All I have to say concerning my long absence from this thread is that right now I am listening to Mozartâs âLeck mich im Arschâ and that isnât even relevant to this.
Fresh new start, letâs look at
Akimine Luna
A CV female bank with 176 samples and the 21.10.11 noted as her first web presence.
This bank sports a clear sample sound with only a little background noise. Occasionally one might hear a light clicking noise at the end of the samples, though this has been fixed in the oto.ini. All samples have an average length of one second as well as a good overall volume. Spaces before and after the samples, I am totally a-okay with this.
The voicebank sports a light, feminine sound which goes into a more childish tone through its relatively high voice. The hardness when rendered is what Iâd consider normal, with only little tendencies to sounding soft â can all be evened out when mixing.
I opened the voicebank with relatively neutral expectations. What surprised me though, was that instead of an extra artwork folder or separate artwork itself, it had a detailed description of the design in the readme.txt. Not as satisfying as actual artwork, since you have to, well, imagine it and hope youâre right with your mental image, but still much much better than nothing. The readme also provides a suggestion for the flags that should be used â fresamp and bkh01 â and a list of resamplers that are known to work well with the voicebank. Additionally, there is a short description of the voicebank sound itself as well as recommended music styles and speeds, along with the terms and conditions of usage.
The samples are encoded in hiragana and provided with romaji aliases. Additions to the standard list are l- and v-combinations, these encoded in romaji and only the v-combos aliased in hiragana. All .frq-files are present and â this I like very much â the voicebank also comes with .pmk-files, the generated files the TIPS-Resampler uses. The oto.ini is well done â I miss my much praised negative overlap but this one works well without. It only might need a little more touch up when used on usts.
In the review request I was given a string of flags that are recommended for usage (Y0BRE0g-4F0) along with a note that the g-4 flag is a must. Of course, when Iâm asked to use special flags, Iâll try them out. After all, when the creator recommends those, they should be best for that, since, you know, theyâre the creator of the voicebank. In this case though I have to note that the necessary g-4 flag could also be left out â the sampled voice is already relatively high, so to me it seems a bit better to not further heighten it with the g-flag. I also find it more pleasant to my ears without the flag, since it causes it to generate more noise with it.
Bkh01 causes the bank to have a very hoars and rispy sound. Fresamps muffles the consonant and causes it to sound more muffled overall. In my opinion TIPS works best, as it produces a close to noiseless result and makes it sound a little softer.
Conclusion:
A very pleasant high sounding CV bank with comparably high quality samples and easy to use, if youâre willing to spend just a little more time on fitting the ust.