Speaker listening position FR targets

Discussion in 'Speakers' started by Serious, Aug 28, 2020.

  1. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Tonight might be the wrong time for this post, however I had it planned for tonight and don't really have the time to do it at another point in time. Plus I kinda want a thread to discuss speaker targets.

    When I set up my the coaxials on my desk and had to resort to EQ anyway, I was free to choose any speaker target I'd like. So which one did I go for? This is what I went for after trying out a couple of them, EQing by ear after I had it EQd flat with measurements:
    32Sat Desk EQ5 vs Serious LP speaker target.jpg
    That's with each channel measured individually, not both driven at the same time. Reading online it seems many people are using a warmer target with somewhat of a downward slope. The 1dB/octave slope seems popular, which is quite similar to the B&K 1974 target. Harman seems to use a bass shelf, then a straight line from 300Hz-2kHz and somewhat of a downward slope past 2kHz. That type of curve (with or without the bass shelf) seems to be somewhat popular aswell. Looking at stereophile's measurements however it appears that many highly-rated and neutral sounding speakers measure considerably flatter than that when spatially averaged around the listening position.

    In the original B&K 1974 article (https://www.bksv.com/media/doc/17-197.pdf) however there's an important detail:
    B&K 1974 target measurement method.jpg
    The measurement method upon which the target was based on is inherently going to result in a warmer frequency response as both channels are driven at the same time. Depending on the distance between the two speakers the effect is going to be more or less pronounced.

    I'm not really a fan of measuring both channels at the same time with a mono source as that is going to mask channel mismatch issues (which hurt the imaging) and IMO it is also much harder to do consistently. Move the mic 2mm from the center between both speakers and your upper octave is going to take a dump due to comb-filtering. Move it even more and the comb-filtering becomes much worse.

    So how much warmer is the response going to look? Turns out flat at the listening position may result in a somewhat B&K-type of frequency response with a typical speaker setup. Compare these two curves:

    First up, the left and right channel of my main speaker rig measured individually:
    VoxOB at LP vs Serious LP speaker target.jpg

    And here's the response with both channels driven at the same time. Black is without B&K target compensation applied:
    VoxOB_at_LP_stereo_averaged_response_with_and_without_BK_target.jpg

    It turns out a response that's flat at the listening position inherently results in something like a B&K curve when measured as intended. The difference really isn't that big, I have attached a plot between an averaged mono response and the B&K-compensated stereo response. The main difference seems to be in the 3kHz region here, which I have some ideas as to why that is, but I'm not yet sure*.

    It's possible I understood it wrong, but it seems to me people are applying a B&K-type curve when measuring each individual channel, which will result in a considerably warmer sounding response than intended. Of course it's a matter of preference. For a year or so I listened to such a response every day, but over time I grew tired of it. Flat at the listening position still sounds somewhat bright to me, which is why I lower the response past 10kHz. In fact the coaxials (which measure close to linear on axis) sounded quite a bit too bright in this use case and I really had to lower the treble quite a bit with EQ.

    Also I think there are good arguments for the popular response targets:
    1. A BBC dip may be beneficial under certain circumstances for two reasons:
      1. A room with too much reverb in the upper midrange (which is somewhat typical of rooms with many bare walls) somehow sounds more natural to me when there's a slight dip in the 2-4kHz range.
      2. In my opinion the phase shift associated with many M-T crossovers places some emphasis on this region and tuning it slightly low often sounds more natural with better integration. Speakers with less phase shift in this region sound more neutral without the dip to me. However many multi-ways have their crossover in the 2-3kHz region. My target would look somewhat warmer if I didn't correct for the excess phase with these speakers.
    2. Similarly to 1) a room with too much reverb will sound emphasized in the upper mid/lower treble region and as such turning down the whole upper mid/lower treble range with a 1dB/octave target may be a good idea.
    3. Somewhat counterintuitively I also prefer smaller speakers to have some bass emphasis. However don't overdo it. This is because with larger speakers we tend to crank them up more due to lower distortion and we can get away with a leaner response due to equal-loudness contours. Or at least that's the case for me.
    Anyway, I'm interested to hear some other thoughts on this topic. I'm sure many of you will prefer a somewhat warmer response than I do. Let me know! It's very possible I'll change my preference in the future, anyway.


