Wave Field Synthesis

Thread Starter

kahnzo

Joined Jan 4, 2022
13
I'm interested in making an inexpensive version of this
https://empac.rpi.edu/program/research/wave-field-synthesis

I saw an article written by Robert Keim which uses PWM and a low pass filter for audio in place of a DAC.
I would like to create an audio device that is recognized by a computer running MAX MSP and SPAT (see link) that will output as many channels of audio as I can. So computer to device - dac - amp - speakers. I also, eventually want to be able to put little modular speakers anywhere in a room and have directional control over those speakers rotate and tilt.

I have the typical budget of some who wants to experiement at home.
 

Audioguru again

Joined Oct 21, 2019
6,674
It produces stereo with left channel, right channel and every space in between with 31 amplifier/speakers.
One sub-woofer produces low frequencies and only two tweeters produce stereo with two amplifiers.
 

DickCappels

Joined Aug 21, 2008
10,152
@kahnzo
The linked page starts out with "The EMPAC Wave Field Synthesis Array is a 558-channel discrete loudspeaker array designed to take advantage of the added sound-stage dimensionality made available through the use of wave field synthesis (WFS)."

Do you have a question?
 

Thread Starter

kahnzo

Joined Jan 4, 2022
13
Thank you audioguru and DickCappels for responding to my post.

Some perspective on me. In 2020 I turned 50, learned I had cancer, went through a year of radiation and chemo, wife had a heart attack and is doing well, so far, and COVID concerns has made my 10 year old resiliant social daughter feel very isolated. A little fear has inspired me to find out if this old dog might be able to learn some new tricks.

I think I understand what audioguru is implying. Are you saying that I'd be better off buying a very nice pair of stereo speakers and getting a nice stereo image that includes the high end than trying to build a WFS array. There is absolutely no doubt that that is true as a listener. I was looking at it from the perspective of someone who would like to experiment with different speaker configurations for applicates like art installations, vr gaming, and for mixing in surround, as well as attempts to replicate a Wave Field Synthesis array. I also want to learn things along the way, while not breaking the bank on the way to a probably disappointing result. Thanks for the warning.

DickCappels, as you said, I didn't really have a question.

My primary question is a follows: The EMPAC array uses DANTE's version of AES67 audio over ethernet standards to feed the speakers. What approach could I take to make a homebrewed version of an AES67 device whose only role is output. The advantage of using AES67 compliant devices is that the virtual soundcards are already available, and I could leverage what EMPAC has already done. The other advantage is that I would learn some valuable knowledge by getting my hands dirty setting up a network.

Or perhaps I can avoid using a network altogether. Given that you can convert PWM into an audio signal, perhaps I set up a number of microcontrollers akin to the link below https://electronics.stackexchange.com/questions/93639/microcontroller-with-lots-of-pwm-outputs

I know these questions aren't very specific. Sorry. I really am just looking for some general ideas and warnings.

Thank you for your time.
 

Audioguru again

Joined Oct 21, 2019
6,674
Kahnzo, I am glad that you and your wife survived your medical ordeals. I am also an old dog (76) and I survived a heart attack 12 years ago, cataracts in my eyes and a couple of other ordeals and now I am young again.

Since we have two ears then we can hear an entire sound stage from only two speakers (stereo) in front of us.
Maybe the many speakers in an array provides a stereo sound stage to people from another planet who have only a single ear?
One of my hearing aids (I have old age high frequencies hearing loss) has failed so I use the remaining working one only half the time so I hear low frequencies in stereo.

Guess why we have two eyes. Imagine a stop sign that actually has 31 stop signs in a row so that a person with a single eye can determine the distance to stop.
 

Thread Starter

kahnzo

Joined Jan 4, 2022
13
Kahnzo, I am glad that you and your wife survived your medical ordeals. I am also an old dog (76) and I survived a heart attack 12 years ago, cataracts in my eyes and a couple of other ordeals and now I am young again.

Since we have two ears then we can hear an entire sound stage from only two speakers (stereo) in front of us.
Maybe the many speakers in an array provides a stereo sound stage to people from another planet who have only a single ear?
One of my hearing aids (I have old age high frequencies hearing loss) has failed so I use the remaining working one only half the time so I hear low frequencies in stereo.

Guess why we have two eyes. Imagine a stop sign that actually has 31 stop signs in a row so that a person with a single eye can determine the distance to stop.
Audioguru, I'm enjoying my foray into mixing surround sound using Ambisonics. I've also wanted to explore Head Related Transfer Functions to try to get our good old two ear system dialed in. Our brains do some amazing things with those two inputs. Glad to hear that you're young again. Keep on truckin.
 

