pattern control below 400Hz

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Looks like REW? I love it but it can't draw directivity graphics or count DI.

Yea, REW. I've really only dabbled with what it can do, but I really like it.

But taking measurements at 10¤ steps and overlaying all of the 0-90¤ (and 90-180¤ separately) would reveal the directivity "shape". Set IR gating to 6, 12, 20, 60 etc. ms and look at same measurements again. I have found 20ms and 1/12 smoothing best and then first reflections can be detected easily.

Do you know if, in REW, is there a way to normalise the on-axis plot, and apply the same amount of correction to it's family of off-axis plots? It would make the comparisons a bit simpler. I've poked around the help files etc. and can't find anything like this.

This is from ARTA and from this thread

Yeah, those plots are a lot more elegant.
 
As far as I know, the earlier "Resistance enclosure" threads didn't have (or didn't show) an identical sealed box to act as a baseline. Having this baseline is amusing / frustrating. The process often goes:

1) I record an off-axis plot for the "Resistance enclosure"
2) I congratulate myself, cos it looks like a pretty good weakening of rearward radiation
3) I then realise it is only slightly better than the sealed box

Other than that, I've mostly noted noted the same things as the earlier threads.


Again, the Adventures in Cardioid is the best resistance thread there is. Pay particular attention to his measurement with and without baffle and that he is doing it outside on a stand of sufficient height for the freq..

https://www.diyaudio.com/forums/multi-way/142691-adventures-cardioid.html#post1808323


Note that the the baffle is almost non-existent (relative to the driver) and the rear/back is NOT absorptive/lossy, AND that it's small (in volume). Of course it's also lossy all the way around its sides.
 
Do you know if, in REW, is there a way to normalise the on-axis plot, and apply the same amount of correction to it's family of off-axis plots? It would make the comparisons a bit simpler. I've poked around the help files etc. and can't find anything like this.
I've had success using "A/B" using the trace arithmetic functions. The first curve would be 1/1, the next 2/1, etc. which produce a series of normalized traces that can be plotted relative to 0 dB. You can offset the traces if necessary to push them back up to the nominal level, but I usually just leave them referenced to 0 dB. OmniMic is a good package to plot the polar diagrams by importing the .frd files--freely. Works like a champ.



Chris
 
MBK's "cookie" midrange cardioid is quite similar to Gradient Helsinki mid, mentioned in the last post of the thread.

The funny thing in that thread is that MBK is working above the actual cardiod response range, and didn't measure backside hardly at all. A cardioid behaves pretty much like a dipole,when discussing loudspeaker drivers with cone membranes. The on-axis response shows typical 6dB/oct rise and very constant directivity until the dipole peak is reached (determined by baffle width). Then response drops to dipole null (determined by baffle width) and there directivity index is zero (rings like a bell to all directions). Above this null, driver membrane diameter is determining directivity instead of the baffle, and it follows same pattern as normal monopole cone driver in frontal hemisphere. MBK looks at that range, starting from transition area around 500Hz. hollowboy is interested in lower midrange, below 400Hz!

The difference of dipole and cardioid is in what happens after 30¤ off-axis. The dipole attenuates rapidly up to null at 90¤ but cardioid keeps on radiating only slightly attenuated. A supercadioid has minimum around 120-150¤, real cardioid at 180¤. And both have poor efficiency at low freq.

adma2.png


I have constructed several cookie cardioid midrange prototypes and measured them. It was a revelation to find the similarity to dipole behaviour. I was aiming for dipole, so I didn't go on or publish my measurements in AINOgradient thread.

So, the benefit of cardioid response in low midrange is to avoid nulling/reflections from the front wall behind the speaker. It is not to have narrow directivity frontally, ie. not to avoid lateral side wall or floor reflections (but they get attenuated)!

My AINOgradient is clone of Gradient 1.3, a design by genius Jorma Salmi. It has cardioid response in low midrange as result of monopole bass and dipole mid. Crossover is around 200Hz which gives cardiod response for almost two octaves around it with LR2 xo. This is the where typical front wall reflection happens! The challenge of the design is to combine different radiation types smoothly, to get balanced room response. Gradient Revolution and Helsinki have dipole bass and cardioid midrange. Now the latest Gradient 1.4 has again monopole bass, but (hyper?)cardioid midrange crossed around 200Hz.

Amphion Krypton is another Finnish speaker to use cardioid midrange, xos 160 and 1600Hz. Kimmosto's first cardioid prototypes were influenced by these.

An externally hosted image should be here but it was not working when we last tested it.
 
