Video Editing Without the Editor

Breaking down computational filmmaking at TRASH

Our Chief Scientist and Co-Founder, Dr. Geneviève Patterson is featured on the Creative Next podcast with Dirk Knemeyer and Jonathan Follett to discuss her work on computational filmmaking at TRASH. Check it out!

▶️ Spotify
▶️ Apple Podcasts

“The big goal of our app is both to make it so that users can have an Instagram or Photoshop type feeling about creating beautiful videos, but not having to really understand how editing works.”

She begins by breaking down the processes behind the training models for the AI which enable TRASH to identify what is happening in your videos and then provide you with an edit. The tl;dr is TRASH wants to automate the most labor-intensive aspects of video editing to leave the artistic direction up to the user. She says the most major aspect of this work is training the AI to understand how objects move in time 🤯

“When you create machine learning systems, those systems can only possibly understand or know the data that is fed to them at the beginning. And it is very easy to feed them the wrong data, to feed them images that don’t apply to the context in which the system will be used.”

In response to Dirk Knemeyer’s question about the implications of computational filmmaking for cases like Deepfakes, Dr. Patterson says that this is more of a concern for the professional film industry which will use Deepfake technology to cut production costs. She says, “I’m not as worried about making unrealistic things happen and more like, how are people going to show the realities of their lives and the reality of the things they are interested in, in a new way that I couldn’t anticipate right now”.

One of her final points is about the future of TRASH where she says, “I am excited and really hopeful that TRASH is going to have a positive economic impact on lots of creative professionals who aren’t filmmakers”.

Check out her episode here to learn more 🧠