Skip to content

Latest commit

 

History

History
106 lines (93 loc) · 4.86 KB

README.md

File metadata and controls

106 lines (93 loc) · 4.86 KB
Raw

Text-to-Speech Radio

This python script converts text to speech using the Google Cloud engine. Furthermore, radio effects (high, low pass filters and white noise) can be added to the speech output. The input text is read from an Excel or csv file.

Installation

You need to have python and pip installed. Furthermore, the script requires the following packages, which can be easily installed via pip:

  • pydub
  • google-cloud
  • pandas

Data Table

The input data is provided as a table in either xlsx or `csv format:

image

The xlsx or csv files need to have the following columns:

  • Text: The text that is converted to speech. This can contain SSML syntax.
  • File Name: The output file name. This will be saved as in .ogg format.
  • Subtitle: An optional subtitle if different from the Text column.
  • Voice Name: The Google voice if other than the default voice en-US-Standard-A.
  • Highpass: The high pass filter frequency in Hz. Default 4000 Hz.
  • Lowpass: The low pass filter frequency in Hz. Default 3000 Hz.
  • Nfilter: Number of times the high and or low pass filters are applied. Default 3.
  • Volume Boost: The volume boost in dB applied to the final normalized output. Default 0 dB.
  • Noise Boost: The white noise boost in [dB]. Default None. If set, white noise is added. Try a negative value of -25 dB first.
  • Emphasis: Add emphasis to the voice, c.f. https://cloud.google.com/text-to-speech/docs/ssml#emphasis.
  • Rate: Adjust speech rate, c.f. https://cloud.google.com/text-to-speech/docs/ssml#prosody
  • Pitch: Adjust the pitch of the voice, c.f. https://cloud.google.com/text-to-speech/docs/ssml#prosody
  • Click In: Add a radio click at the beginning of the sound file.
  • Click Out: Add a radio click at the end of the sound file.

Usage

Change to the directory, where the ttsr.py file is located and type

python ttsr.py --credentials X:\<Path to JSON file>\google-credentials.json

where the parameter --credentials points to your Google cloud credentials file. Alternatively, you can set the environment variable GOOGLE_APPLICATION_CREDENTIALS to point to the credentials file.

This will scan the current directory for all xlsx and csv files and start the conversion process. For each input file, a directory named as the input file is created that contains the sound .ogg files. Additionally, a csv file with the parameters is created that also contains the duration of the sound file as additional column. This csv file can be used as input for various MOOSE classes (e.g. ATIS, RANGE, AIRBOSS) as described below.

Command Line Parameters

The following command line parameters can be used:

  • --filetype: Specify whether you want Excel xlsx or comma-separated value csv files as input.
  • --inputdir: Specify the path of the directory where the input files are located.
  • --inputfile: Specify the path to a single input file (if you do not want to convert all files in a directory).
  • --credentials: Path to your Google credentials file.

Specify File Type

Process only csv files in the current directory

python ttsr.py --filetype csv

Specify Input Directory

Process all xlsx and csv files in a certain directory

python ttsr.py --inputdir <Path To Input Files>

Specify Input File

Process only one specific input file

python ttsr.py --inputfile <Path to Input File>

Specify Voice

Use a specific voice for all input files (overrules setting in input files)

python ttsr.py --voice en-GB-Wavenet-F

Specify Noise

Use a white noise for all input files (overrules setting in input files)

python ttsr.py --noise -25

The white noise volume is reduced by -25 dB.

MOOSE Classes

The generated sound files can of course be used in DCS, when included in a mission miz file. Add the generated sound files to your mission (open the miz with 7-zip and drag & drop the folder). The output of the csv files that contain the duration of the sound file can be used in some MOOSE classes.

ATIS

A simple ATIS script can look like:

local atis=ATIS:New("Batumi", 305)  
atis:SetSoundfilesPath("ATIS-en-GB-Wavenet-F/", "Caucasus-en-GB-Wavenet-F/", "NATO Alphabet-en-GB-Wavenet-F/")
atis:SetSoundfilesInfo("D:/DCS/Scripts/My_Missions/parameters-ATIS-en-GB-Wavenet-F.csv")
-- More custom input...
atis:Start()

The parameters-ATIS-en-GB-Wavenet-F.csv is the output file generated by the script and needs to be located on your local hard drive (i.e. not in the miz file).

RANGE

local range=RANGE:New("Test")
range:SetSoundfilesPath("Range/")
range:SetSoundfilesInfo("D:/DCS/Scripts/My_Missions/_Classes/MSRS/parameters-Range.csv")
-- More custom input...  
range:Start()

AIRBOSS

WIP, not supported yet.

Ackknowledgements

This script is based on https://github.com/adamclmns/DCS_RadioComm_VoiceLineGenerator