Command Line TTS Tool

This is a command line tool that accepts a text input from stdin or a file and converts it into audio using Google Cloud TTS or Azure TTS.

This is a fork of Google Squawk by ruhnet. I just added support for Azure TTS.

Build

Azure TTS support requires Microsoft Cognitive Speech Services SDK to be installed.

Follow the instructions for installing the SDK on your platform.

export SPEECHSDK_ROOT="$PWD/speech-sdk"
mkdir -p "$SPEECHSDK_ROOT"

macOS

wget -O MicrosoftCognitiveServicesSpeech.xcframework.zip https://aka.ms/csspeech/macosbinary
unzip MicrosoftCognitiveServicesSpeech.xcframework.zip -d "$SPEECHSDK_ROOT"
export CGO_CFLAGS="-I$SPEECHSDK_ROOT/MicrosoftCognitiveServicesSpeech.xcframework/macos-arm64_x86_64/MicrosoftCognitiveServicesSpeech.framework/Headers"
export CGO_LDFLAGS="-Wl,-rpath,$SPEECHSDK_ROOT/MicrosoftCognitiveServicesSpeech.xcframework/macos-arm64_x86_64 -F$SPEECHSDK_ROOT/MicrosoftCognitiveServicesSpeech.xcframework/macos-arm64_x86_64 -framework MicrosoftCognitiveServicesSpeech"

Linux

wget -O SpeechSDK-Linux.tar.gz https://aka.ms/csspeech/linuxbinary
tar --strip 1 -xzf SpeechSDK-Linux.tar.gz -C "$SPEECHSDK_ROOT"
export CGO_CFLAGS="-I$SPEECHSDK_ROOT/include/c_api"
export CGO_LDFLAGS="-L$SPEECHSDK_ROOT/lib/x64 -lMicrosoft.CognitiveServices.Speech.core"

Now we are ready to build the application.

go build

Authentication

Google Cloud TTS

To use the program with Google Cloud TTS, specify the path to your Google cloud account credentials with the environment variable GOOGLE_APPLICATION_CREDENTIALS.

export GOOGLE_APPLICATION_CREDENTIALS=/path/to/serviceaccount/credentials_file.json

Azure Cloud TTS

To use Azure Cognitive service, specify your Azure API Key and region using environment varialbles AZURE_API_KEY and AZURE_REGION.

export AZURE_API_KEY="..."
export AZURE_REGION="centralindia"

Usage

By default, input is supplied via standard input, but can also be specified from a file with the -i option. Both plain text and SSML input are supported. Specify -ssml to tell the program to expect SSML input from your text source.

Output is to a file, but you can set the filename to - to send to stdout. This can be useful if you want to convert the file on the fly to a format that the program doesn't support, with sox or ffmpeg or the like.

Specify the path to the Speech SDK library using

Run tts -h to see the help:

Usage of ./tts:
  -db float
    	Volume gain in dB. [-96 to 16]
  -e string
        TTS Engine. ['google', 'azure'] (default "google")
  -f string
    	Audio format selection. MP3 is 32k [mp3,pcm] (default "mp3")
  -g string
    	Gender selection. [m,f,n] 'n' means neutral/don't care. (default "m")
  -i string
    	Input file path. Defaults to stdin. (default "-")
  -l string
    	Language selection. 'en-US', 'en-GB', 'en-AU', 'en-IN',
    	'el-GR', 'ru-RU', etc. (default "en-US")
  -o string
    	Output file path. Use '-' for stdout. (default "./tts.mp3")
  -p float
    	Pitch. E.g. '0.0' is normal. '20.0' is highest,
    	'-20.0' is lowest. (default 1)
  -r int
    	Samplerate in Hz. [8000,11025,16000,22050,24000,32000,44100,48000] (default 24000)
  -s float
    	Speed. E.g. '1.0' is normal. '2.0' is double
    	speed, '0.25' is quarter speed, etc. (default 1)
  -ssml
    	Input is SSML format, rather than plain text.
  -v string
    	Voice. If specified, this overrides language & gender. (default "")

Troubleshooting

If you are getting the error

error while loading shared libraries: libMicrosoft.CognitiveServices.Speech.core.so: cannot open shared object file: No such file or directory

either copy the shared libraries to /usr/lib

cp $SPEECHSDK_ROOT/lib/x64/*.so /usr/lib/

or add the library path to the LD_LIBRARY_PATH environment variable

export LD_LIBRARY_PATH="$SPEECHSDK_ROOT/lib/x64:$LD_LIBRARY_PATH"

JoyceBabu/tts