Announcement

Collapse
No announcement yet.

Taking accurate freq measurements?

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Taking accurate freq measurements?

    Well ... how's it done in the real world?

    Getting in to building my own crossovers. To keep costs and design/rework time down, and avoid anymore frustration than necessary, I need to start with accurate measurements.

    I'll be using REW and a PE UMM-6 usb mic to measure in box freq response/make .frd files, and REW along with a Behirnger UCA222 usb audio interface and the Arta impedance box to measure in box impedance/make .zma files. Can also use Arta, but a bit more familiar with REW.

    There are so many different ways to measure, which one is correct for those of us in the real world? Not many of us have an an anechoic chamber or are willing to plant our speakers in the ground. Some speakers are just too large to easily move to an open spot in our back yard, nor does the weather always cooperate.

    So I'll tell you the way I've been measuring in box freq response, it's stupid simple, and you tell me what needs to change.

    The mic sets 1 meter in front of the tweeter, same height, pointed directly at it. Then REW does its thing. The response then gets gated to the first reflection. If I'm measuring individual drivers in the same speaker, the process is repeated for all of them without moving the mic or making any other adjustments. If the speaker is small enough to move outside, they are set on a tall saw horse and measured in the same manner. REW can also calculate driver offsets for later use.

    Along with measuring at 1 meter, I've read to measure at the listening position, to measure about 10" from the cone, and various other distances. Even tried Jeff B's (white paper) near-field method, which didn't work for me at all; maybe it was the driver's large phase plugs effecting the outcome or something >I< screwed up.

    So what's the take? What is the right way to do this in the real world?



    http://techtalk.parts-express.com/fo...khanspires-but
    http://techtalk.parts-express.com/fo...pico-neo-build
    http://techtalk.parts-express.com/fo...ensation-build

  • #2
    Originally posted by Kornbread View Post
    So what's the take? What is the right way to do this in the real world?
    I don't know, but I'm betting you'll get many different answers.

    Comment


    • #3
      From reading Bagby's comments regarding crossovers on facebook, tt seems he likes to be close to the drivers when designing crossovers to eliminate room issues. The arguments against that method being: why not measure the speaker from desired listening distance to get a more desired FR at listening distance. I guess there's some flex in design styles.

      Comment


      • #4
        Measuring.
        1)Response of an individual driver on a standard baffle
        2) Measuring driver/s in the speaker cabinet and looking at the response and diffraction.
        3)Measuring Speaker including x-over on and off axis
        4) Measuring at the different distances and looking at the first arrival + secondary arrival aka looking at the response of the speaker in the room.
        http://www.diy-ny.com/

        Comment


        • #5
          While sensitivity is referenced to 1 meter it's almost never measured at one meter. If you're measuring near field to take the room of out the equation you do so with the mic about a half inch from each driver and then combine their individual responses. If you're measuring far field that's done from at least 2 meters, preferably more, to allow the wave fronts from the individual drivers to integrate and become one.
          www.billfitzmaurice.com
          www.billfitzmaurice.info/forum

          Comment


          • #6
            Thanks guys.

            Originally posted by r-carpenter View Post
            Measuring.
            1)Response of an individual driver on a standard baffle
            2) Measuring driver/s in the speaker cabinet and looking at the response and diffraction.
            3)Measuring Speaker including x-over on and off axis
            4) Measuring at the different distances and looking at the first arrival + secondary arrival aka looking at the response of the speaker in the room.
            In response to #2. What measurement technique/s have you found works best, for you, measuring raw drivers in the speaker cabinet in order to obtain accurate .frd and .zma files for crossover building? Do you measure nearfield (less than f rom the cone), further away, both?

