Announcement

Collapse
No announcement yet.

Off axis "flares". Why do they happen? How to fix them?

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

  • #31
    I saw someone recently used cardboard with multi-pointed cut-outs over the drivers and this appeared to help smooth diffraction a lot. It probably doesn't reduce the total amount of diffraction, but is smears it across a wide frequency range so it appears to be reduced. I think one would want to make the multi-pointed star shaped cut-outs rather asymmetrical for best effect (some of the cut-outs should be deeper and/or wider than others). If there is a software package that could help you simulate this I don't know of it. I think this would involve a lot of trial and error. The good news is that the material is cheap and easy to cut so it doesn't take very long to try new things

    Comment


    • #32
      Originally posted by dlr View Post
      Neoprene may be good for panel vibrations, but not for a baffle diffraction application. That name in the first article sounds familiar. It uses misinformation and exaggeration to push the product, using "audiophile" wording to hype the claims, but it's not anywhere close to the claims. Too thin and not a good material in my experience. I've tested similar products and found them, in many cases, to make the situation worse.

      dlr
      Yep. It'll probably not work. Further reading leads me to believe it was just an audiophile tweak that has no solid data to back it up.

      Yet, I did found something really interesting. Measurements for the Revel F208 tower speaker, with 2 8" woofers and a single 5.25" midrange. Somehwat similar in configuration to mine. These are its off axis measurements (from audiosciencereview)


      image widget

      I can see almost the same flare I have at 1.8 KHz. Interestingly, it's pretty much behaving the same - at 1.8 KHz, off axis radiated power is higher than on axis. Which is my exact same issue.

      The issue is clearer in this graph.


      It's behaving exactly the same. Off axis, a flare at 1.6 KHz or so.

      So, whatever it is that I'm doing right or wrong... a $5K per pair amazing sounding speaker does it, too.

      Maybe I'm not that wrong after all.
      Last edited by fjhuerta; 02-09-2021, 07:08 AM.
      Line Array: IDS-25 Clone, FE-83.
      2-2.5 Way:
      Zaph Audio's winning entry: ZA5+SB29. - Microliths: RS125+RS28. - Small Bangs: TB W4-1658SB+SEAS 27TBFC/G. - Monoliths: Peerless 830884+SEAS 27TBFC/G.
      3-3.5 Way:Miniliths: SEAS P21/CA21REX+Neo8 PDR+Neo3 PDR. - Megaliths: 2xDayton RS270+2xT-B W4-1337SB+SB29. - ZDT3.5 +: 2xDayton RS180+Dayton RS52+Vifa DQ25. Reflexos: OB 4xDayton RS150 + Neo3 PDR.

      Comment


      • #33
        Maybe not easily avoidable, but still wrong ;-)

        I'm only kinda joking, as the Revels are constrained to being something that will look presentable in the living room for most folks willing to plunk down $5k on speakers. For those of us who don't care so much about what it looks like as what it sounds like you can probably do better. In fact, that's one of the few facets in which DIY beats commercial nowadays. We can build whatever we want irrespective of the marketing department.

        Comment


        • #34
          Just to be clear, if your on-axis dip at 1.6kHz is about equivalent to the off-axis peak you probably won't notice it in most normal rooms (i.e. not overly dead room). In fact, without a direct A to B comparison it would probably be quite difficult for most folks to hear the difference assuming the peaks/dips are only a couple of db.

          Comment


          • #35
            Originally posted by LIDAR View Post
            Just to be clear, if your on-axis dip at 1.6kHz is about equivalent to the off-axis peak you probably won't notice it in most normal rooms (i.e. not overly dead room). In fact, without a direct A to B comparison it would probably be quite difficult for most folks to hear the difference assuming the peaks/dips are only a couple of db.
            Exactly. That's what I was aiming for. And the listening room measurements are pretty even from 500 hz and up.

            I did wonder how Revel got away with no round overs. I have the Concerta2 M16, and I guess that's why I thought chamfers, round overs and felt were overrated. Now I know - they balance the off axis diffraction anomalies with the on axis response.