    *Note the measurements have to be done at slightly different positions as I normally measure the left speaker where my left ear is and vice versa. With the stereo measurement the mic is where the imaginary lines between speakers and ears would meet behind my head.
    **Correction: Unlike what the legend on the attached picture says, the averaged mono response does not have my speaker target applied as a compensation IIRC.
     

    Attached Files:

    Last edited: Aug 28, 2020
  2. nishan99

    nishan99 Friend

    Pyrate
    Joined:
    Dec 1, 2019
    Likes Received:
    1,617
    Trophy Points:
    93
    Location:
    Layla
    Also don't forget about the dispersion patterns and the off-axis response and whether the speakers are designed to be toed in or not.

    FR of speakers is a really alien thing to me, it's hard to trianglate sometimes so I am very excited to read the discussion!.
     
  3. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I’ve always aligned to B&K target with both speakers playing. Yeah, need to be careful of the comb-filtering effect.
     
  4. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    4,276
    Trophy Points:
    113
    My preference usually changes depending on SPL. For low level listening I like the B&K sloping curve, but for louder flat is my preference. Keep in mind that this isn't for a single point measurement and is more for a power response.
     
  5. DigitalMaven

    DigitalMaven Acquaintance

    Joined:
    Nov 13, 2016
    Likes Received:
    25
    Trophy Points:
    18
    Location:
    tucson
    Maybe diffraction from the speaker baffle requires a BBC dip? Also, I believe that because human hearing is really sensitive in the transition from Tweeter to Midwoofer the combined crossover region plus increased baffle boost requires a slight bbc dip not to sound bright and not to have weird interactions between the tweeter and woofer during transition.
     
  6. murphythecat

    murphythecat GRU-powered uniformed trumpkin

    Pyrate Banned
    Joined:
    Apr 4, 2016
    Likes Received:
    1,201
    Trophy Points:
    93
    hmmm, me its almost the opposite. when i listen to low spl, i boost the bass and treble on my ifi retro, when i listen loud i prefer B&K.
     
  7. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    This is just but one tool, one method.

    A flat FR on-axis in an open space, anechoic chamber, whatever will often yield a gentle-downward slope at the listening position in a decently treated room (that is a room with furniture, bookshelves, window treatments, rugs, etc.)

    BBC dip is for classical music, overly reverbant rooms, or off-axis xover peaks around that area.
     
  8. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    2024 UPDATE:

    It seems the -1dB/oct target has become quite popular or sometimes rather 0.6 to 0.8dB per octave. I find such a target somewhat warm sounding and I feel it sounds more realistic leaving the midrange and bass flat. However I gotta admit that the above target was bright sounding.

    This is mostly after listening to my cardioid monitors, Hathor a lot, one of the reasons which is that the OBs, Amon sound somewhat forward, somewhat unforgiving even.

    Amon.jpg

    Per this new target the lower treble to mid-treble region is too emphasized, which I think is more accurate to how I hear it. Also the whole response is still kind of jagged.

    Hathor XO11.jpg

    Hathor is much less jagged, but I do think the 4-10kHz region could use a bit of a boost at this angle. I will soon change the crossover a bit to decrease the upper treble a bit and allow for closer to on-axis positioning.
    If possible I want to find a speaker position with less of a 30Hz dip, since these speakers really do go low.


    My dad's speakers are now biamped and use a DSP to EQ the woofer respone to fight the room modes in the living room. Here's what that looks like:
    Canton Reference 3K Bass EQ2 20-20k 24th.jpg

    Subjectively there's a bit of a 2-3kHz dip and then there's that treble emphasis. This is with the tweeter set to +1.5dB. It also measures somewhat bassier than I'd like, but really the 400W Class D amp used here sounds kind of lean, especially with the DSP so it works out. Also the Aegir softens the treble somewhat. It's past the sibilance region, too.

    The bad news is lowering the tweeter level does two things: Lowers the region where the peak is located, but it also lowers the upper midrange quite a bit. It doesn't do as much for the 3-6kHz region, IIRC. So you can't have less treble without less of an upper midrange dip.

    The Canton Reference 2K is more forward in the upper midrange and less bassy sounding, generally more honest. The newer Reference series with the black drivers is much smoother sounding, to the point of glossing over information. At least it did so the two times I heard the Reference GS, so far.
     
