One method for determining modifications between two versions of the same media asset may involve the detection of common subsequences shared by both media assets. A suitable longest audio common subsequence (LACS) algorithm or longest video common subsequence (LVCS) algorithm may be used to identify common subsequences. By mapping common subsequences shared by media assets, additions, deletions, substitutions, changes to order, etc. may be determined. Various features may be used when comparing the individual subsequences within two media assets. In some examples, for audio assets, spectrograms (e.g., log-mel spectrograms) may be used. In some examples, for video assets, reduced-resolution frames may be used. Comparisons may be computed using a suitable distance function (e.g., a standardized Euclidian distance or a mean squared error (MSE)). Once a pairwise comparison matrix that includes comparisons of the individual subsequences in two media assets has been computed, an optimal path may be read from the table that indicates the longest common subsequences shared by the media assets and/or indications of differences between the media assets.