Averaging Listening Positions when doing multi-channel equalization.

Discussion in 'Archived Threads 2001-2004' started by Larry Chanin, May 29, 2002.

  1. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    In anticipation of receiving my AudioControl Bijou multi-channel equalizer (hopefully tonight), I’ve started to measure the characteristics of my room. I’ve purchased Acoustisoft’s ETF5 software to use as my real-time acoustic analyzer, and last night I attempted my first low frequency response measurements. The equalizer instructions recommends taking individual RTA measurements for each speaker at about four different positions, and averaging them when setting the initial equalizer slider positions. Although I haven't tried it yet, I believe the ETF5 software will do file averaging automatically for me to generate a single composite frequency response curve for the general listening area. The positions that I chose in my first attempt are numbered 1 through 4 in the following diagram.
    Listening Positions
    Position #1 obviously isn't an actual seating position, but it was selected because at the time I thought it would be a central position that might represent a "good" average of the actual seating positions.
    But....
    Here's the Low Frequency Response of Position #1 for my Left Main speaker. (The curve was taken with a Radio Shack Analog Sound Pressure Level Meter using a "correction" file.)
    You'll notice the huge dip at 100Hz. This may be attributable to a room induced null at that position that is not encountered at the real seating positions. So my thinking is that it would probably be better to simply ignore this position when doing the averaging and use just three positions. Does this approach make sense?
    Any other tips, do's and don'ts, and general observations regarding ETF5, the Bijou or multi-channel room equalization would be appreciated.
    Please don't tell me that a rookie like me shouldn't be wading out into the deep water. :b I figured you folks would throw me a life preserver if I got too deep over my head. Right?
    Thanks.
    Larry
     
  2. Pete Mazz

    Pete Mazz Supporting Actor

    Joined:
    May 17, 2000
    Messages:
    761
    Likes Received:
    0
    Trophy Points:
    0
    Was the LF measurement done with only the sub/s, or were the mains on as well? It looks like it might be a cancellation or phase problem instead of a room problem.

    Pete
     
  3. Wayne A. Pflughaupt

    Moderator

    Joined:
    Aug 5, 1999
    Messages:
    6,185
    Likes Received:
    51
    Trophy Points:
    9,110
    Location:
    Katy, TX
    Real Name:
    Wayne
    Congratulations Larry, you’ve chosen a really nice equalizer. Audio Control is about as good as it gets for residential graphic equalizers, IMO.
    First thing first. Now that you have a good equalizer you can put the sub in the corner. That will give you the best pre-EQ performance, SPL and extension.
    Next, the second “surround sub” at its current location will not contribute anything useful to bass response or total bass output. Many respected professional evaluators have concluded this, and my own tests have as well, as have readings I’ve received from people I have helped to EQ and set up subs.
    For instance, adding a second sub in the same corner as the first will result in a 6dB increase in gain, while having the second sub elsewhere in the room will result in the same maximum SPL readings as only the single sub in the corner. In other words the second, remote sub contributes nothing to total low frequency SPL.
    In addition, the rear sub (on a wall 5-ft. from the corner) will have considerably less extension than the one on the front corner. The combined output will have the effect of “dragging down” overall system extension, as two subs (theoretically) operate at the same levels from say, 100Hz - 40Hz while only one “carries the weight” from 40Hz - 20Hz.
    Of course, the best reason to put both subs in the same location is that they can both be equalized by the Bijou’s excellent 1/6-octave sub EQ. Two subs at separate locations require different EQ curves; the Bijou doesn’t have the capability to EQ two subs independently.
    Third, I’ve recently found that many low frequency response problems that appear to be room-induced peaks and nulls can be eliminated with proper phase adjustment on the sub. I don’t readily understand all the “mechanicals” involved in this, but with the subs near the mains like you have, the phase should be adjusted as follows, according to the dB/octave slopes of your crossovers:
    • 6dB/octave (1st order) - set to 45deg.
    • 12dB/octave (2nd order) - set to 90 deg.
    • 18dB/octave (3rd order) - set to 135 deg.
    • 24dB/octave (4th order) - set to 180 deg.
    I don’t know how you would deal with staggered filters, but I’m guessing a point half-way between these recommendations.
    Fourth, the ETF averaging feature is nice, but it will be time consuming taking multiple readings for all six speakers. If I were doing it I would limit my readings and EQing to the primary position. Personally I would rather have “perfect” response at the main position and “good enough” response at the other positions than have “good enough” response everywhere. Does that make sense? Besides, in my experience EQing based on a position in the center of the room will result in excellent and acceptable sound at most other positions in the room, except those against a wall. And it looks like you don’t have any positions against a wall, so there you go!
    Fifth, don’t try to equalize for ruler-flat response. That will not sound good. What sounds best is a “house curve,” where response gradually rises at lower frequencies. Most of the house curve adjustment is needed with the sub. Do a search and you can find plenty of information on this Forum about house curves.
    I suggest first taking independent readings of each speaker from the primary position and focus on smoothing out response. The goal, especially with the mains, is not “perfect” response, but “improved” response. Although the Bijou has a ±6dB range, for the main channels typically you don’t want to use that much boost or cut. Cutting or boosting too severely will affect other, adjacent frequencies that might not need equalizing at all. Thus if the EFT shows a response problem of 6dB, for instance, it will sound audibly better with only a 3dB adjustment.
    After EQing the speakers separately take new readings with:
    • The left and right operating together.
    • The left, right and subs together.
    • The left and right rear speakers together (not sure how you’re going to EQ the center rear..?).
    You may find that running these combinations will show that minor EQ adjustments are needed. I wouldn’t’ worry about other combinations but these.
    A word of warning: While boosts or cuts more severe than 6dB are acceptable for the sub, that might mot be enough; sub response deviations can often be three times that much. Thus the ETF may show that you have low frequency problems that the Bijou can’t address.
    Regards,
    Wayne A. Pflughaupt
     
