Amazon Transcribeでいろんな音源を文字起こし(英語編)

Amazon Transcribeを利用して様々な英語の音源を文字起こししてみました。

Amazon Transcribe サービス検討の参考にしていただければと思います。

2018/04 時点の機能

  • アメリカ英語、スペイン語に対応
  • 話者を認識
  • 語彙の登録

と言った機能が備わっています。

数ヶ月以内に、以下の言語に対応予定であることがアナウンスされています。

  • 日本語
  • ロシア語
  • イタリア語
  • トルコ語
  • チェコ語
  • 繁体字中国語(traditional Chinese)

10秒コマーシャル

  • 言語:アメリカ英語
  • 種類:コマーシャル
  • 長さ:10秒

Amazon が主催する機械学習(machine learning)に関するオンラインカンファレンス"AWS Innovate"の10秒コマーシャルにおけるアメリカ英語の短いナレーションです。

At a w s a ii powered by machine, london drives innovation and unlocks. Tomorrow. Register for a w s. Innovate today and start building.

  • AWS => a w s
  • AI => a ii
  • machine learning => machine, london

というように書き起こされています。

30 秒コマーシャル

  • 言語:アメリカ英語
  • 種類:コマーシャル
  • 長さ:30秒

Nintendo Labo の30秒コマーシャルにおけるアメリカ英語のナレーションです。

Introducing the nintendo lebow variety kid. A new experience for a nintendo switch you can make your own nintendo lebeau creations on. Bring them to life with games on nintendo switch nintendo, lebow. Make play, discover available now rated e for everyone.

  • labo => lebow や lebeau

というように書き起こされています。

昔のスピーチ(アメリカ大統領 JFケネディ)

  • 言語:アメリカ英語
  • 種類:演説
  • 長さ:9分20秒
  • 音源:mp3
  • 44.1kHz/Stereo

ジョン・F・ケネディが1963年6月26日に西ベルリンで行った有名な演説、"Ich bin ein Berliner" を文字起こししてみます。

少し長いです。

I am proud to come to the city as the guest of your distinguished mayor who has symbolized throughout the world oh the fighting spirit of west berlin and i'm proud and i am proud to visit the federal republic with your distinguished chancellor who for so many years as committed germany to democracy and freedom and progress and to come here in the company of my fellow american general play who has been in this city during its great moments of crisis and we'll come again if ever needed two thousand years ago two thousand years ago the proudest boast law's key wis romano some today in the world of freedom the proudest boast is ish being i'd be lena i i i appreciate i appreciate my interpreter translating my german there are many people in the world who really don't understand or say they don't know what is the great issue between the free world and the communist world let them come to berlin there are some who say that communism is the wave of the future let them come to berlin and they say in europe and elsewhere waken work with the communist left them come to berlin and there are even a few who say that it's true that communism is an evil system but it permits us to make economic progress love three knock berlin in common let them ah last seen a trillionth common freedom as many difficulties and democracy is not perfect but we have never had to put a wall up to keep our people in to prevent them from leaving us.

I want to say on the off of my countrymen who lived many miles away on the other side of the atlantic who are far distant from you, that they take the greatest pride that they have been able to share with you even from a distance the story of the last eighteen years i know of no town, no city that has been besieged for eighteen years that still lives with the vitality and the force and the whole and the determination of the city of west berlin wait is the most obvious and vivid demonstration off the failures of the communist system.

For all the world to see.

We take no satisfaction in it, for it is, as your mayor has said on offense not only against history bought an offense against humanity, separating families dividing husbands and wives and brothers and sisters and dividing up people who wished to be joined together.

What is true off this city is true of germany.

Really lasting peace in europe can never be assured as long as one german out of four is denied the elementary right of free men and that is to make a free choice in eighteen years off peace and good faith, this generation of germans as earned the right to be free, including the right to unite their families and their nation in lasting peace with goodwill to all people.

You live in a defended island of freedom.

But your life is part of the main.

So let me ask you, as i closed, lift your eyes beyond the dangers of today to the hopes of tomorrow, beyond the freedom merely of this city of berlin or your country of germany, to the advance of freedom everywhere, beyond the wall, to the day of peace, with justice beyond yourself and ourselves to all mankind, freedom is indivisible.

And when one man is enslaved all enough free when all of free, then we can look forward to that day when this city will be joined as one and this country and this great continent of europe in a peaceful and hopeful globe, when that day finally comes as it will, the people of west berlin can take sober satisfaction in the fact that they were in the front line for almost two decades.

A lot of all free men, wherever they may live, are citizens of berlin.

And therefore, as a free man, i take pride in the words ich bin ein be alina.

