This work experimentally studies the noise features of flying drones under representative working conditions, including hover, spin, climb, vertical flight, and cruise. Being different from outdoor experiments where many uncertain factors, such as gust wind and ground reflections, could affect test results, measurements in this study were conducted in a large anechoic chamber. The noise was measured with a ground-based rectangular microphone array covering an area of 5.6 m x 6.4 m and a 4.6 m linear array on the sidewall to account for emission directivity. A binocular vision tracking system was used to trace drones' motion in real-time. Results show that a flying drone's noise feature is sensitive to its flight altitude, relative distance from observers, and motion status. Interestingly, under hovering conditions, the noise directivity is similar to that of an isolated propeller, and a low-noise region is found underneath the fuselage. While for cruising flight, the region with higher noise levels is observed behind the drone due to the fuselage's pitch angle. More importantly, the synchronized systems for noise and position measurements allow finding of an acoustic response to any change in motion status across flight tests. A coherence between real-time changes in the noise spectrogram and variations of the propeller rotating speed is found, thanks to the short-time Fourier analysis of both noise and motion-tracking data. The present finding extends the authors' previous work, enabling a deeper understanding of general drone noise features under different operating conditions.