Determination of the granule density as a critical quality attribute during roller compaction/dry granulation provides an important parameter for further successful processing of the compacted materials. In this study, a novel approach was developed to determine the granule density using the principle of multi-stage air classifying. A roller compactor was utilized to produce ribbons and granules, composed of varying amounts of different excipients. The dry granules were separated using an inhouse built multi-stage air classifier and analyzed via dynamic image analysis in terms of particle size distribution and shape, resulting in a material independent prediction model for the density of granules based on their separation behavior. Overall, the present study introduced a novel, continuous, non-destructive, and material independent method that provides the suitability to monitor the granule density with high accuracy, that can be used as potential process analytical technology for implementation into the roller compaction/dry granulation process. (C) 2020 Elsevier B.V. All rights reserved.