Ways to identify a frequency

[tonyjeffs]

I know of two ways to identify a frequency:
1. Resonance... eg: Set up an array of tuned resonators, bottles, or piano strings, and see which one resonates to the frequency under investigation.
2. Counter & Timer... Count the peaks per second, mechanically or electronically

Are there any other ways?

Thanks

Tony

 

[GregLocock]

Fourier transform

Cheers

Greg Locock

 

[dubi123]

If there is just one frequency component you may mesure time between zero crossing or do autocorrelation to extract period and then calculate f=1/T.

Regards

Dubi

 

[m777182]

In some older mechanical instruments the frequency of rotation (rpm) was magneticaly coupled with the instrument pointer that due to its mass could not follow the magnitude and the angle in phase so it pointed to some angle that was proportional to frequency.I had it on my bike and so had many cars before the microcomputers era. Very similar principle was used in old multi purpose instruments capable of measuring voltage, current, capacity and frequency too.
m777182

 

[tonyjeffs]

Thanks.
That has helped my thinking....
.....................
I have another thought... A very long pipe with a speaker at one end, and pressure sensors along its length.
For a given wavelength, pressure sensor should detect a minimum at 1/4 wavelength.
-perhaps?
...................
The reason for my question is I'm studying the inner ear.
Von Bekesey's theory, simply, is thus:
the (dead) cochlea is a 32mm long tapered tube divided along almost its entire lenth by the Basilar Membrane.
Sound enters and travels along the top fluid filled chamber. The compression travels a distance dependant on frequency, takes a short-cut accross the Basilar membrane, and returns along the lower fluid-filled chamber.
The distance travelled is short for high frequencies and long for low frequencies.

..I cant see any physical reason for this.

An earlier, similar theory by Helmholtz proposed that cross sections of the membrane under different tensions like resonating guitar strings. This turned out to be wrong, but at least he had some physics that I recognise there.

If it worked like my proposed tube, I calclate the cochlea would have to be 1/4 wavelength long, sveral metres for the lowest notes.

...The point of max distortion of the Basilar membrane is independant of how loud the sound is.

I cant find any physics, resonance or otherwise that explains this simple mechanical frequency analyser.

(In the live cochlea, the accuracy is thought to be refined by the electromechanical action of the muscle-like outer hair cells)

I'm aware of complex mathematical models by people like Lighthill, but I don't think a model explains or justifies the physics. (I could write a mathematical formula for the height of a flying pig!)

I'm starting to doubt the truth of von Bekesey's theory for which he received a nobel prize!


Cheers

Tony

 

[magicme]

it sounds as if you have completely ruled out a computer model ......?

however, even if your ultimate goal is to generate closed form equations for the physics, a finite element model could be a great supplemental analysis.

daveleo

 

[GregLocock]

"I have another thought... A very long pipe with a speaker at one end, and pressure sensors along its length.
For a given wavelength, pressure sensor should detect a minimum at 1/4 wavelength.
-perhaps?"

Not usually. Usually the wave will travel down the tube. It is only stationary at resonance (that's why we call them standing waves).

Given that error I don't think you need to worry about Lighthill and the Nobel prize quite yet, why not find a good book on Acoustics and partial differential equations and start there?



Cheers

Greg Locock

 

[electricpete]

Hmmm. Interesting suggestion that the ear consitutes a frequency detector.

True that the brain can perceive pitch. But it perceives a lot more about sound as well.

The eye can also perceive color which constitutes different frequencies of light. Is there supposed to be a frequency detector in the eye as well?


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[GregLocock]

Vaguely remembering back to psychoacoustics the model that seems to work best is that the inner ear can be modelled as a series of bandpass resonant systems in parallel, the RMS level of each is reported to the brain, but the level detector works at quite a high sampling frequency. This means the brain can get more out of each detector than just a level.

Whatever tricks are used by the ear/brain combination, it is demonstrably possible to hear phase relationships in signals, and this probably also explains why we can detect frequencies so accurately and quickly, even though we don't have very many detectors. I'm specifically thinking about tuning a guitar here.



Cheers

Greg Locock

 

[IRstuff]

There is a big difference between the eye performance and the ear performance. While there is yet to be found someone who can accurately determine the wavelength of a monochromatic source, there are plenty of people capable of perfect pitch hearing.

The eye consists of only 4 frequency selective elements, while the cochlea contains effectively thousands.

TTFN

 

[electricpete]

IRStuff

I can't tell you G, B-flat, Middle-C etc unless I have a known reference pitch to compare it to.

But I can easily tell green, red, blue and many shades betwen by looking at them without any reference. Sounds like the visual equivalent of perfect pitch to me.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[tonyjeffs]

Interesting comparison with the eye.

The eye has detectors for three frequencies, but can't tell us anything about the actual frequencies, or their ratios.

The ear has detectors for >1000 frequencies, and although it can't tell us the exact frequency, it does report the ratio between any two frequencies.

eg 1:2 100hz & 200hz -it clearly sounds like an octave.
1:1.2 100hz & 120hz minor (moody) harmony
1:1.25 100hz & 125 hz major (happy) harmony
1:1.5 100hz & 150 hz dominant (power chord) harmony

This isn't just a learned musical skill like perfect pitch might be; It is something that is unambiguous & inntrinsic to the design, & can be understood and recognised in 1/2 hour to a person who wasn't previously aware of it.

The ear can tell us the frequency ratio between frequencies
The eye can't tell us that.

.................
Another unconnected thing that's odd about musical ratios:
All common harmonies can be played as harmonics on a guitar string, except for the minor third.
I've no idea why.


If the hair cells were themselves tuned and simply passed their signal individually to an auditory neuron, there'd be no reason why a given 2 notes were melodious, and another two notes were discordant. They'd just be like bars on a bar chart.

The ears much more of a puzzle than the eye!
cheers



Tony

 

[EnglishMuffin]

The first frequency analyzer that I used was not an FFT type. It was called a "real time analyzer" and was built in the sixties. It must have used some sort of analog filter approach. It was better than an FFT analyzer in some ways because it gave a continuous real time output - no computer processing delay at all.
Regarding perfect pitch, it has been found that a majority of people, when asked to recall a favorite popular song recording in their minds, are usually able to recall the pitch within 50 cents. They just can't name the pitch. That suggests that perfect pitch could be taught if you could catch people young enough and train them. I have always been a bit skeptical of stories about people like Mozart who could reportedly instantly name any pitch they heard. Which tuning standard was he using ? In his day, orchestras in different countries, and even different towns, used their own pitch standards for any given note, There was no standardization such as we have today with "international concert pitch".

 

[GregLocock]

Ah, your real time analyser had a settling time for each filter very closely related to 1/(bandwidth)

Very, very, closely related!

Heisenberg cannot be tricked. On the other hand there is a good argument that our concept of frequency in non stationary signals is very flaky.

As to whether the human eye can discriminate colours, I'm imagining the equivalent of an ABX test. ie shine a monotonic laser beam into the eye and ask whether it is the same as laser A or laser B.

If your eye detects colour by looking at the ratio of the responses from the three detectors, then I think we /could/ learn to identify light frequencies objectively.

The big problem I have with this argument is that each of my eyes has a different shading (ie one sees a given scene as more blue or yellow) than the other, if I close one for any significant time (say five minutes).

Perhaps visual colour is more akin to loudness than tone in sounds?




Cheers

Greg Locock

 

[tonyjeffs]

"Perhaps visual colour is more akin to loudness than tone in sounds?"
I agree.

If I played "do ray me" to anyone, they'd easily put the tones in frequency order.
If I presented them with red, blue, green, most people wouldn't have a clue.

But in both cases they could say which was the loudest or brightest.

Perfect pitch.
Think of your favourite song. Hum it to yourself. play the track & see if it's the same

...Just tried it. Totally different key :-(

Tony

 

[EnglishMuffin]

tony: I didn't say that everyone can do it - I thought the university research paper that I read on the internet (some time ago - can't remember the link) said it applies to about 50% of people, so maybe "majority" was an exaggeration, and maybe not all of that 50% could get within half a tone - my memory may be faulty on that too. However, I have found to my amazement since reading about this that I can recall middle C within about half a tone. There are some scams out there now, perhaps based on this research, which claim to be able to teach perfect pitch to anyone. But this was genuine research.

 

[EnglishMuffin]

Tony: OK - Here's the link:

http://cogprints.ecs.soton.ac.uk/archive/00000643/00/pitch.HTM

 

[tonyjeffs]

Hi EnglishMuffin
Sorry I didn't mean to imply you'd said that.
I just kinda assumed that I had perfect pitch, and was a bit disappointed to be a tone out.
I do wonder if pp is more common than we imagine.

Cheers

Tony