In the soundcard challenge, I couldn't tell the difference between the reference file, and the same loopbacked five times in the audiophile 2496 card !Yes, the mysterious challenger who took the test was me
David:I don't remember the exact figure, but IIRC my Audiophile max. I/O level was similar to yours, it was quite below 2 Vrms. I'm not really sure this 2 Vrms is a true "standard", but IIRC my father's old Denon cd-player has a similar output voltage.
1.)The forum suggests the use of a passive attenuator(volume control) with the M-Audio 2496, which may not be such a good idea. The version I had used the Inverting Input of the IC...noise would be a function of the source impedance, hence the floor noise would vary with the position of the potentiometer. (This would also explain why input overload measurements on this card vary with the test setup.)The better approach would be to place the balance control ahead of a low-noise preamp/buffer eg AD797,SSM2134
2.)Ripping CD music via the Analog outputs is adding more links in the chain then needed, why not use the Digital out or Optical out from the CD player, the 2496 supports this....No more clipping!
3.) In all cases one has to Zero the DC offset on every file.
I doubt using an adequate simple passive voltage divider in order to avoid clipping would degrade notabily input SNR, but I haven't tried. Have you verified this personally via measurements? The use of an active device as input stage makes things much more complex, and I doubt would give any practical benefit.
We are not talking about ripping here. I believe we are talking about recording and analyzing of the output of a device
Most cards don't have DC offset problems at the input, since they are AC-coupled (this means that they block the DC at their inputs).