During the pandemic, lectures originally held in person were often switched to online video format. Although most of them have reverted to the traditional classroom format, the former created videos continue to serve as supplementary material. While lecture slides and notes are frequently updated, videos often remain unchanged due to the increased resources required. Videos and lecture notes inevitably become out of sync after a few years and may contain varying errors. The here presented workflow facilitates automatic creation of slides, lecture notes, and videos from a single LaTeX file, enabling simultaneous revision without additional effort. This maintains synchronicity across all provided elements.

Videos of lectures often consist of the presentation slides, with an explanation spoken in the background exactly as it would take place in the lecture hall. Normally, such videos are simply recorded with a screen recorder or a conventional camera in the lecture hall. A video edit is limited to inserting opening and closing credits and making corrections if necessary. Even the smallest corrections after completion of the video require a new run through of the editing process, which is very time consuming. To create lecture videos automatically, the spoken word can be transcribed by an artificial intelligence and saved as additional text which can be easily revised and edited. This is then converted back into an audio track via a AI text-to-speech interface and merged with the lecture slides. The length of the generated audio track determines the display duration of each individual slide.

Lecture slides and notes can already be generated in LaTeX from a common file. The very mature package Beamer is used for this purpose. With this it is very easy to specify that certain sections are printed in addition to the lecture slides in the written notes. To get the same for the spoken word of the video, the BeamerVideo package was developed by the authors of this paper. Among other things, this provides the additional LaTeX-command "Speech". The text inserted with this command does not appear on the lecture slides or in the notes, but is used to generate the audio file. Of course, this can also be the same text as in the written notes. The workflow details are presented in the final paper.

The here presented workflow simplifies the creation of tutorial videos if the slides are already created with LaTeX. An automatic speech synthesis creates a high quality video from a single source document without any manual intervention or editing. The created video is very easy to keep in sync with the lecture slides and the notes, as it is created from the same sources and revised simultaneously to with the other media. In addition, subtitles can be created fully automatically to ensure that the videos are accessible for disabled students.

Keywords: LaTeX, Video, Tutorial, AI.