Dispersion - the art of measurement

[siddharthdas]

@Vineethkumar01 recently published a bunch of specs on his DIY build - https://www.hifivision.com/threads/hello-from-a-speaker-builder.91412/post-1023234 - see bottom two measurements for dispersion.



You can see how angular horizontal deviation from the front of the speaker changes response. Around 50 Hz, dispersion is almost 360 degrees (see right diagram right hand axis), i.e. you can hear exactly what you hear in the front, all around the room (at the same distance from the speaker). At 20 khz, as you move away from the direct listening axis, the loudness drops: +/- 30 deg looks great, +/- 60 deg is -24 db, a significant drop. There's also a narrow neck between 200-500 Hz, which is probably the crossover point from bass to mid. While @Vineethkumar01 can tell us further about the actual details, just from the figure alone it appears that the directivity change from bass to upper bass is quite sharp, and somewhat uniform thereafter.

Gainphile has an amazing blog with all sorts of directivity plots. http://gainphile.blogspot.com/2010/11/s15-econowave-dsp.html - see how this speaker has excellent dispersion almost throughout the spectrum - courtesy of horn loading.

Taking these measurements is a bit of an art and I hope @Vineethkumar01 will be so kind as to tell us how he does them. Some still shots of him performing measurements would also be excellent (maybe some of us can get together and document the process with images). I personally thing it will be super helpful to the entire community. There's so much conversation around super subjective stuff. This side of the art needs some more engagement than we see. DiyAudio is great - although there too I'd love to see some simple how to guides. It's trickier than meets the eye to parse these forums form information - not to mention, very slow!!

 

[Vineethkumar01]

@siddharthdas: I was once more active on this forum, trying to spread more awareness and share knowledge, trying to get feedback, engage in more discussions about measurements and hoping to see some interest in the measurements-oriented speaker design process. I was literally fed up of seeing designs solely based on just on-axis measurements only. However, there was little to no interest in such discussions. Some people also were of the opinion that measurements can go to hell and what matters is whether audio equipment sounds nice or not. I completely agree that the end goal of all audio-related pursuits, for most people, is to get a system that sounds best as per one's personal tastes. However, my way to it was through a combination of objective measurements to help with the design process and root causing of potential issues that one may face and using my own subjective preferences to fine-tune and finalize a design. I didn't see much interest in the former part. So I gave up and shut up.
However, it is finally nice to see some activity & interest on the measurements front. With that out of the way, I wouldn't consider the above plot as the best result (horizontal directivity-wise) that I personally achieved in a speaker configuration. I would consider that as following 2 way system. It is a 2 way with Satori woofers and ST260 horn on a compression driver taking over duties above 1kHz.


Take a look at the horizontal directivity on the left-hand side. That is near constant directivity with close to 90 degrees pattern control in the horizontal plane with gentle decrease in directivity index towards the lower frequencies eventually reaching omni directional radiation pattern in the bass region. That is what I would call state of the art. Also the above results will be even better with probably a nice 12inch woofer like the Faital pro 12PR320 or similar.
While I do agree that econowave type designs linked in the previous post were nice for the time they were introduced, we have moved beyond those in terms of achiveble best results.

I would also not worry much about the chicken neck type pattern around 1kHz region in the vertical directivity plot as long as the neck is about 15 to 20 degrees wide and the listening position is about 2+m away from the speaker (which is my personal situation). That is typical of a non coaxial design. Also our hearing is far more sensitive to variations in directivity/responses in horizontal plane compared to the vertical plane.
Now how does this all sound. These speakers image like crazy. One can pick out instruments hanging around in 3D space. The sound stage width and depth have reasonably good dimensions. The only caveat here is that they will call a little bit of attention to the region where the horn is operating due to the crazy amount of details being presented with some compression drivers. And that is an inevitable result of increased acoustic resolution brought in by the horn and may not be to everyone's personal tastes. To improve (not the right word to use here but still) the state of things the only way is to passively or actively increase the directivity at lower end to something like a cardioid/super/hyper cardiod radiation pattern with the woofer-midrange driver. That is what systems like the Dutch & Dutch 8C and the Kii 3 do. Near full range constant directivity...