            I'm still going to test the wool and neoprene.
            Line Array: IDS-25 Clone, FE-83.
            2-2.5 Way:
            Zaph Audio's winning entry: ZA5+SB29. - Microliths: RS125+RS28. - Small Bangs: TB W4-1658SB+SEAS 27TBFC/G. - Monoliths: Peerless 830884+SEAS 27TBFC/G.
            3-3.5 Way:Miniliths: SEAS P21/CA21REX+Neo8 PDR+Neo3 PDR. - Megaliths: 2xDayton RS270+2xT-B W4-1337SB+SB29. - ZDT3.5 +: 2xDayton RS180+Dayton RS52+Vifa DQ25. Reflexos: OB 4xDayton RS150 + Neo3 PDR.

            Comment


            • #36
              Originally posted by fjhuerta View Post
              I'm still going to test the wool and neoprene.
              Since you do have felt, consider my suggestion to use it for design only. It will improve both the on-axis and off-axis. This will allow you to more easily design for power response by going for flat on-axis response. When done, leave the felt off. You'll have a bit of on-axis dips/peaks, but the off-axis will have the opposite which will leave the power response roughly the same. Far easier to design that way, rather than try to guess the optimum on-axis response for balanced power. Then you can do small tweaks according to your perception of the system, since on-axis is generally considered the most important.

              dlr
              WinPCD - Windows .NET Passive Crossover Designer

              Dave's Speaker Pages

              Comment


              • #37
                Originally posted by dlr View Post
                Since you do have felt, consider my suggestion to use it for design only. It will improve both the on-axis and off-axis. This will allow you to more easily design for power response by going for flat on-axis response. When done, leave the felt off. You'll have a bit of on-axis dips/peaks, but the off-axis will have the opposite which will leave the power response roughly the same. Far easier to design that way, rather than try to guess the optimum on-axis response for balanced power. Then you can do small tweaks according to your perception of the system, since on-axis is generally considered the most important.

                dlr
                Thanks Dave. That sounds like a really interesting idea. So basically, I use the felt, design for flat on axis, eliminate peaks off axis and take the felt off afterwards?

                Frankly, it sounds great. The speakers are beautiful, and the felt would do no favors to their looks.
                Line Array: IDS-25 Clone, FE-83.
                2-2.5 Way:
                Zaph Audio's winning entry: ZA5+SB29. - Microliths: RS125+RS28. - Small Bangs: TB W4-1658SB+SEAS 27TBFC/G. - Monoliths: Peerless 830884+SEAS 27TBFC/G.
                3-3.5 Way:Miniliths: SEAS P21/CA21REX+Neo8 PDR+Neo3 PDR. - Megaliths: 2xDayton RS270+2xT-B W4-1337SB+SB29. - ZDT3.5 +: 2xDayton RS180+Dayton RS52+Vifa DQ25. Reflexos: OB 4xDayton RS150 + Neo3 PDR.

                Comment


                • #38
                  Originally posted by fjhuerta View Post
                  Thanks Dave. That sounds like a really interesting idea. So basically, I use the felt, design for flat on axis, eliminate peaks off axis and take the felt off afterwards?
                  If you apply enough felt appropriately, you won't need to consider the off-axis while designing. Felt can smooth on- and off-axis well. Once you verify raw measurements with the felt, simply design on-axis. Never hurts to measure as well, but any small deviations off-axis are usually minimized. After removing it, you may still want to tweak some, but that's when I do it with my ears rather than purely measuring. Experiment.

                  Frankly, it sounds great. The speakers are beautiful, and the felt would do no favors to their looks.
                  I'm willing to have that "problem" as I want the absolute best I can get in a design. They're only for me, though.

                  dlr
                  WinPCD - Windows .NET Passive Crossover Designer

                  Dave's Speaker Pages

                  Comment


                  • #39
                    Originally posted by dlr View Post
                    If you apply enough felt appropriately, you won't need to consider the off-axis while designing. Felt can smooth on- and off-axis well. Once you verify raw measurements with the felt, simply design on-axis. Never hurts to measure as well, but any small deviations off-axis are usually minimized. After removing it, you may still want to tweak some, but that's when I do it with my ears rather than purely measuring. Experiment.


