Extract:Dialogue 1.2 with Improved AI Model
We have updated Extract:Dialogue, to version 1.2. Extract:Dialogue is a plug-in that separates dialogue from common types of background noise such as wind, rustle, traffic, hum, clicks and pops.
Version 1.2 includes a new and improved neural network model. A large set of high quality audio recordings is crucial for the training of neural networks. We have teamed up with the sound effects company Soundly in our pursuit of high quality training data. Soundly is a complete sound effects platform with a user-friendly software package and a huge library of professional quality audio recordings.
Our partnership with Soundly allows us to gain access to a massive library of consistently labelled audio recordings that we can use as high-quality training material for our neural networks. The quality of the training set is reflected in the robust noise suppression that the latest version of our Extract:Dialogue plug-in offers. There were also improvements in the structure of the neural network that lead to better performance and less perceivable artifacts.
The upgrade is free for Extract:Dialogue and Acoustica Premium Edition 7.x owners. Extract:Dialogue 1.2 is included in Acoustica from version 7.4.7.