Rational Acoustics



Niick
April 28th, 2017, 12:02 AM
I was just reading a post ( http://forums.prosoundweb.com/index.php?topic=78045.0 ) where Tom Daley states "Your ear has about a 1/6 octave critical bandwidth" ......

Can anybody elaborate on this, or possibly reference some reading material on the matter.

Arthur Skudra
April 28th, 2017, 12:47 AM
I believe what he's referring to is the ability of the human ear to perceive frequency resolution. Not sure if I'm in agreement with the 1/6th octave resolution, personally I think it's as high as 1/24th octave. Opinions vary widely on this topic. FYI, previous generations of Smaart's transfer function had a FPPO limitation of 24 points per octave. Since version 7, the MTW is effectively 80 FPPO, far greater resolution than the human ear can resolve.

merlijnv
April 28th, 2017, 01:19 AM
I was just reading a post ( http://forums.prosoundweb.com/index.php?topic=78045.0 ) where Tom Daley states "Your ear has about a 1/6 octave critical bandwidth" ......

Can anybody elaborate on this, or possibly reference some reading material on the matter.

https://en.m.wikipedia.org/wiki/Critical_band


Verzonden vanaf mijn iPad met Tapatalk

Niick
April 28th, 2017, 01:53 PM
Thanks Arthur and Merlin,

That's what I figured. Thanks for the confirmations.

So, basically, this seems to be an evolving theory with early documentation stating 1/3, later 1/6, 1/12, Dr. Toole has stated 1/20th (or greater even)....

So this seems to be, like many things, an area of human auditory perception that is still a bit fuzzy, with no hard, definitive absolutes that are unamiously agreed upon by all.

This reminds me of a topic that has interested me for quite some time now. The topic of frequency dependent time windowing, and how this seems to correlate more closely to human subjective perception than the classic "Time blind" RTA.

Smaart calls it MTW, SysTune calls it TFC, a very similar thing is occurring either way. Shorter time constants for higher frequencies.

I asked Bruce Olson if he had any papers or studies on this topic, and he told me that to his knowledge, there aren't any "official experiments" that are documented that explore this topic. It seems to be all anecdotal evidence that supports this claim. That's not to say I don't believe it. I do. I've found Smaart's MTW to be INCREDIBLY useful for what I do, and it has more to do with the way the coherence trace is effected as compared to a fixed size FFT than it does with the constant bandwidth resolution.

One of the things that drives me is the quest to further my own understanding of how to correlate objective data with subjective perception.

The MTW thing intrigues me greatly. I wonder, do you guys think that ideally, differnt MTW TC sizes would be used for smaller environments (like a car) than would be used for larger environments (like a concert venue)? I know that Rational Acoustics isn't likely to introduce an adjustable MTW anytime soon, but hypothetically, IF it could be done, do you guys think it would be useful, or do you think that the same TCs that would be ideal in a larger space would then also apply to a smaller space?


I hope that made sense.

Arthur Skudra
April 28th, 2017, 02:29 PM
The MTW thing intrigues me greatly. I wonder, do you guys think that ideally, differnt MTW TC sizes would be used for smaller environments (like a car) than would be used for larger environments (like a concert venue)? I know that Rational Acoustics isn't likely to introduce an adjustable MTW anytime soon, but hypothetically, IF it could be done, do you guys think it would be useful, or do you think that the same TCs that would be ideal in a larger space would then also apply to a smaller space?

I really am not privy to the "secret sauce" on how Rational does their MTW, but my *assumption* is that the combination of TC's are dependent more on frequency, resolution and size of window regardless of room size being measured. Whether you listen on a pair of headphones, or in a car, or in a large performing arts hall, we hear the same bandwidth of response regardless, and can measure it the same as well. There is a mistaken notion in some of our industry that you need to be physically further away from a subwoofer to be able to "hear" the low frequencies, or to use a large diaphragm mic to pick up low frequencies "better". Nothing can be further from the truth if you understand the physics involved.

