There are several artificial intelligences that can generate images with DALL.E and Stable Diffusion being two of the most well known. Last week Meta showed off Make-A-Video which generates videos from user-supplied text. Not to be outdone today is Google which announced two artificial intelligences with similar capabilities.
Imagen Video and Phenaki developed simultaneously can generate short videos based on text input provided by the user. Imagen Video generates video using the Imagen system which previously only generated images. A total of 16 picture frames at a rate of 3FPS are generated by Imagen Video at a resolution of 24×48 pixels. It was then upgraded to 128 frames, 24PFS at 720P resolution using artificial intelligence.
According to Google, Imagen Video also has an advantage where it can understand instructions when asked to generate text in the video by the user. A feature that artificial intelligence has not yet been able to do is generate images and other videos.
Phenaki, on the other hand, can generate videos of up to two minutes based on longer text inputs. A short story video for example can be generated by Phenaki. The difference between these two artificial intelligences according to Google is that the video quality of Phenaki is lower than that of Imagen Video.
Because Imagen Video and Phenaki are developed by Google, the two teams will now work together to produce artificial intelligence that generates better and higher quality videos.