Uses mimic https://github.com/MycroftAI/mimic1?tab=readme-ov-file to speak text or output wav or mp3 file

text_to_wav.sh

#!/bin/bash

# Uses mimic https://github.com/MycroftAI/mimic1?tab=readme-ov-file to speak
# optionally uses aplay https://linux.die.net/man/1/aplay to speak then delete the output file

# Set default values
VOICE='slt_hts'
WAVFILE="$TEMP/$(date "+%Y%m%d.%H%M%S")-tts.wav"
TEXT=''
TEXTFILE=''
PITCH=110
SPEED=1.25
PLAY=false
CONVERT_TO=""
BITRATE=""

# Function to print usage message and exit
print_usage_and_exit() {
    echo "Usage: $0 -v <voice> -f <textfile> -w <$WAVEFILE> -t <text> -p <pitch> -s <speed> [-l] [-m | -a]"
    echo "If -l is specified, the output file is piped into aplay, then DELETED afterwards"
    echo "Example to speak the date/time to the current user: $0 -t "\"Attention \$USER, \$\(date +\"the date is %A, %B %d, %Y and the time is %I:%M %p.\"\)"\" -l 1"
    echo "Example to read output from fortune and output mp3 file: $0 -t \"\`fortune -a\`\" -w $HOME/Desktop/fortune.wav -m 64"
    echo "Example to read a file and speak into a wav file: $0 -f text.txt -w $HOME/Desktop/text.wav"
    echo "Optional conversion options:"
    echo "  -m <bitrate>  Convert the WAV file to MP3"
    exit 1
}

# Parse command-line options using getopts for better syntax and robustness
while getopts ':v:f:w:t:p:s:l:m:a' opt; do
    case ${opt} in
        v )
            VOICE=$OPTARG
            ;;
        f )
            TEXTFILE=$OPTARG
            ;;
        w )
            WAVFILE=$OPTARG
            ;;
        t )
            TEXT=$OPTARG
            ;;
        p )
            PITCH=$OPTARG
            ;;
        s )
            SPEED=$OPTARG
            ;;
        l )
            PLAY=true
            ;;
        m )
            CONVERT_TO="mp3"
            BITRATE=$OPTARG 
            ;;
        \? )
            print_usage_and_exit
            ;;
        : )
            echo "Invalid option: -$OPTARG requires an argument" 1>&2
            print_usage_and_exit
            ;;
    esac
done
shift $((OPTIND -1))

# Check if required arguments are provided
if [[ -z "$TEXT" && -z "$TEXTFILE" ]]; then
    echo "Error: Either -t or -f is a required argument." >&2
    print_usage_and_exit
fi 

# If TEXT is not provided but TEXTFILE is, read the text from the file
if [[ -z "$TEXT" ]]; then
    if [[ -f "$TEXTFILE" ]]; then
        TEXT=$(<"$TEXTFILE")
    else
        echo "Error: The specified file $TEXTFILE does not exist." >&2
        print_usage_and_exit
    fi
fi

# Validate numeric input for PITCH and SPEED
if ! [[ "$PITCH" =~ ^[0-9]+$ ]] || ! [[ "$SPEED" =~ ^[0-9]+([.][0-9]+)?$ ]]; then
    echo "Error: PITCH and SPEED must be valid numbers." >&2
    print_usage_and_exit
fi

# Run the mimic command with sanitized input
mimic --setf int_f0_target_mean=$PITCH --setf duration_stretch=$SPEED -voice $VOICE -t "$TEXT" -o $WAVFILE

echo "Text to speech conversion completed. Output file: $WAVFILE"

# If play option is specified, play the audio and delete the file
if [ "$PLAY" = true ]; then
    if command -v aplay &> /dev/null; then
        aplay "$WAVFILE"
        rm "$WAVFILE"
        echo "Audio played and output file deleted."
    else
        echo "Error: 'aplay' is not installed. Please install it to play the audio." >&2
    fi
fi

# Convert the WAV file if conversion option is specified
if [[ -n "$CONVERT_TO" ]]; then
    if ! command -v ffmpeg &> /dev/null; then
        echo "Error: 'ffmpeg' is not installed. Please install it to convert the audio." >&2
        exit 1
    fi

    OUTPUT_FILE="${WAVFILE%.*}.${CONVERT_TO}"
    ffmpeg -i $WAVFILE -b:a $BITRATE $OUTPUT_FILE
    rm $WAVFILE
    echo "File converted and original WAV file deleted. Output file: $OUTPUT_FILE"
fi