For my IACD final project I have written some code that uses subtitles to extract dialogue from movies and strings clips of dialogue together by finding rhyming phrases with the same number of syllables.

This Repo contains a single python file main.py which contains a couple variables you should tweak before running

• MOVIE_PATH_BASE_DIR should be set to the directory where all your videos reside
• SUB_PATH_BASE_DIR should be set to the directory where all your subtitles reside
• OUTPUT_FINAL should be set to the base filename (no extension) for your output video(s)
• NUM_SEGMENTS_MAX should be reduced if it is taking too long to process or you are getting a out of memory exception
• SELECTED_NUM_SYLLABLES declares which number of syllables you care about in the output video

Your video files should be in mp4 format and your subtitles should be in SRT format and the filename bases should be the same: ex. movies/Star_Wars_A_New_Hope.mp4, subs/Star_Wars_A_New_Hope.srt.

main.py depends on multiple python libraries that can be easily installed using pip install <dependency>

• moviepy for concatenating clips together
• numpy
• pysrt for reading srt files
• nltk as well as the cmudict corpus

• Rhyming and syllable detection using pronuciation rules from cmudict shout outs to CMU
• Standardizes movie height
• Standardizes loudness of sound clips

• https://vimeo.com/164774426
• https://vimeo.com/165395316