Ocean Sound circuit

[David Pariseau]

I want to build a circuit that simulates the sound of the ocean as a
sleep aid. The idea is to block out background noise, my snoring, for
my wife. Currently we have this huge air purifier in the bedroom that
we have to schlep around with us when we travel.

The units I've seen w/ background noise all have pre-recorded sounds
that loop after some time and the ear gets used to the patterns and
start focusing on them. White noise is irritating and pink noise is
only just a shade better.

I did some exhaustive searches in this forum and online and turned up
some links but nothing to give me a leg up on this. I realize I need
to create a random noise source, several circuits to generate
different parts of the surf sound fed by the noise and then something
to mix them all together and interface with the speaker(s).

Any thoughts on approaching this project? Circuit sources or a plan
for how to go about this from the ground up? I'm very comfortable
with micros, digital and analog electronics though I haven't done much
audio sound simulation stuff, aside from telephony dial-tone and
ringback emulation.

Thanks in advance for all constructive input,
Dave Pariseau.

 

[Ban]

I want to build a circuit that simulates the sound of the ocean as a
sleep aid. The idea is to block out background noise, my snoring, for
my wife. Currently we have this huge air purifier in the bedroom that
we have to schlep around with us when we travel.

The units I've seen w/ background noise all have pre-recorded sounds
that loop after some time and the ear gets used to the patterns and
start focusing on them. White noise is irritating and pink noise is
only just a shade better.

I did some exhaustive searches in this forum and online and turned up
some links but nothing to give me a leg up on this. I realize I need
to create a random noise source, several circuits to generate
different parts of the surf sound fed by the noise and then something
to mix them all together and interface with the speaker(s).

Any thoughts on approaching this project? Circuit sources or a plan
for how to go about this from the ground up? I'm very comfortable
with micros, digital and analog electronics though I haven't done much
audio sound simulation stuff, aside from telephony dial-tone and
ringback emulation.

Thanks in advance for all constructive input,
Dave Pariseau.

Dave,

get an MP3 player and record it at the sea or waterfall...
but I think no noise can cover a good snore. )

But there is one perfect solution to your problem,
let your wife have her own bedroom. Seems there is some space available
because the kids moved out, isn't it?

sorry, but I couldn't resist

ciao Ban

 

[David Pariseau]

Dave,

get an MP3 player and record it at the sea or waterfall...
but I think no noise can cover a good snore. )

But there is one perfect solution to your problem,
let your wife have her own bedroom. Seems there is some
space available because the kids moved out, isn't it?

sorry, but I couldn't resist

ciao Ban

Ban,

I don't want to record natural sounds because the ear quickly
becomes accustomed to the tape and learns to discern patterns.
It's especially annonying when you're trying to go to sleep
and your too-active mind starts looking for these patterns.

As for the good snore, hopefully I'm just south-of-good. But,
I've always found it a pain to share a room (or a bed) with
someone who does snore, so putting some effort into at least
masking the problem is called for.

Thanks,
Dave.

 

[Phil Hobbs]

Ban,

I don't want to record natural sounds because the ear quickly
becomes accustomed to the tape and learns to discern patterns.
It's especially annonying when you're trying to go to sleep
and your too-active mind starts looking for these patterns.

As for the good snore, hopefully I'm just south-of-good. But,
I've always found it a pain to share a room (or a bed) with
someone who does snore, so putting some effort into at least
masking the problem is called for.

Since ocean sounds are chaotic, you could take N samples in one run, M samples in
another, edit the files to make N and M relatively prime, play them both in endless
loops, and have a convincing natural sea sound that wouldn't repeate for N*M
samples. For N and M reasonably large, that's a _long_ time.

Cheers,

Phil Hobbs

 

[David Pariseau]

Since ocean sounds are chaotic, you could take N samples in one run,

M samples in another, edit the files to make N and M relatively prime,
play them both in endless loops, and have a convincing natural sea
sound that wouldn't repeate for N*M samples. For N and M reasonably
large, that's a _long_ time.

Cheers,

Phil Hobbs

Interesting idea though if there are noticeable features in one of
the samples, it might have a similar effect, though perhaps to a
much lesser extent.

Dave.

 

[Rob]

Look at Feb. 1997 copy of Everyday/Practical Electronics mag.
See http://www.epemag.wimborne.co.uk
Or I can scan it in for you.