  9. Poleepkwa

    Poleepkwa Friend

    Pyrate BWC
    Joined:
    Oct 1, 2015
    Likes Received:
    1,557
    Trophy Points:
    93
    Location:
    Finland
    @Serious, interesting looking "house" curve you have now. Looking at it, I wonder if it seems a lot, but it seems every speaker and room is different. My experience with my speakers has been that cd/horns usually start to sound dull if I roll it of that much. They seem to prefer a much flatter tilted towards the treble. My little coaxials on the other sounds great with similar curve to yours.
     
    • Agreed, ditto, +1 Agreed, ditto, +1 x 1
    • List
  10. Ardacer

    Ardacer Friend

    Pyrate Contributor
    Joined:
    Oct 15, 2017
    Likes Received:
    986
    Trophy Points:
    93
    Location:
    Croatia
    I always make my speakers measure flat, and in normal rooms always get the response from your first post, one purrin mentions, with rolled off high end (brilliance) at listening positions. Eq-ing that never seems to help. I do add 3-6db more to bass, as a personal preference, and like extension down to 16hz. Those low tones really give you a sense of existential dread.

    Regarding combing effects - I've given this some thought, and have come to a conclusion (possibly wrong, of course).
    When you cross 2 drivers at a given frequency, especially if it is a steep slope, the more you move away from the center -6db(assuming lr type), the more the single driver takes over so combing isn't as much of an issue. The problematic zone is the center, crossover frequency itself. And really, if you put one driver over the other, in the plane between them, there can never be any combing (not taking reflections into account - just the crossover). The wave path difference from one driver to the ear vs the other driver to the ear will be pretty much the same, allowing only for constructive interference (blending seamlessly - of course, I'm not taking into accound stupid shit like putting the mid into the floor and tweeter into the ceiling). Vertical planes are where you can expect problems.

    Now - if the path difference is equal to pi, or half of the wavelength, that will be the first true cancelling, the first null. The biggest path difference is almost always at the 90 degrees, so direct vertical up and down. So if the difference between acoustic centers is exactly half of the wavelength of the crossover frequency, first null will be at 90 degrees. If it's less, you can make it so that the null never fully happens im the front 2pi space.

    So in general, tldr,
    Put tweeter directly above mid, and as close as possible to it, and cross it as low as possible, if you want to minimize timing error effects.
    2 sources coming from within the length of 1/8th of wavelength behave de facto as single point source.

    Acoustically treating the room in any way (literally putting a carpet on the floor helps a lot) helps keeping the main original incoming sound wave a lot stronger than reflections, helping with clarity of music and speech. But if you can talk normally in your room, it should be fine for normal voice reproduction. For orchestral music, yes, it'd be best to sit in the studio with nearfields (or use headphones)... but you sacrifice some clarity for the common sense and waf and practicality.
     
    Last edited: Feb 20, 2024
  11. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Flat is relative. I think it's best to go for a power response with a downward slope to it, while going for a smooth DI. On-axis might look worse, but in room it's going to be more linear.*

    Two things:
    1. What you are looking at in your example is a constant frequency, the crossover frequency. However the driver blending happens over a much larger range. Even with the 1.5kHz 2nd order XO between my 4" mid and 1" tweeter there are multiple nulls from comb filtering in the vertical plane, despite time aligned acoustic centres. Not even with coaxials you can get rid of that issue entirely (the acoustic center is not the same for every frequency), what you need are small widebanders.
    2. Minimizing timing error effects takes a first order crossover with both drivers connected in the same acoustic polarity, which is very hard to do and quite the opposite goal from minimizing comb filtering. As can be seen with the Hathor 3 way experiments.
    Comb filtering between M and T is a given with multi ways. It just seems the most logical compromise. This is why normally you see vertical off axis measurements maybe to ±15°, never the full frontal 90°.

    There's a certain brand which prouds itself on their 54dB/octave crossovers. Some may like that sound and it goes a long way towards achieving uniform vertical off axis response. However I despise their sound.


    I also realized those graphs above were overly pessimistic with regard to the floor bounce effects. A spatial average around the listening position while actually sitting there looked much better for both speakers.

