Real-time sound cyphering algorithm
Astier Anisse
My problem is quite simple:
I'am willing to find a simple algorithm to cypher a sound:
I need the cyphered sound to be audible, but not understandable,
and that we can decypher it from a recorded sound (verbatim copy not needed).
This seems simple, but in fact, i tried a few algorithms and i had no results.
The real problem is that the input data is a 8-bits encoded buffer
(or more, depending on the encoding).
It represents the movement of the speaker membrane.
I tried classical cypher or modifications on the stream:
-XOR-ing it with a 1 or 4 bytes key. It is or understandable with the 1 byte
keys, or impossible to decypher (because we may not have the same values
due to numerical/analogical conversion, and that we need to know how to sync
the key with the data)
-permutation of amplitude ranges (still understandable)
-modifying harmonics after a fast fourier transform. Increasing the level
of each harmonic (still understandable but strong noise), multiplying all
harmonics by a factor (inverse fourier transform goes out of sound range).
The objective is to have an output sound that seems to be a random sound.
And that we can decypher after recording. "real-time" is achevied by a small
buffer (0.5 sec for example).
Thank you.
John Hadstate
Astier Anisse wrote:
> Hello,
> My problem is quite simple:
> I'am willing to find a simple algorithm to cypher a sound:
> I need the cyphered sound to be audible, but not understandable,
> and that we can decypher it from a recorded sound (verbatim copy not needed).
>
> This seems simple, but in fact, i tried a few algorithms and i had no results.
> The objective is to have an output sound that seems to be a random sound.
> And that we can decypher after recording. "real-time" is achevied by a small
> buffer (0.5 sec for example).
>
Let me suggest the use of a device known as a "modem" to transmit your
encrypted data.
Andrew Swallow
Try -
Record sound in MP3 digital format
Encrypt the MP3, either as a file or a bit stream
Use a modem to transmit the encrypted data
Receive the data using a second modem
Decrypt the incoming data giving the MP3
Play the MP3 producing the original sound.
AES in CTR mode is a good method of encryption.
Andrew Swallow
Jan Panteltje
Anisse <ast...@ece.fr> wrote in <slrnd98uu9...@gandalf.ece.fr>:
hehe, fun project.
Why not play it backwards?
It is a signal processing issue, you could make a single for double sideband
signal, reverse sideband, shift... the works.
Maybe also ask in a comp dsp newsgroup or something.
Analog based encryption systems like that have been used for example for police.
Unruh
>Astier Anisse wrote:
>> Hello,
>> My problem is quite simple:
>> I'am willing to find a simple algorithm to cypher a sound:
>> I need the cyphered sound to be audible, but not understandable,
>> and that we can decypher it from a recorded sound (verbatim copy not needed).
>>
>> This seems simple, but in fact, i tried a few algorithms and i had no results.
>> The objective is to have an output sound that seems to be a random sound.
>> And that we can decypher after recording. "real-time" is achevied by a small
>> buffer (0.5 sec for example).
It is extremely difficult. The sound is converted to analog and then
redigitized. The new digitization may differ significantly from the
original-- especially in amplitude at various frequencies but also phase at
various frequencies. Any encryption thus needs to be such as to be
extremely tolerant of such errors.
Now, you could of course do frequency shifting-- ie take the spectrum say
in 1 sec slices, and shift around the frequencies in bands. Again this is
liable to be useful only for low fidelity sounds.
Joseph Ashwood
news:slrnd98uu9...@gandalf.ece.fr...
If you want real security for the sound with a bitrate that sounds decent
for human speech and the ability to broadcast with a speaker you've got a
very difficult problem. The modem suggestion actually won't work, the
highest frequency that can be safely held by a CD is 22.05KHz, so you'd only
have 22.05Kbps, about 2/3 of the best Ogg Vorbis can offer for sound, and
MP3 would be completely incomprehensible at that level. Your best bet is to
make use of several of the bits in the sample but this provides it's own
(very substantial) difficulties. The best solution I can think of in the
second category is to use AES in CTR mode to directly encrypt the digitized
sound stream. The top bits will be recoverable with >50% accuracy, and with
a good enough filter and mic you should be able to recover the top 6-bits,
this is enough to recreate recognisable human speech, but it won't be great
quality. The bigger difficulty is in syncing up the exact sample rate, for
this I'd suggest stealing a concept from modems where you first begin with a
known sample, the first 1024-bits of sound would be known and allow for a
near perfect sync, you'd also have to sacrifice a few samples every second
for resync as well, this will also enable fwd and rewind later on.
For this problem compression is very much not your friend because you will
lose data, compression has a tendancy to mix bits resulting in you losing
the entirety of every sample instead of just the lower bits.
Joe
Michael Brown
> Hello,
> My problem is quite simple:
> I'am willing to find a simple algorithm to cypher a sound:
> I need the cyphered sound to be audible, but not understandable,
> and that we can decypher it from a recorded sound (verbatim copy not
> needed).
How about something like running a stream cipher through a low-pass filter
and adding it to the recorded sound? Probably want to run the "plaintext"
through the same filter as well to avoid leaking higher frequencies. One
problem with this approach is that you'll probably have to avoid wraparound
(which will introduce high frequencies and possibly get filtered out during
transmission), which means you leak information; using a higher volume on
the keystream will reduce this (eg: output = input * 0.1 + keystream * 0.9).
You'll still have to figure out how to sync the streams (as Joe mentioned),
but this should be quite robust even if you compress the resulting audio
with a lossy algorithm (not recommended though, as you'll be essentially
trying to compress white noise ...).
[...]
--
Michael Brown
Add michael@ to emboss.co.nz ---+--- My inbox is always open
Nicol So
>
> the
> highest frequency that can be safely held by a CD is 22.05KHz, so you'd only
> have 22.05Kbps
You confused bandwidth with channel capacity.
--
Nicol So
Disclaimer: Views expressed here are casual comments and should
not be relied upon as the basis for decisions of consequence.
Astier Anisse
Thanks for all your comments.
Unruh, Jon, Joe you synthetised the problem very well.
I'll try to find a way to sync the signal on order to know when the key begins.
I tryed the playing backwards technique earlier, it works fine : when you
hear it, you know someone is speaking, but do not understand anything.
I can't use any modem, because the purpose is to send the sound
through the air.
I'm working on it, so thank you very much.
Regards, Anisse
John E. Hadstate
"Unruh" <unruh...@physics.ubc.ca> wrote in message
news:d72oh6$nte$1...@nntp.itservices.ubc.ca...
>
> It is extremely difficult. The sound is converted to
> analog and then
> redigitized. The new digitization may differ significantly
> from the
> original-- especially in amplitude at various frequencies
> but also phase at
> various frequencies. Any encryption thus needs to be such
> as to be
> extremely tolerant of such errors.
>
> Now, you could of course do frequency shifting-- ie take
> the spectrum say
> in 1 sec slices, and shift around the frequencies in
> bands. Again this is
> liable to be useful only for low fidelity sounds.
>
You may be on to something here. The first secure
telephones used analog trickery to reflect and shift the
audio spectrum (low frequencies became high, high
frequencies became low). It sort of worked, but history
records that some people could train themselves to
understand the audio-reversed speech. Also, undoing the
translation wasn't very difficult.
However, with digital processing of the audio, and the
ability to generate keyed random permutations of certain
spectral components, I'd bet your idea would work better
than you seem to think.
John E. Hadstate
>
> I can't use any modem, because the purpose is to send the
> sound
> through the air.
>
When you say "send the sound through the air", do you mean,
"transmit it over a radio" or do you literally mean "amplify
it and play it through a speaker"?
If "transmit over a radio", you can still use a modem,
though maybe not an off-the-shelf modem.
Tom Giarmo
> Why not play it backwards?
I once listened to Stairway To Heaven backwards and it fukken made me
fukken my sister.
True story.
Bryan Olson
> I'am willing to find a simple algorithm to cypher a sound:
> I need the cyphered sound to be audible, but not understandable,
> and that we can decypher it from a recorded sound
Crypto has gone digital. Ciphering entirely in the analog domain
has a dreadful security record, and the modern crypto literature
essentially ignores it. For what is available, try Googling
"speech scrambling".
The way we would approach this today is to digitize the sound,
compress (with an audio-specific coder), encrypt, code for
forward error correction, and modulate onto an audio carrier.
This group should be able to help with the encrypt step, and
people here might know something about the others, but each of
the steps is now reasonably well studied on its own.
--
--Bryan
Jan Panteltje
Hadstate" <jh11...@hotmail.com> wrote in
<_Gile.32334$J25....@bignews6.bellsouth.net>:
If 'through the air' you have more bandwidth :-)
You could use a fake channel in the audio range, and an
other channel, for example AM modulated, at 30 kHz.
An other interesting experiment that I want to try some day is
use a 40kHz power oscillator, a diode ring modulator, generate one double
sideband signal and one carrier tone (with directional speakers).
Normal audio amp output should be enough to drive a ring modulator
via a transformer.
Then where the sound crosses (that location both transducers point to)
demodulation will happen if there is a non linear medium (like your ear
\ head).
I think somebody is already selling this as 'speaker less' Hi Fi, using
very small ultrasound speakers.
Maybe it is patented.
Joseph Ashwood
news:6Gble.225$zb.20@trndny04...
> Joseph Ashwood wrote:
>>
>> the
>> highest frequency that can be safely held by a CD is 22.05KHz, so you'd
>> only
>> have 22.05Kbps
>
> You confused bandwidth with channel capacity.
It was my impression that generally they functioned by oscillating between
+1/-1 to send a 1/0, not being an expert in modem technology I could be
wrong. Even going to the far reaches of current standard analog modem
technology (56Kbps) that still puts the bandwidth below the level of
acceptable MP3 (64Kbps). If you can achieve 112.5Kbps (which I would expect
to be extremely difficult with a speaker and microphone) you're now in
acceptable transfer speeds for real time voice. Come to think of it, I
believe there are codecs available that perform better than either MP3 or
Ogg when encoding a human voice (my assumption for the transfer I don't
remember it being originally stated), and even under different assumptions I
would assume there are possible codecs optimized for the various expected
transmissions.
Joe
Andrew Swallow
[snip]
>
> I can't use any modem, because the purpose is to send the sound
> through the air.
You can send the output of the modem to a loud speaker rather than a
telephone wire providing you ensure voltage compatibility.
Andrew Swallow
Unruh
>The way we would approach this today is to digitize the sound,
>compress (with an audio-specific coder), encrypt, code for
>forward error correction, and modulate onto an audio carrier.
>This group should be able to help with the encrypt step, and
>people here might know something about the others, but each of
>the steps is now reasonably well studied on its own.
The problem is how to remove it from the audio carrier (played over a
loudspeaker with its largely random frequency dependent amplitude and phase
changes) and still have it decoded to the original. Ie, you need
huge error correction redundancy to hope to resurect the original signal.