 

[bogax]

About 20 years ago I built something very like what
you seem to be discribing (still use it too)
I don't have the circuit any more.

I built it from 4000 series cmos.
It was (is) two LFSRs clocked by Schmitt trigger oscillators
One of the LFSRs is a noise source and the other modulates
the noise.

I don't recall the exact details but I found that if the noise
source recirculated in less than about 6-8 seconds I could
percieve that it was repeating and I believe I settled on
a clock of around 100kHz, why exactly, I don't recall
except that I think that the higher clocking rates sounded
better and that was high enough with out pushing the LFSR length
up too much.

The thing puts out two channels

I tapped the LFSR in several places and summed them with resistors
I don't recall where the taps were or how many I used per channel
I wanted two channels and I wanted them to be different as much as
possible. I just played around to find something that seemed
adequate, though I don't recall if it really makes that much of
a difference.

The other LFSR was used to modulate the two channels.
The modulation was formed by taking 5 succesive stages and summing
them through resistors of increasing (or decreasing) weight,
with the order of the weighting reversed between the channels,
this produces a sweeping effect. I don't recall what the weighting
was but I think it was roughly binary.

All four summing points (noise and modulation for two channels)
feed capacitors for filtering. The filtering is slightly different
between channels so that one contains a little more of the highs.
The weighting of the modulation is arranged so that it sweeps from
the lower channel to the higher channel.

I started off with diode modulators but ended up substituting
transistors, just why I don't recall. But the transistors are not
operated linearly (it's still basically a diode modulator but
with transistors) (diodes worked just fine for modulation, but
iirc the gain of the transistors allowed substantially less power
to be used)

As I recall, the Modulation LFSR is clocked at about 1Hz and has 11
stages, can't say I was ever really bothered by or even aware of
the repetition (I mean, I know it repeats, it just never was the problem
for me that you seem to expect it to be)

Both LFSRs have an odd number of stages and use one xor gate for
summing taps for feedback.
The feedback is run through the spare xor gates (four in a package)
and a spare Schmitt trigger is used to generate a power on set
(I guess you'd call it) that is, the feedback is inverted at power
on to insure the LFSRs don't get stuck in the zero cycle.

It runs for around a year on four AA cells (but they just power the
noise it goes to a stereo to produce the sound)

It works good, but it's really more suggestive of surf than it is
like surf.

by messing with the timing of the modulation you can get something
that's suggestive of rushing water (increase the clock by about 10
basically)

If I were doing it today, I'd use a PIC, and I'd try to simulate
rain on the roof also.


bogax

 

[The real Andy]

I want to build a circuit that simulates the sound of the ocean as a
sleep aid. The idea is to block out background noise, my snoring, for
my wife. Currently we have this huge air purifier in the bedroom that

Stop drinking so much beer, cause that stuff make you snore real bad. My
wife is always threatening to leave me if I dont do something about my
snoring, maybe I should give up the beer...Burp...

Anyway, how bout some white noise, amplitude modulated with some kind of
random number generator thingy?? I find the Knuth (refer google) algorithm
good for randomness? Perhaps a Knuthed micro driving a digipot or DAC for
AM, or perhaps even some super DSP magic? Hmmm, let me have a few more
beers, i seem to have better ideas that way..Will report back in an hour or
two.

Keep on Beerin' (and snorin'),

Andy

 

[David Pariseau]

About 20 years ago I built something very like what
you seem to be discribing (still use it too)
I don't have the circuit any more.

Your description of the circuit you built is along the
lines of what I have in mind though it was a little
difficult to follow without a schematic. But, it's
kind of what I had in mind.

Thanks,
Dave Pariseau

 

[David Pariseau]

Andy,

I actually don't drink, so imagine what it would be like if I did! : )

I've tried a white noise source and it's not very pleasant. It's an
irritating/grating sound. Pink noise is only mildly better.

Dave.

 

[The real Andy]

Andy,

I actually don't drink, so imagine what it would be like if I did! : )

Ahh, well that is the problem you see. Now if you were to take up drinking
for a few months, then give it away...Well your wife will be giving you some
lovin' that you probably never experienced before!

I've tried a white noise source and it's not very pleasant. It's an
irritating/grating sound. Pink noise is only mildly better.

How bout some DSP white or pink noise with some super AM modulation and some
tricky Knuthed up ramdom filtering to remove the harshness??

Andy