Method for independent dynamic range control

Different dynamic range control values are applied to the 2-channel and m-channel outputs without repeating the inverse transform or the windowing of the audio samples. First, m-channel dynamic range control values are applied to audio samples in the frequency domain ("frequency samples" or "frequency coefficients"). The frequency samples are then inverse transformed to generate audio samples in the time domain ("time samples") and windowed to generate windowed time samples. The windowed time samples are saved and the 2-channel dynamic range control values are applied to the windowed time samples. 2-channel dynamic range control values include 2-channel scale factors that, when multiplied with groups of the windowed time samples, at least partially remove the effects of windowing and the m-ch dynamic range control values applied in the frequency domain and readjust the dynamic range for 2-channel output. Thus, a set of windowed time samples under m-channel dynamic range control values and a set of windowed time samples under 2-channel dynamic range control values are generated without repeating the inverse transform or the windowing, which are computational and/or memory intensive.