HELP WITH OTO

[Tin-Loid]

plz halp me by doing this oto for Tree Sounds (play on IA Rocks) here is his (the letter that comes after beta) /1.0 : http://www.mediafire.com/download/uw57ass47edvzip/Tree+Sounds+Ver.1.0.rar

also if you could fix the weird overtones that would be AWESOME and you will get MAJOR credit for helping. SOLVED!

 

[Tomato Hentai]

You could always use moresampler's auto oto feature, since the voicebank is CV.
EDIT: Also, that "weird undertone" is due to the background noise, and the only way to fix it would be to re-record as noise removal usually screws with the actual audio and can worsen the quality of a sample.

 

[Zoku]

1) @Tomato Hentai Moresampler doesnt support CV yet.. ;;
2) @Tree Sounds The weird undertones are background noise. Re-record in a more isolated/quieter environment. If you have a closet that you can fit in, record in there.
3) The samples are cut at the beginning... It's a flawed and obsolete technique. This will have a NEGATIVE impact on transitions between notes. If you still have the original samples without the cut at the beginning, that would be awesome and will help a lot in otoing.
4) You use the American/English R sound.. ;; If you can't do the Japanese R (it sounds like the 't/d'-ish sound in 'water' and 'butter'), then just record a quick L sound in place of the actual R sound.

That's all I can say before oto-ing it... Of course, I could oto it with all of the imperfections, but it will be very helpful, especially for the oto-er and users, if you fixed the problems before the oto.

 

[Tomato Hentai]

1) @Tomato Hentai Moresampler doesnt support CV yet.. ;;

Oops. I thought it was CV only? I guess I was wrong.

 

[수연 <Suyeon>]

Oops. I thought it was CV only? I guess I was wrong.

Nope. Only hiragana encoded vcv works for now as far as the oto creator.

 

[Lee / Sambutt]

plz halp me by doing this oto for Tree Sounds (play on IA Rocks) here is his (the letter that comes after beta) /1.0 : http://www.mediafire.com/download/uw57ass47edvzip/Tree+Sounds+Ver.1.0.rar

also if you could fix the weird overtones that would be AWESOME and you will get MAJOR credit for helping.

Nope. Only hiragana encoded vcv works for now as far as the oto creator.

A very quick way to OTO is to copy-paste the files into Oremo's "result" folder and use its OTO generation feature. Then you have to edit it as needed (usually not too much), and DONE.

 

[수연 <Suyeon>]

A very quick way to OTO is to copy-paste the files into Oremo's "result" folder and use its OTO generation feature. Then you have to edit it as needed (usually not too much), and DONE.

In this case, the user's bank is a CV - from @Zoku's description a not up-to-standards one at that.

EDIT: I've looked at the bank myself, and I agree with @Zoku that the OP needs to rerecord in a quiet environment (or invest in better recording equipment), retain silence at the beginning and end of the samples, correct their pronunciation for the ら, り, る, れ, ろ, りゃ, りゅ, りょ (ra, ri, ru, re, ro, rya, ryu, ryo) samples, and they also need to maintain consistency. Some samples are mixed down as stereo which may give certain resamplers issues. The best sample rate for wav is 44100, 16-bit, mono - this combination tends to give resamplers little to no problems in most cases.

 

[Zoku]

In addition, the frq's rendered say that it is at C2.... The samples don't sound that low at all. Also when rendered in bkh01, there are samples that sound completely devoiced, as if it was a pure whisper. Like. 100% breath.

 

[Tin-Loid]

In addition, the frq's rendered say that it is at C2.... The samples don't sound that low at all. Also when rendered in bkh01, there are samples that sound completely devoiced, as if it was a pure whisper. Like. 100% breath.

I didn't record them at a c2 it was like at an e or something but not a c or an f

 

[수연 <Suyeon>]

The reason the frequencies are so messed up is....
1. not all of the samples aren't on the same note.
- The tuner I'm using (which is pretty accurate) is reading all the notes as being anywhere from A#2, B2, or C2...

2. noise within the samples are overriding the vocals all together.
- I've sung the more consistent notes from your bank myself to see what note UTAU should be reading and it appears that the proper note should be at least G#3 (the tuner was able to detect this despite my laptop mic being a noisy pos), which means that the noise is overpowering your samples (hence the readings of C2 by UTAU and some samples coming across as whispers, white noise, or unvoiced by certain resamplers)

In any case, you'll need to re-record. May I ask what you're using to record your samples with?

 

[Tin-Loid]

The reason the frequencies are so messed up is....
1. not all of the samples aren't on the same note.
- The tuner I'm using (which is pretty accurate) is reading all the notes as being anywhere from A#2, B2, or C2...

2. noise within the samples are overriding the vocals all together.
- I've sung the more consistent notes from your bank myself to see what note UTAU should be reading and it appears that the proper note should be at least G#3 (the tuner was able to detect this despite my laptop mic being a noisy pos), which means that the noise is overpowering your samples (hence the readings of C2 by UTAU and some samples coming across as whispers, white noise, or unvoiced by certain resamplers)

In any case, you'll need to re-record. May I ask what you're using to record your samples with?

there shouldn't have been any background noise I was at my house by myself and turned off EVERYTHING that made noise except my tower........ it's probably my mic and i'm re-recording again tonight.