> For the complete documentation index, see [llms.txt](https://book.ferro.pro/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://book.ferro.pro/multimedia/audio.md).

# audio

* [Speech recognition](#speech-recognition)
  * [ML - OpenAI](#ml---openai)
  * [Azure](#azure)
* [Text To Speech](#text-to-speech)
  * [Microsoft](#microsoft)
  * [Bing](#bing)
  * [Ali](#ali)
  * [Google](#google)
  * [AWS](#aws)
  * [IBM](#ibm)
  * [Xunfei](#xunfei)
  * [Tencent](#tencent)
  * [Baidu](#baidu)
  * [Offline](#offline)
* [Software](#software)
* [Music](#music)
  * [Apps](#apps)
  * [Basic](#basic)

## Speech recognition

<https://github.com/Uberi/speech\\_recognition>

### ML - OpenAI

<https://github.com/openai/whisper>

### Azure

<https://azure.microsoft.com/en-us/pricing/details/cognitive-services/speech-services/>

```
free per month: 5 audio hours / 0.5 million characters /  10,000 transactions
```

## Text To Speech

My rate: Microsoft Neural > Google > Microsoft Zira Desktop \~ AWS Justin >> IBM/Baidu >> espeak

### Microsoft

Demo: <https://speech.microsoft.com/customvoice\\>
download: <https://greasyfork.org/en/scripts/444347-azure-speech-download>

```
git clone https://github.com/Azure-Samples/cognitive-services-speech-sdk.git
=> quickstart/python/text-to-speech/quickstart.ipynb
```

<https://azure.microsoft.com/en-us/pricing/details/cognitive-services/speech-services/>

```
STT - 5 audio hours free per month
TTS Standard - 5M characters free per month
TTS Neural   - 0.5M characters free per month
```

Portal: <https://portal.azure.com/#blade/HubsExtension/BrowseResource/resourceType/Microsoft.CognitiveServices%2Faccounts>

SDK/REST support: <https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/overview>

<https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/rest-text-to-speech>

Python SDK: <https://github.com/Azure-Samples/cognitive-services-speech-sdk/blob/master/quickstart/python/text-to-speech/quickstart.ipynb\\>
Console: <https://github.com/Azure-Samples/cognitive-services-speech-sdk/tree/master/samples/python/console\\>
Flask app: <https://github.com/MicrosoftTranslator/Text-Translation-API-V3-Flask-App-Tutorial>

### Bing

<https://azure.microsoft.com/en-us/services/cognitive-services/speech/\\>
<https://azure.microsoft.com/en-us/pricing/details/cognitive-services/speech-api/>

| TIER                    | FEATURES                         | UNIT                              | PRICE                     |
| ----------------------- | -------------------------------- | --------------------------------- | ------------------------- |
| Bing Speech API—free    | Transactions                     | 5,000 transactions free per month |                           |
| Bing Speech-to-Text API | Utterances up to 15 seconds long | Transactions                      | $4 per 1,000 transactions |
| Bing Text-to-Speech API | Transactions                     | $4 per 1,000 transactions         |                           |

<https://github.com/westparkcom/Python-Bing-TTS>

### Ali

Demo: <https://ai.aliyun.com/nls/tts\\>
Non-free: <https://help.aliyun.com/document\\_detail/84881.html#h1-u8BA1u8D39u65B9u5F0Fu548Cu62A5u4EF73>

### Google

<https://play.google.com/store/apps/details?id=com.google.android.tts>

<https://github.com/pndurette/gTTS>

```
gtts-cli -f hello.txt -l 'cs' -o hello.mp3

from gtts import gTTS
tts = gTTS(text='Hello', lang='en', slow=True)
tts.save("hello.mp3")
```

### AWS

<https://console.aws.amazon.com/polly\\>
<https://aws.amazon.com/polly/pricing/>

free: 5 million characters/m/first 12m\
$4.00 per 1 million characters

### IBM

<https://text-to-speech-demo.ng.bluemix.net\\>
<https://console.bluemix.net/catalog/services/text-to-speech>

Free: 10,000 characters per month. deleted after 30 days of inactivity.\
Paid: $0.02 USD/THOUSAND CHAR

<https://github.com/watson-developer-cloud/python-sdk>

### Xunfei

<http://www.xfyun.cn/services/offline\\_tts> (99+ RMB/m)

### Tencent

<https://cloud.tencent.com/product/aai> (Price unknown)

### Baidu

<https://developer.baidu.com/vcast>

<http://yuyin.baidu.com/#try\\>
<http://yuyin.baidu.com/docs/tts/136>

### Offline

<https://en.wikipedia.org/wiki/Comparison\\_of\\_speech\\_synthesizers>

<https://pyttsx3.readthedocs.io/en/latest/> (no file saving?)

```
SAPI5 on Windows XP and Windows Vista and Windows 8,8.1 , 10
NSSpeechSynthesizer on Mac OS X 10.5 (Leopard) and 10.6 (Snow Leopard)
espeak on Ubuntu Desktop Edition 8.10 (Intrepid), 9.04 (Jaunty), and 9.10 (Karmic)
```

<http://espeak.sourceforge.net/commands.html>

```
espeak "unnatural man voice"  -w out.wav  # apt install -y libespeak1
setup_espeak.exe #  SAPI5 version
```

Swift: <https://youtu.be/bl6hEBXuv5Y?t=1057>

## Software

<https://en.wikipedia.org/wiki/Comparison\\_of\\_digital\\_audio\\_editors\\>
<https://en.wikipedia.org/wiki/Comparison\\_of\\_free\\_software\\_for\\_audio>

Audacity: <https://github.com/audacity/audacity/releases>

Input/Output equalizer: <https://sourceforge.net/projects/equalizerapo/>

AI Real-Time Voice Cloning: <https://github.com/CorentinJ/Real-Time-Voice-Cloning>

## Music

### Apps

* Coding: <https://sonic-pi.net/#examples>
* Singing Voice Conversion: <https://github.com/svc-develop-team/so-vits-svc>

### Basic

```
♭ flat:  lower in pitch by one semitone (half step)
♯ sharp: higher in pitch by one semitone (half step)
```

Equal temperament / 十二平均律: <https://baike.baidu.com/item/%E5%8D%81%E4%BA%8C%E5%B9%B3%E5%9D%87%E5%BE%8B/592297\\>
![](https://bkimg.cdn.bcebos.com/pic/0b46f21fbe096b6300e5207b09338744ebf8ac19)