>--
>--Bryan
Unruh
>On a sunny day (Thu, 26 May 2005 07:49:39 -0400) it happened "John E.
>Hadstate" <jh11...@hotmail.com> wrote in
><_Gile.32334$J25....@bignews6.bellsouth.net>:
>>
>>"Astier Anisse" <ast...@ece.fr> wrote in message
>>news:slrnd9ba6s...@gandalf.ece.fr...
>>>
>>> I can't use any modem, because the purpose is to send the
>>> sound
>>> through the air.
>>>
>>
>>When you say "send the sound through the air", do you mean,
>>"transmit it over a radio" or do you literally mean "amplify
>>it and play it through a speaker"?
>>
>>If "transmit over a radio", you can still use a modem,
>>though maybe not an off-the-shelf modem.
>If 'through the air' you have more bandwidth :-)
>You could use a fake channel in the audio range, and an
>other channel, for example AM modulated, at 30 kHz.
No you could not. No speaker produces appreciable sound at 30KHz. Most
audio speakers are lucky to get to 10KHz.
>An other interesting experiment that I want to try some day is
>use a 40kHz power oscillator, a diode ring modulator, generate one double
>sideband signal and one carrier tone (with directional speakers).
>Normal audio amp output should be enough to drive a ring modulator
>via a transformer.
>Then where the sound crosses (that location both transducers point to)
>demodulation will happen if there is a non linear medium (like your ear
>\ head).
>I think somebody is already selling this as 'speaker less' Hi Fi, using
>very small ultrasound speakers.
>Maybe it is patented.
Yes, they are. It is even being advocated as a person specific weapon (
focus the ultrasound on one individual, the ringleader) Of course the
effect of that intensity of ultrasound on people might be a bit bad. (The
non-linearity is almost certainly not that high, which means that the UF
sound needs to be far louder than the signal you want to deliver. Just
because the ear cannot hear it, does not mean it cannot do damage.
Unruh
Yes, certainly you can send it. Whether you can decode it is another
matter. Why do modems engage in many seconds of negotiations when a
connection is made? To calculate all the phase shifts etc and correct for
them.
>Andrew Swallow
Bryan Olson
Things have gotten much better.
Trellis Coded Modulation (TCM), introduced by Gottfried
Ungerboeck in 1982, is great for modulating (and demodulating)
discrete signals over noisy analog lines. For error correction,
Turbo Codes, introduced by Berrou, Glavieux, and Thitimajshima
in 1993, are significantly more efficient than the error
correction used in CD's, DVD's and HDTV. Turbo codes require
some buffering; I think the stated half-second window should be
fine.
There are also various multi-media packetizing mechanisms
designed to allow the stream to continue even while
communication is lost for short periods. We'd probably want to
use one, and make sure our crypto doesn't break it.
I have now written most of what I know on the subject, and, I
hope, not too much more.
--
--Bryan
John E. Hadstate
>
> No you could not. No speaker produces appreciable sound at
> 30KHz. Most
> audio speakers are lucky to get to 10KHz.
>
What about those piezo-electric tweeters? I've seen some of
them rated +/- 3 dB from 10 to 80 Khz.
I'm working on a test of your plan to digitally scramble the
audio spectrum.
Unruh
>"Unruh" <unruh...@physics.ubc.ca> wrote in message
>news:d75rjl$1eu$2...@nntp.itservices.ubc.ca...
>>
>> No you could not. No speaker produces appreciable sound at
>> 30KHz. Most
>> audio speakers are lucky to get to 10KHz.
>>
If you have your own speaker and microphone to record that information, why
the hell not just string a wire between the two instead and get rid of the
two transducers. ie, IF the persons wants to convert the stream to audio,
convert it back to electronic, I strongly suspect he does not have control
over those parts of the path. Otherwise the whole exercise seems to be
"academic" in the worst sense of the word.
Jan Panteltje
<unruh...@physics.ubc.ca> wrote in <d75rjl$1eu$2...@nntp.itservices.ubc.ca>:
>Jan Panteltje <pNaonSt...@yahoo.com> writes:
>
>>On a sunny day (Thu, 26 May 2005 07:49:39 -0400) it happened "John E.
>>Hadstate" <jh11...@hotmail.com> wrote in
>><_Gile.32334$J25....@bignews6.bellsouth.net>:
>
>>>
>>>"Astier Anisse" <ast...@ece.fr> wrote in message
>>>news:slrnd9ba6s...@gandalf.ece.fr...
>>>>
>>>> I can't use any modem, because the purpose is to send the
>>>> sound
>>>> through the air.
>>>>
>>>
>>>When you say "send the sound through the air", do you mean,
>>>"transmit it over a radio" or do you literally mean "amplify
>>>it and play it through a speaker"?
>>>
>>>If "transmit over a radio", you can still use a modem,
>>>though maybe not an off-the-shelf modem.
>>If 'through the air' you have more bandwidth :-)
>>You could use a fake channel in the audio range, and an
>>other channel, for example AM modulated, at 30 kHz.
>
>No you could not. No speaker produces appreciable sound at 30KHz. Most
>audio speakers are lucky to get to 10KHz.
You are 100% wrong.
My tweeters are specified to 40kHz (-3db).
Before I blew them up.
I have extensively experimented and designed with ultrasound here in my lab.
If you download the data sheet of ANY mid range speaker, especially the small
ones as used in litte radios, these will go way to 20kHz.
But actually I used piezzo transduceres.
Little piezzo horns go for 3$ a piece here.
In a reasonable speaker box (2 or 3 way) you will find these, or some other
better quality types (ferro fluid, ribbon).
>>An other interesting experiment that I want to try some day is
>>use a 40kHz power oscillator, a diode ring modulator, generate one double
>>sideband signal and one carrier tone (with directional speakers).
>>Normal audio amp output should be enough to drive a ring modulator
>>via a transformer.
>
>>Then where the sound crosses (that location both transducers point to)
>>demodulation will happen if there is a non linear medium (like your ear
>>\ head).
>
>>I think somebody is already selling this as 'speaker less' Hi Fi, using
>>very small ultrasound speakers.
>>Maybe it is patented.
>
>Yes, they are. It is even being advocated as a person specific weapon (
>focus the ultrasound on one individual, the ringleader) Of course the
>effect of that intensity of ultrasound on people might be a bit bad. (The
>non-linearity is almost certainly not that high, which means that the UF
>sound needs to be far louder than the signal you want to deliver. Just
>because the ear cannot hear it, does not mean it cannot do damage.
Oh well, I have used it as weapon against people of cause ;-).
Don't you hear it?
You need a lot of power to even get attention (with a clean sinewave).
The range can be quite big actually (30 meters worked).
You need a good amp to drive these piezzos, (or any other high power tweeters).
This one will do 20 kHz square wave at 75 W RMS (so 40 kHz sine no problem):
http://panteltje.com/panteltje/amplifier/
hehe
David Eather
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
I think this is the hard way to do this - a good way that will get a
perfect result, but still a hard way. The OP wanted to decrypt and
identify the sound by ear, so it doesn't have to be bit perfect.
Encode the sound stream by adding (not xoring) the stream cipher
output. Decode is just receive the sound (normalise it first) and
subtract the same stream cipher output. The result will probably
have bit errors in most bytes but still have the identifiable sound.
(I skipped over the necessity of keeping both stream ciphers in sync
- -If you can make sure there is no similar frequency in the ciphered
sound stream I suggest periodically adding a few bytes of high
frequency as sync bytes . Or if there is enough computational power
superimpose the whole waveform on a low frequency carrier and use the
zero crossings for sync - because of the errors in transmition this
will take a bit of trial and error but you will know when you have
the correct result - a FFT will show the result is no longer white
noise).
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA+AwUBQpfzUJS9Fk5okqe7EQLqUACgmd+313eDHaU+gjulWC5W8scXopsAmMor
t6ZgpyP8vWIwLl1fcn0uomo=
=es3n
-----END PGP SIGNATURE-----
Unruh
>I think this is the hard way to do this - a good way that will get a
>perfect result, but still a hard way. The OP wanted to decrypt and
>identify the sound by ear, so it doesn't have to be bit perfect.
>Encode the sound stream by adding (not xoring) the stream cipher
>output. Decode is just receive the sound (normalise it first) and
>subtract the same stream cipher output. The result will probably
>have bit errors in most bytes but still have the identifiable sound.
>(I skipped over the necessity of keeping both stream ciphers in sync
>- -If you can make sure there is no similar frequency in the ciphered
>sound stream I suggest periodically adding a few bytes of high
>frequency as sync bytes . Or if there is enough computational power
>superimpose the whole waveform on a low frequency carrier and use the
>zero crossings for sync - because of the errors in transmition this
>will take a bit of trial and error but you will know when you have
>the correct result - a FFT will show the result is no longer white
>noise).
This problem of sync is one problem. A worse problem is that of
differential phase shifts. Remember that a phase shift is the same as a
time delay. speakers are notorious for introducing differential phase
shifts into the sound. Thus, let us say you use the low freq carrier idea
and the system introduces a phase shift (time delay) into that low freq sound.
The ear is almost completely insensitive to such phase shifts-- but they
can almost completely alter the shape of the wave form. Your
encryption system would be highly sensitive to such phase shifts.
The problem with his scheme in general is precisely this problem which
arises from the demand that the sound be played over a speaker system
(which I assume he has no control over, and has not been optimised for
minimal phase shifts). If it were audio transmitted over a wire, my worry
would disappear.
David Eather
Yes, my guess is it may still be manageable.
>
> The problem with his scheme in general is precisely this problem which
> arises from the demand that the sound be played over a speaker system
> (which I assume he has no control over, and has not been optimised for
> minimal phase shifts). If it were audio transmitted over a wire, my
> worry would disappear.
But the wire removes the almost uniqueness of the problem
Mike Amling
>
> I think this is the hard way to do this - a good way that will get a
> perfect result, but still a hard way. The OP wanted to decrypt and
> identify the sound by ear, so it doesn't have to be bit perfect.
>
> Encode the sound stream by adding (not xoring) the stream cipher
> output.
This is dubious. If you add white noise to a recorded conversation,
it sounds like a conversation with tape hiss. Noise, which is what I
presume a stream cipher would sound like, has to be overwhelming before
a conversation becomes totally unintelligible.
> Decode is just receive the sound (normalise it first) and
> subtract the same stream cipher output. The result will probably
> have bit errors in most bytes but still have the identifiable sound.
--Mike Amling
David Eather
> David Eather wrote:
>>
>> I think this is the hard way to do this - a good way that will get a
>> perfect result, but still a hard way. The OP wanted to decrypt and
>> identify the sound by ear, so it doesn't have to be bit perfect.
>>
>> Encode the sound stream by adding (not xoring) the stream cipher
>> output.
>
> This is dubious. If you add white noise to a recorded conversation,
> it sounds like a conversation with tape hiss. Noise, which is what I
> presume a stream cipher would sound like, has to be overwhelming
> before a conversation becomes totally unintelligible.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Correct - to be impossible to follow the signal would have to be 20
db below the ambient noise floor (just over 3 bits) and a cipher
output does sound like white noise.
