This article discusses a method for rapid synthesis of digital video of arbitrary length. The algorithm presented in this study allows the user to directly specify the desired video parameters, making the synthesis process more transparent for the end user. The synthesized video adheres to a multi-component model with statistical characteristics defined at the level of autocorrelation functions and brightness distributions. Within the framework of this method, it is important only to preserve characteristics relative to the source video, rather than its semantic content. One key requirement for this method is a small number of parameters to enable quick estimation based on real videos. The paper proposes a five-component model that describes an average frame (including both textured and piecewise-constant elements) of the video. A procedure is also described for estimating scene change intensity—a parameter that captures the dynamic properties of the video. The article provides descriptions and experimental results to evaluate the spatial and temporal characteristics of the source videos. The model parameters can vary widely, as needed for the experimental study of specific video processing algorithms. In the final section of the article, examples are presented of individual components and the combined images that will compose the final video.