I think the major difference between measuring an environment such as a car or small recording studio versus a large hall has more to do with several factors:
Modal behaviour and resonance
Distance from speaker and the type of speakers used
Reflections within the space and their time relationships
And I'm sure there's plenty of other factors as well...

That said, I'm looking forward to optimizing my car audio system and see what similarities and differences I can glean from doing so compared to the larger rooms I do on a regular basis....I got myself a nice 8 channel Match DSP/Amplifier and some Morel Hybrid 402's that I need to find the time to install! Too many "work" projects, too little time for my toys!!! :p

Niick
April 30th, 2017, 01:01 AM
I really am not privy to the "secret sauce" on how Rational does their MTW, but my *assumption* is that the combination of TC's are dependent more on frequency, resolution and size of window regardless of room size being measured. Whether you listen on a pair of headphones, or in a car, or in a large performing arts hall, we hear the same bandwidth of response regardless, and can measure it the same as well. There is a mistaken notion in some of our industry that you need to be physically further away from a subwoofer to be able to "hear" the low frequencies, or to use a large diaphragm mic to pick up low frequencies "better". Nothing can be further from the truth if you understand the physics involved.

I think the major difference between measuring an environment such as a car or small recording studio versus a large hall has more to do with several factors:
Modal behaviour and resonance
Distance from speaker and the type of speakers used
Reflections within the space and their time relationships
And I'm sure there's plenty of other factors as well...

That said, I'm looking forward to optimizing my car audio system and see what similarities and differences I can glean from doing so compared to the larger rooms I do on a regular basis....I got myself a nice 8 channel Match DSP/Amplifier and some Morel Hybrid 402's that I need to find the time to install! Too many "work" projects, too little time for my toys!!! :p Right on. :)

As I've never personally used Smaart (or any dual FFT) for anything other than cars (and measuring raw drivers on a big baffle), I can tell you what some of things you'll encounter will be.

In a car, the low frequencies are SUPER easy to measure, super high coherence, clean phase trace. The HF is difficult. Low coherence (using MTW) HORRIFIC phase trace. The phase trace in the HF in a car is basically all noise...... using any measurement platform OTHER than smaart. With Smaart, somehow, they've optimized the smoothing algorithms to, almost magically, allow one to pick out the driver's phase response from that of the reflections. Of course, heavy smoothing should always be used with caution, and (I personally believe) avoided unless one has good reason. And in a car, using Smaart, one has very good reason to engage heavy phase smoothing from time to time. The only other way I can measure phase in cars in the HF is by using SysTune's TFC window, but that is FAR more tedious, time consuming, and error prone than simply choosing 1/3 smoothing in Smaart. It is my belief that Smaart's phase smoothing algorithm is the "unsung hero" of the software, as one would never know that it was capable of working such wonders until one actually tries to measure speakers in a very small reverberant space. I know this works because I regularly measure tweeters and midrange outside the car before installation, either mounted on a tall pole (tweerers) or large baffle (open back midranges), then, when installed in the car, I can simply engage some phase smoothing and, almost like magic, there's the response, "picked out" of the "noise"

Another thing that happens in cars is the coherence tends to really drop in the HF. Using MTW. Switch to a 32k FFT and the coherence is more or less rock solid at the upper frequencies. Coming from SysTune, this really tripped me up at first, until I learned that the lower TCs of the MTW at higher frequencies was the reason. I simply manually selected small FFT sizes and just like the MTW, the coherence went from rock solid (@32k FFT) to relatively "poor" using one of the very small FFT sizes. So this opened up a whole new can of worms for me.... Which I like :)

In a car, spatial averaging is critical to making useful measurements above 100Hz or so. I use a mic array containing 7 mics in a semi-circle slightly wider than a human head. Polar averaging, coherence weighted.

I use the IR/ETC, coherence, and comparisons to the Quasi-anechoic data acquired from my baffle/pole setup to aid in making decisions regarding speaker placement, tweeter aiming.

I'd love to know how you get along when you do go thru with the project. Be sure to keep us posted on the project.