ゆっくり話しているためか、古い録音であるにもかかわらず、かなりの精度で文字起こし出来ています。

一部ドイツ語・ラテン語が含まれています。

実行結果を確認すると

  • civis romanus sum(ラテン語「私はローマ市民である」) => key wis romano some
  • Ich bin eir Berliner(ドイツ語「私はベルリン市民である」) => ish being i'd be lena

というように書き起こされています。

transcript 全文 http://news.bbc.co.uk/2/hi/europe/3022166.stm

昔のスピーチ(英首相:チャーチル)

  • 言語:イギリス英語
  • 種類:演説
  • 長さ:2分24秒

チャーチルが1940年6月4日に行った有名な演説、We shall fight on the beaches を文字起こししてみます。

1'16" からを抜粋します

wei feel go onto the end fighting brahms little fried on the seat and oceans we shall fight with growing confidence and growing strength in the air we shall defend our island whatever they got maybe, we shall fight on the b trees will fight on the landing grounds we shall fight in the fields and in the streets we shall fight in the hills we shall never surrender

正しい transcript は以下です。

We shall go on to the end, we shall fight in France, we shall fight on the seas and oceans, we shall fight with growing confidence and growing strength in the air, we shall defend our Island, whatever the cost may be, we shall fight on the beaches, we shall fight on the landing grounds, we shall fight in the fields and in the streets, we shall fight in the hills; we shall never surrender

Amazon Transcribe が現時点ではイギリス英語に対応していないこと、チャーチルのこの演説は非常に聞き取りにくく感じること(英語力含めて個人差が大きいと思います)を考慮すると、かなり善戦していると感じました。

transcript 全文 https://winstonchurchill.org/resources/speeches/1940-the-finest-hour/we-shall-fight-on-the-beaches/

コールセンター

ここからは話者の識別機能を試してみます。

  • 言語:アメリカ英語
  • 種類:コールセンター
  • 長さ:2分08秒
  • 話者:2名
  • 音源 : https://www.directlineinc.com/wp-content/uploads/sites/51/2016/03/Plumbing-Call-Recording.mp3 (https://www.directlineinc.com/audio-samples-call-center-services/)
  • 44.1kHz/Stereo

コールセンターでの電話録音です。

話者の数を2名でジョブ実行した結果が以下です。

概ね正確に文字起こしされています("terrace"が物騒な単語に誤って書き起こされていますが、、、)。

また、話者の判別もかなり正確です。話者が2名であり、余計な雑音がないこと、男性と女性で音質が大きくことなる点などが貢献しているのかもしれません。

Speaker 0: Good afternoon. Thank you for covering professional plumbing

Speaker 1: yes ah hi. I need a plumber to come up to my home as possible

Speaker 0: all right, sir, we can assist you with that. May i ask who's calling place?

Speaker 1: My name is joe green

Speaker 0: thank you, sir. That's j o e and then green is a g r e m

Speaker 1: g r e e n e

Speaker 0: thank you, mr green that's g r e e mr greene, may i have your area code in telephone number please?

Speaker 1: My phone number is six to six five five five thirteen thirty eight

Speaker 0: thank you, sir. That's six to six five five five one, three, three eight mr green

Speaker 1: yes that's right

Speaker 0: and once again you need services evening, sir.

Speaker 1: Ah, yes i i do i have three batters five, two and a half bath but two of my toilets are backing up and i don't know so so i need a plumber need appointment come out it is possible

Speaker 0: definitely serve will be able to assist you with that. May i have your address where you need the service please, mr green

Speaker 1: yeah, the address is two, three eight walnut terrorists

Speaker 0: thank you, sir. That's two three a walnut terrorists that's w l n u t

Speaker 1: yes that's right wolf

Speaker 0: and your zip code police sir

Speaker 1: nine one three to eight

Speaker 0: thank you, sir. That's nine one three to eight and i'm sorry, i i neglected to ask. Is there an apartment or unit number there, sir?

Speaker 1: Uh, no, no, tom,

Speaker 0: all right, sir, and you're in northridge, correct?

Speaker 1: Yes, you're

Speaker 0: right, sir. And once again, i'm mark this urgent. So we'll dispatch this out immediately. Certain and once again, that's. Two toilets that are backed up. You

Speaker 1: have two toilets backing up. How soon do you think i can have a plumber? Come out?

Speaker 0: What? We were going to dispatch this out immediately for you, mr green and one of our plumbers should be contacting you within thirty minutes, sir. Okay,

Speaker 1: great. Thank you. Thank

Speaker 0: you, sir. Goodbye. All

Speaker 1: right.

対談

対談での識別機能を試してみます。

  • 言語:アメリカ英語
  • 種類:対談
  • 長さ:6分34秒
  • 話者:2名

TheEllenShow での Ellen DeGeneres と Bill Gates の対談です。

観客の笑い声もありますが、基本的に Ellen と Bill の二人の対話です。 先程のコールセンターと同じく、話者は2名であり、男性と女性ということもあり、音質が大きくことなります(少なくとも、人間にとっては判別しやすい)。

話者の数を2名でジョブ実行した結果が以下です。

概ね正確に文字起こしされています。

発言がかぶっていたり、切り替わるタイミングを中心に、話者の認識精度が落ちています。

Speaker 1: I'm so happy to have you here this is the first time having you on, so thanks so i know you were you were nervous about the entrance he thought i think people feel like they're supposed to dance and so and i was really surprised because that was here earlier today for you rehearsal and then you abandoned it. But we should at least show them the rehearsal because it was really good.

