Ah Xu was not very interested in listening to books at one time （虽然一大部分原因是我看的小说没有有声书）, AI mechanical pronunciation, fixed accent, speech speed, will always lose the immersive emotional experience brought by the ups and downs of the plot
But only recently did I know that Microsoft released the artificial voice of its TTS 2021 version at the //Build 2021 conference, which gave me an innovative view on listening to books
You can listen to the examples on the official website. The 5 new Chinese voices can’t be heard by a machine at all, and they are realistic and scary: Click to view the examples on Microsoft’s official website
For those who like to listen to books , they probably have only one thought in mind like Ah Xu: let me experience it quickly! ! !
In short, Ah Xu is going to introduce how to use Microsoft's latest TTS voices on Android/iOS today!
1 Android terminal
For most people, TTS must still be used to read novels on mobile phones, so here is how to experience it on mobile phones first.
1.1 TTS
Currently the easiest way is this small APP made by @ag2s20150909
The latest version of APP is released on the author's Github: https://github.com/ag2s20150909/TTS/tree/master/release
If you have difficulty accessing Github, it is recommended to read "2022 Github Accelerated Access Tutorial"
After downloading and installing the APP, click "Set TTS" in the upper left corner, and change the default text-to-speech engine of the phone to "TTS"
Then go back to the APP, the default AI is Xiaoxiao, if you want to switch to another AI, you must check "Custom" first, and then click to switch!
Ah Xu himself prefers the AI voice of Yun Xi , and you can also choose your speaking style (news/customer service/assistant/chat, etc.), Ah Xu himself uses the default voice
And because TTS is a system-level function, basically all e-book readers support calling the system TTS to read aloud
You can immediately experience the awesomeness of Microsoft TTS on your favorite apps such as Reading and Jingdu Tianxia !
However, it is reported that this TTS will have problems of lagging and skipping from time to time. The solutions given by netizens are as follows:
- change format
- Change AI speaker
- change time
Changing the format means that in the settings in the upper right corner of TTS, change the format of the audio stream
To put it bluntly, it is to reduce the audio quality, 32kbit/64kbit/128kbit can always be understood
Theoretically, if the reading is stuck due to network problems, reducing the audio quality should significantly improve the problem
The reason for the above problems is probably because everyone uses the author's API (or the author dug it from somewhere) . The disadvantage of public services is that the more people use it, the less easy it is to use...
1.2 Read aloud
If you want a long-term stable and available Microsoft TTS service, it is recommended that you use this app to read aloud
However, the prerequisite for using this APP is that you must meet one of these two conditions: have a foreign edu mailbox or have dual currency cards such as VISA and MasterCard
But in fact there is no way, you can also use the omnipotent Taobao
You can still find a lot of sellers with edu mailboxes, but please note: not all edu mailboxes can subscribe to Azure ⚠️, please consult the seller before buying!
Or you can choose Taobao to find a virtual credit card for Azure payment verification (subsequent use will not be deducted)
In short, if you meet the above conditions, you can register Microsoft Azure service by yourself, and then borrow this app to read aloud, so as to obtain long-term stable and available Microsoft TTS service!
This APP is developed by Coolapk @IIIIII : http://www.coolapk.com/u/453544 (you can open this link with Coolapk, which is the author ID address)
However, since he has hidden his previous dynamics now, I can't find his original dynamics, you can click here to get this APP
The APP now supports automatic updates, and there is no need to worry about the problem of not being able to find the latest version in the future.
If you have a foreign edu mailbox, you can directly search for "Azure for students", and then fill in the school mailbox to register, no credit card is required (provided that your edu mailbox supports Azure subscription)
The reason for asking for foreign edu mailboxes is that the discounts for domestic edu mailboxes have long been wiped out due to the fact that Chinese people have a lot of money from Microsoft...
