Spilling some tea on Kagamine ACT1 and the rest of Crypton's V2s (barring Kagamine ACT2 and as well as the append voicebanks) here.
All of Crypton's V2s suffer from random drops in volume, but Len's ACT1 vb gets it the worst. The best way to fix it (that I've found) is to adjust the dynamics where you hear the volume drop until it's less noticeable (adjusting the brightness helps, too). Luka is very quiet and will need to have both her brightness and dynamics adjusted before you really hear her well.
Rin's ACT1 is surprisingly smooth; about as smooth as a decently OTO'd and tuned CV UTAU for comparison. However, she has issues transitioning between "a" and "i" samples (ex: "hai" sounds like "ha i" rather than "hai") and can be quiet and muffled at times; increasing her brightness, dynamics, and clarity while also decreasing her gen parameter slightly helps. I haven't tested any fixes for vowel connection yet. She also sounds a bit stuffy nosed sometimes. She has a hard time holding long notes sometimes, but works quite well for songs with a lot of short notes. She also handles tempos outside her recommended range quite well (I had her singing at 229 BPM and she handled it very well; she sounded better than in any of my other tests.).
For an example of how she sounds, here's a test cover a Patchwork Staccato I did:
Len is choppy due to his vowels cutting off, mainly on his "a"s. I haven't tested a lot of fixes, but splitting the note and adding "-" to new notes seems to help.
More to come soon. :P
All of Crypton's V2s suffer from random drops in volume, but Len's ACT1 vb gets it the worst. The best way to fix it (that I've found) is to adjust the dynamics where you hear the volume drop until it's less noticeable (adjusting the brightness helps, too). Luka is very quiet and will need to have both her brightness and dynamics adjusted before you really hear her well.
Rin's ACT1 is surprisingly smooth; about as smooth as a decently OTO'd and tuned CV UTAU for comparison. However, she has issues transitioning between "a" and "i" samples (ex: "hai" sounds like "ha i" rather than "hai") and can be quiet and muffled at times; increasing her brightness, dynamics, and clarity while also decreasing her gen parameter slightly helps. I haven't tested any fixes for vowel connection yet. She also sounds a bit stuffy nosed sometimes. She has a hard time holding long notes sometimes, but works quite well for songs with a lot of short notes. She also handles tempos outside her recommended range quite well (I had her singing at 229 BPM and she handled it very well; she sounded better than in any of my other tests.).
For an example of how she sounds, here's a test cover a Patchwork Staccato I did:
Len is choppy due to his vowels cutting off, mainly on his "a"s. I haven't tested a lot of fixes, but splitting the note and adding "-" to new notes seems to help.
More to come soon. :P