I do measurements of drivers as well as the overall speaker as per the guidelines mentioned for generating CEA 2034 spinorama plots.

CEA 2034-A-2015 (ANSI) - Standard Method of Measurement for In-Home Loudspeakers

cea20342015ansi-Standard Method of Measurement for In-Home Loudspeakers-This standard describes how to determine the frequency response, directivity and maximum

webstore.ansi.org

To take measurements at home, I use by ESI sound card, sonarworks soundID reference mic, loopback cable, DIY turntable setup and REW. For post-processing the data I use VituixCAD. Very detailed explanations on how to take polar measurements are explained here: https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_help_20.html

Here is me taking polar measurements of the bass cabinet. I take 19 (0 degrees to 180 degrees in 10 degree increments) measurements in the horizontal plane and 18 measurements in the vertical plane

Here is me taking measurements of the ST260 horn (Note the rotating turntable at the bottom)


In short, I manage with what I have. I have been begging people to help me with making a proper turn table with lazy suzan bearings etc capable of handling about 40 kg load etc.. But no one is interested or they don't want to help me make one..

 

[grindstone]

Speaking only for myself, you're not being ignored, just waiting for you to go through your process and figure out what you like and "season" some more about your preferences. We are all indebted to Earl, Kimmo, Harmon/Sean/Floyd, Marcel, on and on. The thing is the sound. Personally, I'm waiting for you to get done messing around with little speakers and get back to that 15 (at which time I predict a fork in the road, one way or the other). Can't wait

Vcad does a lot of important stuff in a fantastically-integrated manner and Kimmo's approach isn't the debate, but the wonderland of loudspeakers is boundless. Things that are interesting to me, vcad doesn't even do. But it's like when you first get to work as a young person--you ask the old guys what to do and they say "do this" (do the Harmon stuff and use vcad). What matters (to me as a forum member) is the quirks and individual crusades and dogged pursuit of whatever tangent engages a person. In this way, you have enriched the community (and diyA) inasmuch as your crusade is science that uses/overlaps with your background. That's what's interesting to me--people doing things because They Must And the interests change over time. To see how people find their way is the fun--and in that, you have provided much. It's just a different way for everybody. Sometimes people like things for reasons I can't explain (me included)--that's Interesting. We all have biases about what topology we choose, what's practical for us in our own spaces, etc.--how we solve our own problems to find out how much of what is enough for us (ie where the compromises are), etc. I understand that you may achieve the participation you desire elsewhere, but I am glad that you've not bailed entirely. This place has "another gear" that some others don't and allows (fosters?) melding perspectives in a way I've not seen anywhere besides one other mailing list. Happy you are engaged and busy, just hope you get time to listen once in a while Be well!

 

[Vineethkumar01]

Thanks a lot @grindstone. I had started with that 15inch driver. I was about to get a box built for it. But eventually, people who agreed to help me with that project bailed out. After a while of disappointment, I just put those ideas in cold storage where they stay now. But all the time that thought and longingness comes up to hear a bit of a big speaker. Try to put that 15PR400 driver I have in a BLH/TL enclosure and partner it with a good horn for a big format 2 way. I just hope and wish that I get a chance to do that speaker, hopefully soon..

 

[grindstone]

Now we're talking! It's ok, stuff happens. Might be cosmically ordained to be in this order You will have lots of other experiences (and skills!) to draw-upon if and when the time comes. And big stuff isn't cheap (although I see you already have the crazy-steep driver for that top). Your enthusiasm is contagious, keep it up, man (it got me to download the step file for that ugly BLD horn because it's unloved...but it's Really ugly) Congratulations on what you've achieved in a relatively short time--you did a LOT of very good work and I have no doubt it sounds pretty great.

Apologies to Sid for (temporary) diversion.