In audio engineering the 20db figure was widely quoted. +20db(spl)
above the ambient noise to be easily distinguished (by an audience).
+0db compared to the ambient noise can be followed for a short time
by people highly interested in the broadcast. -20db is considered
inaudible.
The constrains given by the OP don't give much room for a "perfect"
result - processing in real time and no feedback and it is probably
safe to assume it is a one time project for an individual. To do the
job properly would be to reproduce the equivalent of something
similar to SigSally - which would not be practical with the above
constraints - or designing a new form of audio signal scrambling from
scratch. I don't see anyone putting their hand up for that.
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA/AwUBQppvspS9Fk5okqe7EQJxcACg+nvuyu8qLqP0wYX4fv47m71sWtcAnjg0
x0srdPfKfcha97p0u96LRDL4
=Xhvx
-----END PGP SIGNATURE-----
Bryan Olson
> Correct - to be impossible to follow the signal would have to be 20
> db below the ambient noise floor (just over 3 bits) and a cipher
> output does sound like white noise.
>
> In audio engineering the 20db figure was widely quoted. +20db(spl)
> above the ambient noise to be easily distinguished (by an audience).
> +0db compared to the ambient noise can be followed for a short time
> by people highly interested in the broadcast. -20db is considered
> inaudible.
So, correct me if I'm wrong: your solution is to transmit the
sum of payload-sound plus a 20dB larger crypto-pseudo-random
noise signal. Then the receiver subtracts off the crypto signal,
to get the plaintext plus any noise on the transmission. For the
sound to "be easily distinguished", we want the plaintext sound
to be 20dB above the noise on the transmission channel. To get
the payload 20dB above the noise, and the crypto-stream 20dB
above the payload, we need an audio carrier with a signal-to-
noise of 40dB, which is 10,000-to-1 in sound power.
Modern digital methods offer vastly better performance. We can
transmit easily-intelligible speech at a fraction of the power,
or high-fi stereo music at the same power. Google "turbo codes"
along with "dB" for more.
The digital method also offers better security than does adding
pseudo-random noise. Both rely on the security of a
cryptosystem, but the analog method also assumes that no useful
information is available in a signal 20dB below the noise.
"Considered inaudible," "in audio engineering," does not imply
the level of confidentiality required in cryptography. -20dB
certainly limits the bit-rate, but leaking slowly isn't good
enough.
> The constrains given by the OP don't give much room for a "perfect"
> result - processing in real time and no feedback and it is probably
> safe to assume it is a one time project for an individual.
Cheap mass-market (or used) PC's can do the digital processing,
and cheap commodity audio systems can handle the analog I/O. In
terms of software, the modern methods are sophisticated, but the
hard parts are already done.
--
--Bryan
David Eather
Hash: SHA1
<snip>
Warning. Sarcastic mode should only be turned on when God mode is
correctly engaged,otherwise the recipient is likely to slip into Slap
Down mode without warning. (Words I may yet regret)
BO wrote>
> So, correct me if I'm wrong:
DE wrote>
Is that sarcastic mode I read?
BO wrote>
your solution is to transmit the
> sum of payload-sound plus a 20dB larger crypto-pseudo-random
> noise signal. Then the receiver subtracts off the crypto signal,
> to get the plaintext plus any noise on the transmission. For the
> sound to "be easily distinguished", we want the plaintext sound
> to be 20dB above the noise on the transmission channel. To get
> the payload 20dB above the noise, and the crypto-stream 20dB
> above the payload, we need an audio carrier with a signal-to-
> noise of 40dB, which is 10,000-to-1 in sound power.
IIRC, when I went to school, audio volume was measured in dB - spl
(sound pressure level, which I noted in the pervious post) which is
20 x log10 and not 10 x log10 which is a measure for comparing power
- - you did not do your basic research. Try -
http://en.wikipedia.org/wiki/Sound_unit
Forty dB is a factor of 100.
BO wrote>
> So, correct me if I'm wrong:
<snip>
You are wrong and you are corrected.
As I also mentioned, people interested in the signal can follow if
the signal to noise is 1 dB. I suspect my suggestion can do a little
better than that (my expectation would be +3 to +10 db, perhaps more
with a bit of fiddling). I made 2 assumptions. The first was that
the receiver was interested in the message (or else there is no point
in sending it), and second, that the message was not going to be the
collected works of Emmanuel Kant. Is that unreasonable?
>
> Modern digital methods offer vastly better performance. We can
> transmit easily-intelligible speech at a fraction of the power,
You have not thought about this at all have you? You've just flopped
out the "correct" answer without regard to the circumstance..
The OP implied the transmition of the sound data should occur in real
time when he said:
OP wrote>
I need the cyphered sound to be audible, but not understandable
DE wrote>
This real-time requirement is further suggested by his (OP's) report
of his experiments which appear to be conducted in real time. (If
real time transmition is not an issue then the simplest solution is
to stretch out the transmition time and frequency double the source
as much as necessary to compensate - I assume this was thought of and
discarded)
How are you going to send your data? fiberoptic, radio, coax, UTP?
As another poster pointed out a speaker used for acoustic transmition
has all sorts of phase shifts which will be unpredictable and
unrepeatable unless you want to specify exactly which speaker (not
which brand - which one), sound card, microphone and which
temperature, pressure and humidity conditions apply. You could only
use the most crude forms of phase shift modulation, if any. You'll
be *lucky* if you get an unreliable 20k bits per second. If you use
all of that bandwidth you could just squeeze in a highly compressed
sound and some serious error correction.
> or high-fi stereo music at the same power. Google "turbo codes"
> along with "dB" for more.
Sure, I am aware of turbo codes, use in space craft etc, etc, never
quite got around the math of them (turbo codes), didn't have so much
trouble with all that error correcting stuff that you can do with
shift registers and such - Real Simple Codes. Useful, even if, like
me, you don't have the mind of Solomon. (Too subtle?)
I didn't think that the OP's requirements warranted a complex or
exact result. The OP seemed to have quite limited goals and
resources in mind - but then again, I do know one person who thought
it a good idea to hunt mice with a 12g shotgun (true).
>
>
> The digital method also offers better security than does adding
> pseudo-random noise.
Yes of course. I said that such an approach would be a "perfect
solution" But how well does it fit within the problem put by the OP?
>Both rely on the security of a
> cryptosystem, but the analogue method also assumes that no useful
> information is available in a signal 20dB below the noise.
> "Considered inaudible," "in audio engineering," does not imply
> the level of confidentiality required in cryptography.
What security level was suggested by the OP? 2**64? 2**128? or
something unintelligible to others hearing the source?
OP wrote>
I need the cyphered sound to be audible, but not understandable
DE wrote>
The OP further defined his security boundary with this -
OP wrote>
I tried classical cypher or modifications on the stream:
-XOR-ing it with a 1 or 4 bytes key.
DE wrote>
The OP specifies "not understandable" with a 1 to 4 byte key
BO wrote>
> So, correct me if I'm wrong:
<snip>
You are wrong and you are corrected.
BO wrote>
>-20dB
> certainly limits the bit-rate, but leaking slowly isn't good
> enough.
>
>
> > The constrains given by the OP don't give much room for a
> "perfect"
> > result - processing in real time and no feedback and it is
> probably
> > safe to assume it is a one time project for an individual.
>
> Cheap mass-market (or used) PC's can do the digital processing,
> and cheap commodity audio systems can handle the analog I/O. In
> terms of software, the modern methods are sophisticated, but the
> hard parts are already done.
Before your post I had already agreed that such a system as you
describe would give a perfect result. It is in every way the
"perfect" and "correct" answer.
The OP talks about security parameters like "not understandable" with
a 1 to 4 byte keys and using classical ciphers.
My roughed out concept is much simpler than yours.
Which one more closely fits the stated requirements of the OP?
Which one is the OP more likely to be able to implement?
OP wrote>
" I tried classical cypher or modifications on the stream:
-XOR-ing it with a 1 or 4 bytes key. "
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA/AwUBQpsiRJS9Fk5okqe7EQKP7gCfdYNb1lpgPIouNmzJrxVSnMdAdY8AoLpU
3WBJqwf4FKqHNe1sIuYQEf6F
=KEIB
-----END PGP SIGNATURE-----
Jan Panteltje
<eat...@tpg.com.au> wrote in <429b...@dnews.tpgi.com.au>:
>http://en.wikipedia.org/wiki/Sound_unit
>
>Forty dB is a factor of 100.
For voltage only.
But 2 x voltage results is 4 x power in same impedance (speaker).
P = (U * U) / R
http://encyclopedia.laborlawtalk.com/Decibel#Acoustics
>OP wrote>
>I need the cyphered sound to be audible, but not understandable
I think this can also be done by using a say 5kHz audio bandwidth
(more then enough for speech), and AM modulating a carrier at 20kHz.
The lower sideband will go to 20 - 5 = 15 kHz.
You will hear some high peaked noises.
An other alternative is to use a FM carrier, and narrow band FM (so
use only one sideband), and use 10 kHz carrier :-)
This requires some filtering at the rx side, but any dsp board should do.
FM has the advantage that it is not so sensitve to amplitude changes.
The possibilities are endless.
John E. Hadstate
"David Eather" <eat...@tpg.com.au> wrote in message
news:429b...@dnews.tpgi.com.au...
>> To get
>> the payload 20dB above the noise, and the crypto-stream
>> 20dB
>> above the payload, we need an audio carrier with a
>> signal-to-
>> noise of 40dB, which is 10,000-to-1 in sound power.
>
> IIRC, when I went to school, audio volume was measured in
> dB - spl
> (sound pressure level, which I noted in the pervious post)
> which is
> 20 x log10 and not 10 x log10 which is a measure for
> comparing power
> - - you did not do your basic research. Try -
>
> http://en.wikipedia.org/wiki/Sound_unit
>
> Forty dB is a factor of 100.
40 dB is a factor of 10,000 in power. Power increases with
the square of the voltage (or pressure). It takes a factor
of 100 in voltage (or pressure) to cause a 40 dB change in
power.
If memory serves, binary (two-state) data can always be
correctly decoded if the signal-to-noise ratio in the
channel is greater than 3 dB. For every level you want to
distinguish, 3 dB additional S/N is required. One bit
requires 3 dB, two bits requires 12 dB, three bits requires
24 dB, etc. I think this is a result of one of Shannon's
theorems.
Bryan Olson
>> To get
>>the payload 20dB above the noise, and the crypto-stream 20dB
>>above the payload, we need an audio carrier with a signal-to-
>>noise of 40dB, which is 10,000-to-1 in sound power.
>
> IIRC, when I went to school, audio volume was measured in dB - spl
> (sound pressure level, which I noted in the pervious post) which is
> 20 x log10 and not 10 x log10 which is a measure for comparing power
What did you think "10,000-to-1 in sound power" meant?
