이번시간에는 파이썬으로 텍스트-음성 변환하는 라이브러리들을 살펴보도록하겠습니다. 본강의는 맥북기준으로 진행됩니다.
pyttsx3
우선 대표적인 것으로는 pyttsx3가 있는데, 단 버젼은 pyttsx3==2.99로 설치하셔야합니다. 최근버젼에 버그가 있어서 현재시간으로는 아직 개선중인것 같습니다(참고). 2.99버젼은 pip에는 없고 https://test.pypi.org/simple/에서 가져다 설치하셔야해요.
pip install --no-cache-dir --extra-index-url https://test.pypi.org/simple/ pyttsx3==2.99코드는 아래와 같이 사용합니다.
import pyttsx3
engine = pyttsx3.init()
text = "Hello, world!"
engine.save_to_file(text, "output_pyttsx3.mp3")
engine.runAndWait()추가적인 옵션으로는 속도를 변경하려면 rate속성을 바꾸면 됩니다.
engine.setProperty('rate', 125)볼륨을 변경하시려면 volume의 값을 변경하시면 되는데, 0에서 1사이의 값을 넣습니다.
engine.setProperty('volume',0.9)그리고 목소리도 변경할 수가 있는데, voice속성에 변경할 voice ID를 할당하시면 됩니다.
voices = engine.getProperty('voices')
engine.setProperty('voice', voices[66].id)참고로 pyttsx3에서 지원하는 voice들은 아래와 같습니다(배열방번호, ID, name, 언어, 성별, 나이).
0,com.apple.speech.synthesis.voice.Agnes,Agnes,en_US,VoiceGenderFemale,35
1,com.apple.speech.synthesis.voice.Albert,Albert,en_US,VoiceGenderNeuter,30
2,com.apple.speech.synthesis.voice.Alex,Alex,en_US,VoiceGenderMale,35
3,com.apple.speech.synthesis.voice.alice,Alice,it_IT,VoiceGenderFemale,35
4,com.apple.speech.synthesis.voice.allison.premium,Allison,en_US,VoiceGenderFemale,35
5,com.apple.speech.synthesis.voice.alva,Alva,sv_SE,VoiceGenderFemale,35
6,com.apple.speech.synthesis.voice.amelie,Amelie,fr_CA,VoiceGenderFemale,35
7,com.apple.speech.synthesis.voice.anna,Anna,de_DE,VoiceGenderFemale,35
8,com.apple.speech.synthesis.voice.ava.premium,Ava,en_US,VoiceGenderFemale,35
9,com.apple.speech.synthesis.voice.BadNews,Bad News,en_US,VoiceGenderNeuter,50
10,com.apple.speech.synthesis.voice.Bahh,Bahh,en_US,VoiceGenderNeuter,2
11,com.apple.speech.synthesis.voice.Bells,Bells,en_US,VoiceGenderNeuter,100
12,com.apple.speech.synthesis.voice.Boing,Boing,en_US,VoiceGenderNeuter,1
13,com.apple.speech.synthesis.voice.Bruce,Bruce,en_US,VoiceGenderMale,35
14,com.apple.speech.synthesis.voice.Bubbles,Bubbles,en_US,VoiceGenderNeuter,0
15,com.apple.speech.synthesis.voice.carmit,Carmit,he_IL,VoiceGenderFemale,35
16,com.apple.speech.synthesis.voice.Cellos,Cellos,en_US,VoiceGenderNeuter,50
17,com.apple.speech.synthesis.voice.damayanti,Damayanti,id_ID,VoiceGenderFemale,35
18,com.apple.speech.synthesis.voice.daniel,Daniel,en_GB,VoiceGenderMale,35
19,com.apple.speech.synthesis.voice.Deranged,Deranged,en_US,VoiceGenderNeuter,30
20,com.apple.speech.synthesis.voice.diego,Diego,es_AR,VoiceGenderMale,35
21,com.apple.speech.synthesis.voice.ellen,Ellen,nl_BE,VoiceGenderFemale,35
22,com.apple.speech.synthesis.voice.fiona,Fiona,en-scotland,VoiceGenderFemale,35
23,com.apple.speech.synthesis.voice.Fred,Fred,en_US,VoiceGenderMale,30
24,com.apple.speech.synthesis.voice.GoodNews,Good News,en_US,VoiceGenderNeuter,8
25,com.apple.speech.synthesis.voice.Hysterical,Hysterical,en_US,VoiceGenderNeuter,30
26,com.apple.speech.synthesis.voice.ioana,Ioana,ro_RO,VoiceGenderFemale,35
27,com.apple.speech.synthesis.voice.joana,Joana,pt_PT,VoiceGenderFemale,35
28,com.apple.speech.synthesis.voice.jorge,Jorge,es_ES,VoiceGenderMale,35
29,com.apple.speech.synthesis.voice.juan,Juan,es_MX,VoiceGenderMale,35
30,com.apple.speech.synthesis.voice.Junior,Junior,en_US,VoiceGenderMale,8
31,com.apple.speech.synthesis.voice.kanya,Kanya,th_TH,VoiceGenderFemale,35
32,com.apple.speech.synthesis.voice.karen,Karen,en_AU,VoiceGenderFemale,35
33,com.apple.speech.synthesis.voice.Kathy,Kathy,en_US,VoiceGenderFemale,30