Below, Ah Xu only introduces how to register and use the dual-currency card
Don't worry: the dual-currency card is only used for verification, and will not actually deduct money, and subsequent use will not deduct money
The first step is to open this address: https://azure.microsoft.com/zh-cn/free/cognitive-services/
Similar to Google Cloud, the first time you sign up for an Azure account, that is, Microsoft Cloud, you will be provided with 12 months of popular services for free, and there is also a service payment limit of 200 US dollars, but the AI voice service we will use later is provided by a free version , so don’t worry about charging after one year
Log in to your Microsoft account at the above address "Click to start using for free", and fill in the relevant account information + check the agreement + fill in the dual-currency credit card information
Wait patiently for credit card verification and register an account, open this address: https://azure.microsoft.com/zh-cn/services/cognitive-services/text-to-speech/
Click "Already using Azure? Try this service for free now"
On the resource creation page, proceed as shown in the figure below, just pay attention to the free option at both the subscription and the Pricing tier
Then click "Review and Create" at the bottom, wait for the review to pass, and continue to click "Create"
After creating, click "Go to Resource"
Click Manage Keys
You will see two keys, just use one
Just fill in the key in the app that reads aloud, and then click "SSML Speech Synthesis Markup Language"
Enter the code below and save it, then you can use it as you like~
Regarding this code, you can adjust these 3 parts as needed:
- name="zh-CN-YunxiNeural", that is, the above picture of Ah Xu uses the AI voice of Yunxi (云希)
- style="assistant", which means that the speaking style is set to assistant (speak to the digital assistant in a warm and relaxed tone)
- styledegree="2", this value can specify the strength of the speaking style, the default value is 1, the minimum value is 0.01, and the maximum value is 2
<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis"
xmlns:mstts="https://www.w3.org/2001/mstts"
xml:lang="zh-CN">
<voice name="zh-CN-YunxiNeural">
<prosody rate="${(rate-100)?c}%" pitch="${(pitch-100)?c}%">
<mstts:express-as style="assistant" styledegree="2">${text}
</mstts:express-as>
</prosody>
</voice>
</speak>
If you want to replace the above configuration by yourself, you need to refer to Microsoft's official documentation: https://docs.microsoft.com/zh-cn/azure/cognitive-services/speech-service/speech-synthesis-markup
What you need to know is that the style of each AI is different . For example, Yunyang currently only has the style of customerservice
And Yunye has calm, cheerful, sad, angry and many other styles to choose from
Finally, it should be noted that the current free quota of Microsoft Azure is 500,000 text-to-speech characters per month
See the latest free quota here: https://azure.microsoft.com/zh-cn/pricing/details/cognitive-services/speech-services/
2 iOS terminal
2.1 Source reading
On iOS, there is no app that can be used after installation like TTS on Android
At present, the only app that Ah Xu knows is "Yuandu" which supports customizing and adding Microsoft AI voice
Specifically, you must register a Microsoft Azure account and apply for a Key before you can use it. Please refer to Section 1.2 of this article for specific application steps, and I will not repeat it here.
But here is a very important step to note: the current version 1.4.8 of iOS Source Reader cannot customize the TTS service area, so if you want to use it at present, you must select East Asia when you first create the voice service. ⚠️
That's the only thing you need to pay attention to, and it won't matter if you follow up.
Open the source reading APP, find Settings » voice management » create cloud voice
The service provider chooses Azure, enter the voice name casually, and then fill in your own Key. For the setting method of the speaker, please refer to section 1.2 of this article.
Then you can directly select and use it when listening to the book
In addition, due to some reasons, this APP was removed from the AppStore in China in January 2022. If you want to download this APP, you need to download it from the AppStore in foreign regions.
For details on how to register a foreign Apple ID, please refer to my original tutorial: https://zhuanlan.zhihu.com/p/49754407
2.2 Cloud phone
Another feasible method is to spend money to use cloud mobile phones as mentioned when I wrote iOS WeChat dual opening .
https://mp.weixin.qq.com/s/fDtU19-QNB3tdDCaPaKOFw
The cloud phone is an Android system , so you only need to go back and refer to the content of the first section of this article.
In addition, if you don't want to use it for listening to books, but want to use Microsoft AI voice for self-media dubbing, please refer to the original text for related software and websites
