Files
audiogame-manager/speech/set-voice.sh

443 lines
15 KiB
Bash
Executable File

#!/usr/bin/env bash
# Set Speech
# Configure SAPI voices and speech rate for wine32 bottle
# License header
# The contents of this file are subject to the Common Public Attribution
# License Version 1.0 (the "License"); you may not use this file except in
# compliance with the License. You may obtain a copy of the License at
# https://opensource.org/licenses/CPAL-1.0. The License is based on the Mozilla Public License Version
# 1.1 but Sections 14 and 15 have been added to cover use of software over a
# computer network and provide for limited attribution for the Original
# Developer. In addition, Exhibit A has been modified to be consistent with
# Exhibit B.
#
# Software distributed under the License is distributed on an "AS IS" basis,
# WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
# for the specific language governing rights and limitations under the
# License.
#
# The Original Code is audiogame manager.
#
# The Original Developer is not the Initial Developer and is . If
# left blank, the Original Developer is the Initial Developer.
#
# The Initial Developer of the Original Code is Billy "Storm Dragon" Wolfe. All portions of
# the code written by Billy Wolfe are Copyright (c) 2020. All Rights
# Reserved.
#
# Contributor Michael Taboada.
#
# Attribution Copyright Notice: Audiogame manager copyright 2020 Storm Dragon. All rights reserved.
#
# Attribution Phrase (not exceeding 10 words): A Stormux project
#
# Attribution URL: https://stormgames.wolfe.casa
#
# Graphic Image as provided in the Covered Code, if any.
#
# Display of Attribution Information is required in Larger
# Works which are defined in the CPAL as a work which combines Covered Code
# or portions thereof with code not governed by the terms of the CPAL.
# Detect dialog interface type BEFORE potentially setting DISPLAY
if [[ -z "$DISPLAY" ]]; then
dialogType="dialog"
else
dialogType="yad"
fi
# Source dialog interface wrapper
source "${0%/*}/../.includes/dialog-interface.sh"
# Settings to improve accessibility of dialog
export DIALOGOPTS='--insecure --no-lines --visit-items'
# Turn off debug messages
export WINEDEBUG="-all"
# Set DISPLAY if needed
if [[ -z "$DISPLAY" ]]; then
export DISPLAY=:0
fi
# Set wine prefix to wine32 - this is the only supported bottle for SAPI
export WINEPREFIX="$HOME/.local/wine32"
export bottle="$HOME/.local/wine32"
help() {
echo "${0##*/}"
echo "Released under the terms of the Common Public Attribution License Version 1.0"
echo -e "This is a Stormux project: https://stormux.org\n"
echo -e "Usage:\n"
echo "With no arguments, open the voice configuration menu."
echo "-h: Show this help screen."
echo "-r <rate>: Set voice rate (0-10, where 10 is fastest)."
echo "-v <voice>: Set voice by name."
exit 0
}
msgbox() {
agm_msgbox "Set Speech" "Set Speech" "$*"
}
infobox() {
local timeout=3
agm_infobox "Set Speech" "Set Speech" "$*"
read -r -n1 -t $timeout
read -r -t 0.01
}
yesno() {
if agm_yesno "Set Speech" "Set Speech" "$*"; then
echo "Yes"
else
echo "No"
fi
}
voice_menu() {
declare -a menuList=()
for i in "${@}"; do
menuList+=("$i" "$i")
done
agm_menu "Set Speech" "Set Speech" "Please select a voice:" "${menuList[@]}"
}
rate_menu() {
declare -a rateList=()
for i in {0..10}; do
local desc="Rate $i"
if [[ $i -eq 0 ]]; then
desc="$desc (Slowest)"
elif [[ $i -eq 5 ]]; then
desc="$desc (Default)"
elif [[ $i -eq 10 ]]; then
desc="$desc (Fastest)"
fi
rateList+=("$i" "$desc")