A bel is, by definition, a factor-of-10 in power. Sound
pressure increases as the square root of power, so a bel implies
a factor of the square root of 10 in pressure. They are not two
different measures.
'SPL' is an absolute scale, fixed by specifying dB values from a
0dB reference approximately at the edge of human hearing. Saying
+20db(spl) above the ambient noise
doesn't make sense. It's 20dB above the ambient noise, which is
a factor of 10 in pressure and a factor of 100 in power. If we
want to say how loud the ambient noise is in an absolute sense,
there we'd use SPL.
> - - you did not do your basic research. Try -
>
> http://en.wikipedia.org/wiki/Sound_unit
>
> Forty dB is a factor of 100.
Forty dB is a factor of 100 in sound pressure and a factor of
10,000-to-1 in sound power. Click the "Decibel" link on the page
you cited and read the paragraph under "Definition".
The way the notation works, one does not specify decibels of
pressure or decibels of power. A decibel figure implies both the
pressure multiple and the power multiple. When one goes to say
what the multiple is, then one must say power or pressure,
because the power multiple is the square of the pressure
multiple.
[...]
> The OP implied the transmition of the sound data should occur in real
> time when he said:
Which step do you think a cheap PC cannot do in real time?
> As another poster pointed out a speaker used for acoustic transmition
> has all sorts of phase shifts
Not all sorts, mostly just one sort: speakers send different
frequencies to different physical sound-generators.
> which will be unpredictable and
> unrepeatable unless you want to specify exactly which speaker (not
> which brand - which one), sound card, microphone and which
> temperature, pressure and humidity conditions apply.
Why should those have phase-shift problems?
> You could only
> use the most crude forms of phase shift modulation, if any. You'll
> be *lucky* if you get an unreliable 20k bits per second.
A made-up figure, right?
--
--Bryan
Mike Amling
> BO wrote
>> your solution is to transmit the
>> sum of payload-sound plus a 20dB larger crypto-pseudo-random
>> noise signal. Then the receiver subtracts off the crypto signal,
>> to get the plaintext plus any noise on the transmission. For the
>> sound to "be easily distinguished", we want the plaintext sound
>> to be 20dB above the noise on the transmission channel. To get
>> the payload 20dB above the noise, and the crypto-stream 20dB
>> above the payload, we need an audio carrier with a signal-to-
>> noise of 40dB, which is 10,000-to-1 in sound power.
>
>
> IIRC, when I went to school, audio volume was measured in dB - spl
> (sound pressure level, which I noted in the pervious post) which is
> 20 x log10 and not 10 x log10 which is a measure for comparing power
> - - you did not do your basic research. Try -
>
> http://en.wikipedia.org/wiki/Sound_unit
>
> Forty dB is a factor of 100.
A factor of 100 in amplitude (which is what "sound pressure level"
sound like) is a factor of 10,000 in power.
--Mike Amling
David Eather
Hash: SHA1
John E. Hadstate wrote:
> "David Eather" <eat...@tpg.com.au> wrote in message
> news:429b...@dnews.tpgi.com.au...
>
>
>>> To get
>>> the payload 20dB above the noise, and the crypto-stream
>>> 20dB
>>> above the payload, we need an audio carrier with a
>>> signal-to-
>>> noise of 40dB, which is 10,000-to-1 in sound power.
>>
>> IIRC, when I went to school, audio volume was measured in
>> dB - spl
>> (sound pressure level, which I noted in the pervious post)
>> which is
>> 20 x log10 and not 10 x log10 which is a measure for
>> comparing power
>> - - you did not do your basic research. Try -
>>
>> http://en.wikipedia.org/wiki/Sound_unit
>>
>> Forty dB is a factor of 100.
>
The relevant measure for acoustic intensity (volume) is dB(spl) and
40dB(spl) = 100
BO defined the wrong measurement, and then wrongly applied it. Tough
luck,
everyone get over it.
> 40 dB is a factor of 10,000 in power. Power increases with
> the square of the voltage (or pressure).
As above
>It takes a factor
> of 100 in voltage (or pressure) to cause a 40 dB change in
> power.
Audio volume as measured by our ears and microphones is a measurement
of
pressure (or to split a hair, a pressure differential). I dealt with
that in
the previous post - which I know you didn't see (I haven't posted it
yet).
If you want a change in signal level there are two other ways to deal
with
the problem which have nothing to do with power - move the microphone
closer
to the speaker or use a more efficient speaker.
BO defined the wrong measurement, and then wrongly applied it. Tough
luck.
>
> If memory serves, binary (two-state) data can always be
> correctly decoded if the signal-to-noise ratio in the
> channel is greater than 3 dB. For every level you want to
> distinguish, 3 dB additional S/N is required. One bit
> requires 3 dB, two bits requires 12 dB, three bits requires
> 24 dB, etc.
dB's are a log unit are you sure about the above? I suspect the
answer you
want is 3, 6, 9, 12 dB(p) respectively - but if you are talking about
the
signal at the microphone end of this project (which is a pressure
differential) you might mean 6, 12, 18, 24 dB(spl).
>I think this is a result of one of Shannon's
> theorems.
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA/AwUBQpvQRZS9Fk5okqe7EQKznwCghKYXyQqVMJbel/uhy91Vuz2kzyEAoLnt
HxfdkdNiASL2N5SUQ93ZvC2J
=/jxC
-----END PGP SIGNATURE-----
David Eather
Hash: SHA1
"Mike Amling" <nos...@nospam.com> wrote in message
news:E8Ime.16271$26.1...@chiapp18.algx.net...
Talking about power when the subject is volume (acoustic intensity)
is irrelevant. If I
have a speaker that produces 86dB(spl) at one meter with one watt
(typical test parameters - typical domestic speaker). What are you
going to tell me about the
volume at 2 meters? One watt or 80dB(spl). One answer is nonsense,
one answer has meaning.
To illustrate the mistake of using power levels to try and describe
volume consider these options.
I can also change the sound intensity by moving the microphone
relative to the speakers, use a more efficient speaker, change the
room acoustics, use multiple speakers in an array or use a more
efficient speaker enclosure. All change the sound level but have
nothing to do with power.
I read Brian's stuff because he is usually interesting. In this case
Brian defined and used the wrong units. He was in error.
>
> --Mike Amling
Just going over the stuff at the top of the post which seems to have
sparked all this. Was it understood that addition of noise to a
signal would/could/should be done like most other addition in crypto
(mod something appropriate)?
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA/AwUBQpvVFJS9Fk5okqe7EQLMewCggAQ5iv/5INuILsP3BZazQ+DRGr4An364
kISsbvCE6lBdL/OSA+k8OBuV
=r6uM
-----END PGP SIGNATURE-----
David Eather
Hash: SHA1
Jan Panteltje wrote:
> On a sunny day (Tue, 31 May 2005 00:28:50 +1000) it happened "David
> Eather" <eat...@tpg.com.au> wrote in <429b...@dnews.tpgi.com.au>:
>
>> http://en.wikipedia.org/wiki/Sound_unit
>>
>> Forty dB is a factor of 100.
> For voltage only.
No. Your own entry into the reference stakes:
http://encyclopedia.laborlawtalk.com/Decibel#Acoustics
quote:
Acoustics
The decibel unit is often used in acoustics to quantify sound levels
relative to some 0 dB reference. *The reference may be defined as a
sound pressure level (SPL)*, commonly 20 micropascals (20 µPa). To
avoid confusion with other decibel measures, the term dB(SPL) is used
for this. *The reference sound pressure* (corresponding to a sound
pressure level of 0 dB) can also be defined as the sound pressure at
the threshold of human hearing, which is conventionally taken to be
*2×10-5 newton per square metre, 2×10-5 N/m² or 20 micropascals*.
That is roughly the sound of a mosquito flying 3 m away. *The ears
are only sensitive to sound pressure deviations.*
end quote:
As you can see all the references to volume (acoustic intensity) and
its perception are
measured and defined in terms of pressure. The correct unit is
dB(spl)
> But 2 x voltage results is 4 x power in same impedance (speaker).
> P = (U * U) / R
> http://encyclopedia.laborlawtalk.com/Decibel#Acoustics
I don't wish to embarrass, but I am tertiary qualified in electrical
engineering. Ohms law and derivatives of it are not exactly new to
me. I could have defined Pi at the beginning and claimed it gave me
some sort of authority. Everyone would then rightly point out its
irrelevance to measuring a sound's volume.
We are talking about recording a sound with a microphone.
The microphone is like out ear only sensitive to pressure variations.
Talking about power when the subject is volume is irrelevant. If I
have a speaker that produces 86dB(spl) at one meter with one watt
(typical test parameters). What are you going to tell me about the
volume at 2 meters? One watt or 80dB(spl). One answer is nonsense,
one answer has meaning.
A conversation with the same error as BO's could go like this (sorry
I can't think of the crypto equivalent)
B1 "Power equals volts multiplied by current"
B2 "Yeah great. What's the car battery voltage?"
B1 "Five thousand watts"
>> OP wrote>
>> I need the cyphered sound to be audible, but not understandable
> I think this can also be done by using a say 5kHz audio bandwidth
> (more then enough for speech), and AM modulating a carrier at
> 20kHz. The lower sideband will go to 20 - 5 = 15 kHz.
> You will hear some high peaked noises.
> An other alternative is to use a FM carrier, and narrow band FM (so
> use only one sideband), and use 10 kHz carrier :-)
> This requires some filtering at the rx side, but any dsp board
> should
I like that idea. I would just suggest that it will not suit the OP
in terms of resources and ease of implementation
> do. FM has the advantage that it is not so sensitve to amplitude
> changes.
> The possibilities are endless.
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA+AwUBQpvOj5S9Fk5okqe7EQLlxACYz1tGROB0GRb5d1wx1PJDWwlAdgCfdkx+
kfmHzAwlhnh3qZUquheb5Yk=
=MoNc
-----END PGP SIGNATURE-----
Bryan Olson
> We are talking about recording a sound with a microphone.
> The microphone is like out ear only sensitive to pressure variations.
>
> Talking about power when the subject is volume is irrelevant.
I was talking about the transmitted carrier's signal-to-noise
ratio:
So, correct me if I'm wrong: your solution is to transmit
the sum of payload-sound plus a 20dB larger crypto-pseudo-
random noise signal. Then the receiver subtracts off the
crypto signal, to get the plaintext plus any noise on the
transmission. For the sound to "be easily distinguished", we
want the plaintext sound to be 20dB above the noise on the
transmission channel. To get the payload 20dB above the
noise, and the crypto-stream 20dB above the payload, we need
an audio carrier with a signal-to-noise of 40dB, which is
10,000-to-1 in sound power.
Transmission strength is universally expressed as power, not
pressure or voltage. More fundamentally, the carrier's
theoretical capacity in digits base d is:
bandwidth * log_d(1 + S/N)
where S/N is signal power over the (Gaussian) noise power. The
10,000-to-1 signal-to-noise power ratio is vastly more important
than the 100-to-1 audio volume ratio.