  4. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
     
  5. Mark Seaton

    Mark Seaton Supporting Actor

    Joined:
    Oct 10, 1999
    Messages:
    599
    Likes Received:
    0
    Trophy Points:
    110
    Real Name:
    Mark Seaton
     
  6. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    Hi Wayne:
    I must say I greatly appreciate the comprehensiveness, of not only this response, but of all of your responses to others. Before posting this “open-ended” question, I made a point of searching for many of your previous postings dealing with equalization, and I found them very helpful.
     
  7. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    Hi Mark:

    Thanks for your response. Wow, I'm sure getting my money's worth on this thread!

     
  8. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    Hi Mark:


    Quote:



    What I am probably interested in is to see the compared response at location #4 to see which problems are associated with the width of the room, along with a response at #2 or #3 for further comparison.





    Here's an overlay of the low frequency response of my Left Main speaker showing all four positions on one graph.
    I've switched to a 1/6 octave logarithmic frequency scale. Also on the same overlay is the frequency response of the LFE subwoofer (the magenta curve Sub1).

    The graphs abruptly drop off at 30 Hz. This is a function of the software. When plotting the greater detail of a 1/6 octave display, the program trades off drawing the lower frequencies. If I select a 1/3 octave display the program will plot the response curves down to 20 Hz, but some of the peaks and dips will no longer be visable with the same degree of detail. For example here's two response curves of the same subwoofer data plotted on the same overlay with 1/3 and 1/6 octave scales.

    In comparing the two overlay plots its interesting to note that at 35 Hz and higher my vintage Polk Audio SDA-1C speaker has higher low frequency responses for two of the four positions than the subwoofer. The subwoofer has better performance in the 20 - 30 Hz range.

    I believe this tells me that I have correctly set up my crossovers in my Lexicon MC-1 processor. I've set the main as small and crossed them over at 40 Hz. I've set the LFE subwoofer at 40 Hz.


    Quote:



    Finally, what equipment in terms of pre/pro/receiver/amp, speakers and subs do you have? This comes down mostly to determining the best use of the speakers' capabilities.





    Equipment:

    LEXICON MC-1 SURROUND PROCESSOR
    I recently purchased a Lexicon MC-1 surround processor, and frankly, although it sounds great to me, I'm not sure that I have set it up optimally.

    It has a ton of adjustments, but here are the ones I think might be relevant to setting-up the room equalization:

    Bass Management
    The Lexicon can adjust the cross over frequency for all 8 channels. The options are 40 Hz, 80 Hz, and 120 Hz. (I'm using 40 Hz for the mains and LFE subwoofer, 80 Hz for the center channel, with the side surrounds set to small and crossed at 40 Hz. (I described this at greater length in my previous posting.)