            While sensitivity is referenced to 1 meter it's almost never measured at one meter. If you're measuring near field to take the room of out the equation you do so with the mic about a half inch from each driver and then combine their individual responses. If you're measuring far field that's done from at least 2 meters, preferably more, to allow the wave fronts from the individual drivers to integrate and become one.
            So it's either nearfield, as in Bagby's white paper, or measure from at least 2 meters? I've tried to follow Bagby's white paper on measuring near field. It was a total failure. How would the rather large phase plugs on the Dayton RS line, effect that type of measurement when the phase plug is right where the mic needs to be?

            I'll increase the next measurements to 2 meters and compare them to their 1 meter counterparts.









            http://techtalk.parts-express.com/fo...khanspires-but
            http://techtalk.parts-express.com/fo...pico-neo-build
            http://techtalk.parts-express.com/fo...ensation-build

            Comment


            • #7
              I would recommend this https://www.amazon.com/Testing-Louds...4351534&sr=1-1
              to start. You can find it on ebay for a good price. It covers measuring loudspeaker.
              After you done with D'
              read this: https://www.amazon.com/Sound-Reprodu...ED4QWT56Q38T98
              http://www.diy-ny.com/

              Comment


              • #8
                The answer is... it depends :-)

                ​The results of my own tests suggest that the closer the mic is to the speaker, the worse the measurement is likely going to be for x-over design, if you follow the usual approach of measuring on the tweeter's axis. I've attached two images indicating the results I obtained when performing measurements on my MS3.2 speakers, where the center-to-center distance between the two drivers is about 11cm. Minor differences started showing up just going from a 100cm to a 90cm measuring distance, and the differences started to get pretty significant above 3.5kHz at the mic started to get closer to the speaker.
                Brian Steele
                www.diysubwoofers.org

                Comment


                • #9
                  Originally posted by Brian Steele View Post
                  The answer is... it depends :-)

                  ​The results of my own tests suggest that the closer the mic is to the speaker, the worse the measurement is likely going to be for x-over design, if you follow the usual approach of measuring on the tweeter's axis. I've attached two images indicating the results I obtained when performing measurements on my MS3.2 speakers, where the center-to-center distance between the two drivers is about 11cm. Minor differences started showing up just going from a 100cm to a 90cm measuring distance, and the differences started to get pretty significant above 3.5kHz at the mic started to get closer to the speaker.
                  Do you feel the measurements taken @100cm were accurate, or does the mic distance need increased from there? Would it be likely that a speaker with multiple drivers, say a wmtmw, or where a woofer is placed far from the tweeter for floor reinforcement, would be more accurately measured at distances more than 100cm, while something like a small bookshelf, with driver centers fairly close, would be ok at 100cm? How do you do it and why?

                  Mr, Carpenter, thanks for the reading materiel, I've been looking for something like that.
                  http://techtalk.parts-express.com/fo...khanspires-but
                  http://techtalk.parts-express.com/fo...pico-neo-build
                  http://techtalk.parts-express.com/fo...ensation-build

                  Comment


                  • #10
                    If you're measuring your drivers in a box that's already built (which then already incorporates baffle-step and diffraction effects - so you don't have to estimate (or model) them), MANY would take "gated" measurements (like, w/ARTA) at a meter (could be more, could be less) on the tweeter-axis. In a room w/an 8' high ceiling, you'd get the mic and tweeter 4' off the floor and take measurements gated to approx. 3.3msec (assuming that the floor, ceiling, and closest walls were no less than 4' from the mic or tweeter) to "gate out" 1st reflections (effectively taking your room out of the data). When you do that, your measurements are really only good down to around 300Hz. THAT's why a woofer (then) needs to be close-mic'd, although, for a 2-way (crossing at say, 2kHz) you'd probably get accecptable results by using a box model below 300 and building in roughly HALF your BSC from 300 on up into your XO. Your measurements around the Fc (plus probably +/-2 octaves) will be adequate for XO design work.

                    If you measure (your woofer) nearfield (which would then use NO gate, or a very long one - 33msec is good down to 30Hz or so), you then need to look at a box model and adjust the SPL of those measurements up/down to match your 1m measurements. Having a port adds another level of estimation (guessing?) into your calculations. Acceptable practice is to use a box model (massaged for baffle-step losses) below 300Hz and "splice" it on to your actual in-box measurements.

