Has anything much changed in R-to-R design recently? If you look at the specs for a modern machine you can see that it's still far from a transparent medium. SNR say 70dB, distortion 0.3%, frequency response +/-2dB.
It will be difficult, if not impossible, to attain 'the best' from any medium until loudspeaker imperfections and loudspeaker/room interaction are resolved