    EDIT:
    *The simple answer would be to just go for a response that matches the desired in-room response at the listening position. Or to not be too focused on FR linearity in the first place.
     
    Last edited: Feb 20, 2024
  12. Ardacer

    Ardacer Friend

    Pyrate Contributor
    Joined:
    Oct 15, 2017
    Likes Received:
    986
    Trophy Points:
    93
    Location:
    Croatia
    Ah, we probably think about different things when we say timing error. I was not precise enough (my typical error, apologies).
    You probably mean how the crossover mangles up phase, and indeed 1st order is best at that respect. I meant acoustic path difference. Space and time is really interchangeable there since with constant velocity one is the other.

    You care about the phase which would explain your viewpoint, I don't really bother with phase much at all. That's why I said, steep crossovers, small acoustic centers distance, smart crossover freq. choice = almost no combing. But yes, a lot of mangled phase. My reasoning is that, since inner ear is basically a constant spectrum analyzer, the audible effects of phase (other than things like out of phase bass cancelling and obvious stuff like that) are difficult for me to understand. I tried to blind test the effects of phase, I personally can't hear it. Doesn't mean it's not there, just that it doesn't bother me and I also don't see how it could actually be heard - I'm not saying it's impossible, just that I have no idea. If you have an explanation for that, I'd be happy to know (not sarcastic, really).

    But if you want to use low orders like 2nd, 1st, cause they sound better to you, I don't really see much of a way out, other than optimizing seating positions, speaker position and trying to deal with reflections if you care about that. It's inevitable. Especially with big woofers crossing directly to tweeters.

    With regards to linearity - yes, look, if it's within 3db it's never going to be a problem. Hell, within 6 db is possibly fine too if it isn't too bizzare. I don't bother with it much, and it's relatively easy to manipulate so too each his own, do whatever. One possible problem is phase change induced with filters. I forgot the difference between iir and fir, something might be possible in such regard with fir perhaps, I just don't remember. Or care much.
    When you say flat is relative - you are correct. I always make speakers to be flat in anechoic 1m test conditions, stick them in the room and forget about linearity at the sitting position (other than bass, ok, or if something is really weird).

    Bit of offtopic, sorry - we have quite a different approach to audio, or speakers, but that's what I like actually. I really enjoyed browsing through your posts and took a look at your webpage. I had pure silver cables at one point made by my friend, lots of twisted silver strands in a transparent tubing. Glorious. I might make some for the kef build, just to look at them. Too bad you put your silver cables in white protective cloth, you can't see the silver, but at least you know it's there.
     
    Last edited: Feb 20, 2024
  13. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Ehh, 3dB in the wrong place can be quite bad :p
    Another way to put it would be to optimize the speaker for a linear response at 15-20° off-axis, where diffraction effects start to even out. So you may sometimes need a dip in the upper midrange for a normal 6+1 two way, for example.
    But really, it depends. I think the only proper way is to measure the whole 360° horizontally and vertically for every driver and then simulate and tweak from there. At least that's what I did and I think it worked.

    Well, the path difference and a phase difference both have the same result: One signal is delayed compared to another, which is why I didn't differentiate between them.
    In the case of a 2nd order crossover the time delay is such that the drivers cancel at the crossover frequency, hence why one driver is inverted. I know you know that.

    But here comes the part where I likely f**k up somewhere, cause my math is bad:
    • Half of the 2500Hz wave takes 2ms. That's a delay of 6.88cm.
      So even with a very small midrange you will get lobing.
    • A typical driver spacing for a small midrange might be twice that.
      Its nulls will be around 45(?) degrees off axis. if the drivers were DSPd to have no relative delay. That's what we see with the D&D 8C. The delay introduced by the crossover will just make things worse.
    Well, the LS50 driver coaxials will solve that problem. My point above was that with both drivers of a coax run full-range you would still get lobes, even due to driver size differences alone. And that the "Uni-Q" only works for a small range around where they then put their XO frequency.

    I agree: For DSPd speakers it makes so much sense to just use steep filters and a FIR filter to cancel the introduced phase difference, with the resulting delay being the price to pay. That's what I do with my LS50 Meta at my desk. That's how the Dutch & Dutch 8c is designed, or the Kii Three and many other speakers nowadays.


