Time Phase alignment

Discussion in 'Official REW (Room EQ Wizard) Support Forum' started by graham, Feb 6, 2018.

  1. graham

    graham New Member
    Thread Starter

    Joined:
    Dec 21, 2017
    Messages:
    4
    Likes Received:
    2
    Hi all,
    I occasionally supply sound at free outdoor parties.
    Normally i run 2 mono-ed/ clustered 15" subs with 2 12" stereo tops at around 1.6 to 2K RMS (driverack 260 or ultradrive speaker management)
    Crossover between the subs and "high" i normally set to around 100Hz and in the past i have just phase aligned at the crossover point for maximum output (100hz sinewave/ SPL meter)

    Question, should i be phase AND time aligning? and in what order? Is there an easy and quick way to do it/ them in REW?

    Tonight i experimented with time delays and got no real positive results- first i muted the subs and done a 10- 20000hz sweep noting the reference time signal (respect to loopback)
    I then ran the same sweep across the subs (with the tops muted)
    I then split the time difference and added this delay to the tops. For a memory the tops were about 15ms to reference and the subs were around 50ms to reference so i delayed the tops by 35ms which i would have thought would be correct.. I then looked at the phase graphs to find that none of the graphs lined up- i thought they would!
    Am i sweeping too much range? am i getting room issues (experimenting in my lounge)?

    Otherwise loving REW for EQing!
    I run multiple sweeps from different areas of the "dancefloor" then combine the results and set EQs acordingly, if i could do phasing and time alignment i would be even happier.

    Another quick question, i wonder if the sine sweeps could be optional "jungle noise" or something a little nicer (i'm doubting it is possible)- by the time i'm setting up sound at parties there is usually a few there and the sine sweeps and pink noise isnt doing much for the atmosphere :)


    Looking forward to some replies

    Thanks for a great program and will very likely be donating in the next few days to help the cause :)

    cheers, Graham
     
    Joerg Schulte likes this.
  2. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    I saw lot of Videos runnung Smaart V8. I also would try to do time-align the Subs.....- so still waiting for help, I think it would be helpfull to see SPL of Pink Noise in Realtime. :-)
     
    graham likes this.
  3. MAD

    MAD Member

    Joined:
    Jan 14, 2018
    Messages:
    17
    Likes Received:
    3
    You can use acoustic reference in REW to determine the delay between two systems. Set it to use the top as acoustic reference (one channel) and measure both with one sweep (two channel), the delay will be shown then in the info tab of the sub measurement. After setting the delay in your DSP, measure again and double check the impulse response to see if the slope (not the maximum) of both IRs is aligned now, if not,add the delay value shown in "estimate IR delay" for the sub to your delay in the DSP and measure again. Now the impulse responses should be aligned and your systems too, meaning the phase alignment shall now be the same as originally (both speakers at the same distance and no delay).

    And yes, when measuring in a room you need to set a measurement window to remove reflections from the measurement based on the distance of the Mic to the speakers and to walls/ceiling/floor (whatever is nearest). But this also means that the measurement data below a specific frequency is useless. I'm using this calculator (German) http://www.jobst-audio.de/tools-akustik/messabstand
    I'm sure there are some in English on the web too...
     
    graham and Joerg Schulte like this.
  4. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    411
    Likes Received:
    60
    REW doesn't have a 2-channel FFT mode, but it is somewhere on my lengthy todo list :)

    Joerg, you can use the RTA window if you would like to see real time SPL of pink noise. Use Pink PN rather than the random Pink Noise, means you don't have to do any averaging. These settings are suitable for viewing pink PN on the RTA:

    rtasettings.jpg
     
    graham likes this.
  5. graham

    graham New Member
    Thread Starter

    Joined:
    Dec 21, 2017
    Messages:
    4
    Likes Received:
    2
    Hi John,
    Thanks heaps for this great software, i appreciate the huge amount of work that has gone into this free software..
    I sent through a donation last night. (i hope others are)

    So are you saying that without dual channel FFT i'm not really able to do phase and time alignment like i could maybe do on smartV8 ect (if i was a rich man :( ) ??

    And thanks for pointing out the "real time analyzer" option, very handy for phase alignments using my "old fashioned" way.. (saves me running back and forths with an SPL meter :) ) Will it also work with say a 100hz tone? I cant try tonight with visitors.
     