                    I'm willing to have that "problem" as I want the absolute best I can get in a design. They're only for me, though.

                    dlr
                    I wish I could say the same, but I'm married :D

                    I'm curious about your approach; unfortunately, the felt will arrive in at least two weeks. So I'll have to wait for a while.

                    Meanwhile, I found the spinorama measurements for the Revels. Knowing that they have the same issue as my speaker, I was wondering how Revel's on axis FR looks like., and I saw that they have 1 less dB from 1.5 to 2 KHz, approximately. The rest, from 300 Hz to 5 KHz looks almost identical, and the rising FR is, I guess, due to my RS28F's natural FR, and the fact that Revel is using a waveguide.

                    Which seems to coincide with my overall idea of depressing the presency region, in order to tame first reflections, and achieve a flatter in room response.

                    Felt can't come soon enough.

                    Line Array: IDS-25 Clone, FE-83.
                    2-2.5 Way:
                    Zaph Audio's winning entry: ZA5+SB29. - Microliths: RS125+RS28. - Small Bangs: TB W4-1658SB+SEAS 27TBFC/G. - Monoliths: Peerless 830884+SEAS 27TBFC/G.
                    3-3.5 Way:Miniliths: SEAS P21/CA21REX+Neo8 PDR+Neo3 PDR. - Megaliths: 2xDayton RS270+2xT-B W4-1337SB+SB29. - ZDT3.5 +: 2xDayton RS180+Dayton RS52+Vifa DQ25. Reflexos: OB 4xDayton RS150 + Neo3 PDR.

                    Comment


                    • #40
                      Originally posted by fjhuerta View Post
                      So, whatever it is that I'm doing right or wrong... a $5K per pair amazing sounding speaker does it, too.

                      Maybe I'm not that wrong after all.
                      I haven't read the thread closely and the figures in the OP are missing but if you wish to get a smoother horizontal directivity you will need to consider the local baffle width around the drivers (and a waveguide on the tweeter). This is what the models higher up the Revel range do and it seems to be fairly common in German DIY circles. Here is an example but I have seen others. This is what you are chasing?

                      Comment


                      • #41
                        Originally posted by andy19191 View Post

                        I haven't read the thread closely and the figures in the OP are missing but if you wish to get a smoother horizontal directivity you will need to consider the local baffle width around the drivers (and a waveguide on the tweeter). This is what the models higher up the Revel range do and it seems to be fairly common in German DIY circles. Here is an example but I have seen others. This is what you are chasing?
                        That's exactly what I wanted to do: a WMTMW with such controlled horizontal directivity. I even started with a pair of SEAS DXT on my cabinets. Unfortunately, one tweeter seems to be damaged. :(
                        Line Array: IDS-25 Clone, FE-83.
                        2-2.5 Way:
                        Zaph Audio's winning entry: ZA5+SB29. - Microliths: RS125+RS28. - Small Bangs: TB W4-1658SB+SEAS 27TBFC/G. - Monoliths: Peerless 830884+SEAS 27TBFC/G.
                        3-3.5 Way:Miniliths: SEAS P21/CA21REX+Neo8 PDR+Neo3 PDR. - Megaliths: 2xDayton RS270+2xT-B W4-1337SB+SB29. - ZDT3.5 +: 2xDayton RS180+Dayton RS52+Vifa DQ25. Reflexos: OB 4xDayton RS150 + Neo3 PDR.

                        Comment


                        • #42
                          I can't see any pictures of the speaker, but here's what I'd try.

                          The mid may be influenced by the woofer nearby, or the other mid may be influenced by the top edge of the cabinet. I'd measure each individually, and see which is to blame. Then, I'd notch the response in that mid in a very narrow band. I don't know if this would work, but it's the direction I'd take.