                    Comment


                    • #11
                      Originally posted by Chris Roemer View Post
                      If you're measuring your drivers in a box that's already built (which then already incorporates baffle-step and diffraction effects - so you don't have to estimate (or model) them), MANY would take "gated" measurements (like, w/ARTA) at a meter (could be more, could be less) on the tweeter-axis. In a room w/an 8' high ceiling, you'd get the mic and tweeter 4' off the floor and take measurements gated to approx. 3.3msec (assuming that the floor, ceiling, and closest walls were no less than 4' from the mic or tweeter) to "gate out" 1st reflections (effectively taking your room out of the data). When you do that, your measurements are really only good down to around 300Hz. THAT's why a woofer (then) needs to be close-mic'd, although, for a 2-way (crossing at say, 2kHz) you'd probably get accecptable results by using a box model below 300 and building in roughly HALF your BSC from 300 on up into your XO. Your measurements around the Fc (plus probably +/-2 octaves) will be adequate for XO design work.

                      If you measure (your woofer) nearfield (which would then use NO gate, or a very long one - 33msec is good down to 30Hz or so), you then need to look at a box model and adjust the SPL of those measurements up/down to match your 1m measurements. Having a port adds another level of estimation (guessing?) into your calculations. Acceptable practice is to use a box model (massaged for baffle-step losses) below 300Hz and "splice" it on to your actual in-box measurements.

                      Since I'm also trying to learn how the take accurate speaker measurements for crossover design, I found the above post to be immensely helpful. I cannot thank fellow forum veterans (such as yourself) enough for taking the time to respond to our questions.

                      -D

                      Comment


                      • #12
                        Originally posted by Kornbread View Post

                        Do you feel the measurements taken @100cm were accurate, or does the mic distance need increased from there? Would it be likely that a speaker with multiple drivers, say a wmtmw, or where a woofer is placed far from the tweeter for floor reinforcement, would be more accurately measured at distances more than 100cm, while something like a small bookshelf, with driver centers fairly close, would be ok at 100cm? How do you do it and why?
                        ​At 100cm the measurements were accurate enough for my needs. There was little variation going from 100cm to 90cm, so I suspect there would be even less variation going from 100cm to a greater distance. And yes, I suspect that for speakers with a greater distance between the tweeter and the driver or drivers handling the rest of the spectrum, a greater measuring distance might be required.

                        Basically I follow the procedure outlined by Chris for frequency response measurements.
                        Brian Steele
                        www.diysubwoofers.org

                        Comment


                        • #13
                          for multiple drivers on a baffle / box , i measure each from 3-4ft on tweet/design axis using the same amplitude .i set up the amp , mic. and speaker and try to take all needed measurements at the same time i don't use gating , but will blend near field woofer measurements below 200 .
                          Paper Towers
                          RS180P/28F surrounds
                          Boombox

                          Comment


                          • #14
                            What is optimal near-field mic placement when the woofer has a phase plug that sits right where the mic is supposed to go?
                            http://techtalk.parts-express.com/fo...khanspires-but
                            http://techtalk.parts-express.com/fo...pico-neo-build
                            http://techtalk.parts-express.com/fo...ensation-build

                            Comment


                            • #15
                              Originally posted by Kornbread View Post
                              What is optimal near-field mic placement when the woofer has a phase plug that sits right where the mic is supposed to go?
                              1 cm from the phase plug. You can go just to one side of the phase plug as that does not effect the FR much, if any. Most of the higher frequencies of the near field you are throwing away when splicing/ blending anyway.
                              John H

                              Synergy Horn, SLS-85, BMR-3L, Mini-TL, BR-2, Titan OB, B452, Udique, Vultus, Latus1, Seriatim, Aperivox,Pencil Tower

                              Comment

                              Working...
                              X