    Equalization controls
    Although recommended only for adjusting for source material rather than room conditions, it has an equalization menu that permits adjusting front speaker's Bass, Treble, and Tilt. The bass can be adjusted +- 6 dB below 250 Hz, the treble +- 6 dB above 1.5 kHz. The tilt can adjust tonal balance by shifting +- 3 dB centered on 1 Hz. The menu also turns on Loudness to boost low volume bass.

    A Subwoofer Peak Limiter control is available. I have mine set at the factory preset of +15dB. The ranges are +35 dB to -5 dB. There is a calibration routine that can be run to set the optimum peak level of protection for your equipment I have not used it yet.

    SPEAKERS

    Mains: Polk Audio SDA-1Cs
    These are vintage full range, floor standing speakers. They have a somewhat unique design in that they use opposing mid range driver arrays to cancel inter-aural crosstalk and enhance imaging. Because of this they are designed to point straight out without being toed-in. (Which is a good thing since they fit tightly in the cabinets and can't be toed-in anyway.) Stereo magazine measured their -5 dB frequency limits to be 20 Hz and 18kHz. I'm going to take an other set of measurements using a very low gate time and with the microphone a meter from the speaker. Hopefully this will permit me to duplicate a typical speaker response plot that ignores room effects. From that I should be able to approximate the -3 dB points on the speaker.

    LFE Subwoofer: Carver Knight Shadow
    This subwoofer is identical to Sunfire's smaller Architectural subwoofer (which they say is designed to be installed in cabinets) except it has an amplifier with an RMS wattage of 500 watts, instead of 1,200 watts. Although its only an 11 inch cube, Carver claims that its frequency response is 18 Hz to 100 Hz. They didn't specify the dB points, again this could be measured.

    Center Channel: Polk Audio CS400I
    Polk specs this speaker's -3 dB limit points at 50 Hz and 25kHz.

    Side Surrounds: Polk Audio f/x500I
    Polk specs this speaker's -3 dB limit points at 50 Hz and 25kHz.

    Surround Subwoofer: Miller & Kreisel VX-4
    The low frequency -3 dB limit is 30 Hz.

    AMPLIFIERS

    I have 3-Parasound amplifiers providing 7 channels of amplification. Each channel is rated at 200 watts @ 8 ohms and has a peak current capacity of 60 amperes.

    Thanks again for your very helpful comments.

    Larry

    PS I forgot to mention that I received delivery of the equalizer late tonight. However, I've got to wait for the silver interconnects that I've ordered. It's probably just as well, this will force me to spend more time analyzing my acoustic environment (and picking more pointers from you folks) before I go messing around with slider controls.
     
  9. Wayne A. Pflughaupt

    Moderator

    Joined:
    Aug 5, 1999
    Messages:
    6,185
    Likes Received:
    51
    Trophy Points:
    9,110
    Location:
    Katy, TX
    Real Name:
    Wayne
    If this is the corner sub, it should respond nicely to the Bijou. The peaks and dips correspond to its fixed 1/6-octave frequency centers and bandwidth. You’re very fortunate!
    Regards,
    Wayne A. Pflughaupt
     
  10. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    Hi Wayne:

    Thanks for taking the time to provide such a helpful response.

     
  11. Wayne A. Pflughaupt

    Moderator

    Joined:
    Aug 5, 1999
    Messages:
    6,185
    Likes Received:
    51
    Trophy Points:
    9,110
    Location:
    Katy, TX
    Real Name:
    Wayne
    Larry,
     
  12. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    Hi Wayne:
     
