My definition of accurate, in this case, is to reproduce the waveform that's represented on the CD (or file) at the transducer (speaker cone) as closely as possible. If I feed my 861v6 a pure / clean (call it what you will) waveform - that's what I want my system to attempt to produce. Even better if that waveform's what finally hits my ears after room-effects & potential induced corrections (e.g. MRC) - but then of course there's the (sometimes coloured) filter that sits between my ears !
What happens before that - before the file / waveform reaches my processor - is of course a whole other can of worms....

Trinnov Altitude CI-12; Meridian 861v8 + ID41 + UHD722v2 + 271, 2*SE-update DSP8.2Ks, SE-update DSP7200HC, 4*DSP3200, 4*DSP33, 218, 251
JVC DLA-NZ8; Panasonic TX-P60ZT60E; Lumagen Radiance Pro 5348
Roon [13i7 Win.11 NUC]; SKY Q; Oppo UDP-203EU; AppleTV / Fire Cube / Roku / NVidia