Last edited:
The funny thing in that thread is that MBK is working above the actual cardiod response range, and didn't measure backside hardly at all.


"1- let's see the rear radiation (full polars) - it could look like anything from omni transitioning to directional regime, to open baffle, to cardioid-like..

To address 1- see pic attached: sonogram shows rear radiation is indeed mostly absent, not perfectly so, but largely 6+ dB down in the range that interests me, i.e., 300-2000 Hz at most."



https://www.diyaudio.com/forums/multi-way/142691-adventures-cardioid-2.html#post1808702


It's not great for < 400 Hz :eek: , but it does show full polar. :)


Again though, I wouldn't use a resistance design if I had several woofers and DSP. The better method is an opposed design (at first dipole) with front time-delay (relative to the out-of-phase rear) to cardioid.
 
Last edited:
I exported REW measurements to Dayton Omnimic which is easy to use and directivity graphics works also in free version. You just can't save measurements for free.

..OmniMic is a good package to plot the polar diagrams by importing the .frd files--freely. Works like a champ.

Very cool! :cool: :)

There is at least one directivity plot that OmniMic does that Arta does not, I wanted it.. but was to cheap to pay just for that. :eek: (..and thought it was just "pay" software.)
 
Oh yes MBK had also many 0-180¤ plots. But the main point is that in my opinion, he didn't quite catch the idea how baffle width /diameter affects in each frequency range.

gainphile followed his studies closely and adopted lots of it in his projects. They both used a waveguide tweeter above mid, like Gradient's many models (coaxial) and Amphion Krypton (wg). That is the way to make frontal directivity pretty smooth over wide range. D&D 8C does the same.

Princeton University's acoustic lab has measured 3D didrectivityy of gainphiles R16, Gradient Helsinki among many other type of speakers including GedLee Nathan. In most measurements scale starts around 200-500Hz, sadly, because of their too small anechoid chamber.

gainphile R16 shows the problem of dipole+wg - rear output disappears around 1,5kHz and this makes sound in a room strange. Cardioid mirange+wg tweeter would be more balanced. Very wide baffle with normal monopole limits backside radiation quite well too, like Sonus Faber Stradivari and planet10's flat 2-way.

This is R16
Gainphile%20R16%20H%20Contour%20Plot.png


at 3D3A page, you can list the 26 measured speakers in many ways. When crition is "Constant directivity front horizontal", winners on tie are two totally different speakers - Gradient Helsinki and Genelec 8351A! The difference is on the backside and I bet they sound different in a normal room, I have heard both.

ScottG, I have tested that delayed double bass system, but it was a terrible mess to measure indoors. In pro audio it is used commonly to reduce spl on-stage. The nice thing is that you can adjust it's comfort zone by changing delay. It can also be done in a reasonably small box, about half the size of an acoustic perforated cardioid bass. With modern multiway dsps like minidsp 2x4HD it would make a very nice active 3-way cardioid using two channels for double bass cardioid, acoustic cardioid mid and wg tweeter. Actually this might be my next challenge...
 
Last edited:
Yeah, Princeton's 425 Hz is enough for head-shading effects, but really doesn't meet the <400 Hz requested here. MBK does state that his data is good to somewhere between 150-200 Hz though, and that there is the grass/ground reflection around 300 Hz that results in that 300 Hz dip/bump in some of the graphs.

I actually think that MBK got the interaction with baffle better than anyone. Not only is it an evolutionary process that he presents with and without baffle, he specifically makes a comment with full polars that a larger baffle INCREASES rear output as seen here with no baffle, 12" wide baffle, and 16" wide baffle. Basically ANY substantive baffle increases rear output to a fairly low freq. (when considering the small width of the baffles tested), the larger the baffle, the more rear-output to a lower freq.:

https://www.diyaudio.com/forums/multi-way/142691-adventures-cardioid-2.html#post1808747


This lets me know that IF I were doing a <400 Hz resistance design (or perhaps most any cardioid design), that I'd want to minimize the baffle's size to almost nothing to reduce the output to the rear.


BTW, I personally am not a fan of most resistance designs (moreso as freq.s increase into the lower midrange and up). It's not the directivity, rather it's what it does to low-level excursion and it's why I'd go for a DSP "double-driver" design (if having that as an option). Note the various subjective descriptions/results of all of Kimmo's designs - he loved that large bass cardioid, but when he got into the midrange: not so much. (..for me resistive mid.s tend to kill sound-stage depth).
 
Last edited:
I've had success using "A/B" using the trace arithmetic functions.

