Auxiliary data is transported in a conventional audio signal by hiding the data in the form of colored noise. The colored noise has a spectrum that simulates the spectrum of the primary audio signal. The data to be transported is first converted to a spread spectrum signal. The primary audio signal is analyzed to determine its spectral shape. The same spectral shape is imparted to the spread spectrum signal, which is then combined with the primary audio signal for transmission. The spectral shaping can be performed using time domain modeling and synthesis such as linear predictive coding or by using subband coding techniques such as fast Fourier transforms. A plurality of different auxiliary information streams can be transported on the audio signal. By adjusting the gain of individual spread spectrum signal carrier(s) and the power of the colored noise, the auxiliary information stream(s) can be rendered inaudible in the primary audio signal, or at any desired level below or above an audible threshold.