Speaker 1: Your daughter is looking at you like i've never seen you dance like that.

Speaker 1: Hi. So the last time we saw each other, it was at the white house, we both were receiving the medal of freedom, and that was quite a day, wouldn't it

Speaker 0: was amazing group? Yeah, really,

Speaker 1: really fun. So you were here with your daughter, who is twenty one, right? And you were twenty one when you became a billionaire. Is that right?

Speaker 0: Almost all

Speaker 1: right. So around that age, when you're like the youngest person to become a billionaire, is that right

Speaker 0: for you? Yeah, in terms of my own earning it on my own. Yeah, right.

Speaker 1: Ok, so what is the most important thing? Yeah, so did you when you were a kid, did you did you care about money or you just cared about technology and that's just happened?

Speaker 0: Mostly i love software i do remember the private school i went to there were other kids whose families were better off like they had a porch or something but it wasn't that that big of a deal my thing was that i just love doing software i loved hiring people and i was stunned when it and that being so valuable really

Speaker 1: yeah you that surprised you yeah,

Speaker 0: because i always had to be careful that we weren't hiring too many people i was always worried because i was people work for me were older than me and they had kids and i always thought what

Speaker 1: if we don't get paid

Speaker 0: will i be able to meet the payroll so i was always very conservative about the finances and then when we did go public uh was i uh thirty by then uh then i was kind of stunned what it multiplied out too right?

Speaker 1: So when you became a billionaire at what point did you start relaxing? Were you still nervous when you became a billionaire like i got to watch this? Uh well, i always want to have

Speaker 0: enough money in the bank so that even if our customers stand pace for here we could still keep paying everybody and do the r and d so i still be viewed as conservative you know? I don't have that many things that are extravagant tastes so uh didn't change too much so

Speaker 1: nothing really changed. You didn't say oh, i'm going to buy a portion.

Speaker 0: I did okay,

Speaker 1: yeah,

Speaker 0: that was that was an indulgence. And then eventually for my travel, i got a plane, which is a huge indulgence. So those are my two? Well,

Speaker 1: not really, because you travel all the time. So that's important too, that you have a plane. So you have a portion, a plane and that's it

Speaker 0: well, in terms of crazy things. Yeah, yeah, there's.

Speaker 1: Not like any, like, wild. Like you didn't build like an aquarium with sharks in it or something like that. We

Speaker 0: have a trampoline room in our house.

Speaker 1: Wow, the

Speaker 0: kids like that indoor trampoline recommended

Speaker 1: just one giant trampoline. Yeah,

Speaker 0: yeah. It's a room with a very high ceiling. Well,

Speaker 1: yeah, yeah, that would. That would be cool if you didn't put a high ceiling. Go on, kids.

Speaker 1: All right, so let's, talk about this. So you have you already put forty billion dollars of your own money into your foundation? Forty billion and on dh you've i think kind of encouraged other billionaires to do this as well, because it really is kind of to the people to fix the problems in the world, it seems right, so what is your main focus right now?

Speaker 0: Um my wife ellen and i picked global health as our big thing the fact that still have five million kids who die under the age of five now it was over ten million when we got started. So there's been huge progress over the last eighteen years? Uh some things like malaria, diarrhea coming up with new drugs and vaccines and getting them out to all the kids in the world. That's our our main thing, our second biggest thing is all in the u s which is trying to help improve the education system here yeah,

Speaker 1: that which needs

Speaker 0: on and how

Speaker 1: do you do that? I always think you get what you pay for, so if you don't pay teachers because most teachers air paying out of their own pocket to take care of these students. So how do you do that? Well,

Speaker 0: there's, some really phenomenal teachers. And so the dream is that you could take that top ten percent and have them help the others to get best practices, the best teaching ideas to spread all over the country and

Speaker 1: and how you know, we're listening to you obviously forty billion dollars does a lot and there are other people better helping, but how? What can what can we do? What what is the best thing that you could say that that just one person can do to help? Well

Speaker 0: particularly schools the ability to go through with the local public, school or charter school and engage with the kids mentor kids talk about the kind of work you do there's, …

まとめ

素直なアメリカ英語が正しく録音されている限りは、かなりの高精度で文字起こし出来ています。

機械的に文字起こし後、人間が補助的に誤認識を直したり、句読点を追加すると、一気に自然な文章になります。

面白い例が見つかり次第、随時更新したいと思います。