Hello all 
I have had a strange idea for an audio delay effect. Most now use digital storage or tape delays, but I am keen to try a real acoustic delay. I am aware of the big limitation of a fixed delay period, but it's an experimental thing and I change the playback speed of my samples a lot anyhow.
My idea is to use a flexible hose, I have seen one 4mm internal, 6mm external. I will say around 150m will be used for a 437ms delay time, but certainly over 100m. I plan to "drive" the line with an inner ear/ear canal phone and recieve at the end with a small electret capsule (WM-61A or similar). Electronics to adjust feedback levels etc will be incorporated.
My concerns are firstly, can a small ear piece drive a 150m hose? How would you find out/work this out? It should act as a type of transmission line so I would expect it to be efficient?
Secondly, I do not want the sound to be a perfect replica at the other end (I'd use my existing digital delays for that!), but I also don't want it to be too seriously degraded. I'm expecting to experience a loss of treble, mabye some resonance effects and some reverb/ghosting from sound reaching the mic early through the hose walls etc. Any ideas how much treble loss I am likely to experience? How significantly will bending of the hose effect this?
I know this is pretty obscure but any insight is appreciated!
I have had a strange idea for an audio delay effect. Most now use digital storage or tape delays, but I am keen to try a real acoustic delay. I am aware of the big limitation of a fixed delay period, but it's an experimental thing and I change the playback speed of my samples a lot anyhow.
My idea is to use a flexible hose, I have seen one 4mm internal, 6mm external. I will say around 150m will be used for a 437ms delay time, but certainly over 100m. I plan to "drive" the line with an inner ear/ear canal phone and recieve at the end with a small electret capsule (WM-61A or similar). Electronics to adjust feedback levels etc will be incorporated.
My concerns are firstly, can a small ear piece drive a 150m hose? How would you find out/work this out? It should act as a type of transmission line so I would expect it to be efficient?
Secondly, I do not want the sound to be a perfect replica at the other end (I'd use my existing digital delays for that!), but I also don't want it to be too seriously degraded. I'm expecting to experience a loss of treble, mabye some resonance effects and some reverb/ghosting from sound reaching the mic early through the hose walls etc. Any ideas how much treble loss I am likely to experience? How significantly will bending of the hose effect this?
I know this is pretty obscure but any insight is appreciated!
From my experience (limited) there would be severe attenuation and I don't see how you would get any drive into the lows with one of those. For an experiment I would couple a real speaker with a funnel or something and I'm sure you have an unused computer mic in the junk box that would yield a small capsule. Intuition says that no matter what there will be severe degradation.
Thanks for replying!
The only test I've managed to set up is with about 1m of slightly larger hose (7mm internal mabye) and an earphone and electret mic. The sound recieved at the mic was perfectly clear, the only issue being that with a ~1m hose I got a large lift to the bass frequencies akin to loading a speaker in a transmission line (I presume).
This is clearly too short a length to judge with though, I can't imagine the results multiplied up by a factor of 100 or more. I didn't notice significant treble loss but possibly some, again multiplied by 100 it could be significant; an EQ to boost the earphone treble level could be employed up to a point. I need a longer hose to test with, around 10m should be enough to guage it.
Severe degredation sounds a bit drastic! I certainly want there to be some, but if the recieved sound is indistinguishable it won't be very satifactory. Ideally there will be some, which after several feedback repeates becomes very noticeable, to create an interesting effect
The only test I've managed to set up is with about 1m of slightly larger hose (7mm internal mabye) and an earphone and electret mic. The sound recieved at the mic was perfectly clear, the only issue being that with a ~1m hose I got a large lift to the bass frequencies akin to loading a speaker in a transmission line (I presume).
This is clearly too short a length to judge with though, I can't imagine the results multiplied up by a factor of 100 or more. I didn't notice significant treble loss but possibly some, again multiplied by 100 it could be significant; an EQ to boost the earphone treble level could be employed up to a point. I need a longer hose to test with, around 10m should be enough to guage it.
Severe degredation sounds a bit drastic! I certainly want there to be some, but if the recieved sound is indistinguishable it won't be very satifactory. Ideally there will be some, which after several feedback repeates becomes very noticeable, to create an interesting effect
I would think that a hose would act as a straight pipe, & you get resonance at 1/4 wavelength; but I wonder what effect coiling the hose would have on this???
Treble loss is part of the trad 'echo' effect anyway, because you'd be mixing the result back into the original signal as an effect I don't think it's too important; To be practical I think you'd need more input power than an earpice could provide; How about multiple paths - several different lengths of hose driven by a large driver???
Treble loss is part of the trad 'echo' effect anyway, because you'd be mixing the result back into the original signal as an effect I don't think it's too important; To be practical I think you'd need more input power than an earpice could provide; How about multiple paths - several different lengths of hose driven by a large driver???
Been done US4701952. They needed a way to pipe music to patients during NMR with no metalics allowed. The patent uses a 4" speaker in a small box coupled to the tube. I don't know if it would reach 300' but they say "very" long tube. Google has it as a free download. Keep us informed, I wanted to do an "eternal" echo sound installation a few years ago but never tried it.
A possible improvement a freind and I came up with. Gaining the advantages of using a more compact hose, variable delay time and presumably no issues driving the hose etc. Hopefully such a length provides enough "degredation", that's yet to be found out.
Unfortunately, I have no real idea about PICs, but if I can get enough help from some people I know it might be possible, mabye.
Unfortunately, I have no real idea about PICs, but if I can get enough help from some people I know it might be possible, mabye.
