Aino gradient - a collaborative speaker project

Room response of AINOgradient

A single bass box and two towers (LM,HM,T) playing at WAF-accepted speaker positions in my living room (60cm from front wall, 2,5m apart, 2m from side walls, mic at 3m distance. Mic centerpoint a my listening spot and moved at max 1m radius in 3 dimensions, 13 measurements "center-weighted" and also shown the effet of moving in just one plane. 200ms rear window and 1/6 oct smoothing in all measurements.

I think that the power response in close to what is considered good for small rooms (my room 3,6 x 5m with many openings and a tilted ceiling, lots of windows. The measurements with low level of highs are from floor level!
 

Attachments

  • ainogice v62 room  aver of 13 200ms.png
    ainogice v62 room aver of 13 200ms.png
    45.7 KB · Views: 425
  • ainogice v62 room  all of 13 200ms.png
    ainogice v62 room all of 13 200ms.png
    124.6 KB · Views: 674
  • ainogice v62 room var dist same h and lat.png
    ainogice v62 room var dist same h and lat.png
    98.3 KB · Views: 421
  • ainogice v62 room var later same h same d 200ms.png
    ainogice v62 room var later same h same d 200ms.png
    81.2 KB · Views: 414
  • ainogice v62 room var height same lat same dist 200ms.png
    ainogice v62 room var height same lat same dist 200ms.png
    91.5 KB · Views: 416
That dip above at 250Hz was a phase issue. I carried the proto woofer and MTM pole outside today and set W to LM xo lower. I had to boost LM's low end a little. LM still has a dip around 220Hz, it is a floor bounce cancellation at my measuring distance 150 to 150cm.

there is a difficulty with phase matcing of the lowest xo of a 4-way speaker having Duelund(ish) xo. I have managed that by using LR4/LR2/LR4 orders instead of all being similar. But in the room - monopole radiation of the woofer is practically impossible to match flawlessly with a dipole lower mid. A sacrifice of my concept. Generally speaking (sub)woofer matching is very difficult below the Schroeder frequency of the room, because reflective modes make phase go all around.

I didn't save outdoor measurements, but here are the indoor measurements with speakers at listening position, like in the previous post. The last one is with LR playing together.

AINOgradient ice v7 indoor
 

Attachments

  • ainogice v7 room lr aver 5 200ms 16.png
    ainogice v7 room lr aver 5 200ms 16.png
    66.4 KB · Views: 68
  • ainogice v7 room l r lr 200ms.png
    ainogice v7 room l r lr 200ms.png
    51 KB · Views: 60
  • ainogice v7 room l decay spectr 6ms.png
    ainogice v7 room l decay spectr 6ms.png
    462.1 KB · Views: 60
  • ainogice v7 room l gd 6ms.png
    ainogice v7 room l gd 6ms.png
    49.2 KB · Views: 67
  • ainogice v7 room l 0¤ all indiv 200ms 13.png
    ainogice v7 room l 0¤ all indiv 200ms 13.png
    65.7 KB · Views: 402
Last edited:
Here I visualize, why I wanted speakers with high and even directivity index (DI) for my living room setup. The animation show sound power gradients interfenrence/summation in stereo and we notice the wider listening area (relative L and R power ratio being constant). Of course time coherence is maintained only at exactly same distance to L and R.

We can also see why it is important to toe-in high DI speakers!

The second major benefit of high directivity is that speakers make less floor, sidewall and ceiling reflections. These reflections degrade soundfield and are perceived mainly as deterioration of stero image and irregulaties in amplitude response.
 

Attachments

  • high vs low di stereo ani.gif
    high vs low di stereo ani.gif
    66.5 KB · Views: 118
Last edited:
Here I visualize, why I wanted speakers with high and even directivity index (DI) for my living room setup. The animation show sound power gradients interfenrence/summation in stereo and we notice the wider listening area (relative L and R power ratio being constant). Of course time coherence is maintained only at exactly same distance to L and R.

We can also see why it is important to toe-in high DI speakers!
If it were as easy as that in the real world. :eek:
Does your wider sweet spot keep its quality over the colored area? It is more like three (or five) single spots in my case where time/intensity trading works.
 