[...]
> A conversation with the same error as BO's could go like this (sorry
> I can't think of the crypto equivalent)
>
> B1 "Power equals volts multiplied by current"
>
> B2 "Yeah great. What's the car battery voltage?"
>
> B1 "Five thousand watts"
Except that's the wrong subject. A talk with DE (who probably
likes his initials better than I like mine) might go:
E1: "How powerful is that radio station?"
E2: "5000 volts"
>>>OP wrote>
>>>I need the cyphered sound to be audible, but not understandable
>>
>>I think this can also be done by using a say 5kHz audio bandwidth
>>(more then enough for speech), and AM modulating a carrier at
>>20kHz. The lower sideband will go to 20 - 5 = 15 kHz.
>>You will hear some high peaked noises.
>>An other alternative is to use a FM carrier, and narrow band FM (so
>>use only one sideband), and use 10 kHz carrier :-)
>>This requires some filtering at the rx side, but any dsp board
>>should
>
> I like that idea. I would just suggest that it will not suit the OP
> in terms of resources and ease of implementation
In WWII, the U.S. and Briton used a slightly more sophisticated
idea along the same lines. Not only was it broken, with
training, operators could understand the signals directly. See:
http://en.wikipedia.org/wiki/Single-sideband_modulation
--
--Bryan
Unruh
>http://encyclopedia.laborlawtalk.com/Decibel#Acoustics
>quote:
> Acoustics
>end quote:
Actually the definition of dB is in terms of energy rate per square meter.
the zero reference is 10^-12 watts/m^2. (this is the threshold of hearing
from some standard good listener.) Every factor of 10 increase in energy is
10dB (or 1Bel).
For air, with a velocity of sound of 340m/sec, and density of about
1.4Kg/m^3, this corresponds to a pressure of about 2 10^-5 P. Note that
because the energy rate in a sound wave is proportional to pressure^2, this
means that every factor of 10 in pressure is a factor of 100 in energy or
20dB. But the fundamental definition of dB is in terms of energy not
amplitude.
Note that in water, with its much higher density and velocity of sound, the
relation between energy and pressure is very different. This has caused
immense confusion in the definition of dB in water (beause of the
difference in reference energy rate per unit area and pressure).
Ie, 120dB in water can mean either 1 watt/m^2 or a pressure of 20P which
corresponds to an energy rate/m^2 of about 3 10^(-4) watts/m^2 (ie 33dB
less).
>> But 2 x voltage results is 4 x power in same impedance (speaker).
>> P = (U * U) / R
>> http://encyclopedia.laborlawtalk.com/Decibel#Acoustics
>I don't wish to embarrass, but I am tertiary qualified in electrical
>engineering. Ohms law and derivatives of it are not exactly new to
>me. I could have defined Pi at the beginning and claimed it gave me
>some sort of authority. Everyone would then rightly point out its
>irrelevance to measuring a sound's volume.
>We are talking about recording a sound with a microphone.
>The microphone is like out ear only sensitive to pressure variations.
>Talking about power when the subject is volume is irrelevant. If I
>have a speaker that produces 86dB(spl) at one meter with one watt
>(typical test parameters). What are you going to tell me about the
>volume at 2 meters? One watt or 80dB(spl). One answer is nonsense,
>one answer has meaning.
The volume (intensity) of a sound IS power per m^2. Your example is
irrelevant. The 1 watt here is the INPUT to the speaker. Speakers are
incredibly inefficient. In your example, 86dB is about 4 10^{-4}w/m^2.
Since a sphere of 1m radius has about 12m^2 area this would mean a total
sound energy of about 5 10^-3 w. (Ie, the efficiency is only about .5%
conversion from input energy to sound, assuming uniform angular
radiation).
>A conversation with the same error as BO's could go like this (sorry
>I can't think of the crypto equivalent)
>B1 "Power equals volts multiplied by current"
>B2 "Yeah great. What's the car battery voltage?"
>B1 "Five thousand watts"
>>> OP wrote>
>>> I need the cyphered sound to be audible, but not understandable
>> I think this can also be done by using a say 5kHz audio bandwidth
>> (more then enough for speech), and AM modulating a carrier at
>> 20kHz. The lower sideband will go to 20 - 5 = 15 kHz.
>> You will hear some high peaked noises.
Actually due to the non-linearity in the ear you are liable to hear the
original sound. As cats-whisker radio recievers found, a very simply
non-linear elemnt ( the diode of the cat's whicsker) allowed one to hear
the radio broadcast very well.
>> An other alternative is to use a FM carrier, and narrow band FM (so
>> use only one sideband), and use 10 kHz carrier :-)
again a phase shifter (which loudspeakers are really great at) and a
non-linear element ( the ear) would again allow the sound to be heard.
Unruh
>John E. Hadstate wrote:
>>>> signal-to-
>>>> noise of 40dB, which is 10,000-to-1 in sound power.
Perfectly correct.
>>>
>>> IIRC, when I went to school, audio volume was measured in
>>> dB - spl
>>> (sound pressure level, which I noted in the pervious post)
>>> which is
>>> 20 x log10 and not 10 x log10 which is a measure for
>>> comparing power
Yes, and he discussed power.
>>> - - you did not do your basic research. Try -
>>>
>>> http://en.wikipedia.org/wiki/Sound_unit
>>>
>>> Forty dB is a factor of 100.
No it is NOT 100 in power. It is a factor of 100 in amplitude.
>>
>The relevant measure for acoustic intensity (volume) is dB(spl) and
>40dB(spl) = 100
Intensity is related to power not amplitude. Read your physics texts.
Unruh
>> > BO wrote
>> >> your solution is to transmit the
>> >> sum of payload-sound plus a 20dB larger crypto-pseudo-random
>> >> noise signal. Then the receiver subtracts off the crypto signal,
>> >> to get the plaintext plus any noise on the transmission. For the
>> >> sound to "be easily distinguished", we want the plaintext sound
>> >> to be 20dB above the noise on the transmission channel. To get
>> >> the payload 20dB above the noise, and the crypto-stream 20dB
>> >> above the payload, we need an audio carrier with a signal-to-
>> >> noise of 40dB, which is 10,000-to-1 in sound power.
>Just going over the stuff at the top of the post which seems to have
>sparked all this. Was it understood that addition of noise to a
>signal would/could/should be done like most other addition in crypto
>(mod something appropriate)?
Yes, the idea was to add a random stream to the digitized sound mod the max
amplitude. This would produce a white noise stream with no speach
identifiable in the output. IT would be pure white noise. The decryption
would be to subtract (mod that max) that same random stream.
However this would have real problems with both amplitude but more
crucially differential phase shifts in the sound. I do not believe this
technique would work, unless you had complete control over the whole sound
chain, in which case why bother with the conversion to sound.
Jan Panteltje
<eat...@tpg.com.au> wrote in <429b...@dnews.tpgi.com.au>:
>Jan Panteltje wrote:
>> On a sunny day (Tue, 31 May 2005 00:28:50 +1000) it happened "David
>> Eather" <eat...@tpg.com.au> wrote in <429b...@dnews.tpgi.com.au>:
>>
>>> http://en.wikipedia.org/wiki/Sound_unit
>>>
>>> Forty dB is a factor of 100.
>> For voltage only.
>
>No. Your own entry into the reference stakes:
>
>B1 "Power equals volts multiplied by current"
>
>B2 "Yeah great. What's the car battery voltage?"
>
>B1 "Five thousand watts"
hey! Not completely fair.
I merely stated that doubling the voltage in the same load, results
in 4 x the power.
Since our ear is logarithmic (in volume) we do not percieve 4 x more volume.
For a car battery the energy content is measured in ampere-hour.
So 100 Ah is 100 ampere for 1 hour (or 50 for 2).
There is a practical limit as the battery internal resistance will limit Imax.
So, I dunno, maybe I misunderstood your original question or was it statement.
Here is a better, more clearly written link:
http://www.phys.unsw.edu.au/~jw/dB.html#definition
>>> OP wrote>
>>> I need the cyphered sound to be audible, but not understandable
>> I think this can also be done by using a say 5kHz audio bandwidth
>> (more then enough for speech), and AM modulating a carrier at
>> 20kHz. The lower sideband will go to 20 - 5 = 15 kHz.
>> You will hear some high peaked noises.
>> An other alternative is to use a FM carrier, and narrow band FM (so
>> use only one sideband), and use 10 kHz carrier :-)
>> This requires some filtering at the rx side, but any dsp board
>> should
>
>I like that idea. I would just suggest that it will not suit the OP
>in terms of resources and ease of implementation
Dunno, for FM 2 4046 PLL chips, some opamps....
I have done that.
PC with soundcard and some soft would do too.
I was wondering where OP was getting this idea from, then I found
an article on nytimes.com about a device that makes telephone
conversations un-intelligible for other people (standing next to it, so
they cannot hear what you say).
It does that by making some bogus sounds if I did read it right.
Maybe that is what made him ask?
Jan Panteltje
<unruh...@physics.ubc.ca> wrote in <d7h1hj$apa$1...@nntp.itservices.ubc.ca>:
>Actually the definition of dB is in terms of energy rate per square meter.
>the zero reference is 10^-12 watts/m^2. (this is the threshold of hearing
>from some standard good listener.) Every factor of 10 increase in energy is
>10dB (or 1Bel).
>For air, with a velocity of sound of 340m/sec, and density of about
>1.4Kg/m^3, this corresponds to a pressure of about 2 10^-5 P. Note that
>because the energy rate in a sound wave is proportional to pressure^2, this
>means that every factor of 10 in pressure is a factor of 100 in energy or
>20dB. But the fundamental definition of dB is in terms of energy not
>amplitude.
>
>Note that in water, with its much higher density and velocity of sound, the
>relation between energy and pressure is very different. This has caused
>immense confusion in the definition of dB in water (beause of the
>difference in reference energy rate per unit area and pressure).
>
>Ie, 120dB in water can mean either 1 watt/m^2 or a pressure of 20P which
>corresponds to an energy rate/m^2 of about 3 10^(-4) watts/m^2 (ie 33dB
>less).
Interesting ;-)
>>> I think this can also be done by using a say 5kHz audio bandwidth
>>> (more then enough for speech), and AM modulating a carrier at
>>> 20kHz. The lower sideband will go to 20 - 5 = 15 kHz.
>>> You will hear some high peaked noises.
>
>Actually due to the non-linearity in the ear you are liable to hear the
>original sound. As cats-whisker radio recievers found, a very simply
>non-linear elemnt ( the diode of the cat's whicsker) allowed one to hear
>the radio broadcast very well.
Yes, what you want to 'demodulate' an AM (amplitude modulated) carrier
is detection.
You need the carrier and at least one sideband.
The litte detector (I actually had one as a kid) you describe, is in
fact a diode (you need to find a spot where it conducts electricity
one way).
