# **Audiovisual Stimuli Generation Pipeline** This OSF page contains resources and scripts for an automated pipeline that can rapidly generate hundreds of animated audiovisual stimuli. For information on how the files on this OSF page are organized, see the [File Organization page][1]. The pipeline is described and evaluated in [this preprint][2]; please cite the preprint if you use the pipeline. This OSF page also contains the supplementary materials for the preprint, as well as data and analysis scripts for some of the results reported in the paper. ---------- If you encounter any issues or need additional information about the pipeline, please email Halie Olson (holson@mit.edu). ---------- The pipeline was created to generate language stimuli for a pediatric event-related fMRI study on language processing, but it can generate audiovisual stimuli for a variety of applications. This page contains documentation for three separate processes: - **audio generation:** this portion of the pipeline generates audio files from text stimuli using a custom script that interfaces with Google Cloud Text-to-Speech software. - **video generation:** this portion of the pipeline generates animated videos from audio files using Windows Power Automate and Adobe Character Animator. - **wordlist generation:** while separate from the main stimuli generation pipeline, we are also providing scripts and documentation for how to generate random word-lists that contain words from from a list of text stimuli and minimize accidental coherence, as well as how to format them for optimal audio generation. ---------- ##### **Credit statement:** Audio pipeline scripts: Bianca Santi, Greta Tuckute, Aalok Sathe. Video pipeline scripts: Matthew Soza, Halie Olson, Bianca Santi. Documentation: Bianca Santi, Matthew Soza. Testing and stimuli generation: Bianca Santi, Halie Olson, Nika Ananieva, Sophia Seitz-Shewmon. PI: Evelina Fedorenko. [1]: https://osf.io/x49gu/wiki/File%20Organization/ [2]: https://doi.org/10.31234/osf.io/8gcn7_v1