Attachments
for the adjustable delay part you could just get one of these. the delay on the chip is not mixed with the original on the board, so you can mix it as you please.
i have a few on order to try out with a DIY synth i am building, and eventually hope to get the development board and make some effects and active crossovers with it.
The samples sound pretty good. Certainly a speaking tube works but I'll bet there is a lot of acoustic power at the input.
Long time I know, but I had a bit of an experiment with this idea again recently. I built a transducer to drive the tube from a small speaker I had on hand:
Visaton - Lautsprecher und Zubehör, Loudspeakers and Accessories
Mounted into a small cylindrical MDF enclosure, with the acoustic output via a 5mm brass tube.
Issue with it is very little treble output. See these measured responses, Blue with the microphone directly on the output, Green with 1m of hose:
Basically perfect from very low till 800Hz, then a strong resonance at 1kHz before rapidly loosing energy and just nasty peaky response being left at the top. Listening to it, the sound is muffled and the 1kHz peak can be heard.
So, the poor response seems to be inherent in the transducer, so how I can I make a better one? The patent for that audio system used in MRI machine rooms says about loss of high frequencies and required electronic compensation, but that's oweing to the tube itself it seems, so starting off with weak treble isn't ideal!
I will try to add a photo of it but don't have a decent camera with me now.
Visaton - Lautsprecher und Zubehör, Loudspeakers and Accessories
Mounted into a small cylindrical MDF enclosure, with the acoustic output via a 5mm brass tube.
Issue with it is very little treble output. See these measured responses, Blue with the microphone directly on the output, Green with 1m of hose:
Basically perfect from very low till 800Hz, then a strong resonance at 1kHz before rapidly loosing energy and just nasty peaky response being left at the top. Listening to it, the sound is muffled and the 1kHz peak can be heard.
So, the poor response seems to be inherent in the transducer, so how I can I make a better one? The patent for that audio system used in MRI machine rooms says about loss of high frequencies and required electronic compensation, but that's oweing to the tube itself it seems, so starting off with weak treble isn't ideal!
I will try to add a photo of it but don't have a decent camera with me now.
Attachments
Yes, strangely I've used such a system myself once on a plane! Much more recent than the 70's though as I'm only 23! Must have been using an old model of aircraft for that flight I guess. It sounded alright as I recall, so I don't know why my transducer has such poor high frequency output. There's only so much I can boost it before hiss will become an issue.
Just thinking out loud...
Perhaps the cylindrical enclosure is acting similar to a compression chamber in a BLH.
Low frequencies are passed up to the cut-off frequency then the high frequencies are attenuated.
It's quite feasible. I built another one using the Monacor SPH-30X driver and it has a similar response overall. I had hoped that driver's strong peak at 10kHz would help to even out the top end, but it's still weak above 1-2k.
How would on go about minimising the treble loss? Is it unavoidable because lower frequencies are transmitted as pressure waves but higher frequencies have to 'bounce along' the tube? I just find it odd that even with no tube length the high frequencies are minimal. It's not a good starting point really, needing 30dB of correction before it even travels down any hose!
How would on go about minimising the treble loss? Is it unavoidable because lower frequencies are transmitted as pressure waves but higher frequencies have to 'bounce along' the tube? I just find it odd that even with no tube length the high frequencies are minimal. It's not a good starting point really, needing 30dB of correction before it even travels down any hose!
I would try using a larger diameter tube.
I had I similar idea one day when I was in home depot and yelled into a rolled up 100' piece of coreagated drain field tubing.
Higher frequency's attenuate at a faster rate than low frequency's do over the same distance.
It is possible that a smaller diameter tube could be imposing a higher impedence to the higher frequency's. jer
I had I similar idea one day when I was in home depot and yelled into a rolled up 100' piece of coreagated drain field tubing.
Higher frequency's attenuate at a faster rate than low frequency's do over the same distance.
It is possible that a smaller diameter tube could be imposing a higher impedence to the higher frequency's. jer
Could also be very true. I can't use any bigger than the 4mm internal diameter tube due to physical size/cost/availability issues though.
Hmm, this project could very well be doomed, if I want a full acoustic delay of 0.5s or so which needs 170 metres of hose, might end up needing an unachievable 100dB of treble boost . A transducer with flat response acoustically to start with would give me some more confidence.
Hmm, this project could very well be doomed, if I want a full acoustic delay of 0.5s or so which needs 170 metres of hose, might end up needing an unachievable 100dB of treble boost
. A transducer with flat response acoustically to start with would give me some more confidence.
I wonder if an alternative material for the enclosure would help?
Not sure that MDF is ideal for preserving high frequencies.
Both good suggestions! Yes, the MDF could well be absorbing much of the high frequency energy. Plastic or metal are the only real alternatives, unless sealing the MDF on the inside surfaces has a really dramatic effect.
A compression driver might work, I expect with <1W input it can cope with signals down to 20Hz without bottoming out? They certainly have great efficiency so it might make it easier to apply any required treble boost without pushing too much power.
Unfortunately looks like any reasonable unit is about £17, which isn't too bad if it works, but like everything in this project it's so experimental that I literally don't know what it'll do until I have physical parts to test and measure (including expensive hose!)
A compression driver might work, I expect with <1W input it can cope with signals down to 20Hz without bottoming out? They certainly have great efficiency so it might make it easier to apply any required treble boost without pushing too much power.
Unfortunately looks like any reasonable unit is about £17, which isn't too bad if it works, but like everything in this project it's so experimental that I literally don't know what it'll do until I have physical parts to test and measure (including expensive hose!)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.