT HP-41CV, 2-METER AMPS, 200CH SCANNER
    http://www.geocities.com/SiliconValley/Monitor/4710/
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-