It did not work very well (the one I had, or I had no patience), but
once I got a germanium diode (OA79 IIRC) my xtal radio worked really well!
I am not sure if the carrier is at 20 kHz, enough of it will be present
for detection by the ear (would have to try).
The lower sideband would interfere in the audio range big time anyways,
it is 'reversed' so a modulating tone of say 1000 Hz would appear in AM
in this system as a lower sideband at 19kHz, a carrier at 20kHz, and an
upper sideband at 21 kHz.
But for say 3 kHz, the lower sideband will be at 17 kHz! So the higher
the modulating frequency, the lower the tone you hear....
The higher sideband would be 23kHz, and out of audible range anyways.
But you are right, almost ANY nonlinear medium will detect....
Old FM transmissions could be received by an AM receiver tuning to the
slope (side) of the tuning curve, then any change in frequency caused a
change in amplitude.
Great fun stuff to play with.
But if you are worried about 'detection by non linear ear', let's use
double sideband surpressed carrier, easier to make too.
That was my first proposition, diode ring modulator.
Who needs DSP! I grew up with analog electronics ;-)
>>> An other alternative is to use a FM carrier, and narrow band FM (so
>>> use only one sideband), and use 10 kHz carrier :-)
>
>again a phase shifter (which loudspeakers are really great at) and a
>non-linear element ( the ear) would again allow the sound to be heard.
You'd hear a howling tone.
But indeed perhaps with some training....
Phase shift has no influence on FM.
Only detection would be non linear frequency of ear, but with a 10kHz
loud beep, not so sure you could make out words... but you could try :-)
I say: double sideband surpressed carrier :-)
David Eather
Strange. The signal in (acoustic intensity) is measured in dB(spl) - and it
is, and the background noise is also measured in dB(spl) and it is. ????
Sound volume (acoustic intensity - the signal discussed here) is measured in
terms of pressure. I did not make it so. I just point out that Wiki,
Britannica etc agree with me. I'm sorry.
If your talking about RF signals, which are measured in terms of power input
/ output, I would agree with you. The units used for what you *hear* are
dB(spl)
>
>
> [...]
> > A conversation with the same error as BO's could go like this
> (sorry > I can't think of the crypto equivalent)
> >
> > B1 "Power equals volts multiplied by current"
> >
> > B2 "Yeah great. What's the car battery voltage?"
> >
> > B1 "Five thousand watts"
>
> Except that's the wrong subject. A talk with DE (who probably
> likes his initials better than I like mine)
OK - I'll move away from initials. (B1 and B2 weren't references to Brian -
they are from children's TV. It just happen to be on at the time.)
might go:
>
> E1: "How powerful is that radio station?"
>
> E2: "5000 volts"
Not likely.
>
>
> >>>OP wrote>
> >>>I need the cyphered sound to be audible, but not understandable
> >>
> >>I think this can also be done by using a say 5kHz audio bandwidth
> >>(more then enough for speech), and AM modulating a carrier at
> >>20kHz. The lower sideband will go to 20 - 5 = 15 kHz.
> >>You will hear some high peaked noises.
> >>An other alternative is to use a FM carrier, and narrow band FM (so
> >>use only one sideband), and use 10 kHz carrier :-)
> >>This requires some filtering at the rx side, but any dsp board
> >>should
> >
> > I like that idea. I would just suggest that it will not suit the
> OP > in terms of resources and ease of implementation
>
> In WWII, the U.S. and Briton used a slightly more sophisticated
> idea along the same lines. Not only was it broken, with
> training, operators could understand the signals directly. See:
>
> http://en.wikipedia.org/wiki/Single-sideband_modulation
SigSaly has already had a mention in this thread - I misspelt it as SigSally
David Eather
So the above (top) reference,(provided by someone else) that states clearly
that acoustic intensity is a pressure, not a power measurement is wrong?
Wiki says the same thing - it's wrong too? And Britannica?
quote:
>> The decibel unit is often used in acoustics to quantify sound levels
>> relative to some 0 dB reference. *The reference may be defined as a
>> sound pressure level (SPL)*, commonly 20 micropascals (20 µPa). To
>> avoid confusion with other decibel measures, the term dB(SPL) is used
>> for this. *The reference sound pressure* (corresponding to a sound
>> pressure level of 0 dB) can also be defined as the sound pressure
endquote:
>Your example is irrelevant.
No.
>The 1 watt here is the INPUT to the speaker.
Yes.
>Speakers are incredibly inefficient.
Yes.
>In your example, 86dB is about 4 10^{-4}w/m^2.
Yes. 86 dB(spl) at 1 watt input measured at 1 metre is typical result for a
mid-sized domestic speaker.
> Since a sphere of 1m radius has about 12m^2 area this would mean a
> total sound energy of about 5 10^-3 w. (Ie, the efficiency is only
> about .5%
Yes.
Example -
http://www.wharfedale.co.uk/model.php?model_id=17
Larger and 3 db more efficient
http://www.wharfedale.co.uk/model.php?model_id=1
Believe me now?
Drivers used in public venue work, where flat smooth response from 20 - 20k
is not required are more efficient - about 96db(spl) 1watt @.1 metre
wouldn't be a surprise. You further boost the output volume by using horns
or stacks/arrays.
David Eather
> On a sunny day (Tue, 31 May 2005 12:43:30 +1000) it happened "David
> Eather" <eat...@tpg.com.au> wrote in <429b...@dnews.tpgi.com.au>:
>
>> Jan Panteltje wrote:
>>> On a sunny day (Tue, 31 May 2005 00:28:50 +1000) it happened "David
>>> Eather" <eat...@tpg.com.au> wrote in <429b...@dnews.tpgi.com.au>:
>>>
>>>> http://en.wikipedia.org/wiki/Sound_unit
>>>>
>>>> Forty dB is a factor of 100.
>>> For voltage only.
>>
>> No. Your own entry into the reference stakes:
>>
>> B1 "Power equals volts multiplied by current"
>>
>> B2 "Yeah great. What's the car battery voltage?"
>>
>> B1 "Five thousand watts"
> hey! Not completely fair.
> I merely stated that doubling the voltage in the same load, results
> in 4 x the power.
Sorry B1 and B2 weren't aimed at you - I was just pointing to nonsense that
results when the wrong units of measure are used.
David Eather
> "David Eather" <eat...@tpg.com.au> writes:
>
>> John E. Hadstate wrote:
>>>>> signal-to-
>>>>> noise of 40dB, which is 10,000-to-1 in sound power.
>
> Perfectly correct.
>
>>>>
>>>> IIRC, when I went to school, audio volume was measured in
>>>> dB - spl
>>>> (sound pressure level, which I noted in the pervious post)
>>>> which is
>>>> 20 x log10 and not 10 x log10 which is a measure for
>>>> comparing power
>
> Yes, and he discussed power.
>
>>>> - - you did not do your basic research. Try -
>>>>
>>>> http://en.wikipedia.org/wiki/Sound_unit
>>>>
>>>> Forty dB is a factor of 100.
>
> No it is NOT 100 in power. It is a factor of 100 in amplitude.
Yes that is what I said a factor of 100 is equivalent to 40dB spl.
dB(spl) is the correct measure for sound pressure, dB(p) (and dBm etc) is
not. I only ever spoke about volume. I used the correct units. Brian
introduce the power unit and that unit is not the correct one for volume.
>
>>>
>> The relevant measure for acoustic intensity (volume) is dB(spl) and
>> 40dB(spl) = 100
>
> Intensity is related to power not amplitude. Read your physics texts.
OK. I see your point - my use of the word "intensity" is incorrect - sound
pressure, sound pressure level are the correct terms for discussing volume -
I would have used "volume" but I wanted to avoid a discussion of how you
could make more volume (as in size) by just turning a nob.
Unruh
>You'd hear a howling tone.
>But indeed perhaps with some training....
>Phase shift has no influence on FM.
A phase shift which changes with frequency changes FM to AM. That is how FM
demodulators work.
>Only detection would be non linear frequency of ear, but with a 10kHz
>loud beep, not so sure you could make out words... but you could try :-)
differential phase shift, non-linearity, 10KHz fitering and there you are.
Clear signal.
Unruh
>Strange. The signal in (acoustic intensity) is measured in dB(spl) - and it
No, it is not. That is a derived quantity. It is measured in power per unit
area. Why do you think, in terms of pressure, it is 20 log(<Prms>) why the
20? Because the Bel was defined in terms of power, in which case it is
10 log(power)
Now some people liked to talk about amplitudes instead and since power is
proportional to amplitude squared that geve the factor of 20.
>is, and the background noise is also measured in dB(spl) and it is. ????
>Sound volume (acoustic intensity - the signal discussed here) is measured in
>terms of pressure. I did not make it so. I just point out that Wiki,
>Britannica etc agree with me. I'm sorry.
The reason for the spl subscript is precisely to let the reader know that
these are not the usual dB defined in terms of 10^-12 watts/m^2, but the
"sound Pressure Level" dB defined in terms of 2 10^-5 pascal. (why in the
world would that 2 be there if it were the fundamental definition?)
Bryan Olson
> Strange. The signal in (acoustic intensity) is measured in dB(spl) -
and it
> is, and the background noise is also measured in dB(spl) and it is. ????
> Sound volume (acoustic intensity - the signal discussed here) is
measured in
> terms of pressure. I did not make it so.
And 40dB is a factor of 10,000 to one in power, just as I and
others have been telling you, and you cannot stop it from being
so. Your personal refusal to consider power is not going to make
it go away, make your system any more efficient, nor make other
people's statements about power wrong or irrelevant.
> I just point out that Wiki, Britannica etc agree with me. I'm sorry.
Not on the points at issue they don't. This Britannica
explanation is terrible, but it certainly doesn't agree with
you:
http://www.britannica.com/ebc/article?tocId=9362398&query=decibel&ct=
The Wikipedia explanation of decibels:
http://en.wikipedia.org/wiki/Decibel
Agrees with my explanation in:
<7SHme.307$IE7...@newssvr21.news.prodigy.com>
http://groups-beta.google.com/group/sci.crypt/msg/19d49cbcc14f8c50
--
--Bryan
Jan Panteltje
<unruh...@physics.ubc.ca> wrote in <d7i8l0$onc$1...@nntp.itservices.ubc.ca>:
>Jan Panteltje <pNaonSt...@yahoo.com> writes:
>
>>You'd hear a howling tone.
>>But indeed perhaps with some training....
>>Phase shift has no influence on FM.
>
>A phase shift which changes with frequency changes FM to AM. That is how FM
>demodulators work.
Well, eh, some Foster Seeley and ratio detector, I have designed with these.
Here is a good link on the ratio detector:
http://www.tpub.com/neets/book12/51d.htm
Note you need 2 AMPLITUDE detectors (the diodes).
Old school stuff, I had to do once....
