From c2275d09f2b30fbb7a5606f87ac46297df8881d4 Mon Sep 17 00:00:00 2001 From: Hiroshiba Kazuyuki Date: Wed, 6 Dec 2023 18:41:15 +0900 Subject: [PATCH] =?UTF-8?q?=E6=99=82=E7=B3=BB=E5=88=97=E3=82=92=E5=88=97?= =?UTF-8?q?=E3=81=AB=E3=80=81"=20"=E3=82=92"=E3=80=80"=E3=81=AB?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- test/test_synthesis_engine.py | 12 ++++++------ voicevox_engine/synthesis_engine/synthesis_engine.py | 6 +++--- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/test/test_synthesis_engine.py b/test/test_synthesis_engine.py index c01ea15cc..17f0ed4a5 100644 --- a/test/test_synthesis_engine.py +++ b/test/test_synthesis_engine.py @@ -179,9 +179,9 @@ def test_pad_with_silence(): # Expects true_moras_with_silence = [ - _gen_mora(" ", None, None, "sil", 2 * 0.01067, 0.0), + _gen_mora("　", None, None, "sil", 2 * 0.01067, 0.0), _gen_mora("ヒ", "h", 2 * 0.01067, "i", 4 * 0.01067, 100.0), - _gen_mora(" ", None, None, "sil", 6 * 0.01067, 0.0), + _gen_mora("　", None, None, "sil", 6 * 0.01067, 0.0), ] # Outputs @@ -195,13 +195,13 @@ def test_calc_frame_per_phoneme(): # Inputs query = _gen_query(speedScale=2.0) moras = [ - _gen_mora(" ", None, None, " ", 2 * 0.01067, 0.0), # 0.01067 [sec/frame] + _gen_mora("　", None, None, "　", 2 * 0.01067, 0.0), # 0.01067 [sec/frame] _gen_mora("コ", "k", 2 * 0.01067, "o", 4 * 0.01067, 0.0), _gen_mora("ン", None, None, "N", 4 * 0.01067, 0.0), _gen_mora("、", None, None, "pau", 2 * 0.01067, 0.0), _gen_mora("ヒ", "h", 2 * 0.01067, "i", 4 * 0.01067, 0.0), _gen_mora("ホ", "h", 4 * 0.01067, "O", 2 * 0.01067, 0.0), - _gen_mora(" ", None, None, " ", 6 * 0.01067, 0.0), + _gen_mora("　", None, None, "　", 6 * 0.01067, 0.0), ] # Expects @@ -220,13 +220,13 @@ def test_calc_frame_pitch(): # Inputs query = _gen_query(pitchScale=2.0, intonationScale=0.5) moras = [ - _gen_mora(" ", None, None, " ", 0.0, 0.0), + _gen_mora("　", None, None, "　", 0.0, 0.0), _gen_mora("コ", "k", 0.0, "o", 0.0, 50.0), _gen_mora("ン", None, None, "N", 0.0, 50.0), _gen_mora("、", None, None, "pau", 0.0, 0.0), _gen_mora("ヒ", "h", 0.0, "i", 0.0, 125.0), _gen_mora("ホ", "h", 0.0, "O", 0.0, 0.0), - _gen_mora(" ", None, None, " ", 0.0, 0.0), + _gen_mora("　", None, None, "　", 0.0, 0.0), ] phoneme_str = "pau k o N pau h i h O pau" phonemes = [OjtPhoneme(p, 0, 0) for p in phoneme_str.split()] diff --git a/voicevox_engine/synthesis_engine/synthesis_engine.py b/voicevox_engine/synthesis_engine/synthesis_engine.py index bd6e5d73f..9673e0742 100644 --- a/voicevox_engine/synthesis_engine/synthesis_engine.py +++ b/voicevox_engine/synthesis_engine/synthesis_engine.py @@ -89,9 +89,9 @@ def pre_process( Returns ------- flatten_moras : List[Mora] - モーラ時系列（前後の無音含まない） + モーラ列（前後の無音含まない） phoneme_data_list : List[OjtPhoneme] - 音素時系列（前後の無音含む） + 音素列（前後の無音含む） """ flatten_moras = to_flatten_moras(accent_phrases) @@ -112,7 +112,7 @@ def pre_process( def generate_silence_mora(length: float) -> Mora: """無音モーラの生成""" - return Mora(text=" ", vowel="sil", vowel_length=length, pitch=0.0) + return Mora(text="　", vowel="sil", vowel_length=length, pitch=0.0) def pad_with_silence(moras: list[Mora], query: AudioQuery) -> list[Mora]: