DSP Plugin Test & Verification — Honest THD/CPU Benchmarking

The problem

DSP unit test failures are often caused by flawed test harnesses rather than the audio processing code itself. Developers waste hours debugging implementation logic when the actual issue is an impossible threshold or a measurement helper that clobbers the internal state.

What it does

• Identifies harness bugs like measurement helpers that overwrite plugin configurations before testing.
• Implements differential testing for click-free profile swaps to distinguish harmonic content from genuine pops.
• Provides recipes for measuring THD and aliasing using Goertzel filters and fold-back frequency analysis.
• Verifies real-time safety by benchmarking actual CPU load against hardware block budgets.
• Corrects parameter driving logic to use raw choice indices instead of normalized values.

Frameworks & tools

Optimized for JUCE and C++20 audio plugins, with portable methodologies for any real-time DSP environment.

Why this beats prompting it yourself

General AI often suggests relaxing thresholds to make tests pass. This skill enforces physical reality, ensuring you report honest metrics like -90 dBFS aliasing only when physically achievable, preventing "fake green" builds.

Use cases

• Debug failing assertions in processBlock sequences.
• Verify oversampling suppression at 2x, 4x, and 8x rates.
• Benchmark CPU consumption across different plugin states.
• Audit parameter listeners for index mapping errors.

Known limitations

Not intended for GUI layout testing or general non-audio debugging. Requires an existing C++ test harness to apply fixes.