Hey, Claude from 112dB here - somebody told me there was a little buzz about Morgana here, so I thought I'd jump in.
I'll try to clarify a few things if I may.
Indeed, we feel that Morgana is very warm and punchy. I can't vouch for every vintage digital device, but at least for the Mirage (which you may have inferred our synth is *ahem* inspired by), it's A/D stage definitely had some deficiencies, but bass response wasn't one of them! That, combined with the preamp stage (which was originally in place for directly plugging in a mic into the Mirage - I found that it imparts a really nice distortion onto samples so we simulated it) and the VCFs makes for a sound that is definitely not thin. Of course, it's all subjective, and far be it from me to discourage a healthy dose of skepticism.
As far as the bit reduction goes, that's a very small part of what Morgana does. The difference is that Morgana performs the bitreduction happens on the sample itself, which is then shaped by the VCF and the envelopes, etc. whereas a bitreduction effect is usually applied on the output as a channel insert. We're also doing a lot more subtle things, things like low-resolution envelopes, emulating the crosstalk that occurs at the input stage, and a lot of other little touches. Also at the sample level, we use an add/drop interpolation scheme for the transposition, which adds a lot of artifacts similar to aliasing (but also very distinct from simple sample-rate reduction - which also occurs at the sample level), really coloring the sound in a way that we think is quite pleasing. Suffice to say, there is absolutely no way to get anything like the sound from Morgana just by using insert effects.
As far as each copy of Morgana sounding slightly different - it's just a little gimmick, really. Basically, in the real world due to component tolerances and differences in the rate of aging, different units of the same make and model can sound somewhat different - even the different voices within one sampler can sound different if there is analog processing going on. In a sense this feature is modeling component drift by randomizing certain parameters slightly upon installation. That way your copy of Morgana will sound just slightly different from your friend's (for example). It's just something we noticed when doing research and collecting data, and we realized we could model this behavior pretty easily, so we did.
Well, I hope that clears a few things up - please feel free to ask if you guys have any other questions, I'll be watching this thread.
Cheers,
Claude@112dB Software