Then there is the quadrature detector, this is a reasonable link:
http://hem.passagen.se/communication/quadcoil.html it is basically a tuned
circuit, uses phase shift, but you need a multiplier!
But I wanted to use 4046 chips, these are phase locked loops.
An oscillator is locked to the incoming frequency, and the control voltage
by the frequency detector (xor gate for example) is proportional to
the frequency change, so in this case contains the audio.
A phase shift that changes with frequency IN ITSELF does nothing.
Any tuned circuit exibits this, say a parallel LC with a current f fed into it.
What the old detectors did is SUBSTRACT the shifted signal from the original,
then when in phase you get zero, when the phase shifts you get some signal.
You need to still recify and filter that (or use a multiplier).
There is also the 'slope detector' for FM, as I described in an other post,
here you tune so the carrier is at the side (slope) of the curve, now you
any frequency change causes an amplitude change, now you can use an AM
detector.
There are several more systems, for example in digital you can make small fixed
length pulses with a one-shot, every zero crossing of the incoming signal.
Then average these pulses with a simple low pass RC, and you have a FM detector.
(the higher the frequency - the more pulses - the higher the average DC value.
(affected by pulse width and pulse height).
For very fast logic (or low frequency signals) you could count how many
pulses fit in a period of the incoming signal (use signal as gate) etc etc...
PLL and averaging pulses is very linear, last system was for example used in
the video demodulator of the first 'home' video recorders (Umatic).
Problem here is that if you missed a pulse (dropout on tape), you'd get a
huge spike in the output.
You can make a PLL so it ignores missing pulses (freewheels).
I invented a nice system in the eighties that did just that,
half digital half analog.
Facinating stuff, and still used daily.
>>Only detection would be non linear frequency of ear, but with a 10kHz
>>loud beep, not so sure you could make out words... but you could try :-)
>
>differential phase shift, non-linearity, 10KHz fitering and there you are.
>Clear signal.
Yes, but the ear has no 10kHz filter!
So, I think we agree basically, given some electronics we can of cause always
demodulate / crack this system.
But if a normal listener can make out what is said, I am not so sure.
Mike Amling
> Just going over the stuff at the top of the post which seems to have
> sparked all this. Was it understood that addition of noise to a
> signal would/could/should be done like most other addition in crypto
> (mod something appropriate)?
If the OP's goal is to make some encrypted sound and recover a
plaintext from a recording of that audio ciphertext, I think he'd be
better off trying to digitize his message, encrypt it using normal
cryptographic methods and modulate the ciphertext into the sound wave
using robust analog techniques and error correcting codes, something
similar to what must be used by radio modems (or what's that
IP-over-ham-radio called?) that can bounce a signal off a
none-too-stable ionosphere and still recover the digital data.
--Mike Amling
David Eather
> "David Eather" <eat...@tpg.com.au> writes:
>
>
>> Strange. The signal in (acoustic intensity) is measured in dB(spl)
>> - and it
>
> No, it is not.
I explained why and apologised for my error of using the word "intensity"
when volume, sound pressure and sound pressure level were the correct term.
Once again I am sorry for the error.
David Eather
> David Eather wrote:
> > Strange. The signal in (acoustic intensity) is measured in
> dB(spl) - and it
> > is, and the background noise is also measured in dB(spl) and it
> is. ???? > Sound volume (acoustic intensity - the signal discussed
> here) is measured in
> > terms of pressure. I did not make it so.
>
> And 40dB is a factor of 10,000 to one in power, just as I and
> others have been telling you, and you cannot stop it from being
> so.
I agreed that is correct but sound volume is a measure of pressure and is
measured in db(spl) a unit that is 20 x log10 (X/Y)
Your personal refusal to consider power is not going to make
> it go away, make your system any more efficient, nor make other
> people's statements about power wrong or irrelevant.
>
> > I just point out that Wiki, Britannica etc agree with me. I'm
> sorry.
>
> Not on the points at issue they don't. This Britannica
> explanation is terrible, but it certainly doesn't agree with
> you:
>
>
> http://www.britannica.com/ebc/article?tocId=9362398&query=decibel&ct=
>
> The Wikipedia explanation of decibels:
>
> http://en.wikipedia.org/wiki/Decibel
>
> Agrees with my explanation in:
>
> <7SHme.307$IE7...@newssvr21.news.prodigy.com>
>
> http://groups-beta.google.com/group/sci.crypt/msg/19d49cbcc14f8c50
Volume as in the sound pressure perceived by our ears IS measured in
dB(spl) - show me any reference that says what our ears perceive as sound
volume is power and not pressure.
David Eather
Hash: SHA1
<snip - and no disrespect intended>
My number one bitch - Not one of those people with high tech, big
implementation ideas actually bothered to ask the OP if their
approach would be suitable or even implementable. And without any
further concern for the OP they built on their fantasy of what the
problem was.
Clap!......Clap!......Clap!
Problem solving and implantation according to the OP specifications
and the irrelevant implementations posters have come up with. How
can people be so dense on Sci.Crypt?
The OP specified how he wanted to transmit the data - by sound, not
radio, not fibre optic or anything else. If use other than audio you
are outside the OP's specification and your solution is not relevant
to the original post.
The OP also specified a low level security requirement and his post
shows that he is venturing into new territory and in that area did
not have a vast pool of specific knowledge for him to draw on. A
complex implementation is useless to the OP as he will not be able to
implement it.
Yes there are "better" ways to solve this problem - If you thought
they were necessary you should have asked the OP if they were
permissible, but no one did. So if you want to solve the OP's
problem you have to stick to his specifications - which none of you
did.
Then you had the ignorance to criticise my roughed solution because
it is "inelegant" and "imperfect". Well my concept is that. It is
also the only one put forward that addresses the requirements put by
the OP and may be possible for him to implement.
When you ignore the OP's specification, you ignore his problem and
come up with an ivory tower solution - totally useless because the
person who needs it will never implement it, because he can't.
Your solution can be perfect in every way because it will never be
troubled by the reality of the situation and the real, external
constraints put on the job. The your solution is just a fantasy.
You can't use radio because the OP specified audio.
You can't use something complex to implement because the OP does not
have the skills.
You can't use something expensive because the OP does not have a
budget.
But not one of you though about the OP - you just wanted to pat
yourselves on the back for your fictional, never to be built, never
to be used perfect solution to the completely re-defined problem that
no longer is useful or usable by the OP.
If I ask a student to design and analyse a single transistor,
transformer coupled, speaker amplifier such as you would find in a
cheap radio and the student submits a design of a bi-polar, direct
couple audio amp, using nested differential feedback loops to achieve
distortion level measured in parts per million (there are papers on
this design) - guess what? He fails. The student did not solve the
problem. (The problem is useful as there will be many times he will
have to find a quick, simple and cheap way to put an audio output on
something)
A liquid fuel rocket is in all ways superior to a Chinese firework.
Better specific impulse, higher speeds, more payload, orbital
insertion capability, higher reliability, more controllability etc.
And it remains superior, right up until the moment the requirement is
to make a five year old laugh and clap.
If you want to help the OP then solve the problem in front of you,
not the problem you want to solve.
I am so sick of pricks and their "creative" solutions. If you want
to be creative then at the end of the day you have to have created
something.
If you don't think my solution can work then say so right now. Its
time for us to put up
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA/AwUBQp0ivpS9Fk5okqe7EQJa1ACfYTptoSYv5SfKGCKCv2Mh+8GYsFwAoNw/
lgZ2ggXuUV26a+Shb9iocoDF
=XE4c
-----END PGP SIGNATURE-----
.
John E. Hadstate
"David Eather" <eat...@tpg.com.au> wrote in message
news:429d...@dnews.tpgi.com.au...
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> <snip - and no disrespect intended>
>
> My number one bitch - Not one of those people with high
> tech, big
> implementation ideas actually bothered to ask the OP if
> their
> approach would be suitable or even implementable. And
> without any
> further concern for the OP they built on their fantasy of
> what the
> problem was.
>
> Clap!......Clap!......Clap!
>
If you go back to the original thread, the whole discussion
about transmitting the encrypted signal over speakers
started when I asked the OP if by "over the air" he meant by
radio or literally, through the air. So far as I'm aware,
the OP never responded, and the entire thread took on a life
of its own.
I'm still not sure what the OP wanted or what problem he was
trying to solve. I suspect that he might have been toying
with the idea of some kind of acoustic coupler to interface
his encrypted voice to a telephone or a radio transmitter.
I would put the "through the air" scenario at a very low
probability, though it is an interesting problem in its own
right. I'm not sure why you'd want to do that as a
practical matter.
I still think that permuting the audio spectrum has the best
chance of working in that scenario because I think the ear
will smooth-out most problems with synchronizing the
de-permutation of the spectrum.
Unruh
>Bryan Olson wrote:
>> David Eather wrote:
>> > Strange. The signal in (acoustic intensity) is measured in
>> dB(spl) - and it
>> > is, and the background noise is also measured in dB(spl) and it
>> is. ???? > Sound volume (acoustic intensity - the signal discussed
>> here) is measured in
>> > terms of pressure. I did not make it so.
>>
>> And 40dB is a factor of 10,000 to one in power, just as I and
>> others have been telling you, and you cannot stop it from being
>> so.
>I agreed that is correct but sound volume is a measure of pressure and is
>measured in db(spl) a unit that is 20 x log10 (X/Y)
dB(spl) is a measure for pressure.
dB is a measure of power per unit area.
For sound, dB is 10 log the ratio of the actual power per unit area to
10^(-12) watts/m^2. Again, why do you think they would use 20 in the
definition of dB(Spl)? Why would the reference be 2*10^-5, rather than just
say 10^-5 (and do not say that it is the minimum SPL the ear can hear-- it
isn't. It is a nomial value. The actual minimum varies by more than 10dB
amongst people.)
>Volume as in the sound pressure perceived by our ears IS measured in
>dB(spl) - show me any reference that says what our ears perceive as sound
>volume is power and not pressure.
??? the power per unit area is the pressure(RMS) squared divided by the density
of air times the velocity of sound. Ie, either is convertible to the other.
See Science of Sound-- Rossing and Wheeler as an example.
Unruh
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
><snip - and no disrespect intended>
>My number one bitch - Not one of those people with high tech, big
>implementation ideas actually bothered to ask the OP if their
>approach would be suitable or even implementable. And without any
>further concern for the OP they built on their fantasy of what the
>problem was.
>Clap!......Clap!......Clap!
>Problem solving and implantation according to the OP specifications
>and the irrelevant implementations posters have come up with. How
>can people be so dense on Sci.Crypt?
>The OP specified how he wanted to transmit the data - by sound, not
>radio, not fibre optic or anything else. If use other than audio you
>are outside the OP's specification and your solution is not relevant
>to the original post.
The OP has not told us what he really wants and what he can do.
>The OP also specified a low level security requirement and his post
>shows that he is venturing into new territory and in that area did
>not have a vast pool of specific knowledge for him to draw on. A
>complex implementation is useless to the OP as he will not be able to
>implement it.
