Hmmm, thinking as a consumer - I think Windows 100% had the right idea of packing in a bunch of voicebanks of professional and/or capable singers into a single pack, and charging for that. Conceptually, I think that's a great idea. Of course, as I'm sure you're all aware, the voicebanks themselves are... less than amazing, with often poor recording quality, weird oto.ini files, almost all but one voicebank being monopitch, etc. It really feels like the people involved with making these voicebanks weren't all that familiar with recording equiptment, or at least with UTAU development. That said, with a competent team involved who knows what they're doing, I think it might be cool to, like, make packs of professional-grade UTAU voicebanks in packs of like 2-6 different voices catered to a particular genre or song type ( Like you could have like a "Rock Pack" with two masculine and two feminine specialized rock singers, etc. ). Marketed kind of like a sample pack or vst.
If it's a single UTAU vb though - there's the obvious. Multipitch, high-quality samples, competent recordings; and for me personally, it'd have to be CVVC or VCV, I personally simply wouldn't pay money for a CV voicebank unless it was like, really, really unique and interesting. I also think it should probably provide something not often seen in UTAU voicebanks, such as, for example, being voiced by a professional singer, an uncommon voice type, catered to a particular genre, etc. I think if you're going to sell something that's typically free, you need a real strong "hook" to incentivize purchases, because it's not enough to simply make a good UTAU seeing as you can get plenty of those for free.
Their recent controversies aside, I think VOICE-MiTH's old UTAU voicebanks are good examples of what I would consider commercially viable ( minus ZETA. Sorry, I like her voice, but I would not pay for her Chinese DeepVocal. It's just way too limited, and there's no reason it should be monopitch with ZETA being a primarily Chinese vocal, while her Japanese vb has 3 pitches, and her predecessors all also had 3 or more ). They're high-quality, have distinctive voices, voiced by capable singers, competent oto.ini files, they're just generally good voicebanks, and it makes me sad that VOICE-MiTH felt silly one day and decided to stop being a vsynth company with integrity. Also, when ALYS was on Alter/Ego, I'd always said that I would've payed money for her to be on UTAU instead - and now that we have her recordings on UTAU, I totally stand by that, with her vocals I would have absolutely thrown a few dollars her way ( with a spruced up oto.ini at least, since her final vbs didn't proper oto.inis, and their previous UTAU versions were lower-quality recording-wise ).