Rudolf, mys post #221 covers your question with an overay of 13 measurement made in a hemisphere around "the spot", radius 1m. This shows just amplitude response of two speakers playing same signal simultaneously (thus lots of lobing). I don't trust my measuring skills enough to analyze other aspects.

But, believe me - I and my friends can verify that the sound and image "hold together" exceptionally well with AINOs. I still don't have two bass boxes, so this is not final! I will report back when I have measured and auditioned a full pair. The pianist Petrescu is on tour now and my wife-in-law's son who is an EE (specialized in dsp) is coming here next week. I hope to get boxes ready for them.
360247d1373876103-aino-gradient-collaborative-speaker-project-ainogice-v62-room-all-13-200ms.png
 
Last edited:
Another simplistic presentation of benefits of dipole speakers.

A dipole radiates evenly front and back, but sideways is attenuated more than 10dB. A monopole radiates evenly front and sideways, but backways is attenuated more than 6-10dB.

For good sound perception in stereo, early reflections should be minimal and attenuated. Late reflections add "spaciousness" to the sound. Reflections naturally continue their paths and meet crossed walls, attenuate by distance and start to blend. The bigger and "softer" the room is, the better.

In these graphics, red arrows represent strong spl and blue attenuated spl. The lenth of an arrow represents delayed time for hearing the signal. Close o back wall is no-no for both (unless on-wall for monopoles).

Positioning of speakers in these graphs is ideal. Both surfer from placing them closer to walls, but in a different manner. Dipoles suffer most from placing too close to front wall (behind the speakers) and not toe-in'ing them. Monopoles suffer most from placing too close to side walls.
 

Attachments

  • wide room reflections d vs m.png
    wide room reflections d vs m.png
    47.8 KB · Views: 107
  • narrow room reflections d vs m.png
    narrow room reflections d vs m.png
    34.6 KB · Views: 114
Last edited:
Juhazi,
What do you think about the Fountek ribbon (with horn)?
What other tweeters have you used in the past and how does this horn ribbon compare to them?
You seem to have liked the phase match better at 4000Hz crossover; why did you end up going back to 3000Hz?

I am thinking of purchasing 3.5H to pair with Neo8s. Design will be similar to Greg's studio monitor in the picture. However my baffle will be much wider (think JBL 4345 enclosure), so I am thinking of using a horn/waveguide tweeter to avoid edge diffraction in high frequencies.
 

Attachments

  • DSCN0996-001.JPG
    DSCN0996-001.JPG
    89.4 KB · Views: 129
Most of my speakers have SEAS tweeters, mostly in waveguides (also non-diy ones). My friend has Fountek CD3 with diy-waveguides. I like these 3.5H a lot! sound is very clear and uncoloured - better than SEAS. Those who have listened to them, say the same. The measured response is extemely even and power response falls off nicely "by the book"

Recommended xo is 3500 or more, but waveguide helps. I lowered xo frequency because of the high-mid driver going above it's dipole-null. I am mostly concerned with directivity, distortion is however not a problem by my measurements and by listening. In a baffle like in your attachment, I can't say a word about diffractions and directivity. Perhaps you should do some simulations?

A Warning - so far I have spoiled two ribbons by messing with my amplifiers connections. I recommend a protective high-pass capacitor in series at least when you are just testing these!
 
Most of my speakers have SEAS tweeters, mostly in waveguides (also non-diy ones). My friend has Fountek CD3 with diy-waveguides. I like these 3.5H a lot! sound is very clear and uncoloured - better than SEAS. Those who have listened to them, say the same. The measured response is extemely even and power response falls off nicely "by the book"

Think most are unaware of how critical this actually is. Not just in OB designs but in all others too :)

Recommended xo is 3500 or more, but waveguide helps. I lowered xo frequency because of the high-mid driver going above it's dipole-null. I am mostly concerned with directivity, distortion is however not a problem by my measurements and by listening.