DickCappels

Joined Aug 21, 2008
10,152
Welcome to allaboutcirucits.com!

The travails of growing old. My grandmother used to warn me about these things and now I know that her warnings were true. I am glad that your family and you are still intact

I'm up there in years too and retired from computer industry and then electronics was fun again! No deadlines except those I chose, complete freedom to spend time looking into things that would not have been related to my work.

We will have to wait for somebody who know more about ethernet, but I have had some experience generating waveforms with PWM with a microcontroller.

I think you will be hard pressed get quality audio using pulse-width modulation and even with digital-to-analog converters you would have to be very careful in order to get good audio out of a microcontroller.

For the PWM approach you might not be able to find a fast enough PWM system -that depends a lot on how big your output filter will be but I think you would find a good sampling rate of over 48 kHz to be very hard to do, especially given the variable latency of interrupts at this kind of speed.

An MP3 chip (with which I have had no experience) can run much faster and control the fetching of data from memory and the decoding process faster and with more regularity of timing than you could expect from a microcontroller. For more than twenty years there have been examples on the web that might give you some insight into how easy it can be to control MP3 players.

https://www.sparkfun.com/search/results?term=mp3
 

Thread Starter

kahnzo

Joined Jan 4, 2022
13
Welcome to allaboutcirucits.com!

The travails of growing old. My grandmother used to warn me about these things and now I know that her warnings were true. I am glad that your family and you are still intact

I'm up there in years too and retired from computer industry and then electronics was fun again! No deadlines except those I chose, complete freedom to spend time looking into things that would not have been related to my work.

We will have to wait for somebody who know more about ethernet, but I have had some experience generating waveforms with PWM with a microcontroller.

I think you will be hard pressed get quality audio using pulse-width modulation and even with digital-to-analog converters you would have to be very careful in order to get good audio out of a microcontroller.

For the PWM approach you might not be able to find a fast enough PWM system -that depends a lot on how big your output filter will be but I think you would find a good sampling rate of over 48 kHz to be very hard to do, especially given the variable latency of interrupts at this kind of speed.

An MP3 chip (with which I have had no experience) can run much faster and control the fetching of data from memory and the decoding process faster and with more regularity of timing than you could expect from a microcontroller. For more than twenty years there have been examples on the web that might give you some insight into how easy it can be to control MP3 players.

https://www.sparkfun.com/search/results?term=mp3
Thanks so much, that really seems like an interesting approach. The reason that I suggested PWM is actually from this very site. https://www.allaboutcircuits.com/technical-articles/turn-your-pwm-into-a-dac/

I plan on having a lot of control over the signal that gets sent to the microcontroller so it don't have to do much work. Use software crossovers in the PC so that each PWM is working on only one part of the signal and pre-apply any low pass filter.

Again, I really have no idea what the possibilities are, but I have heard some pretty nice results. In fact, people have used HiFiBerry Amp2 with a Raspberry Pi to make audiophile quality music competitive with far more expensive gear.
 

Audioguru again

Joined Oct 21, 2019
6,674
The HIFI Berry Amp2 produces only 14W per channel into 4 ohms with distortion at 0.1% or less when its power supply is 12VDC.
They do not say the distortion over all frequencies so it might produce intermodulation artifacts at high audio frequencies.

A class-D amplifier IC uses PWM.
There are many higher power class-D amplifiers with better specs available that are already designed and made. An example is the TPA3116 by Texas Instruments.
 

drjohsmith

Joined Dec 13, 2021
852
This does sound , no pun intended , interesting,
Have you looked at the Teensy
https://www.pjrc.com/
he has a great set of audio libraries,

Multiple speakers were an idea I remember from the 70's,
I just cant seem to find a link

The idea was rather like a phased array radar
Multiple drivers are user, to make a wave front

control the phase and amplitude and content sent to each speaker,
 

Thread Starter

kahnzo

Joined Jan 4, 2022
13
This does sound , no pun intended , interesting,
Have you looked at the Teensy
https://www.pjrc.com/
he has a great set of audio libraries,

Multiple speakers were an idea I remember from the 70's,
I just cant seem to find a link

The idea was rather like a phased array radar
Multiple drivers are user, to make a wave front

control the phase and amplitude and content sent to each speaker,
Ambisonics - I used that as a VST plugin in Reaper - it works very well. It's making a comeback due to it being nice to mix down to headphones for gaming applications.