>Yes there are "better" ways to solve this problem - If you thought
>they were necessary you should have asked the OP if they were
>permissible, but no one did. So if you want to solve the OP's
>problem you have to stick to his specifications - which none of you
>did.
One assumes that the OP does not need you to talk for him, to guess equally
ignorantly what his desire is or what his level of sophistication is.
>Then you had the ignorance to criticise my roughed solution because
>it is "inelegant" and "imperfect". Well my concept is that. It is
>also the only one put forward that addresses the requirements put by
>the OP and may be possible for him to implement.
>When you ignore the OP's specification, you ignore his problem and
>come up with an ivory tower solution - totally useless because the
>person who needs it will never implement it, because he can't.
He did not give specifications except in very loose terms. Eg, does he have
control over the speakers and microphone, or does he have to make do with
what someone else selects for him? What ability to train his system does he
have (eg modems train each other in a very sophisticated manner. Can he do
so?) Etc.
>Your solution can be perfect in every way because it will never be
>troubled by the reality of the situation and the real, external
>constraints put on the job. The your solution is just a fantasy.
You can tell us what the constraints are?
>You can't use radio because the OP specified audio.
>You can't use something complex to implement because the OP does not
>have the skills.
You know this how?
>You can't use something expensive because the OP does not have a
>budget.
You knwo this how?
Talk about hubris.
>But not one of you though about the OP - you just wanted to pat
>yourselves on the back for your fictional, never to be built, never
>to be used perfect solution to the completely re-defined problem that
>no longer is useful or usable by the OP.
>If I ask a student to design and analyse a single transistor,
>transformer coupled, speaker amplifier such as you would find in a
>cheap radio and the student submits a design of a bi-polar, direct
>couple audio amp, using nested differential feedback loops to achieve
>distortion level measured in parts per million (there are papers on
>this design) - guess what? He fails. The student did not solve the
>problem. (The problem is useful as there will be many times he will
>have to find a quick, simple and cheap way to put an audio output on
>something)
You think a low noise low distortion amp is expensive? Sheesh.
>A liquid fuel rocket is in all ways superior to a Chinese firework.
>Better specific impulse, higher speeds, more payload, orbital
>insertion capability, higher reliability, more controllability etc.
>And it remains superior, right up until the moment the requirement is
>to make a five year old laugh and clap.
>If you want to help the OP then solve the problem in front of you,
>not the problem you want to solve.
And you know what that problem is? Please tell us, since you apparently can
read his mind.
Bryan Olson
> Bryan Olson wrote:
>> 40dB is a factor of 10,000 to one in power, just as I and
>> others have been telling you, and you cannot stop it from being
>> so.
>
> I agreed that is correct
You disagreed, to the extent that you claimed to correct it.
Look:
http://groups-beta.google.com/group/sci.crypt/msg/2eb4cedc2b33d6ab
<429b...@dnews.tpgi.com.au>
> but sound volume is a measure of pressure and is
> measured in db(spl) a unit that is 20 x log10 (X/Y)
You seem to have "but" confused with "and" (though since you
didn't define X and Y, you're still talking nonsense). Did
someone say sound volume is not a measure of pressure? Didn't I
carefully explain how a dB value gives both the pressure and
power factors?
<7SHme.307$IE7...@newssvr21.news.prodigy.com>
http://groups-beta.google.com/group/sci.crypt/msg/19d49cbcc14f8c50
Did I mis-use 'SPL'? You did when you wrote:
In audio engineering the 20db figure was widely quoted.
+20db(spl) above the ambient noise to be easily
distinguished (by an audience).
+20dB(SPL) means 20dB (a factor of 10 in pressure and 100 in
power) above a constant 0dB reference. "+20db(spl) above the
ambient noise," means 20dB above the zero reference above the
ambient noise, which doesn't mean much of anything.
> Volume as in the sound pressure perceived by our ears IS measured in
> dB(spl) - show me any reference that says what our ears perceive as sound
> volume is power and not pressure.
Show me a statement I made that was incorrect and I'll correct
it.
On the question of relevance: the sound identified as +40dB
above ambient noise is *not* intended for our ears. How is
"Volume as in the sound pressure perceived by our ears" relevant
to a signal that is not intended for our ears?
It is, rather, a carrier signal. If you think power is
irrelevant, that just shows that you were wrong when you claimed
competence in E.E.
--
--Bryan
Bryan Olson
> My number one bitch - Not one of those people with high tech, big
> implementation ideas actually bothered to ask the OP if their
> approach would be suitable or even implementable. And without any
> further concern for the OP they built on their fantasy of what the
Sci.crypt has improved over the years in that questions are
receiving better, more accurate, and more relevant answers. It
has simultaneously been getting worse in that more and more
nonsensical and wrong answers also appear, often written as if
the author understands what he does not.
David, what relevant operation do you think not to be
implementable? Ten dollar sound cards do digital-to-analog and
analog-to-digital conversion. Free, open-source audio codecs do
compression and decompression. The crypto -- no problem. Are
there cheap easy-to-use libraries for error correction coding
and modulation onto an analog carrier? I'd guess so, but how
should I know; I'm a cryptographer.
> Yes there are "better" ways to solve this problem
So don't eliminate the best solution out of ignorance. If the
cost of a solution is too high, argue that. There's no such
thing as a quality overrun.
This is a "sci" group. It's about advancing knowledge, not
dumbing things down to someone's comfort level.
--
--Bryan
David Eather
Hash: SHA1
Apologies for a late reply.
Bryan Olson wrote:
> David Eather wrote:
> > My number one bitch - Not one of those people with high tech,
> big
> > implementation ideas actually bothered to ask the OP if their
> > approach would be suitable or even implementable. And without
> any
> > further concern for the OP they built on their fantasy of what
> the [...]
>
> Sci.crypt has improved over the years in that questions are
> receiving better, more accurate, and more relevant answers. It
> has simultaneously been getting worse in that more and more
> nonsensical and wrong answers also appear, often written as if
> the author understands what he does not.
Careful. You could come out on the wrong side of that. What if what
you had decided was "nonsensical" worked. Wouldn't you then be the
fool, and then twice so, as you not only didn't understand the answer
but also jeered and made fun of it?
>
> David, what relevant operation do you think not to be
> implementable?
The answer to your question is "none" and I suspect all suggestions
are implementable, but it is not about what someone on sci.crypt can
do.
If it was "my" problem, before posting on sci.crypt, I would have
defiantly tried some FSK/PSK with simple EDEC as a first stab. If
that didn't work adequately I probably would have gone with something
like JH first posted -
JH> "Let me suggest the use of a device known as a "modem" to
transmit your encrypted data."
That fits into my skills, but that is not the issue.
The more relevant question is, "what operations may/may not be
implementable by the OP". The OP gives quite a few clues by the way
phrases his question and by the information he provides as to what
skills he does and does not have.
>Ten dollar sound cards do digital-to-analog and
> analog-to-digital conversion. Free, open-source audio codecs do
> compression and decompression. The crypto -- no problem. Are
> there cheap easy-to-use libraries for error correction coding
> and modulation onto an analog carrier? I'd guess so, but how
> should I know; I'm a cryptographer.
>
>
> > Yes there are "better" ways to solve this problem
>
> So don't eliminate the best solution out of ignorance. If the
> cost of a solution is too high, argue that. There's no such
> thing as a quality overrun.
Maxims for the practical solution to problems -
Ockham's razor. The simplest solution is often the correct one -
(this of course concedes that the simple solution must still be a
sufficient answer)
Kiss - Keep it simple. A long used design principle.
Schneier's mantra of simplicity. One of the Magnificent Seven
There are millions of problems that remain unresolved for want of an
efficiently implementable solution, but there are no problems that
remains unresolved because the solution is too efficient.
I suggested the simplest scheme that seemed to meet the requirements
of the OP. I didn't cut / reject anything for my sake.
The OP has already tried to solve this problem. It is reasonable to
assume he used his skill set at close to the best of his ability.
Adding more complexity and increasing the skill level may be
necessary to reach a solution, but unnecessary complications just
make it harder for the OP to implement. As I have maintained, a
solution that the OP can't implement is no solution at all.
>
> This is a "sci" group. It's about advancing knowledge, not
> dumbing things down to someone's comfort level.
I didn't dumb anything down for my comfort. I kept my suggestion as
simple as possible so the OP may have a solution he could implement
while still meeting the things he specified.
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA/AwUBQqUmqJS9Fk5okqe7EQLxMgCfSBq1lZVrHmsPXmnwocn78a+R0vEAoMd5
s5z+V22btAhmqnz9VaaRJjtE
=YjY8
-----END PGP SIGNATURE-----
David Eather
Hash: SHA1
No. But do you know the reasons why you might specify transformer
coupling?
>
>
>
>> A liquid fuel rocket is in all ways superior to a Chinese
>> firework. Better specific impulse, higher speeds, more payload,
>> orbital
>> insertion capability, higher reliability, more controllability
>> etc. And it remains superior, right up until the moment the
>> requirement is to make a five year old laugh and clap.
>
>> If you want to help the OP then solve the problem in front of you,
>> not the problem you want to solve.
>
> And you know what that problem is? Please tell us, since you
> apparently can read his mind.
I read his post. He is a proto-something from ECE (Ecole Centrale
d'Electronique)
His quote gives a general specification:
/OP
"The objective is to have an output sound that seems to be a random
sound.
And that we can decypher after recording. "real-time" is achevied by
a small
buffer (0.5 sec for example)."
OP/
He is not a cryptographer. he does not "talk the talk".
/OP
I tried classical cypher or modifications on the stream:
-XOR-ing it with a 1 or 4 bytes key. It is or understandable with
the 1 byte
keys, or impossible to decypher (because we may not have the same
values
due to numerical/analogical conversion, and that we need to know how
to sync
the key with the data)
OP/
The lack of details on any hardware suggests he is naive in that
area, so probably he is not an electronics student, but he has tried
a bit of signal processing i.e.
/OP
- -modifying harmonics after a fast fourier transform. Increasing the
level
of each harmonic (still understandable but strong noise), multiplying
all
harmonics by a factor (inverse fourier transform goes out of sound
range).
OP/
My guess, inferred from the information provided, is that he is a
proto-programmer. I would make the good will assumption that he is
working at near the best of his current abilities, and hence my
position that a solution should not be much more complex than what he
is doing now.
As to budget, you noticed he specified none. When this happens there
are just three possibilities. The budget is practically unlimited or
there is no budget at all. Which do you think is more likely?
The third possibility is that no thought was given to budget - common
to students, but note that none of the OP's efforts required spending
any money. A zero budget is reasonable.
-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1
iQA/AwUBQqVeA5S9Fk5okqe7EQI5vQCfZQJEBqWyENvCsxQeyg+HG8Dl2SsAnjxy
IO084jOLfDyQ808A2myPBYRf
=kc1U
-----END PGP SIGNATURE-----