    For woofer to mid crossover it's pretty much a non-issue, even 1st order. Wavelengths at XO are 1m or longer.
    As for M-T XO I found neither option satisfactory, which is why I ended up preferring the 4" widebander on its own compared to adding a tweeter. And a first order will be so difficult to implement that one may aswell just give up with most drivers. It only really sort of worked in that case since the tweeter covered the range down to 1000Hz with a 500Hz Fs and the "midrange" covered the range up to 20kHz. And then only a 4-5kHz XO point was viable.
    Compared to that the 2nd order was a breeze with many different crossover points being an option, all the way down to 1000Hz with good linearity. Between 1500 and 2000Hz is what I tried.

    As for the audibility of phase differences, I don't find it very difficult at all. Hence why the first thing I did after hooking up the 2nd order XO was to correct for its excess phase.

    Some years back I made a thread on that topic:
    https://www.superbestaudiofriends.org/index.php?threads/crossover-phase-shift-audibility.9313/
    I should maybe update the thread with an actual test track :D
     
  14. Ardacer

    Ardacer Friend

    Pyrate Contributor
    Joined:
    Oct 15, 2017
    Likes Received:
    986
    Trophy Points:
    93
    Location:
    Croatia
    I have proven to myself that I'm unable to discern a 1dB difference from a cellphone speaker. I should try it on headphones, to be fair, it might be a different story. 3db I can reliably tell 100% of the time, but it's such a minute thing, I really have to concentrate. If the entire response is inside 3dB range, I think that's good. 6dB is a lot, but the whole range inside 6dB band without sudden shifts could probably still be passable (to me?) although I'd try to fix that.

    Thanks for assuming I know that (I do?:p ) but you never know, it's a good reminder. Naturally, you're correct here. 4th order should be "in phase", 8th order too. Unless I'm mistaken in remembering (it happens).
    2nd order is inverted, 1st is in phase (actually the only one that's truly in phase), and other odd orders are shifting around. I've toyed with 2nd order filters a long time ago. The notch they create when not inverted properly was hard for me to notice. Narrow notches are hard to notice with normal material, like music.

    What I meant by - I don't care for phase - is exactly in fact what you just said: it's the same thing as the path difference. It all comes down to amplitude distributions, that's to say, directivity, and on-axis response. I also think you said this. Crossover designs consider that already (or they should).
    So yes, in 2nd order crossovers, phase is important as it's an integral component of the design.
    In general, if everything sums up nicely on-axis, and directivity is good - any excess phase elsewhere is inaudible to me. If its effects are audible, it means it's shitting up frequency responses somewhere (in quite the same way a path difference would) and something isn't set up properly. It's just that - very high-order crossovers shit up such a tiny amount of the spectrum compared to lower-order.


    I see no problems here, tbh, you're spot on.

    Yes.. delay is shit for real-time aplications (it's really big at times), no gaming on those :/
    But it's not an issue for media reproduction.
    That's why I plan to do a smaller full analog (electronic xover) build for my PC. The ls50 project is really just for fun.

    Well, at crossovers and directivity stuff, especially lobing, etc, sure. But there's no way I'm myself able to tell phase difference that results in no fr change somewhere. :)


    tldr;
    Yes, phase is important around crossover points, very much so. But proper crossovers should already take care of that. Otherwise, I'm not so sure.
     
    Last edited: Feb 20, 2024
  15. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Actually it's a constant 90° phase shift between the drivers, which is why it's so hard to do. It should be 90° across the whole spectrum. So if one driver rolls off too quickly that easily makes it 180° somewhere and you get a suck-out in your response. Ideally they sum by 3dB at the XO frequency, instead of 6dB.
    Due to the 90° phase shift the FR magnitude also shouldn't change between inverting one driver and having both in the same polarity. After all, 270° is just -90° in this case. But the sound does change, IME.

    You'd be surprised, as that's what that other thread is about. You can generate such filters in RePhase or use the ones I provided in that thread. You'll need something which accepts a convolution with an IR and preferably something to blind test it, too. Seriously, give it a try!


    I'll see if I can get good measurements with me sitting in my chair later today, to update this thread and to show you guys what a difference it makes.
     
