Maybe theoretically but for our purposes doesn't the Nyquist–Shannon sampling theorem mean it's essentially lossless to our ears?