  6. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    Schönes Programm. Danke für den Hinweis...
     
  7. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    411
    Likes Received:
    60
    No, not saying that, dual channel FFT is just what allows a system to use an arbitrary test signal like music instead of predefined test signals like sweeps and pink noise. There is little in the way of acoustic measurement that can't be done with REW, although I always have a long list of things to add and improve :)
     
    Joerg Schulte likes this.
  8. graham

    graham New Member
    Thread Starter

    Joined:
    Dec 21, 2017
    Messages:
    4
    Likes Received:
    2
    Ok, certainly a very nice piece of software, especially like being able to do quick single sweeps (which i dont think can be done as well on other packages)
    Thankyou also to Mad and his input to this post- i will try that method (when i have a chance)


    (being an electronics person myself i wondered if the sweeps could be broken up into bursts that sounded musical (i dont know anything about FFT processing mind-you :) ))
    probably a whole heap of work to do as well- understand..

    I remember when i was running my own electronics "business" the design documentation was a real time killer- hats off to John :)
     
  9. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    I tried to check it in my Home-Office. Left Speaker has 1,29 ms(0,43M) to acoustic refercence(i use X32 Desk from Behringer. One channel Output and One Channel to measure Acoustic reference) The right Speaker has -7,31 ms(-2,508 M) - The Listening Point ist not triangular for purpose. Now the Question: Delay on Both Channel or only to the left one(closer to Mic)?

    2nd Question: For Sub-Mearement set the Mic @Listening Position in the Venue or @Ground? What are your thoughts on this?

    Unfortunately the Files are to big for Upload.... :-(
     
  10. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    411
    Likes Received:
    60
    Not sure I follow what the setup is Joerg, do you have a 3rd speaker acting as the acoustic reference? From your data the left is arriving later than the reference and the right is arriving earlier.

    For sub measurements would measure at the listening position. Can be useful to also make a nearfield measurement of the sub for reference, to act as a target if you are applying any EQ in the sub path.
     
  11. graham

    graham New Member
    Thread Starter

    Joined:
    Dec 21, 2017
    Messages:
    4
    Likes Received:
    2
    I had another play tonight, starting to get results.

    Tried using acoustic reference and didnt really get an desired result- plus thinking that i'm asking for trouble if i start playing around with input routing setting whilst setting up at parties (too many distractions)

    I did have some success with using loop-back though, i narrowed the sweep to cover the cross over region and a bit more (50 to 150HZ) for a 100Hz crossover. Much better results :)
    I'm sure ive read about this over other posts- REW (and other programs) will have trouble getting good time references in the lower frequencies, make the data similar and the result look to be much better.

    I am finding with this and other programs, that there is lots of information- problem is that it can be difficult for us "new but maybe more advanced " to find the appropriate information without going through many hours of non-related (and maybe wrong) info on the forums.. John has put amazing information together but it is more documentation related than it is a "how to do something easily manual"

    REW has many "beginner" type u-tube tutorials but not much for us slightly more advanced, maybe i can put something together when ive worked it out myself! :) .. im the meantime, please dont be shy (others)
     
    Joerg Schulte likes this.
  12. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    Still not happy with Setup:
    My gear: Soundcard is Mixing Desk Behringer X32 with EQs built in....Aux5 Left Speaker Aux6 right Speaker.
    ECM 8000 on Channel 8
    Ouput: Channel 30.
    Loopback Mixbus Output1 to Channel 7.
    Tried loopback reference and acoustic reference.

    For trainig purpose i put the Mic not triagular to Speaker. The left speaker is closer than the right one to the Mic.

    I guess 6 ms Delay, put Delay on left speker and boom, got phase align. ----but nowhere to read..... on second Screenshot you`ll see nearly same delay.......
    upload_2018-2-9_14-50-36.png

    upload_2018-2-9_15-0-56.png


    upload_2018-2-9_14-18-29.png
     
  13. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    411
    Likes Received:
    60
    If you bring up the Info window you can see the system delay and the combined amount of any IR shifts applied for the currently selected measurement.
     
  14. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    I think it's still going wrong. Delay times are similar despite delay in the channel. I tried to measure Left and Right Channel separate. I used acoustic timing reference and also loopback. Where is the mistake? See pictures in Post #12....
     