  13. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    As I continue to post the results of my foray into the ETF software, I sort of feel like a grade school student behind the wheel of a powerful dragster. Hopefully I won't crash and end up demonstrating how dangerous a rookie can be when armed with powerful machines (ETF5 & the AudioControl Bijou).
    If you'll notice on my Room Layout there are two sets of doors on the right wall (When facing the front of the theater). Not shown on the diagram is a utility room that runs the width of the home theater and is 6 foot deep. The rear wall of the home theater has two sets of folding doors that provide a 10 foot opening to the utility room. With this information as background I'd like to describe an experiment I conducted.
    I orginally ran frequency response curves with all the doors closed with just my Main Right speaker exciting the room. For this experiment I ran additional frequency response curves with 1)all the doors open and 2)with all home theater doors closed, but with the utility room door open. All the measurements were taken at position #1 as shown on the layout diagram. The results seem to show that open doors generally improved the frequency response. However, the best results were obtained with the home theater doors closed, but the theater open in the rear to the utility room. Perhaps the 12'x6' utility room is serving as a large bass trap. (I haven't rigged any acoustic foam in the doorway of the utility room like you would normally do to create a bass trap.)
    Here's an overlay showing the Low Frequency Response (20 Hz - 200 Hz)for all three situations. The blue curve on the bottom is with all doors closed. The green curve in the middle is with all doors opened. The aqua curve on the top is with the theater doors closed and the utility room doors open.
    Here's an overlay of the same data but displaying the entire frequency range (20 Hz - 20,000 Hz).
    Assuming I haven't messed up somewhere, I guess I'll start to do my serious listening with the theater doors closed and the utility room doors open. Who knows, maybe I'll get some acoustic foam and mount it in the opening of the utility room to see if I can enhance the effect.
    Larry
     
  14. Wayne A. Pflughaupt

    Moderator

    Joined:
    Aug 5, 1999
    Messages:
    6,185
    Likes Received:
    51
    Trophy Points:
    9,110
    Location:
    Katy, TX
    Real Name:
    Wayne
     
  15. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    Hi Wayne:
     
  16. Mark Seaton

    Mark Seaton Supporting Actor

    Joined:
    Oct 10, 1999
    Messages:
    599
    Likes Received:
    0
    Trophy Points:
    110
    Real Name:
    Mark Seaton
    Larry,

    I'll respond to some of the other comments later, but could you trace the graph on your computer and find out for sure what frequency that peak is at? What is the ceiling height?

    Also, in your comparison of the 3 plots with the doors, could you add a 4th graph to the comparison next time you measure showing the response with the room's doors open and the utility room's doors closed? What are the utility room doors made of or what style are they?

    As a general comment about the measurements (note I haven't really used ETC much) you will want to be looking at a shorter time window for higher frequency ranges. A good starting point would be to look at 300Hz and up where the direct energy is prioritized.
     
  17. Wayne A. Pflughaupt

    Moderator

    Joined:
    Aug 5, 1999
    Messages:
    6,185
    Likes Received:
    51
    Trophy Points:
    9,110
    Location:
    Katy, TX
    Real Name:
    Wayne
     