 

[Passive_audio_enthusiast]

@siddharthdas: I was once more active on this forum, trying to spread more awareness and share knowledge, trying to get feedback, engage in more discussions about measurements and hoping to see some interest in the measurements-oriented speaker design process. I was literally fed up of seeing designs solely based on just on-axis measurements only. However, there was little to no interest in such discussions. Some people also were of the opinion that measurements can go to hell and what matters is whether audio equipment sounds nice or not. I completely agree that the end goal of all audio-related pursuits, for most people, is to get a system that sounds best as per one's personal tastes. However, my way to it was through a combination of objective measurements to help with the design process and root causing of potential issues that one may face and using my own subjective preferences to fine-tune and finalize a design. I didn't see much interest in the former part. So I gave up and shut up.
However, it is finally nice to see some activity & interest on the measurements front. With that out of the way, I wouldn't consider the above plot as the best result (horizontal directivity-wise) that I personally achieved in a speaker configuration. I would consider that as following 2 way system. It is a 2 way with Satori woofers and ST260 horn on a compression driver taking over duties above 1kHz.
View attachment 73764

Take a look at the horizontal directivity on the left-hand side. That is near constant directivity with close to 90 degrees pattern control in the horizontal plane with gentle decrease in directivity index towards the lower frequencies eventually reaching omni directional radiation pattern in the bass region. That is what I would call state of the art. Also the above results will be even better with probably a nice 12inch woofer like the Faital pro 12PR320 or similar.
While I do agree that econowave type designs linked in the previous post were nice for the time they were introduced, we have moved beyond those in terms of achiveble best results.

I would also not worry much about the chicken neck type pattern around 1kHz region in the vertical directivity plot as long as the neck is about 15 to 20 degrees wide and the listening position is about 2+m away from the speaker (which is my personal situation). That is typical of a non coaxial design. Also our hearing is far more sensitive to variations in directivity/responses in horizontal plane compared to the vertical plane.
Now how does this all sound. These speakers image like crazy. One can pick out instruments hanging around in 3D space. The sound stage width and depth have reasonably good dimensions. The only caveat here is that they will call a little bit of attention to the region where the horn is operating due to the crazy amount of details being presented with some compression drivers. And that is an inevitable result of increased acoustic resolution brought in by the horn and may not be to everyone's personal tastes. To improve (not the right word to use here but still) the state of things the only way is to passively or actively increase the directivity at lower end to something like a cardioid/super/hyper cardiod radiation pattern with the woofer-midrange driver. That is what systems like the Dutch & Dutch 8C and the Kii 3 do. Near full range constant directivity...

I do measurements of drivers as well as the overall speaker as per the guidelines mentioned for generating CEA 2034 spinorama plots.

CEA 2034-A-2015 (ANSI) - Standard Method of Measurement for In-Home Loudspeakers

cea20342015ansi-Standard Method of Measurement for In-Home Loudspeakers-This standard describes how to determine the frequency response, directivity and maximum

webstore.ansi.org

View attachment 73766

To take measurements at home, I use by ESI sound card, sonarworks soundID reference mic, loopback cable, DIY turntable setup and REW. For post-processing the data I use VituixCAD. Very detailed explanations on how to take polar measurements are explained here: https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_help_20.html

Here is me taking polar measurements of the bass cabinet. I take 19 (0 degrees to 180 degrees in 10 degree increments) measurements in the horizontal plane and 18 measurements in the vertical plane

Here is me taking measurements of the ST260 horn (Note the rotating turntable at the bottom)
View attachment 73767

In short, I manage with what I have. I have been begging people to help me with making a proper turn table with lazy suzan bearings etc capable of handling about 40 kg load etc.. But no one is interested or they don't want to help me make one..

how do you remove the room from your measurements ?

 

[Vineethkumar01]

how do you remove the room from your measurements ?

For each driver/speaker, I take gated farfield measurements in a reasonably large living room. Then I take nearfield measurements. After baffle step correction of the nearfield, VituixCad can blend the two curves in an appropriately chosen band of frequencies to create a quasi-anechoic measurement.

 

[Kannan]

I take gated farfield measurements in a reasonably large living room

this is the right way, even better in an open area.

 

[keith_correa]

how do you remove the room from your measurements ?

How to make quasi-anechoic speaker measurements/spinoramas with REW and VituixCAD

Quasi-anechoic measurements are a way to capture the 'real' frequency response of a speaker without the influence of a room. They are called 'quasi-anechoic' because they have the potential to get you really close to the results one might achieve with a non-reflective environment like an...

www.audiosciencereview.com