NovaSonica

From my own experience as a mastering engineer there is a huge difference between 44.1kHz and 48kHz sample rate. It noticeably increases the transparency of the signal - even when simply upsampling a 44.1kHz signal. In the same gist you'll notice that EQs are a lot more precise, especially if you are searching for feedback frequencies and other small frequency ranges. While it is definitely true that analog transfer reduces a signal's quality, you should keep in mind that destructive audio editing (e.g. normalization in WaveLab) in a digital environment will have the same effect. The reason is simple: when applying effects to your "numbers" a new result will come out and that needs to be fitted in your sampling rate and resolution, causing rounding errors. Another factor is your CPU: it cannot calculate an infinite amount of decimals, when getting close to zero or one rounding will occur. Unfortunately this rounding is not always correct and will sometimes yield the opposite result. In case of bits this means that rounding a number close to zero (i.e. off) will incorrectly result in one (i.e. on). To those who think this is negligible, take a look at fractal arts (e.g. http://upload.wikimedia.org/wikipedia/commons/2/21/Mandel_zoom_00_mandelbrot_set.jpg) - these are the result of rounding errors as described above. Greetz, Tox, NovaSonica.com