Signal Convolution and Correlation Workbench

How to use Signal Convolution and Correlation Workbench

Convolution sums x[i]y[k−i]. Real cross-correlation sums x[i]y[i−lag], with full lags from −(M−1) to N−1 before origin shifts. Full, first-input-sized centered and valid crops follow one-dimensional SciPy conventions, including the even-kernel left-center rule.

1. Enter first sequence x, second sequence y, index of first x sample, index of first y sample.
2. Choose the stated convention and check the units and supported model.
3. Calculate, inspect the result, then copy or download it.

Good to know

• Real finite sequences only,1–1024 samples each and at most1000000 products. The direct algorithm uses compensated per-output summation; results remain floating-point approximations. No circular convolution, complex conjugation, normalization, sampling-rate inference or waveform upload. Origins are integers from−1e9 to1e9; they label samples without changing values.
• Runs locally without input persistence. Text/JSON fields have a30000-character cap unless a stricter page limit applies; output has a1000000-character cap. Exact-grid workflows preserve six-decimal arithmetic, while approximate workflows use finite floating-point values. JSON downloads retain complete stored numeric values and exact fractions where provided. Unsupported formats and detected numeric or work-budget excess return an error or a clearly unresolved classification.