Hello everybody! As you may know, I've been working on an update to Arachne's TARANTULA voicebank for the past 6 months or so, dubbed TARANTULA II. This update aims to add new expressions, pitches and POSSIBLY features to her voicebank. However, given the current size of her voicebank (51,800 oto lines split across 12 pitches), opening it hardlocks UTAU, rendering the program fully unresponsive. This is an issue since I primarily use UTAU as I find it easier and faster for me to use than OpenUTAU, and I know many UTAU users who feel the same. As such, I want to present a variety of options that will allow users of UTAU to use the voicebank, and then I will be running a poll on Twitter to determine what I should do. I would like to stress that I will distribute the voicebank in a way that allows those who use OpenUTAU, UTAU-Synth or THAT one Mine Laru plugin to use the bank to its fullest abilities, regardless of what happens, and I will also distribute it unaltered should a workaround be found
Option A is to cut down the planned whisper expression to one pitch in its default state (though allowing the user to place the otos of the pitches in their folders if they so choose). This would bring it well below the oto limit, cutting about 8,620 lines from the VB. This would also allow me plenty of space to add a feature I've been considering adding to TARANTULA II, which is full support for syllabic m, n, and ng. This can be emulated in the current voicebank pretty well however and would affect the range of this expression pretty severely since F4 and A3 pitches do help a LOT with making Arachne's English sound as good as it does.
Option B is to render the vocal fries, glottal stops, syllabic Ls and breaths to, by default, only be rendered with their D4 recordings, cutting the oto lines for them from the A3 and F4 pitches. This would cut the oto size by 2,384 lines, and probably have a minimal effect on most bank use, since the features WILL be usable, it'll just only use the D4 samples. All 12 pitches recorded would be usable by default. The downside of course, is that these extras do lose a bit of their range.
Option C is to delay the release of the voicebank until I can figure out how to make that damn Mine Laru plugin that bypasses the oto limit work with Arachne's English bank and include all the resources needed with the VB. This would allow the voicebank to be used to its full extent within UTAU, but would be a massive fucking pain to learn and set up, but I do think it's fully possible. This would NOT render the voicebank unusable in UTAU-Synth, since the oto limit is actually an issue exclusive to the Windows releases of UTAU. The downside is that this would render the voicebank entirely reliant on a plugin, a notoriously finicky one at that, and require Mac users and OpenUTAU users to use a separate download link. The massive upside is that this would allow me to update Arachne's English voicebank infinitely (in theory) and add as much stuff as I wanted to, including the aforementioned syllabic ms, ns, and ngs, while still having it work within UTAU. Biggest pros, but also some pretty big cons.