From 2da51752d2afa04ecc94346a5cd2b163bb0dcce9 Mon Sep 17 00:00:00 2001
From: Echo-Nie <157974576+Echo-Nie@users.noreply.github.com>
Date: Wed, 16 Apr 2025 15:41:30 +0800
Subject: [PATCH] =?UTF-8?q?=E3=80=90PaddleSpeech=20No.5=E3=80=91=20(#4042)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

* run.sh修改：为 synthesize 和 synthesize_e2e 添加 --stage 参数控制 vocoder 模型选择，REAMDE.md修改：补充 stage 参数说明，明确 vocoder 选择逻辑

* 添加run.sh中stage参数相关的注释

* HiFiGAN改为MultiBand MelGAN

* cmsc文件改回原位（No.15不修改），这里只对No.6做修改

* fix Speech No.5

* update README.md

* fix the README.md
---
 examples/aishell3_vctk/ernie_sat/README.md | 6 ++++--
 examples/aishell3_vctk/ernie_sat/run.sh    | 5 +++--
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/examples/aishell3_vctk/ernie_sat/README.md b/examples/aishell3_vctk/ernie_sat/README.md
index 57dcd472f..e73291add 100644
--- a/examples/aishell3_vctk/ernie_sat/README.md
+++ b/examples/aishell3_vctk/ernie_sat/README.md
@@ -15,7 +15,7 @@ In ERNIE-SAT, we propose two innovations:
 Download all datasets and extract it to `~/datasets`:
 - The aishell3 dataset is in the directory `~/datasets/data_aishell3`
 - The vctk dataset is in the directory `~/datasets/VCTK-Corpus-0.92`
- 
+
 ### Get MFA Result and Extract
 We use [MFA](https://github.com/MontrealCorpusTools/Montreal-Forced-Aligner) to get durations for the fastspeech2 training.
 You can download from here:
@@ -97,6 +97,8 @@ hifigan_aishell3_ckpt_0.2.0
 ```bash
 CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name}
 ```
+synthesize, vocoder is `hifigan`
+
 ##  Speech Synthesis and Speech Editing
 ### Prepare
 
@@ -150,7 +152,7 @@ You can check the text of downloaded wavs in `source/README.md`.
 ```bash
 ./run.sh --stage 3 --stop-stage 3 --gpus 0
 ```
-`stage 3` of `run.sh` calls `local/synthesize_e2e.sh`.
+`stage 3` of `run.sh` calls `local/synthesize_e2e.sh`. `synthesize_e2e.sh` is a script for end-to-end speech synthesis, supporting cross-language speech synthesis tasks, including English-to-Chinese (en → zh) and Chinese-to-English (zh → en).
 
 You can modify  `--wav_path`、`--old_str` and `--new_str` yourself, `--old_str` should be the text corresponding to the audio of  `--wav_path`, `--new_str` should be designed according to `--task_name`, `--source_lang` and `--target_lang` should be different in this example.
 ## Pretrained Model
diff --git a/examples/aishell3_vctk/ernie_sat/run.sh b/examples/aishell3_vctk/ernie_sat/run.sh
index 8cd9d8d1b..d29f0b6e8 100755
--- a/examples/aishell3_vctk/ernie_sat/run.sh
+++ b/examples/aishell3_vctk/ernie_sat/run.sh
@@ -27,10 +27,11 @@ if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
 fi
 
 if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # synthesize, vocoder is pwgan
+    # synthesize, vocoder is hifigan
     CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi
 
 if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
+    # synthesize_e2e, default speech synthesis from Chinese to English, use stage1 to switch from English to Chinese
+    CUDA_VISIBLE_DEVICES=${gpus} ./local/synthesize_e2e.sh --stage 0 ${conf_path} ${train_output_path} ${ckpt_name} || exit -1
 fi