Thanks, will try it out. I was considering exporting the SPL plot as text and normalising / plotting it myself in a spreadsheet.

This is mainly for nerdy gratification. Looking at the bigger picture, I'm not sure how much time I should put into plotting, tweaking and getting a 'perfect' pattern. It seems as if moving the box from outside to inside gives a pattern so different from the outdoors measurement, that a crude measure might be just as good as a sophisticated one.

My current rough plots look like I have OK rear reduction <300Hz and good side reduction 300-500Hz (relative to the sealed box). If my patience runs out, that will have to be good enough.

hollowboy is interested in lower midrange, below 400Hz!

Yea. That's my target, but not set in stone. If 350 or 500Hz works better in practice, that's how I'll roll - hooray for DSP & easy Xover changes.

The difference of dipole and cardioid is in what happens after 30¤ off-axis. The dipole attenuates rapidly up to null at 90¤ but cardioid keeps on radiating only slightly attenuated. A supercadioid has minimum around 120-150¤, real cardioid at 180¤.

Looking at prior work / plots, it looked like 0, 90 and 150 degree measurements were quickest way to spot whether I was getting dipole(ish) or cardioid(ish) pattern. Quick = good for getting through a couple of tweak-and-test cycles in an afternoon.

BTW, I personally am not a fan of most resistance designs (moreso as freq.s increase into the lower midrange and up). It's not the directivity, rather it's what it does to low-level excursion and it's why I'd go for a DSP "double-driver" design (if having that as an option).

The prototype seemed OK for excursion with one OB 18" playing 50-400Hz. In the final (hopefully) version, I intend to use sealed boxes for LF (2x the cabinet pictured in post 28), so I can set the sub/midbass crossover to whatever seems best.

Note the various subjective descriptions/results of all of Kimmo's designs - he loved that large bass cardioid, but when he got into the midrange: not so much

Are you referring to this subjective description from the KS-2125 page?

"There's not too much resolution at midrange. Mild directivity step from midrange to tweeters keeps harshness away, but sometimes midrange sounds pronounced. Height of speaker increases this feature because room height is only 248 cm and roof is reflecting (painted pine panel)."

My goals are probably different enough that this is N/A. I'm aiming to use my perforated box 100-400Hz (ish), whereas the KS-2125 used them 20-220 and 220-1800Hz.

Also, I'm not fancy enough to worry about "sound-stage depth" or any of that stuff. For most of the music I like, that's not a thing I even think about.
 
I haven't heard Kimmo's speakers, but like I said in my previous post, a large waveguide (and for the most a horn) gives such a focused sound that transition is noticed (too?) easily. Ear/brain is very sensitive around 1-4kHz. In a Finnish hifi forum K has said that he didn't like the sound of cardioid+horn and that room is the king anyway. He values high clarity and sharp imaage, definition very much. But still, in one version there is a backside horn added obviously for "ambience")

In this respect we listeners are very different. Many hate and some love dipoles and panels. I get easily tired when listening to conventional or highly directive speakers, they sound unnatural and obtrusive to me.
 
..Are you referring to this subjective description from the KS-2125 page?

All of them in the aggregate with respect to midrange, though he loved that bass bass design (KS-2125) in particular. Also look at the evolution of his designs (ie. look at the date vs. the design).

Note that the KS-2125 has a crossover around 250 Hz.


Read his description for the KS-585 (..with resistance bass and mid.s):

KS-585
 
Making slow progress:

-I have complexity addiction
-I'm trying to build this from existing materials / clear my shed / use up scraps. Marie Kondo would approve :)
-I fail at the "measure twice cut once" thing

The thick trans-laminated cabinet edges were largely made from of plywood offcuts. The multi layered section is only a few cm deep: I want just enough material to allow for some rounding of the edges. Any more would be extra weight.

I also tweaked the plan, making the horn closer to square (90 x 80 degrees). The mouth is now 500mm high, with 300mm of baffle above it (the section covered by tools in the picture), so it should have ~symmetric pattern control.

The face sections are a laminate [ bamboo || ply || plastic grid ] to try for a high amount of strength and damping for a relatively low weight. I'm using a double layer of bamboo in key spots, to allow me to grind + sand more of a curve into the mouth-baffle transition.

Just visible on the right of the pic is one of the holes for the leaky bass.

TL;DR: I'm trying to smoosh a lot of my recent design ideas and experiments into one project, and make it pretty.
 

Attachments

  • square progress.jpg
    square progress.jpg
    645.3 KB · Views: 439
I think I 'understand' the leaky bass section, need to read up a bit more on that.

