Like a prism for patterns
White light looks like one thing. A prism reveals its colors. Fourier does the same for signals: it separates a complicated shape into simple repeating waves.
No equation required
Everything complicated
can be heard as simple waves.
A Fourier transform is a different way of looking at the same thing. Instead of asking “what happens where?”, it asks “which patterns are hiding inside?”
See it happenONE COMPLEX SHAPETHREE SIMPLE WAVES
FOURIER→
01 · The big idea
Same information.
Different point of view.
Nothing gets lost
It is not a summary or approximation. With all the ingredients, you can perfectly rebuild the original—like switching between a recipe and the finished cake.
Each ingredient is everywhere
A frequency is not tied to one instant or one pixel. It is a pattern spread across the whole signal or image. Every ingredient votes on every part of the result.
02 · Build a signal
Three pure waves.
One new shape.
Turn each ingredient up or down. The top view is what you see; the bars below are what Fourier sees.
THE RESULT · TIME VIEW
FOURIER'S VIEW · INGREDIENTSTaller bar = more of that wave
03 · Let Fourier investigate
Draw any signal.
Reveal its ingredients.
YOUR SIGNALDrag across the canvas
HIDDEN INGREDIENTSSlow → fast
Smooth curves mostly need slow waves.
From 1D to 2D
A picture is just a signal
that extends up and sideways.
BROAD SHAPES
+
MEDIUM DETAIL
+
FINE EDGES
=
THE IMAGE
In two dimensions, Fourier asks which stripe patterns—at every size, strength, and direction—combine to make the picture.
04 · Fourier inside an MRI scanner
The scanner does not
take a photograph.
It listens to radio signals from excited hydrogen, while gradient coils label the signal with spatial patterns. The measurements land in a frequency map called k-space.
- 1Wake up a slice
A brief radio pulse tips hydrogen signals into a state where the receiver can hear them.
- 2Give places different rhythms
Gradient coils make position affect frequency and timing—the scanner's way of labeling “where.”
- 3Listen to one mixture
The receiver hears all locations together. That mixture becomes a line of k-space, not a row of image pixels.
The reconstruction desk
Watch an image emerge
from pattern measurements.
Move through the scan. Every new k-space line adds information across the whole image.
FOURIER
REBUILDS→
REBUILDS→
One cell under the microscope
Each k-space cell is
one pattern vote.
A cell does not point to one place in the body. It records how strongly one wave pattern appears across the entire slice.
K-SPACE MAPHover or tap any cell
faster pattern →
faster pattern →
ITS IMAGE-WIDE PATTERNbroad horizontal change
POSITION MEANSVery broad pattern
Near the center means a slow change spread across the image.
BRIGHTNESS MEANSStrong vote
The cell’s magnitude tells us how much of this pattern was heard.
PHASE MEANSPattern alignment
Phase slides the pattern into the right alignment so all the votes combine correctly.
ONE K-SPACE CELL≠ONE IMAGE PIXELBUTALL CELLS TOGETHER=EVERY IMAGE PIXEL
Why the mirrored green cell? In an ordinary real-valued image, opposite k-space cells form a matched pair. Together they describe the same stripe pattern with the alignment needed to make it real.
05 · What lives where?
Center gives the impression.
Edges give the precision.
=
Center only: You recognize the subject, but it is soft. The center carries broad shapes, brightness, and much of the contrast.
06 · A different angle
K-space as a
frequency landscape.
Drag to rotate. Height shows how strongly each 2D pattern appears in the image. Most energy gathers near the center because anatomy is mostly made of broad, smooth regions.
↔ DRAG TO ROTATE
FOURIER
Take this with you
Complex things can be understood as combinations of simple patterns.
Music, images, radio, medical scans—Fourier gives us a reversible bridge between what something looks like and the patterns that build it.
Explore again ↑