Waveguide helps by extending the tweeters power band. Acoustically correcting for an issue normally addressed with equalization, albiet incorrectly done because it can't resolve the off axis without adversly affecting the on axis. Nice

A Warning - so far I have spoiled two ribbons by messing with my amplifiers connections. I recommend a protective high-pass capacitor in series at least when you are just testing these!
Have 100uf in series with my AMT's crossed @2.2k active. Takes nothing but once and they are toast without. It was a handy size laying around ;)
 
I got the boxes back from the painter, but the colour went too blueish because of pearl-fillers used with lacquer. However we decided to keep them as such, and the painter didn't charge much for the work because of the colour mismatch.

One speaker is now wired finally, except for the protective cap for the tweeter. I fiddled with bass peq and xo in the room (final listening position) After these measurements I changed LM high-pass,eq between 200-500Hz and level for even better results. Outdoor and indoor sessions will follow this, but I think that I'm close...

Room mode bumps at 60, 120, 240Hz

Single speaker 1,5m response at 60ms and room response at 200ms (each line average of 6 measurements at 1m sphere)
 

Attachments

  • ainogice v7 room l r  lr aver 200ms 16.png
    ainogice v7 room l r lr aver 200ms 16.png
    54.6 KB · Views: 417
  • ainogice 81 room r 0¤ all indiv 60ms 13.png
    ainogice 81 room r 0¤ all indiv 60ms 13.png
    74.7 KB · Views: 430
Last edited:
I have exhaged some PMs with Greebster -thank you Mike! We are figuring out/trying to understand and improve dipole behavior of the mids. I chose the mtm arrangement after chosing to use 3" high-mids because of their smaller diameter. This creates some other problems - vertical lobing, interferences from the tweeter's body and the baffle/frame all these are mounted on. Frame is 8mm thick. The lower mid suffers from the nearness of high-mids, top of bass box and the post behind it. Post's cross section is elliptic to minimize diffractions.

Poor quality photos of ready-made AINOgradients attached!

I also attach graphs that may help to analyze the behaviour of my mids. All measured outdoors, version AINOgradient Ice v61. All elements attached all the time, just different electrical connections and rotation.

From these graphs we can also see what parametric eq and xrossover setting do. Dipole corrections loses a lot from efficiency! I use minimal boost and only in the region of highpass xo, to avoid digital clipping in minidsp.

My selection of amplifiers is working well. B&O Icepower 125ASX2 are drivin W and LM, 50ASX2 are driving HM and T. I use a Yamaha RXV-630 AV-receiver as preampifier and "mixer". SPL with AINOs is a bit louder than a pair of MarkK ER18DXTs at same volume setting. I don't have a meter but max spl is far above 100dB (my ears started hurting at -10dB) and there was no clipping, distortion or resonaces. LM membrane was moving some 4mm. I use LT type eq in minidsp for SEAS L26ROY woofers, something like -7dB between 30 and 70Hz.

Sorry, I don't want to say anything new about the sound, because I want to get more opinions.

W SEAS L26ROY in sealed 25 liter, polyfill (monopole)
LM Beyma 12MWNd nude, tilted (dipole)
HM Vifa NE95-4 nude, pair in series (dipole)
T Fountek NeoCD3.5H (monopole)
minidsp 4x10HD, acoustic xo's 150HZ LR4 800Hz LR2 3100Hz LR4
Bang&Olufsen Icepower 125ASX2+50ASX2 modules (total 4 channels/speaker)
 

Attachments

  • AINOgradient ready.jpg
    AINOgradient ready.jpg
    276.5 KB · Views: 863
  • AINOgradient ready back.jpg
    AINOgradient ready back.jpg
    309.8 KB · Views: 454
  • ne95 pair vs single 0 90¤ 3ms nosmo  text.png
    ne95 pair vs single 0 90¤ 3ms nosmo text.png
    102.9 KB · Views: 426
  • beyma 12 0¤ raw vs eq xo 90¤ 12ms nosmo.png
    beyma 12 0¤ raw vs eq xo 90¤ 12ms nosmo.png
    92.1 KB · Views: 116
Last edited:
More evaluation of MTM vs. MT arrangement

I played a little more with my old measurements, v61 outdoors. Here I use 6ms IR window. Pair of M is always darker line, single(lower) M is pale color.

There are differencies, but no consistency to my eyes.

