A way of indicating a risk for coronary artery disease is disclosed. A first plurality of first sound recordings of heartbeats and second plurality of second sound recording of the ambient background are obtained. A filtering of each first sound recording is performed by using a simultaneously recorded second sound recording. The filtering of each first sound recording comprises involves a determining a diastolic period of the heartbeat of the first sound recording and performing an adaptive filtering of the first sound recording based in the diastolic period of the first sound recording and the simultaneously recorded second sound recording. This is followed by a determining of an indication of the risk for coronary artery disease based on the filtered first sound recordings.
