https://liatas.com/tags/signal-processing/Recent content in Signal Processing on Chris LiatasHugoen-usTue, 09 Jun 2026 15:30:00 +0300https://liatas.com/posts/savitzky-golay-smoothing/Tue, 09 Jun 2026 15:30:00 +0300https://liatas.com/posts/savitzky-golay-smoothing/<div class="headerclaim">A walk through a smoothing technique and where it earns its keep. The market-flavoured examples later are illustrative <strong>quantitative analysis on synthetic data</strong> — for education, not financial advice.</div> <p>When dealing with noisy data, many of us reach for the same tool first: a <strong>moving average</strong>. It is one line of code, it is intuitive, and it does remove noise. The catch is that it tends to do a bit more than that. A moving average is a fairly blunt instrument: alongside the jitter, it also tends to soften <em>structure</em> — flattening peaks, filling in troughs, and smearing the shape of the features we may actually care about. If all we want is a calmer-looking line, for example to visualize trends over a long period, that is usually fine. If we care about <em>how tall</em> a peak is, <em>how wide</em> a response is, or <em>how fast</em> something is changing, a plain average can quietly discard the part we wanted.</p>