  15. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    411
    Likes Received:
    60
    Depends what happens to your output (out 30) on its way to drive your speaker. With a loopback connection REW drives the output (30) and the loopback (31) with the same signal. Are you using your mixer to route Out 30 to either left or right independently? Where are the delays being applied?
     
  16. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    Hi John,
    yes I´m routing the Out first to LR, than via Matrix1(Left) and Matrix2(right). On Matrix1 i put the Delay, because the speaker is more close to Mic.....- For Subs i use Matrix3 and 4.
    This Scenario on Post #12 is for Training Porpose.....-
    Can I expect Latency while i am using timing reference (route Channel 29 to Bus1 and back to Channel 7)? Or is acoustic reference the only way with that small sweep.? Where is the difference?

    In Smaart they use on Measurement fix f.E. for the left one, and do the right one in Realtime with RTA.

    Thanks for feedback. I am a little bit confused about the results.....

    Greets from Germany.....Joerg
     
    #16 Joerg Schulte, Feb 13, 2018
    Last edited: Feb 13, 2018
  17. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    411
    Likes Received:
    60
    The acoustic timing reference sends a high frequency sweep to the reference channel and uses the peak of its impulse response as a reference for the measurement impulse response. The measurement sweep is emitted a known time after the acoustic timing sweep, REW takes that into account. The timing reference always needs to go to the same speaker.

    Loopback timing reference sends the measurement signal to both the measurement and loopback channels at the same time and uses the peak of the loopback channel impulse response as the timing reference. With a full range sweep both should give much the same results, with a limited bandwidth sweep the loopback method will behave differently, since the reduced range sweep will usually have a broader, delayed peak.
     
  18. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    Not yet solved. I still get similar results. I use acoustic Reference. I think the problem is, that i use same Speaker for timing sweep. I take Channel 30 for left Speaker and 31 panned to the left for Reference. Channel 32 for right Speaker(more away from mic). Calculate the delay after 2 Measurements and set Delay on Output for Channel 30 left Speaker. So the Delay for timing sweep goes delayed too?? Maybe Mistake?
    View attachment 6626
     

    Attached Files:

  19. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    411
    Likes Received:
    60
    Could be, depends on exactly where the delay sits in the path and so whether the right is being compared to a left timing reference with delay or without. If you are using the acoustic reference can just route one output to the left channel and another to the right. Set the one used for left as the timing reference then make one measurement with the measure output as the left and another with the measure output as the right, without making any change to the timing reference or panning anything back and forth.
     
  20. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    We are getting closer to the problem. See 1st and 2nd Measurement. There we can find the Delay. I think my Problem in Understanding the next step. Put the Delay on the left Speaker, because he is closer to the Mic. After i chanched the Speaker from left to right for Timing reference(CH31) i got same Results as in Post #18. The Problem is the Reading Measurement after Delay on Output for the left Speaker. It is not possible for me to align the curve to the right Speake....maybe something in the routing....
    upload_2018-2-15_11-25-53.png
     
  21. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    Any Thoughts how to get the left Speaker aligned to the right Spaeker(purple Curve #3 above)? The Measurement #4 looks like #1in face of Delay-Settings....
     
    #21 Joerg Schulte, Feb 21, 2018 at 4:28 AM
    Last edited: Feb 21, 2018 at 4:34 AM
  22. John Mulcahy

    John Mulcahy REW Author

    Joined:
    Apr 3, 2017
    Messages:
    411
    Likes Received:
    60
    If you mean how to see the effect of a delay, you can use the controls on the IR graph to offset the IR start time and see the effect that has. How to apply it within your configuration and get a good measurement of the result I'm not sure, I still don't really understand the routing that happens to the signals.
     
  23. Joerg Schulte

    Joined:
    Dec 7, 2017
    Messages:
    10
    Likes Received:
    1
    I use a digital Mixing Console(Behringer X32). Behringer ECM8000 in Channel 8 as Input. Channel 30 as Output for Pink Noise R(Channel 32 for Pink Noise Left). Out 5 for left Speaker and Out 6 for right Speaker. The 2 outs to an Amp L/R. The Channel 30 is also for Acoustig timing referece, because more far to the Microphone. After i noticed Delay on CH32, I put Delay on Out 5 to align both Speakers to the Listening Position. For Purpose LP is not triangulated.
     

Share This Page