Skip to content
This repository has been archived by the owner on Mar 3, 2023. It is now read-only.

Commit

Permalink
update readme for 1.2.5
Browse files Browse the repository at this point in the history
  • Loading branch information
luoyi committed Jan 6, 2023
1 parent fc4e9e0 commit 39a943b
Show file tree
Hide file tree
Showing 3 changed files with 27 additions and 4 deletions.
19 changes: 15 additions & 4 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,15 @@ Speech synthesis model repo for galgame characters based on Tacotron2 , Hifigan

## 近期更新

1.2.5:

> 1. 规范diff_svc import
> 2. 精简VITS与Tacotron2代码并移除了matplotlib, tensorflow依赖
> 3. 删除不必要导入及依赖,加快启动速度
> 4. diff svc增加Crepe轻量模型
> 5. 使用两种编译器(pyinstaller, nuitka)
> 6. diff svc音频处理采样类型换为fft
1.2.4:

>1. 增加最近模型与输出路径记录,方便下次打开继续。
Expand Down Expand Up @@ -110,14 +119,16 @@ VITS特殊说明

### diff-svc

![diff-svc gui](./assets/diff_svc.png)
![diff_svc_gui](./assets/diff_svc.png)

参数说明:

1. 升降半音:默认为0,支持正负整数输入,单位为半音
2. 启用Crepe:该选项可降噪音频,启用后CPU耗时较高,越为原音频时长8倍,建议合成最终版本再开启,干净的音频无需开启。
3. 加速倍率:默认为20,耗时约1:3,预览可使用100,耗时约1:1(该设置会影响音频质量)
4. 待转换音频:wav或ogg纯人声音频,转换后为模型角色音色。
2. 启用Crepe:该选项可降噪音频,启用后CPU耗时较高,约为原音频时长8倍,建议合成最终版本再开启,干净的音频无需开启。
3. Crepe轻量模式:在启用Crepe的前提下,勾选此选项后Crepe使用Tiny模型,耗时更短,约为原音频时长1/4。
4. 加速倍率:默认为20,耗时约1:3,预览可使用100,耗时约1:1(该设置会影响音频质量)
5. 加速倍率:默认为20,耗时约1:3,预览可使用100,耗时约1:1(该设置会影响音频质量)
6. 待转换音频:wav或ogg纯人声音频,转换后为模型角色音色。

## 在线Demo

Expand Down
12 changes: 12 additions & 0 deletions README_en.md
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,17 @@ Speech synthesis model repo for galgame characters based on Tacotron2, Hifigan a

## Latest Updates

1.2.5:

> 1. Improve diff svc code.
> 2. Simplify VITS and Tacotron2 code and remove matplotlib, tensorflow dependencies.
> 3. Remove unnecessary imports and dependencies.
> 4. Crepe can toggle between full and tiny models.
> 5. Added nuitka compiled distribution.
> 6. Modified the resampling type used by diffsvc when processing audio sample rates and remove numba dependency. (For nuitka to work.)
>
> A standalone language file and related settings will be available later to support other languages.
1.2.4:

>1. Add recent use history.
Expand Down Expand Up @@ -115,6 +126,7 @@ Setting Instructions:
2. 启用Crepe(Enable Crepe):Improved audio quality when enabled, but will take longer.
3. 加速倍率(Acceleration ratio):Default is 20, higher values will infer faster, but may affect quality.
4. 待转换音频(Input audio):wav or ogg file with vocals only.
5. Crepe轻量模式(Crepe Tiny): With Crepe enabled, Crepe uses Tiny models when this option is checked and takes less time.

## Online Demo

Expand Down
Binary file modified assets/diff_svc.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.

0 comments on commit 39a943b

Please sign in to comment.