However I can hear lobing effects with MTM when playing sine waves. More listening test will be done with pink noise, sine and music when I have an opportunity. SPL, distortion etc. -wise a serial pair and single are equivalent. A benefit from MTM is symmetric vertical directivity and perhaps some more "attack" in transients.
 

Attachments

  • ne95 pair vs single 30 90¤ 6ms 16.png
    ne95 pair vs single 30 90¤ 6ms 16.png
    66.8 KB · Views: 63
  • ne95 pair vs single 0 60¤ 6ms 16.png
    ne95 pair vs single 0 60¤ 6ms 16.png
    78.2 KB · Views: 63
  • ne95 single 0 90¤ 6ms nosmo.png
    ne95 single 0 90¤ 6ms nosmo.png
    92 KB · Views: 56
  • ne95 pair 0 90¤ 6ms nosmo.png
    ne95 pair 0 90¤ 6ms nosmo.png
    110 KB · Views: 63
After sleeping over previous post I noticed that I was mixing indoor and outdoor measurements. A new outdoor measuring session will come some day... and critical listening with friends. I don't want to cut off the upper 3" mid unless I am confident that it is better than mtm.

The problem with lobing and many other intermodulation effects is that they are difficult to catch with measurements. Even more difficult is to identify them (what is causing each) to be able to do something about them.

A dipole multiway speaker never behaves perfectly dipolic. Interferences can't be avoided, only compromized and minimized.
 
The weather doesn't permit outdoor measurements, so indoor... I made some changes to xo, its now all LR4 acoustic, xos 200,1000,3300. Horizontal response is the same as before, nothing new.

Front wall distance affects response a lot. Pseudoanechoid 6ms vs 60 and 100 cm to front wall 200ms - first attachment.

I measured vertical directivity of mtm and mt arrangement. Now I got some sensible results and I must say that mt has perhaps better vertical directivity. However mtm is reasonably even till +/- 10¤ which means +/- 50cm at my listening distance. I'll stick with mtm. I also mesurend +10¤ at 30¤ horizontal and it is just like 0¤, not shown here.

Room response average of 5 mic positions, R speaker at 60cm to wall and bass too heavy. Just to show how much this means.

Horizontal directivity, this time short IR only 4ms. This is mtm. Please notice that 2-3kHz is a bit hot here, but same range is cool at vertical offset - compensates total power response..
 

Attachments

  • ainogice mtmv91 mor rooml 60 100.png
    ainogice mtmv91 mor rooml 60 100.png
    62.8 KB · Views: 74
  • ainogice v9 mt vs mtm 0¤-5 -10 -15¤ 4ms nosmo.png
    ainogice v9 mt vs mtm 0¤-5 -10 -15¤ 4ms nosmo.png
    104.9 KB · Views: 72
  • ainogice v9 mt vs mtm +5 +10 +15 4ms 16.png
    ainogice v9 mt vs mtm +5 +10 +15 4ms 16.png
    76.4 KB · Views: 75
  • ainogice v9 room r 5 aver 200ms 16.png
    ainogice v9 room r 5 aver 200ms 16.png
    66.6 KB · Views: 72
  • ainogice mtm v9 in 0 15 30 45 60¤ 4ms 112.png
    ainogice mtm v9 in 0 15 30 45 60¤ 4ms 112.png
    77.6 KB · Views: 81
Last edited:
Front wall distance affects response a lot. Pseudoanechoid 6ms vs 60 and 100 cm to front wall 200ms - first attachment.

Hello!

60cm? You won't be able to get good midrange response that close IMO. It shows on the graphs, response flattens only around 1k. In my experience, dipoles are a b.... to place, where the midrange is great, bass is crap. Theoretically best place for bass is near side walls/corners. Further away one might manage to excite only one or two modes and use EQ. Or sit very, very close to the speakers..
 
Juha,
Why don't you ditch your dual cone mids and get yourself a pair of B&G Neo8S?
Then you would have the same set of drivers as myself (my Founteks are on the post).
Well, almost same: my 12" PA is Faital 12PR300.
You can do a quick OB vs sealed cabinet experiment and go from there.
I really like the driver placement on Greg's new studio monitor.