  18. Greg Monfort

    Greg Monfort Supporting Actor

    Joined:
    May 30, 2000
    Messages:
    884
    Likes Received:
    0
    Trophy Points:
    0
    WRT the 100Hz null, what are the room dims?
    ====
    >So my thinking is that it would probably be better to simply ignore this position when doing the averaging and use just three positions. Does this approach make sense?
    ====
    All measurements should be taken from the actual seating positions and at an average ear height IMO.
    ====
    >Please don't tell me that a rookie like me shouldn't be wading out into the deep water. I figured you folks would throw me a life preserver if I got too deep over my head. Right?
    ====
    As God asked Noah when he wanted to abandon the building of the ark, "How long can you tread water?" [​IMG]
    ====
    >Next, the second “surround sub” at its current location will not contribute anything useful to bass response or total bass output.
    ====
    It's not meant to and should be BW limited to >20Hz. It's for when the surrounds can't handle the required dynamic range. It's not wired to the LFE ckt, or shouldn't be anyway.
    ====
    >Fifth, don’t try to equalize for ruler-flat response. That will not sound good. What sounds best is a “house curve,” where response gradually rises at lower frequencies. Most of the house curve adjustment is needed with the sub.
    ====
    Well, it's a lofty goal, and I don't agree that it doesn't sound good, but +/-3dB 20-20kHz is 'good enough'. [​IMG] WRT the 'house curve', for proper music soundtrack playback, any boost for special effects needs to be limited to Also, two subwoofers playing at the same level from separate locations will be measured as 3dB more than just one, not the same.
    ====
    Only in the BW where they're within ~1/3WL of each other or if they sum at the listening position.
    ====
    >An RTA displays averaged energy per octave, which includes all the reflections in the room. Ideally the direct response is mostly flat, where an RTA should better indicate the power response in the room (total energy). With most speakers and rooms, a flat RTA display will sound bright. Traditionally people have EQ'd for approximately a 3dB/octave downward slope as frequency rises when viewed on an RTA.
    ====
    Oh?! Since it's measuring total energy using pink noise (whose energy falls at 3dB/octave), how can it be 'bright'? Now if you should use white noise with an RTA, which has equal energy/octave, then EQing at "....3dB/octave downward slope as frequency rises".... makes sense.
    For sure no one's ever called any system I've EQ'd with an RTA 'bright', though I admit that doesn't mean I'm right. [​IMG]
    ====
    >I have a spectrum analyzer connected to my system to give a response display of real time program content. From what I see, the bass in action flicks is broadband, similar to pink noise. Thus it has everything from higher frequency bass to and often below 25Hz.
    ====
    Exactly! The lower the frequency, the wider its BW, ergo DC has infinite BW.
    ====
    >The slope on my processor’s crossovers are all 12 dB/octave, except the subwoofer which is 24 dB/octave. What would your guess be for this combination?
    ====
    Are you saying the mains are high passed at 12dB and the sub is low passed at 24dB? If so then in theory you'd calculate the midpoint within the bandpass, but our hearing acuity isn't too good down low WRT phase so averaging them is 'close enough' IMO, or 18dB/octave - 135deg..
    ====
    >For instance, one set of readings at a placement of 3ft. from a corner should have produced a compounded 6dB null at 113dB (because of two walls being 3ft. away). If it did, it must have been precisely centered and very narrow; a reading at 100Hz showed no hint of cancellation.
    ====
    What formula did you use? Also, the driver must be either parallel or perpendicular to the boundary. I get ~94Hz.
    ====
    >Even more surprising, for another test I put the sub centered on a 16ft. wall, which put it equidistant from three boundaries – left and right walls, and ceiling (8ft. x3). According to the calculations, there should have been a 9dB null at 42Hz,
    ====
    ~35.3Hz.
    ====
    >What I have a problem with is that neither AudioControl or Lucasfilm provide any guidance whatsoever regarding setting the side surrounds even though the THX equalizer has a set of sliders for each surround channel.
    ====
    ???? EQ them ~flat.
    ====
    >With all due respect to Mr. Lucas, this seems silly. You certainly can’t EQ for both scenarios – at least not the sub. I think anyone who has watched more than one movie will agree that overwhelmingly when you have bass action happening, you also have the mains active with it. How many scenes have you seen that where with only the center channel and sub were in operation (excluding mono movies)? You can probably count them on one hand – if any.
    ====
    If both the CC and mains are set to small, then ideally you need to average them out as at least some of the sub's BW will be directional unless you're sitting close.
    ====
    >An other question about equalizing side surrounds relates to dipole speakers. This concern was voiced on an other forum in which the poster, who is also beginning to use a Bijou, questions whether there are any specific issues related to equalizing a speaker whose design relies mainly on reflected sound, not direct sound.
    ====
    Just EQ them from the listening position. How the signal gets there from a dipole isn't an issue.
    ====
    >If I find an uncorrectable problem with my current physical layout I might try relocating the subwoofer behind the entertainment center and take measurements in various positions, left corner, mid-point, etc. However, I consider this a drastic action, and I would welcome opinions on the necessity of this type of approach.
    ====
    If the FOH speakers are all set to large then you may get by with it back there, but if any are set to small then it's not a good plan as any WLs smaller than the distance to listening position will be directional. Let's say you're 12ft from the sub and the room is >12ft wide, then in theory ~60Hz has enough directivity to be localized. I know I can in my room.
    ====
    >I'm going to be able confirm the fact that the Main's individual responses will be about the same as the surround subwoofer, which of course means I really have four subwoofers in the room, not two
    ====
    Yep, with mass quantities of comb filtering and room interactions that will make a high res response look like Chop Suey. Fortunately, our hearing acuity down low will tend to sum it all since we are predisposed by nature to cue in on the loudest sounds, and why nulls must be very low Q to be noticed.
    ====
    > (or one subwoofer and three "woofers" depending on how critical your definition of a "subwoofer" is ).
    ====
    Technically, a sub is for infrasonic BW (
     
  19. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    Hi Mark:

     
  20. Larry Chanin

    Larry Chanin Stunt Coordinator

    Joined:
    Jul 24, 2001
    Messages:
    218
    Likes Received:
    0
    Trophy Points:
    0
    Hi Wayne:

     

Share This Page