34,com.apple.speech.synthesis.voice.kyoko,Kyoko,ja_JP,VoiceGenderFemale,35
35,com.apple.speech.synthesis.voice.laura,Laura,sk_SK,VoiceGenderFemale,35
36,com.apple.speech.synthesis.voice.lekha,Lekha,hi_IN,VoiceGenderFemale,35
37,com.apple.speech.synthesis.voice.luca,Luca,it_IT,VoiceGenderMale,35
38,com.apple.speech.synthesis.voice.luciana,Luciana,pt_BR,VoiceGenderFemale,35
39,com.apple.speech.synthesis.voice.maged,Maged,ar_SA,VoiceGenderMale,35
40,com.apple.speech.synthesis.voice.mariska,Mariska,hu_HU,VoiceGenderFemale,35
41,com.apple.speech.synthesis.voice.meijia,Mei-Jia,zh_TW,VoiceGenderFemale,35
42,com.apple.speech.synthesis.voice.melina,Melina,el_GR,VoiceGenderFemale,35
43,com.apple.speech.synthesis.voice.milena,Milena,ru_RU,VoiceGenderFemale,35
44,com.apple.speech.synthesis.voice.moira,Moira,en_IE,VoiceGenderFemale,35
45,com.apple.speech.synthesis.voice.monica,Monica,es_ES,VoiceGenderFemale,35
46,com.apple.speech.synthesis.voice.nora,Nora,nb_NO,VoiceGenderFemale,35
47,com.apple.speech.synthesis.voice.paulina,Paulina,es_MX,VoiceGenderFemale,35
48,com.apple.speech.synthesis.voice.Organ,Pipe Organ,en_US,VoiceGenderNeuter,500
49,com.apple.speech.synthesis.voice.Princess,Princess,en_US,VoiceGenderFemale,8
50,com.apple.speech.synthesis.voice.Ralph,Ralph,en_US,VoiceGenderMale,50
51,com.apple.speech.synthesis.voice.rishi,Rishi,en_IN,VoiceGenderMale,35
52,com.apple.speech.synthesis.voice.samantha.premium,Samantha,en_US,VoiceGenderFemale,35
53,com.apple.speech.synthesis.voice.sara,Sara,da_DK,VoiceGenderFemale,35
54,com.apple.speech.synthesis.voice.satu,Satu,fi_FI,VoiceGenderFemale,35
55,com.apple.speech.synthesis.voice.sinji,Sin-ji,zh_HK,VoiceGenderFemale,35
56,com.apple.speech.synthesis.voice.tessa,Tessa,en_ZA,VoiceGenderFemale,35
57,com.apple.speech.synthesis.voice.thomas,Thomas,fr_FR,VoiceGenderMale,35
58,com.apple.speech.synthesis.voice.tingting,Ting-Ting,zh_CN,VoiceGenderFemale,35
59,com.apple.speech.synthesis.voice.Trinoids,Trinoids,en_US,VoiceGenderNeuter,2001
60,com.apple.speech.synthesis.voice.veena,Veena,en_IN,VoiceGenderFemale,35
61,com.apple.speech.synthesis.voice.Vicki,Vicki,en_US,VoiceGenderFemale,35
62,com.apple.speech.synthesis.voice.Victoria,Victoria,en_US,VoiceGenderFemale,35
63,com.apple.speech.synthesis.voice.Whisper,Whisper,en_US,VoiceGenderNeuter,30
64,com.apple.speech.synthesis.voice.xander,Xander,nl_NL,VoiceGenderMale,35
65,com.apple.speech.synthesis.voice.yelda,Yelda,tr_TR,VoiceGenderFemale,35
66,com.apple.speech.synthesis.voice.yuna,Yuna,ko_KR,VoiceGenderFemale,35
67,com.apple.speech.synthesis.voice.yuri,Yuri,ru_RU,VoiceGenderMale,35
68,com.apple.speech.synthesis.voice.Zarvox,Zarvox,en_US,VoiceGenderNeuter,1
69,com.apple.speech.synthesis.voice.zosia,Zosia,pl_PL,VoiceGenderFemale,35
70,com.apple.speech.synthesis.voice.zuzana,Zuzana,cs_CZ,VoiceGenderFemale,35음성출력은 아래와 같이 say()함수에 말할 텍스트를 넘겨주면 됩니다.
engine.say("Hello World!")음성파일을 저장하시려면 아래와 같이 save_to_file()함수에 텍스트와 저장할 파일경로를 넘겨주시면 됩니다.
engine.save_to_file("Hello World!", "output_pyttsx3.mp3")어떤 이유에서인지 pyttsx3는 mpg123로는 출력이 안되네요.
gtts
이번에는 gtts를 이용해서 TTS를 구현해 보도록 하겠습니다.
gtts를 설치해주세요.
pip install gtts그리고 tts_with_gtts.py를 생성해서 아래 코드를 저장해주세요.
from gtts import gTTS
tts = gTTS("Hello, world!", lang="en")
tts.save("output_gtts.mp3")실행해 보시면 output_gtts.mp3가 생성되어 있을거에요.
python tts_with_gtts.pygtts로 만든 파일은 mpg123를 이용해서 재생할 수 있습니다.
mpg123 output_gtts.mp3그런데 얘는 다양한 목소리 지원이나 속도, 볼륨조절등의 옵션은 없어요.