fast Fourier Transform by Zedzero on Vimeo.http://vimeo.com/zedzeroUnderstanding that the Fourier transform is based on the general concept of " periodicity" . Now this is a "mash up " concept. The underlying Spaciometry is the circle in the plane. Fourier was interested in describing how " heat" flowed in a conductor and he chose a metal ring! The temperature at a fixed point was increased and the pattern of temperatures in the ring was modelled at various time intervals.
The parameters are the quantities usually independent , that are measured by some Metron. Position was therefore measured by polar coordinates r,ø which can be transformed into orthogonal coordinates rcodø,rsinø which might be denoted as trig line segments x,y. Temperature was measured by mercury expansion. Thus this " temperature" is in fact a heat pressure, comparable to atmospheric pressure also measured in height of mercury!
So for each position a mercury height could be plotted. The fourth parameter is again an independent measure, usually the duration of the experiment measured by a metronome.
By recording these data sets in tabular form Fourier presented the scientific world with a technique of statistically managing sets of data. We have to review the work of Gauss and Boltzmann to find similar statistical management. But it was reputedly an idea of Laplace that Fourier developed extensively. The idea is based on the representation of a polynomial form by simpler polynomials that sum to it.
The purpose of representing a polynomial by a sum of polynomials was and is to be able to handle discrete information by a general polynomial that contains it. What this means is best illustrated by the function or polynomial y= x
2.
Typically we calculate a few discrete values, thrn plot them and join the plots with a smooth , but estimated curve. In one sense we have no way of knowing if that curve is right or rather " true" without calculating the intermediary points. This calculation is called interpolation. Thus it became clear that if one only used straight line equations one could approximate the curve by summing line segments from many lines,
The process clearly is based on difference algorithms and consequently in the limit melds with the calculus methods of Descartes , Leibniz and Newton . Thus the curve can be considered as the envelope "sum" of the tangents at each point.
The question that arose is what was the best fit for any number of discrete points ? Could you recover the " true" polynomial by this approach? In general the newer is no. But under certain constraints you could converge onto the true curve.
Thus it was shown that you could model any complex curved line by a sum of appropriate line segments, and eventually by a minimal choice of polynomials of a lower degree. Within the given range the convergence could be as close as desired, but outside the range the summed polynomials were seen Togo " crazy" . Extrapolation was definitely not possible!
A Fourier transform was based on Fouriers use of the trig functions of Euler. It was known that these functions could disappear if the amplitude was at the right phase in the rotational cycle. This meant and Fourier demonstrated that over the unit circle any complex form could be modelled by a summation of sine and cosines at different frequencies of rotation. The frequency at which the angle was considered to turn determines the phase of the amplitude .,these would thus cancel or add accordingly creating complex patterns of highs lows and blanks.this was called interference either constructive or destructive. Eventually it came to be called interference by superposition.
Now it was clear that in a ring if this dynamic rotation was occurring, then time would be trapped in the ring, and the angular rotation of these sine and coine waves would also be trapped in the ring by polar coordinates. Thus any amplitude would show up as interesting patterns in the ring.
Nobody knew or could think how heat was transmitted or conducted, but it seemed it flowed from hot to cold. Fourier wondered if it rotated or oscillated through the medium. If so the ring would show this by a travelling amplitude of oscillation which he could measure by temperature.
Actually his ideas for heat were misconceived, but his analytical method revealed how any complex pattern, conformed to a ring could be precisely modelled by sine and cosine summations.
His heat idea was misconceived because he eliminated the rotation concept, replacing it by the sine or cosine concept. Typically with the thought of his day and even today, the imaginary component of the exponential forms were always thrown away.
Heat flows from a higher temperature to a lower temperature Newton observed. Fourier never explained why, but he did show how a temperature form was represented ar different times in the process.
The reason heat flows is explained by Newtons laws of motion. The heat indeed increases the quantity of motioning a region. However, unlike a blow, the heat Los changes the bulk properties of the material. Thus the increase motion is not uni directional, the expansion is not an elastic deformation, except at the edges and the viscosity of the material leads to a longer relaxation time, a Thermo relaxation called thermal contraction. It takes time! Thus a pressure wave does travel in the ring, but with all these constraints varying it moves so slowly governed by other factors like latent heat and specific heat rather thn bulk material properties of compression.
That is not to say these transmissions do not happen. They do, but typically in a heat conduction experiment the parameters measured are not the acoustic ones!
The Fourier transform is sometimes confusingly presented as converting time into phase. This video shows precisely what is happening.a spatial forms modelled by several parameters only one of which is time.once the spatial form os modelled as an interpolation of some fixed positions and amplitudes at those positions. Then a time argument is devised to add to the spatial form so as to move it along.
Of course spatial form cn alo change with time. In this case the additional time argument is combined ith a time dependent form of the spatial disposition. The form however it is disposed in time cannot exceed its boundaries or the fixed point values by which it is defined. If these values do vary with time a massive compute is entailed and that is why the fast Fourier transform algorithm has been so important to the study of Dynmics.
The wave concept is a misnomer, in my opinion. The rotation or swivelling function is what was meant by undulation, and the progression of a deformation in a medium as a consequence of resident forces in the medium , ie inertial forces , is what deformation propagation is.
Newton in proposing light was a ballistic corpuscle was certain that it's collision with the retina produced the sensation of colour .he did an experiment to confirm this! However later he discovered colour could be described by angle of di
Refraction. This did not alter his opinion that sensation required collision with the retina. There were other details he chose yo overlook because he believed his model was the most straightforward explanation. Without a clear understanding of superposition of deformations in space, Huygens could only suggest that wave amplitudes joined to make a wave front, as circles do in a Moire pattern. By the time of Fresnel, Arago and others had a speed for light , Fourier had a method for describing complex forms that demonstrated superposition, and Young had shown interference patterns in ligh. Newtons ballistic concept had to go. But what replaced it was an arcane question bout form!
It is irrelevant whether a wave is shoed like a corpuscle or like a sea wave! What is and always has been the issue is the existence of a viscous medium for the propagation of light, whether as wave or corpuscle. That medium was called the luminiferous aether and it was thought to exist, until some disputed it. Today it is again believed to exist but it is called quantum energy etc.
Light is a strain that travels in this medium by Newyonial principles of motion and transmits energy throughout this viscous space. My best guess is that the form of this strain is a complex trochoidal rotation in the medium, modelled by Fourier transforms in 3d. The motion is generated by the restoration of equilibrium to this luminiferous aether or quantum energy .
http://www.youtube.com/v/r4Pc-rGBRJA&rel=1&fs=1&hd=1