1. WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously.

3 way DSP XO

Discussion in 'DIY Audio Forum' started by BE718, Nov 11, 2016.

  1. RayDunzl

    RayDunzl Major Contributor Central Scrutinizer

    Joined:
    Mar 9, 2016
    Messages:
    3,258
    Likes Received:
    459
    Location:
    Riverview, Fl
    Why?
     
  2. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    Short speaker cables. The amp would have to reside in the AV cupboard and would require about 15m of speaker cable. I wouldnt have the amp next to the speakers. Much better to run the balanced input signal over that distance to the speakers.
     
  3. Fitzcaraldo215

    Fitzcaraldo215 Addicted to Fun and Learning

    Joined:
    Mar 4, 2016
    Messages:
    677
    Likes Received:
    131
    Makes sense from an economic packaging point of view, and I do believe in short speaker cables where possible. Might need a long power cord to the speakers, but so what.
     
  4. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    Ok, some measurements. This was a quick and dirty, Acourate corrected measurements were taken at the listening position without having linearised the drivers first. Mainly I was being impatient :) , but this way I can see if there are significant differences when I do get to linearise the drivers. Measured with an Earthworks M23.
    XO at 600Hz and 5kHz with Acourates default Neville Theile 2nd order.

    xo response.PNG

    Frequency Response - Im tilting down from 1kHz to about -5dB 24kHz

    The intention is to use 2 subs crossed over at 50Hz, but this is just the speaker.

    FR.png

    Distortion - need to check out the Woofer distortion, higher than expected. Seems odd, wondering if REW is ok with the delay the convolver is making and tracking harmonics through the sweep.

    Distortion.png

    Impulse response - wow. might look at fine tuning with the pre ringing

    IR1.png

    IR2.png

    Group delay

    GD.png

    RT60 - Pretty much within EBU studio recomendations

    RT60.png

    reverb2.PNG

    Clarity - Still havent got to grips with this measurement

    Clarity.png

    Waterfall

    waterfall.png

    waterfall2.png

    IACC

    IACC.PNG

    How does it sound? Clean, clear and neutral, but not in a boring way, just no emphasis on any part of the frequency range. Very dynamic with instruments separated out in the mix very easily. Big improvement in soundstage/envelopment. Very pleased so far, some of the best sound I have had at home.
     
    Last edited: Jul 27, 2017
    Thomas savage and mitchco like this.
  5. Brad

    Brad Member

    Joined:
    Nov 8, 2016
    Messages:
    50
    Likes Received:
    9
    You may want to use acourate for the distortion measurements, letting the log-sweep recorder apply the filters. See Mitch's book for a good description.
    I'll have to go and look up what this clarity measurement is now...
     
  6. RayDunzl

    RayDunzl Major Contributor Central Scrutinizer

    Joined:
    Mar 9, 2016
    Messages:
    3,258
    Likes Received:
    459
    Location:
    Riverview, Fl
    Use single tone on woofer/low frequencies to verify. RTA works for that.

    Music rarely "sweeps" anyway.

    ---

    There's a new Stepped Sine Test (not swept), also, available from the RTA window (newest REW versions). I tried it like once.

    ---

    Let's see the phase display, too. Maybe L/R/Both.
     
    Last edited: Jul 27, 2017
  7. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    Will try it out. Thanks

    Phase left

    phase.png

    right

    phase right.png

    GD

    gd 2.png
     
    Last edited: Jul 27, 2017
  8. RayDunzl

    RayDunzl Major Contributor Central Scrutinizer

    Joined:
    Mar 9, 2016
    Messages:
    3,258
    Likes Received:
    459
    Location:
    Riverview, Fl
    Phase - ok, similar to my JBL at listening position. It straightens out a lot if measured nearfield. I blame Wide Dispersion reflections, since the panels can throw a flat phase to the listener.

    At the couch. Blue JBL, Green MartinLogan

    upload_2017-7-26_23-34-31.png

    Somtimes the first kink is as low as 200Hz or so, passive cross to woofer is 180Hz
     
  9. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    I will measure the corrected system tonight with Acourate but here is the pre correction phase and GD (just the XO) measurement.....mmmm....

