Here is a link to one page of many in a series of lectures published by the Royal Swedish Academy of Music.
Although it relates to ‘real’ pianos, it is easy to infer the issues that relate to latency when using ‘plug-ins’ and indeed the design of the keybed in the instrument or controller.
Key bottom and hammerstring contact
https://www.speech.kth.se/music/5_lectures/askenflt/keybott.html
Wild. I’ve never consciously noted that (and frankly I don’t put my hands on acoustic pianos often these days) but it makes sense and I imagine we know that’s going on instinctively. I would bet that has never been implemented in any sampled piano; the programmers would automatically align all velocity samples to attack at the same time probably. I wonder how a piano library that preserved that delay would feel. It might be so subtle as to make no real difference.
I believe that is why we have such trouble with finding the perfect keybed or piano plugin.
It’s a subliminal thing - although only if you trained on real pianos I suspect.
There is nothing so satisfying as a good graded hammer action keybed and a plugin that accounts for velocity properly.
This is why I have considered buying the casio PX-5S, not so much for the sounds as the keyboard action.
Tri-sensor Scaled Hammer Action Keyboard
http://www.casiomusicdealer.co.uk/home/tech/
String contact duration and dynamic level
https://www.speech.kth.se/music/5_lectures/askenflt/stricont.html
Mind you, this is strictly for piano type instruments or sounds. When it comes to synth leads / pads or ‘drum kits’ etc I can wear the ‘semi weighted’ abominations they use for keybeds. In fact they might be better for that purpose.
Sounds like a build up for a Casio PX-5 commercial …lol…I can even hear the announcer’s voice.
Sorry guys…I couldn’t resist!
And organs. Playing organ on a weighted keyboard is, if anything, more hideous than playing piano on one that’s not. I occasionally have to use a Nord Stage for Hammond at church (as opposed to a real B3) and I want to stab myself in the head. In fact, as far as I’m concerned, EVERYTHING is better on a semi-weighted or unweighted keyboard- except piano! Which is why, through the years, I’ve come to simply use non-weighted action. And then of course I sit down at an actual piano and I can’t play the damn thing. (That’s not entirely true really… but it certainly suffers). I do wholeheartedly agree that playing a piano on a non-weighted action sucks balls. But since 90% of what I do is synth and organ it’s the trade off I make. And at least the controllers are cheaper lol
BTW playing drums on a weighted keyboard isn’t just bad- it’s literally impossible. For me at least. Too slow to react. It actually can’t be done.
Absolutely! Get deep in their pockets!
@FredProgGH
Absolutely. I don’t do much wurli but funky clav stuff or Bach on my moXF6 is the bomb. The hammer is only used for piano.
Interesting article. I play a real piano almost every day in my job as a choir teacher, and I’ve never noticed that. It explains why a lot of piano players that specialize in late romantic literature sound horrible when they play Bach or Mozart. I was very lucky my college a Beethoven era PIanoforte that I was lucky enough to be able to play a few times. Now that was an entirely different keyboard experience.
Being an inquisitive soul, I opened up the lid on my teachers old upright when I was 8 and experimented with slow and fast notes. Felt (pardon the pun) the hammers and strings and watched the mechanism. (as you do)
Once you get into the issues of the sound board, bridge, three string choir, hammer material etc not to mention standing waves and harmonics etc you realise just what a huge job a piano plug-in designer faces - especially a modelled sound.
… and why it’s soooo hard to find that Goldilocks piano!
Of course twiddling with a complex synth is another black art as well …
I played a fair amount of pipe organ as a kid and there you are well aware of latency differences, On one hand you had the trackers, which had a slow mechanical action but had to be located more or less under the pipes so you didn’t have a long delay from distance. Then you had faster pneumatic types but the consoles might be located on the opposite end of the church and have delays of a couple hundred milliseconds. And of course Hammonds are, in comparison, fairly instant but even they have small key contact latencies- but vst modellers are aware of that and do factor it into some of the organ clones available.
That’s nucking futz lol I’ve heard of that organ but never actually seen it before. Amazing.
I always wanted one of those! 
Terry
Dang, you gotta love the interwebs! I actually found a database with the organ I was taught on. Not quite 400+ ranks, but it seemed pretty awesome at the time 
http://database.organsociety.org/OrganDetails.php?OrganID=38416
That’s wild, Fred - I learned on an M. P. Möller Opus 9810 (1963), same builder as yours!
http://database.organsociety.org/OrganDetails.php?OrganID=38239
Small world!!! 
Terry
That’s crazy! They date from the same time too- he might have built the one at our church right after yours…
…Amen to that.
Here is a new (to me) wrinkle that goes beyond latency that might affect us slightly – it is MIDI Jitter!
This is more an issue for DAWs I imagine than hosts like Cantabile, I would think, but perhaps not.
Also, here is his post concerning his testing methodology:
Terry
Wow. At the end of the day though, I feel if this was happening I’d hear it. If the end result is that my playing sounds good to me than even if jitter (judder?) technically is a problem… it’s not a problem, ya know? If there are people out there having spontaneous “why do I suck!!!” issues this would be good to know about though! Interesting that Mainstage is an offender.