I have definitely checked out Teensy. They have a great community of makers.
 

BobTPH

Joined Jun 5, 2013
8,813
@kahnzo,

Can you explain, in your own words what this paragraph means:

Wave field synthesis (WFS) is a spatial audio rendering technique that places virtual sound sources in real space. The synthesis is precise and physically correct. You cannot distinguish the synthesized source from a “real” wave field at frequencies below the upper frequency limit of the system. Human auditory perception is the same as it would be in a real sound field.
Bob
 

Thread Starter

kahnzo

Joined Jan 4, 2022
13
@kahnzo,

Can you explain, in your own words what this paragraph means:



Bob
I have never experienced it, so what I think it means and what they have actually accomplished could be vastly different. A few years ago, several years in a row, I'd attend the Rocky Mountain Audio Fest in Denver Colorado. You could go into different rooms and listen to audiophile equipment that was extraordinary. Unfortunately, I don't think they'll be having it any more. If you sat in a sweet spot, you could experience truly amazing stereo imagery, where you felt that you could determine the precise location of the sources of the makers of the sound.

Recently, I've been using Reaper with and Ambisonics plugin and my inexpensive home theater system to acheive a similar effect. It is a very wonderful experience. Sit in your sweet spot, and move the musicians around you putting them in specific locations. Harmonizing vocalists, a string quartet, a drum track, as if I was sitting on the kit. Unfortunately, the sweat spot is very narrow.

What I think Wave Field Synthesis does is a tradeoff - you lose high frequency information, but gain very precise location information and you don't need to sit in a specific sweet spot. I'm familiar with MAX/MSP and was interested in testing out SPAT to get a better understanding of what is really going on and experience this myself. Ultimately, I just want to learn some skills that may have other applications, so having a project like this provides that incentive to learn those skill.

Keep in mind, what I really want to do is be the artist creating the sounds, not the engineer creating the device - but I want to know what the engineer knows too.

Thank you for taking the time to respond.
 

BobTPH

Joined Jun 5, 2013
8,813
Okay. I believe they are talking about re-creating the exact acoustic waves of an ensemble on stage by using a phased array of speakers. The content sent to each speaker would need to be precomputed with some serious compute power and played with perfect synchronization, down to the microsecond to maintain the exact phase relationships. Not an easy task, and not feasible using separate micros, it all need to be run off a single clock.

Bob
 

Audioguru again

Joined Oct 21, 2019
6,674
Kahnzo, you say that you want to create the sounds then you will be making synthesized noises, not live or recordings of musical instruments and real people singing.
The array of speakers simply positions sounds left to right in front of you. If you wish to reposition an instrument or singer then you need a separate recording of it and a recording of the group without it.
 

Thread Starter

kahnzo

Joined Jan 4, 2022
13
Okay. I believe they are talking about re-creating the exact acoustic waves of an ensemble on stage by using a phased array of speakers. The content sent to each speaker would need to be precomputed with some serious compute power and played with perfect synchronization, down to the microsecond to maintain the exact phase relationships. Not an easy task, and not feasible using separate micros, it all need to be run off a single clock.

Bob
The idea was networked microcontrollers with a master clock. I had assumed that this issued wouldn't be as big of an issue as you're suggesting. Please give me your suggestions on things to consider. And yes, a beefy computer, or the rental of some cloud computing power is anticipated for preprocessing. As I said, I'm excited about all the crazy issues that will arise. Obviously I wasn't taking the issue of master clock as seriously as I should have.
 

Thread Starter

kahnzo

Joined Jan 4, 2022
13
Kahnzo, you say that you want to create the sounds then you will be making synthesized noises, not live or recordings of musical instruments and real people singing.
The array of speakers simply positions sounds left to right in front of you. If you wish to reposition an instrument or singer then you need a separate recording of it and a recording of the group without it.
I have a home studio. I have no interest in live recordings.
I do have taper friends who use blumlein pairs for beautiful stereo recordings.
I have recorded people each isolated from one another at the same time as well.
 
Last edited:

Audioguru again

Joined Oct 21, 2019
6,674
How can a computer identify and select all the frequencies including all the harmonics of one instrument or singer in an ensemble then cancel those sounds from all the other channels??

Blumlein pairs with two matched front and rear microphones is used for stereo with only two speakers not used with an array of speakers.
 
Last edited:
Top