Tools for Open Science
Bershadskyy, D., Ghadwal, Sunil, & Greif, Jannik. (2024). MTVE: Magdeburg tool for video experiments. Journal of the Economic Science Association 2024, 1–11. https://doi.org/10.1007/S40881-024-00165-7
Bershadskyy, D., Ghadwal, Sunil, & Greif, Jannik. (2024). MTVE: Magdeburg tool for video experiments. Journal of the Economic Science Association 2024, 1–11. https://doi.org/10.1007/S40881-024-00165-7
I am a big supporter of Open Science. For a large Third-Party (DFG) Project we needed a certain communication software. Since such software was very expensive, we had to develop it on our own, and in doing so, we made it open source.
The Magdeburg Tool for Video Experiments (MTVE) is an open-source software designed to facilitate video-based communication in laboratory and online experiments. It ensures high-quality video and audio recordings that can be analyzed using machine learning techniques. MTVE is browser-based and integrates with experimental platforms like z-Tree and oTree, offering researchers control over parameters such as resolution and participant numbers.
Cocktail Party Problem Solution: MTVE records individual audio streams separately, allowing for clearer speaker distinction.
Transcription Support: It enables speech-to-text transcription with models like Whisper, merging individual transcriptions into a coherent dialogue structure.
Ease of Use: Researchers can set up sessions with a user-friendly interface, ensuring data security by storing recordings locally on lab servers.
MTVE simplifies the process of capturing and analyzing video data while adhering to data privacy standards. Although we have successfully used it, we know of its current limitations. These include restricted combined resolution and frame rate options. Yet, as it is an open-source project, we encourage community contributions. See our GitHub repository https://github.com/MaXLab-OVGU/MTVE.
In another project (supporting the open science principles), we currently applied for funding, we aim to develop a tool that facilitates researchers to generate replication packages. Stay tuned, if we receive the funding.