Hey Everyone,
I work at a Cognitive Science lab and we'd like to run auditory experiments using a 7x7 wall of speakers all of which can individually be controlled.
I am tasked with designing this but I am completely new to the topic - I have done some DIY with supervision (e.g. a cracklebox and an Atari Punk Console), so I have indeed soldered in my life but I have no idea where to start with this.
I imagine I'll need some scaffolding to put the speakers on, a bunch of quite small speakers and a method to control them. Can you please point me into directions where to start?
I work at a Cognitive Science lab and we'd like to run auditory experiments using a 7x7 wall of speakers all of which can individually be controlled.
I am tasked with designing this but I am completely new to the topic - I have done some DIY with supervision (e.g. a cracklebox and an Atari Punk Console), so I have indeed soldered in my life but I have no idea where to start with this.
I imagine I'll need some scaffolding to put the speakers on, a bunch of quite small speakers and a method to control them. Can you please point me into directions where to start?
What is the budget? Does the test wall have to be movable, or can it be attached to the floor or building wall?
What is the necessary loudness level? What are the test signals? Will the speakers just be on/off, or will the
loudness level change? Does this have to be controlled by a computer and take data, or can you use a simple
box with knobs? If you need to use a computer, can you write software?
What is the necessary loudness level? What are the test signals? Will the speakers just be on/off, or will the
loudness level change? Does this have to be controlled by a computer and take data, or can you use a simple
box with knobs? If you need to use a computer, can you write software?
Last edited:
Thank you for the questions!
Budget: we'd like to see the price tag for several options going from the cheapest possible to maybe the twice of that amount.
Movability: once it is assembled it can stay in place for a while but when we're not using it for a time we'd be disassembling it. It cannot be attached to the floor but it can to the wall.
Loudness: its going to be located in a soundproof room so it doesn't have to be particularly loud. I'm not sure how to put it into more exact terms.
Test signals: they will mostly be just 'noises' and everyday sounds (like a door thud, a dog bark, stuff like that) but later on we plan to use pure tones as well. So this means the loudness level does change with time.
It needs to be controlled from a computer, we're hoping to use MATLAB & PsychToolbox. We can program those or other interfaces as well if needed.
I appreciate your help!
Budget: we'd like to see the price tag for several options going from the cheapest possible to maybe the twice of that amount.
Movability: once it is assembled it can stay in place for a while but when we're not using it for a time we'd be disassembling it. It cannot be attached to the floor but it can to the wall.
Loudness: its going to be located in a soundproof room so it doesn't have to be particularly loud. I'm not sure how to put it into more exact terms.
Test signals: they will mostly be just 'noises' and everyday sounds (like a door thud, a dog bark, stuff like that) but later on we plan to use pure tones as well. So this means the loudness level does change with time.
It needs to be controlled from a computer, we're hoping to use MATLAB & PsychToolbox. We can program those or other interfaces as well if needed.
I appreciate your help!
You'll first have to decide how wide a frequency range is necessary.
I would think that 100Hz to 5kHz should be the minimum.
Will this involve perceiving directionality to different sources on the wall?
Will more than one of the speakers be making sound at the same time?
If not, this is a very significant and useful simplification.
Will there only be one particular sound used in each test?
How many different identical sources do you need on the wall,
is it 7 sources across, and 7 sources high?
I would think that 100Hz to 5kHz should be the minimum.
Will this involve perceiving directionality to different sources on the wall?
Will more than one of the speakers be making sound at the same time?
If not, this is a very significant and useful simplification.
Will there only be one particular sound used in each test?
How many different identical sources do you need on the wall,
is it 7 sources across, and 7 sources high?
Last edited:
I see, thank you. From what I know that freq range sounds reasonable.
Yes, the directionality is key - it is going to be an important feature of the stimulus.
Yes, at least two, maybe even more speakers should be emitting sound at the same time. In a given test these sounds coming from different sources may not be the same.
Yes, we plan it to be 7 across and 7 high - but this is not a strict requirement, can be more or can be somewhat less as well. But around this range.
Yes, the directionality is key - it is going to be an important feature of the stimulus.
Yes, at least two, maybe even more speakers should be emitting sound at the same time. In a given test these sounds coming from different sources may not be the same.
Yes, we plan it to be 7 across and 7 high - but this is not a strict requirement, can be more or can be somewhat less as well. But around this range.
Last edited:
Probably the closest toic to your need are line arrays (25 drivers/side) is common.