  16. Ardacer

    Ardacer Friend

    Pyrate Contributor
    Joined:
    Oct 15, 2017
    Likes Received:
    986
    Trophy Points:
    93
    Location:
    Croatia
    Shows what I know. Nothing much apparently. It's constant phase, not - in - phase. It never did make much sense why everyone spoke about 1st order being the best with regards to phase and blend and everything and most "in phase". I never bothered to check myself as I'm not usually using anything below 4th order. Thanks for clearing this up. I might have known it at some time, or not, but I do now. I'll refresh my memory on those low order crossovers and phase with the ls50 build once I get to it. (I'll need to)

    Edit: Right - gave it some thought now - it makes perfect sense. Thanks.

    I'll give it a try once I catch some time for sure, thanks! :)
     
    Last edited: Feb 20, 2024
  17. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Well, that didn't age very well. After I took these measurements I realized that by measuring without me sitting at the listening position and instead positioning myself behind and to the side of the microphone the results would get noticeably worse.
    With me actually sitting where I sit while listening the response ended up looking a lot smoother in the lower midrange, less of a floor bounce effect. So these are the updated graphs:

    AMON
    Amon in-room RTA with me at LP.jpg
    LR difference in the lower midrange could be real. The walls aren't the same and the plants aren't symmetrical.

    HATHOR
    Hathor in-room RTA with me at LP.jpg

    Unfortunately I get a suckout at 30Hz. The REW room sim didn't help much, other than to tell me that I will, indeed, get a suckout at 30Hz in every feasible position. Oh well.
     
  18. Ardacer

    Ardacer Friend

    Pyrate Contributor
    Joined:
    Oct 15, 2017
    Likes Received:
    986
    Trophy Points:
    93
    Location:
    Croatia
    Amon is your OB? That probably explains the lack of 30hz drop, right?
    Cause it should happen no matter what, unless you use dipoles (or full range cardioids). Hathor is cardioid and very nice in the normal range, but in bass it's just omni. You might try placing bass traps in the room corners (and middles of walls), it might help a bit, if it's allowed (waf, practicallity etc). But you probably considered that. Buy a big carpet depicting sexual acts and hang it on all your walls. Tell whomever asks it's for audio and it was cheapest
     
  19. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Hahaha.
    Yeah, it's the OB. But it's also somewhat from the speaker and listening position. I don't even think a full range cardioid would work for the null, only OB or something in between, like with a smaller rear lobe could work.

    I guess the best option is to add a subwoofer somewhere where there's no resulting suckout. A Ripol if it has to be OB.

    Regarding OB woofers: I've seriously considered building a rotary woofer with a three phase AC motor and running the three phases into my room to plug it directly into the wall. But then it also has to be safe and quiet, so...
    (The motor does have to be powerful so the pitch signal doesn't modulate the motor speed, which in turn would modulate amplitude and frequency! And it has to be quiet aswell.
    However you generally don't want unstoppable rotating blades in your room.)

    Anyway, building a good rotary woofer is not cheap at all. Nothing feasible for now.


    Back to reality:
    The room is quite dampened in the lower frequencies. Both the ceiling and floor aren't very solid and it helps a lot for the bass that it's halfway in the tilted roof of the house. Yeah, it's less space, but it also acts a bit like a giant bass trap. The 42Hz room mode I get is just 7dB in this room. In the living room one floor below (same length) it's more like 17dB! Hence the need for EQ for the Reference 3K bass section.
    I used to have the 10kg of sheep's wool I use for measuring speakers under my bed, in a corner of the room. That helped by another dB or so. But since I need quick access to it I store it elsewhere now.
     
  20. Ardacer

    Ardacer Friend

    Pyrate Contributor
    Joined:
    Oct 15, 2017
    Likes Received:
    986
    Trophy Points:
    93
    Location:
    Croatia
    Now I'm just spewing nonsense maybe, but how about a helmholz absorber somewhere, well positioned? Usually they are used to reduce the peaks, but in theory it might do something to the room to move the null around.
    You could always put another sub very close to you.
    I've pondered rotary woofers too, and came to the same conclusion - not worth it all things considered. Also potentially dangerous like you said.
    I have 2 big Dayton 15" with fs of 17.5hz. These things are brutal, way more than enough.
     

Share This Page