Adobe recently published a video with a sneak peek at an upcoming feature that will be released in a future version of its Premiere Pro software and will use Adobe’s Sensei machine learning technology to perform seamless automatic reframing of aspect ratios of clips and sequences.
The upcoming Auto Reframe feature will support different aspect ratios including square, vertical or 16:9 and if it will indeed work as shown in the video above it might be able to save editors significant amount of time reframing videos for social media.
Auto Reframe is said to analyzes the video it’s applied to, crops it according to the aspect ratio of your choosing and pan the video to keep track of the subject matter.
The user will be able to choose to automatically reframe a clip or the entire sequence and to do some based on faster or slower motion. Titling and motion graphics are also resized and reframed based on the specific aspect ratio the user chooses to apply.
One feature mentioned in the Adobe video has do do with the option to preserve some of the original adjustments made by the user before applying the Auto Reframe tool (this option will automatically nest some of the aspects of your original edit).
Adobe also mentioned that if the automated result isn’t where it should be, editing the composition can be as easy as changing a few of the keyframes that are automatically created with the tool.
Interestingly all of the examples in the Adobe video show people/faces and not products or animals. At this stage, it is not clear if the Adobe Sensei machine learning algorithm is only programmed to identify, track and recompose based on human faces or if there are more options/capabilities to the algorithm.
Adobe did not mention when the new feature will be available, however, there is at least some chance (based on past releases) that we might see it in the next Adobe Premiere Pro release sometime around October 2019.