Speaker
Description
Measuring the quality of high-performance waveforms with sub-dB accuracy is a technically demanding task. Subtle missteps in alignment, scaling, or statistical assumptions can lead to misleading results—even when using familiar tools. This workshop offers a clear and practical path through those challenges, equipping participants with the insight and techniques needed to measure signal quality with confidence and precision.
Drawing from real-world applications in software-defined radio and high-performance data converters, we’ll explore FFT-based spectrum analysis, Welch PSD, Error Vector Magnitude (EVM), and Allan Deviation (ADEV) for analyzing both stationary and non-stationary signals.
The session begins with a guided technical presentation covering these essential waveform analysis methods and how to properly apply them to real world signals. Demonstrations will include techniques for accurately measuring the ENOB (Effective Number of Bits) of high-speed, high dynamic range A/D converters, along with SNR and EVM analysis of 4096-QAM signals. We’ll also show how EVM can be applied at multiple points in a signal processing chain to isolate and identify dominant noise and distortion sources—providing a practical framework for validating each stage of a system.
In the second half, participants will follow live Python demonstrations using Jupyter notebooks and custom-developed tools that automate alignment, normalization, and measurement workflows. These tools are designed for both transparency and teaching value, revealing each analysis step while helping users avoid common pitfalls in post-processing and simulation environments.
Attendees will leave with reusable Python code, complete installation instructions, and the hands-on knowledge to perform precise, repeatable waveform measurements across a wide range of SDR and DSP applications. Whether you’re debugging a signal chain, characterizing a converter, or verifying system performance, this workshop will help you replace ambiguity with rigor—and produce signal metrics you can trust.
| Talk Length | N/A |
|---|---|
| Link to Open Source Code | Open source code will be provided. |
