[E2EEd]

While I never investigated it personally, my understanding is that the quantization on the MPC60/3000 series was a big part of the feel. By definition, a quantized sequence is not a sample-accurate reproduction of a live MIDI perfomance.

Also, such idiosyncratic quantization is faithfully simulated in software nowadays, anyway.

Contrarily:

I challenge the notion that it is impossible to model the A/D, any DSP, and the D/A of the MPC3000 in software. While modeling non-linear response dependent on factors such as gains and impedances is not trivial, it is certainly feasible. Much of the pro audio world has moved on to plugins for their ability to reproduce the desirable aspects of analogue hardware while removing the undesirable aspects (such as noise).

A hardware sequencer/drum machine offers things that can't be modeled: tactile feel, low(er than some computers, still to this day) midi latency. Even the limitations such as a slow UI synchronously coupled to slow offline processes (which could be conceivably be modeled) affect the creative process in non-intuitive ways.

Similarly: reel to reel creates a smell that affects the vibe. Limited tape and no undo button affect the creative process as well.

DAW's are great but are essentially unlimited. This opens up horizons, but it does remove musicianship as a requirement in a recording context.