29-7=22 (unrealistic but so was my thesis)

[fouriernotfurry]

A lot of manipulations in circuits (I’ve seen) surround just preserving the signal shape. Amplifying the signal? Easy. Amplifying it to a certain level? Fixing distortions? Perfecting the shape? Yes, client, 47 more components

[fouriernotfurry]

Based on what I experienced in grad school so far, I headcanon that the Joker has PhD in electrical engineering and in signal processing

[fouriernotfurry]

Building bridges, blank paper recalling, targeted learning. The reason my thesis defence went so well, despite me being lazy as hell is because I (unknowingly) followed these steps. I was able to practice my speech without the slides not because I had them memorised, but because I understood the structure, how every step affected the result and lead to the conclusion.

[fouriernotfurry]

Any system for which response to sum of inputs is equals to sum of responses to every input is a linear system. But what does it mean? I think I’m starting to understand why math is essential, because, I have a feeling, the only way to understand the point of linearity is picking apart its math and, ugh, examples.

[fouriernotfurry]

I will understand you, asshole

[fouriernotfurry]

Unlock your full engineering power by learning how math translates into physics

[fouriernotfurry]

Expecting the initial signal to be pure is unrealistic. A real signal always comes with all kinds of garbage noise.

Imagine a graph, where ordinate is amplitude and abscess is time. All sound has known speed, 340 m/s or something, the signals start and arrive at the same points of time. The important part, which really characterises the signal, is frequency/period between cycles. Looking at the graph, figuring out which part of it is the noise is pretty much impossible, or at least too much hassle when Fourier transformation already exists.

So what Fourier transform does is translate amplitude-time graph of a signal into amplitude-frequency visualisation. There are filters that can be used to filter the noise, but even visually, we can better see which parts (frequencies) cause the impurities, and, consequently, clean the good signal.

When reverting it back to time-domain, we get a noise-free signal. This is magic.

Filtering sounds like an easy fit when you already have all sorts of filters modelled and ready to use, but DSP is about understanding those filters, why and how they were created. This will make you understand which filters with what properties to seek in different cases, instead of slapping Butterworth on every system.