I don't fully understand it either. I can visualise how a figure 8 pattern can form but not the more complex shapes.

Here are progress shots, showing I how I'm setting the bamboo/plywood panels slightly into the shell, then grinding the shell back to get the edges flush.

This grinding step means it doesn't matter that I jigsawed together a bunch of wonky scraps and offcuts for the trans-laminated part - it makes me a bit sad when I see (otherwise amazing) projects that do CNC translam from whole sheets, generating enormous amounts of offcuts and waste.

The blobby bits in the photo are where I'm filling in chips and gaps with a slurry of glue + wood dust. The bamboo 'heals' remarkably smoothly. I guess it is because the fibres of the normal panel are so smooth / tightly packed that the fixed up section doesn't stand out. In real wood, filler sections look oddly blurry, like they've been photoshopped IRL.
 

Attachments

  • grind before.jpg
    grind before.jpg
    855.7 KB · Views: 398
  • grind after.jpg
    grind after.jpg
    628.2 KB · Views: 383
Last edited:
I don't fully understand it either. I can visualise how a figure 8 pattern can form but not the more complex shapes.

Exactly, the figure eight _mostly_ makes sense to me. By mostly I mean, nearly, of course. :D


Here are progress shots, showing I how I'm setting the bamboo/plywood panels slightly into the shell, then grinding the shell back to get the edges flush.

This grinding step means it doesn't matter that I jigsawed together a bunch of wonky scraps and offcuts for the trans-laminated part - it makes me a bit sad when I see (otherwise amazing) projects that do CNC translam from whole sheets, generating enormous amounts of offcuts and waste.

The blobby bits in the photo are where I'm filling in chips and gaps with a slurry of glue + wood dust. The bamboo 'heals' remarkably smoothly. I guess it is because the fibres of the normal panel are so smooth / tightly packed that the fixed up section doesn't stand out. In real wood, filler sections look oddly blurry, like they've been photoshopped IRL.

They can be very wasteful unfortunately. I really like your approach to building, it's awesome to see you utilize every bit you can.

The wood dust, glue trick is an awesome one. I use that for all sorts or repairs, gap filling, etc.
 
I can see some discussion about KS-585, KS-2125 and KS-1804.
Room acoustics in our living room was very bad at mid...highs a decade ago when those speakers were designed and built. No acoustic panels on the walls ceiling. 700 ms humps at mid...treble in EDT spectrum due to flutter echo between windows in front and back, and reflective ceiling did not help a bit. Resolution was low and sound dead and sometimes harsh. That time I was in war with bad room acoustics, and especially KS-1804 was close to final statement how this war is possible to win with quite small speaker, though changes are still very low. Bad room acoustics could beat you anyway depending on personal preferences, so acoustic treatment targeting low and flat EDT is more final solution giving freedom to select speaker design and drivers by overall quality and preferred sound nuances - not just strong directivity.

Leaking cardioid bass is very flexible application, but also very ineffective sucking all power and excursion available. Bass unit with 2x18" (KS-2125 and KS-1804 mk I+II) has been tested for example in 6m2 bathroom with brick walls, 15m2 concrete "bunker" and large living rooms total area >40m2. Room EQ and level adjustment for bass range is not required though the lowest axial mode could be boosted in small sturdy rooms and some narrow or small dips exist anywhere. Speaker location is not so critical than with omni or dipole to get balanced sound, but locations are worth to optimize anyway because these are not headphones either.

I haven't done much diy designs since acoustics of our living room was improved. Mostly commercial speakers without special directivity tricks. Therefore the latest history of my design does not reveal any progress or trend. I can and will design anything if someone else purchases components and builds cabinets :)
 
I can see some discussion about KS-585, KS-2125 and KS-1804.

Thank you for your comments. I didn't realise (until now) that you had responded.

Room acoustics [...] acoustic treatment targeting low and flat EDT is more final solution giving freedom to select speaker design and drivers by overall quality and preferred sound nuances - not just strong directivity.
That seems valid. I'd like a dedicated room. I currently rent a house with my GF, so there are lots of limits to what I can do with the room.

Leaking cardioid bass is very flexible application, but also very ineffective sucking all power and excursion available. Bass unit with 2x18"
I'm using my 18" as midbass drivers. I'm not trying to EQ the leaky box for flat response to 30Hz.

When testing, the 18" has no visible excursion.

I haven't done much diy designs since acoustics of our living room was improved.
Very interesting. Thank you for the info.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.