Long-duration waveform similarity analysis identifies common waveform patterns among a group of signal recordings and uses these common waveform patterns to quantify the similarities of waveform signals. This quantification process of waveform similarity is an important analysis device for navigating, organizing, and interpreting long-duration waveform recordings. This paper proposes a computational framework for quantifying the timeline similarity relationships employing the sequential pattern analysis of long-duration recording signals at multiple analysis scales. Our proposed computational framework extends the scope of the conventional waveform similarity analysis toward more sophisticated pattern-matching approaches and enables in-depth explorations of larger datasets while allowing more quantitative interpretations and user interactions. The proposed similarity measurement framework is applied to two long-duration recording datasets to demonstrate its usage scenarios.