It provides more accurate math. ALL digital adjustments that occur within Cubase require math… even fader changes. Imagine a 44khz project. You record one single track. You turn the fader down by 1db. Cubase now has to recalculate the value of every single sample - 44100 of them every second… just to make a 1db gain change. The least significant bits in the math are rounded off (for EVERY SINGLE SAMPLE)… There is a small amount of “error” for every single rounding. Now imagine that track is also being panned… more calculation… and it’s going through effects… more calculation… now you record more tracks, and sum some tracks together on a group track - more calculation… then it goes to the master channel - more calculation for summing there - or level adjustment… MORE CALCULATION. Now - think about how many insert effects may be running on a track… think about effect sends… there is a TON of math happening inside Cubase. In a huge project with large track counts and tons of fader gain changes, automation, insert effects, send effects, summing groups, etc., the amount of math (and rounding) happening in Cubase is HUGE. With 64-bit floating point processing, Cubase maintains more accuracy with EVERY SINGLE PROCESS in the chain… gain, summing, VST FX (the ones that support 64-bit), etc. It’s debatable how much of a difference it makes, but technically, it IS more accurate than 32-bit floating point math.