pipAT If you don't want to make something like Garritan CFX, consider making piano tools like EZkeys, Session Keys Y&S, Addictive Keys, or those aim at song writing, chords detection, scales,.midis of different genres, metronome, ear training or something helpful for piano learning. Then, you don't have to sample pianos with dozens of layers or multi-mics.
The entire music making industry is pretty fascinating to me. There are products that require the user to be a competent pianist on one end of the spectrum, and products that can make an interesting beat or chord progression merely by turning a knob or pressing a colored button. I, and probably others on this forum, fall toward the "let pianists who can play, play".
Basically, I make the libraries according to what excites and motivates me to do the work. It's not money, and I can't think of a business case that would motivate me much in that direction. My ability to optimize my instruments is limited by my playing ability and attention span. I've tried to optimize as best I can, and for most people, I think it's more than adequate. For advanced players, its probably ok, but not the cream of the crop. Right now the thing that is preventing me from releasing a Bosendorfer Imperial is making a nice GUI. I'm OK at it, but I'd rather optimize layer transitions than make graphical control animations. I just can't get myself to cross the finish line by finishing the parts where I don't have the ability to meet my own expectations.
I also think that rather than spend another 90% time increasing the quality of my products 10%, I'd rather make more instruments so I have more diversity rather than higher quality. The higher quality part demands too much time.
For example, now that China has opened up, I have to spend more time with my company there, and I might ask Hailun to record one of their concert grands, which is about a 3 hour drive from my office. Wouldn't it be interesting to see how a Hailun concert grand prepared at their factory compares to the Fazioli/BI using the same recording technique? You could compare in your own home. At this point, I'd rather do that then code musical tools and such. Coding also, is a necessary evil for me. I don't love doing it - I love tuning the EQ's and such to get better playability out of the recordings I do. But tuning 19 layer transitions would become a chore. 6-7 is enough that I can feel real progress. More would feel like diminishing returns.
Thanks for the suggestions though. Even showing some interest in the subject helps motivate me to keep going as I've been lagging lately with work and travel.