The price leader here is the Vifa TC9, a driver i do not personally like. A bit more money but still relativly inexpensive would be things like the Fountek FE85 or the Mark Audio CHN-50. Both these drivers are way better than the Vifa. But i am coming from a hifi perspective. You may find some diret cheap surplus drivers that suit your need, small FRs are much, much more common these days due to TVs, computers, and other tech)
You will need one amplifier module per speaker with a separte control feed (ie 49 channels). That is likely the hardest part. A big *** semi-pro i/o interface might work, but they usually have lots of ins, not as many out as necessary.
How low and how high you need to go is important, from your suggestion of test signals, low level resolution (DDR) will be very important.
dave
The price leader here is the Vifa TC9, a driver i do not personally like. A bit more money but still relativly inexpensive would be things like the Fountek FE85 or the Mark Audio CHN-50. Both these drivers are way better than the Vifa. But i am coming from a hifi perspective. You may find some diret cheap surplus drivers that suit your need, small FRs are much, much more common these days due to TVs, computers, and other tech)
You will need one amplifier module per speaker with a separte control feed (ie 49 channels). That is likely the hardest part. A big *** semi-pro i/o interface might work, but they usually have lots of ins, not as many out as necessary.
How low and how high you need to go is important, from your suggestion of test signals, low level resolution (DDR) will be very important.
dave
Will the test sound be the same from each speaker that is on at the same time, or different sounds?
Or would it be the same sound, but at different volume levels?
What is the maximum number of speakers that would be on at a time?
If low budget is important, we need to know the specific needs, and not overbuild the system.
Last edited:
How close you can get the drivers will play a role in how well you can do. Yamaha, i believe it was) did a DSP controlled soundbar where they could steer direction. They uses many very small (35mm?) drivers in a closely spaced XY array. Google did not help me find a ref, i likely have a picture on my HD but no idea what to search for namewize.
dave
What size does the array need to be (i.e. how wide a sound field is required)? And in terms of construction, is it going to be flat or curved? I imagine the speakers will each have to be pointing directly at the test subject and at a similar distance for best results.
Dave, well remembered - the original Yamaha sound projector was the YSP-1100 IIRC. The newer version is the YSP-5600.
Dave, well remembered - the original Yamaha sound projector was the YSP-1100 IIRC. The newer version is the YSP-5600.
Last edited:
The small amplifier modules from Marichino have been recently used in a MA Pluvia line array. Might be too pricey, but decent chip Class D amp boards are available in a large selection that might be suitable. You likely won’t need a whole lot of power (maybe 10w enuff). I have some REALLY smal Dayton branded Class D amps that are a bit less than half that IIRC. One could make a quite small 50 channel amp, or just hang them off the back of the speaker.
Speakers & amps should be fairly easier, feeding th esignal to 49 individual amps is the hardest part to my mind.
dave
Speakers & amps should be fairly easier, feeding th esignal to 49 individual amps is the hardest part to my mind.
dave
Dave's suggestion of a pro audio interface is likely the best way to go. Odds are you'll need more than one - the most analogue outputs I've seen on a single unit is 24.
It's probably a good idea to look at the various audio-over-ethernet options instead of USB. Something like Dante. This will make routing channels much easier.
I'll do some digging for suitable interfaces tomorrow, but in the meantime I suggest asking the folks over at the Gearslutz forums - they know more about this than I do: Gearslutz
It's probably a good idea to look at the various audio-over-ethernet options instead of USB. Something like Dante. This will make routing channels much easier.
I'll do some digging for suitable interfaces tomorrow, but in the meantime I suggest asking the folks over at the Gearslutz forums - they know more about this than I do: Gearslutz
Last edited:
If N channels are sufficient then the problem is building a Nx49 crossbar switch?
Its rare to find multi-position relays outside an old telephone exchange though, and double-throw isn't particularly amenable to cross-bar use!
You obviously need to switch after the power amps of course.
I can see photo-FETs being used as SSR's in a more flexible manner than mechanical relays, but it still feels quite complex.
Perhaps the dumb approach is best, N relays per speaker and a bus of the N signals. And daisy-chain the control signals though shift registers? It would mean a minimum number of contacts in the audio path to the speakers, and you could drive several in parallel if the amp's output and speaker impedances permit it, which is the most flexible arrangement (though you could also misconfigure the system to overload the amps, so well-protected amps would be good).
Its rare to find multi-position relays outside an old telephone exchange though, and double-throw isn't particularly amenable to cross-bar use!
You obviously need to switch after the power amps of course.
I can see photo-FETs being used as SSR's in a more flexible manner than mechanical relays, but it still feels quite complex.
Perhaps the dumb approach is best, N relays per speaker and a bus of the N signals. And daisy-chain the control signals though shift registers? It would mean a minimum number of contacts in the audio path to the speakers, and you could drive several in parallel if the amp's output and speaker impedances permit it, which is the most flexible arrangement (though you could also misconfigure the system to overload the amps, so well-protected amps would be good).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.