Open Patchethium opened 1 year ago
I was wondering what features you are imagining are similar to TALQu's Pitch Editor.
It also sounds a bit close to the Guided Synthesis
you were trying before!
https://github.com/VOICEVOX/voicevox_engine/issues/231
Well, I think the need for VOICEVOX is very low... (because other tools can do it and the difficulty is very high).
I made the issue because we need AudioQuery
at the frame level for anyway.
(I plan to use this for other purposes as well!)
https://github.com/VOICEVOX/voicevox_engine/issues/528
想像されている機能はTALQuのピッチエディタに似ているのかなと思いました。
以前挑戦されていたGuided Synthesis
ともちょっと近そうですね!
https://github.com/VOICEVOX/voicevox_engine/issues/231
まあ、VOICEVOXにとって必要度はとても低いと思いますが・・・(他のツールでできるし、難度がとても高いので)。
兎にも角にもフレームレベルのAudioQuery
が必要だと思ってissueを作ってみました。
(これは別の目的でも使う予定です!)
https://github.com/VOICEVOX/voicevox_engine/issues/528
以前挑戦されていた
Guided Synthesis
ともちょっと近そうですね! VOICEVOX/voicevox_engine#231
That's right, btw I'm still looking for an alternative for Julius...
他のツールでできるし、難度がとても高いので
Vocalshifter
can absolutely do this job, but users need to find the audio file in explorer, drag it into the software window, create a new project, do the pitch editing, then export the audio. The need to do it every time makes it exhausting. I think there is some point in building our own.
I see! To easily use the audio generated by VOICEVOX with other tools, there is a way to make VOICEVOX itself a VST plug-in. https://github.com/VOICEVOX/voicevox/issues/619#issuecomment-1001360234
I'm not familiar with it, so I have no idea how to do it...
なるほどです! VOICEVOXで生成した音声を他のツールで簡単に使うには、VOICEVOX自体をVSTプラグインにするという手段があります。 https://github.com/VOICEVOX/voicevox/issues/619#issuecomment-1001360234
詳しくないのでやり方は全くわかりませんが・・・。
内容
We have
pyworld
at the engine side, so we can directly alter the pitch/duration of the exported raw audio, extending user's control, just like a built-invocalshifter
.Pros 良くなる点
vocalshifter
is not ideal, building it in the software can help.vocalshifter
is Windows only, butpyworld
supports other platforms.Cons 悪くなる点
Need to write the code
実現方法
Send audio to engine > engine returns frame level pitch info > edit the pitch curve on GUI > send altered pitch to engine > engine synthesizes audio with new pitch info
その他
Furthermore, by stretching the acoustic features we can alter the duration too.