    phase.PNG

    gd.PNG


    Convolution test output

    phase
    conv phase.PNG

    GD
    conve gd.PNG
     
    Last edited: Jul 27, 2017
  10. Cosmik

    Cosmik Addicted to Fun and Learning

    Joined:
    Apr 24, 2016
    Messages:
    817
    Likes Received:
    201
    Location:
    UK
    A few thoughts:
    1. Measuring phase and therefore frequency response at the listening position is like using a camera without a lens to measure the performance of a TV - reflections bouncing around the room fall on it. The image looks 'random' but it isn't - it is consistent with the room. Frequency response and phase response are tied together. A lens can make sense of it, as can deconvolution.
    2. Our hearing seems to be the audio equivalent of a lens - we "hear through the room" as Floyd Toole says, or we can consider our hearing to be deconvolving what falls on our ears to isolate the source; we can identify a neutral speaker regardless of the room. If we attempt to 'correct' the no-lens image, then when we fit the lens, we will reveal a now-unnatural image. I think this is one of the reasons that DSP does not yet have a great reputation - people haven't thought it through and, armed with the tools, are mistakenly going into flat-line frenzy because they can.
    3. Near field measurements work until the wavelengths of the frequencies you are measuring get short compared to the diameter of the transducer.
    4. Our audio antecedents built anechoic chambers for a reason - if only the measurement at the listening position matters, why did they bother? We can get the same benefits in an ordinary room by combining low frequency near field measurements with pseudo-anechoic gated measurements at higher frequencies.
    5. Non-ideal dispersion will interact with the room, and cannot be totally corrected for. There is always going to be some subjective element to this. However, dispersion can be improved 'at source' by various measures such as going three-way instead of two, etc. The Kii Threes of this world do it actively by design. The remaining effects can be predicted and corrected for with theoretical curves whose depth is set by ear.
     
    Last edited: Jul 27, 2017
  11. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    OK, here are the acourate measure results of amplitude Phase and group delay

    Not consistent with REW, BUT REW may be upset by the massive convolution delay (0.7secs) even though I did use the "wait for timing reference" option.

    Acourate / REW experts, ideas? :)

    upload_2017-7-27_18-1-28.png

    upload_2017-7-27_18-2-24.png
     
    Last edited: Jul 27, 2017
    Thomas savage likes this.
  12. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    Cant seem to get the distortion calculation to work, just says the selection is too short??????

    Anyhoo re-measured with REW, direct and close. Bugger all wrong with this.

    distortion1.png

    distortion2.png
     
    Last edited: Jul 27, 2017
  13. RayDunzl

    RayDunzl Major Contributor Central Scrutinizer

    Joined:
    Mar 9, 2016
    Messages:
    3,258
    Likes Received:
    459
    Location:
    Riverview, Fl
    Single tone?

    Maybe I misunderstand, but, I meant:

    Use the Generator tool to make a sine wave
    Monitor with the RTA window and its Distortion Panel (numerical display)

    example:

    upload_2017-7-27_15-54-35.png
     
    Last edited: Jul 28, 2017
  14. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    Thanks, Ray, understood, just didnt have time last night to try out. I did the sweep just to check there wasnt anything untoward going on, which there appears not to be. I will try the single tone tonight.

    BTW the "too short" error was from acourate not REW.
     
  15. RayDunzl

    RayDunzl Major Contributor Central Scrutinizer

    Joined:
    Mar 9, 2016
    Messages:
    3,258
    Likes Received:
    459
    Location:
    Riverview, Fl
    Wasn't sure.

    Looks like your speaker project is going well.
     
  16. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    Hi Ray, performed the single tone distortion measurements. Mic was at about 3cm with volume set to about 85 dB at 1 metre, so not blasting. Distortion seems nice and low, increasing at very low frequency but that is to be expected and will be a non issue when the sub is integrated.

    Woofer. XO will be 50Hz to 700Hz
    20hz.jpg 30hz.jpg 40hz.jpg 50hz.jpg 60hz.jpg 70hz.jpg 80hz.jpg 90hz.jpg 100hz.jpg 200hz.jpg 300hz.jpg 400hz.jpg 500hz.jpg 600hz.jpg 700hz.jpg 800hz.jpg 900hz.jpg 1000hz.jpg stepped.jpg
     

    Attached Files:

    • 20hz.jpg
      20hz.jpg
      File size:
      139.8 KB
      Views:
      12
    Last edited: Jul 29, 2017
  17. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    MID. XO will be 700Hz to 5000Hz
    300.jpg 400.jpg 500.jpg 600.jpg 700.jpg 800.jpg 900.jpg 1000.jpg 1500.jpg 2000.jpg 2500.jpg 3000.jpg 3500.jpg 4000.jpg 4500.jpg 5000.jpg 5500.jpg 6000.jpg 6500.jpg 7000.jpg
     
  18. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    Tweeter. XO will be 5000Hz
    3000.jpg 3500.jpg 4000.jpg 4500.jpg 5000.jpg 5500.jpg 6000.jpg 6500.jpg 7000.jpg 8000.jpg 9000.jpg 10000.jpg 11000.jpg 12000.jpg
     
  19. BE718

    BE718 Major Contributor forum experimenter

    Joined:
    Mar 1, 2016
    Messages:
    1,007
    Likes Received:
    165
    Hardly seems worth performing the additional driver linearisation in Acourate. Their basic response from close in measurements are generally +- 1dB within the ranges I will be using them. I think Mitchs advice in the book said +-2 dB.

    fr1.png fr2.png
     
    Last edited: Jul 29, 2017
    mitchco likes this.
  20. RayDunzl

    RayDunzl Major Contributor Central Scrutinizer

    Joined:
    Mar 9, 2016
    Messages:
    3,258
    Likes Received:
    459
    Location:
    Riverview, Fl
    That all looks pretty satisfactory to me...

    Good job.
     

Share This Page