                          Comment


                          • #43
                            Originally posted by fjhuerta View Post
                            That's exactly what I wanted to do: a WMTMW with such controlled horizontal directivity. I even started with a pair of SEAS DXT on my cabinets. Unfortunately, one tweeter seems to be damaged. :(
                            A bit more information on baffle width and chamfer angle given in this write-up. There is likely to be more but I have never searched. The source of the link was a German chap called Rese66 in a thread here that may be worth contacting for more detailed information on the German DIY scene w.r.t. to controlled horizontal directivity.

                            It is an interesting topic that requires the accurate 3D simulation of the sound radiation from drivers in strongly shaped baffles. I have used what you want to know to help flesh out one of my example simulations here. Thanks for that.

                            There are other factors apart from baffle width and shape such as the the driver size and directivity. If you are using FIR filters one can use custom variable slopes to set the level for the overlapping frequency range of the drivers to control the required horizontal beamwidth. One can even go a step further like B&O and place 3 similar drivers in a triangle and control directivity within the passband of the driver not just between drivers of different sizes. All good stuff and I strongly suspect one of the DIY areas to be explored when the mainstream of the hobby finally gives passive crossovers the boot.

                            Comment


                            • #44
                              fjhuerta,

                              I used the RS100 in my CORONA design, which is posted in the Project Gallery section of TechTalk. I took some off-axis measurements, but not past 40 degrees.

                              Of more interest, I have a treatment there of edge diffraction, since CORONA is a worst-case design for edge diffraction, so it really shows what to look for when measuring. I notice, for instance, that in your off-axis measurements there is not only a bump around 1800 Hz, there is also a dip at around 900 Hz. This pattern of effects in even multiples is a sign of an edge diffraction issue.

                              Also, when tuning to take it into account, remember that listening distance matters. The frequencies at which an edge-diffraction problem shows up change with the distance from the baffle. So if you are going to listen at 1.5 meters, you have to measure at 1.5 meters.

                              -Meredith

                              Comment


                              • #45
                                Originally posted by mcargill View Post
                                fjhuerta,

                                I used the RS100 in my CORONA design, which is posted in the Project Gallery section of TechTalk. I took some off-axis measurements, but not past 40 degrees.

                                Of more interest, I have a treatment there of edge diffraction, since CORONA is a worst-case design for edge diffraction, so it really shows what to look for when measuring. I notice, for instance, that in your off-axis measurements there is not only a bump around 1800 Hz, there is also a dip at around 900 Hz. This pattern of effects in even multiples is a sign of an edge diffraction issue.

                                Also, when tuning to take it into account, remember that listening distance matters. The frequencies at which an edge-diffraction problem shows up change with the distance from the baffle. So if you are going to listen at 1.5 meters, you have to measure at 1.5 meters.

                                -Meredith
                                Yep.. I figured that one out with measurements at my sittin position.

                                Here are the results of applying no felt, 1/4", and 3/4", 2" wide, F13 wool felt in my sitting chart.

                                I guess this is how I learned not to worry that much about going more than 45° off axis. Or about using felt.
                                Last edited by fjhuerta; 02-24-2021, 03:44 PM.
                                Line Array: IDS-25 Clone, FE-83.
                                2-2.5 Way:
                                Zaph Audio's winning entry: ZA5+SB29. - Microliths: RS125+RS28. - Small Bangs: TB W4-1658SB+SEAS 27TBFC/G. - Monoliths: Peerless 830884+SEAS 27TBFC/G.
                                3-3.5 Way:Miniliths: SEAS P21/CA21REX+Neo8 PDR+Neo3 PDR. - Megaliths: 2xDayton RS270+2xT-B W4-1337SB+SB29. - ZDT3.5 +: 2xDayton RS180+Dayton RS52+Vifa DQ25. Reflexos: OB 4xDayton RS150 + Neo3 PDR.

                                Comment

                                Working...
                                X