done
agm_menu "Set Speech" "Set Speech" "Please select speech rate:" "${rateList[@]}"
}
get_voices() {
# Get list of available voices using wine reg query
declare -a voices
declare -a voiceKeys
# First, get Microsoft SAPI voices
local allOutput
allOutput=$(WINEPREFIX="$WINEPREFIX" "$wine" reg query "HKLM\\SOFTWARE\\Microsoft\\Speech\\Voices\\Tokens" /s 2>/dev/null)
# Parse output to find voice keys and names
local currentKey=""
local currentName=""
local inMainVoiceToken=false
while IFS= read -r line; do
# Clean up the line first
line=$(echo "$line" | tr -d '\r\n' | sed 's/[[:space:]]*$//')
# Check if this is a main voice token registry key line (exactly one level deep)
if [[ "$line" =~ ^HKEY_LOCAL_MACHINE\\SOFTWARE\\Microsoft\\Speech\\Voices\\Tokens\\([^\\]+)$ ]]; then
# Save previous voice if we found a name and had collected both key and name
if [[ -n "$currentKey" && -n "$currentName" ]]; then
voices+=("$currentName")
voiceKeys+=("$currentKey")
fi
# Start new voice - clean up any carriage returns or whitespace
currentKey="${BASH_REMATCH[1]}"
currentName=""
inMainVoiceToken=true
# Check if this is a subkey (contains additional parts after token name)
elif [[ "$line" =~ ^HKEY_LOCAL_MACHINE\\SOFTWARE\\Microsoft\\Speech\\Voices\\Tokens\\[^\\]+\\ ]]; then
# This is a subkey, not a main voice token
inMainVoiceToken=false
# Check if this is a default value line with voice name and we're in main token
elif [[ "$line" =~ ^[[:space:]]*\(Default\)[[:space:]]+REG_SZ[[:space:]]+(.+)$ ]] && [[ "$inMainVoiceToken" == true ]]; then
# Only capture voice name if we haven't already got one for this token
if [[ -z "$currentName" ]]; then
currentName="${BASH_REMATCH[1]}"
# Clean up the voice name as well
currentName=$(echo "$currentName" | tr -d '\r\n' | sed 's/[[:space:]]*$//')
fi
fi
done <<< "$allOutput"
# Don't forget the last voice
if [[ -n "$currentKey" && -n "$currentName" ]]; then
voices+=("$currentName")
voiceKeys+=("$currentKey")
fi
# Now search for L&H TTS voices
local lhOutput
lhOutput=$(WINEPREFIX="$WINEPREFIX" "$wine" reg query "HKLM\\SOFTWARE\\L&H\\TTS\\V6.0\\Voice" /s 2>/dev/null)
if [[ -n "$lhOutput" ]]; then
while IFS= read -r line; do
# Clean up the line first
line=$(echo "$line" | tr -d '\r\n' | sed 's/[[:space:]]*$//')
# Look for voice name entries like "Carol REG_SZ {227A0E40-A92A-11d1-B17B-0020AFED142E}"
if [[ "$line" =~ ^[[:space:]]*([A-Za-z]+)[[:space:]]+REG_SZ[[:space:]]+\{([^}]+)\}$ ]]; then
local lhVoiceName="${BASH_REMATCH[1]}"
local lhVoiceGuid="${BASH_REMATCH[2]}"
# Add L&H prefix to distinguish from SAPI voices (escape ampersand for YAD)
voices+=("L&amp;H $lhVoiceName")
voiceKeys+=("LH_$lhVoiceGuid")
fi
done <<< "$lhOutput"
fi
if [[ ${#voices[@]} -eq 0 ]]; then
return 1
fi
# Export arrays for use in other functions
export voiceList=("${voices[@]}")
export voiceKeyList=("${voiceKeys[@]}")
return 0
}
get_current_voice() {
# Get current default voice
local currentVoice
currentVoice=$(WINEPREFIX="$WINEPREFIX" "$wine" reg query "HKCU\\SOFTWARE\\Microsoft\\Speech\\Voices" /v "DefaultTokenId" 2>/dev/null | grep "REG_SZ" | sed 's/.*REG_SZ[[:space:]]*//')
if [[ -n "$currentVoice" ]]; then
# Extract just the voice key from the full registry path
# The path format is: HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech\Voices\Tokens\MSMike
if [[ "$currentVoice" =~ Tokens\\([^\\]+)$ ]]; then
local voiceKey="${BASH_REMATCH[1]}"
# Clean up any whitespace or special characters
voiceKey=$(echo "$voiceKey" | tr -d '\r\n' | sed 's/[[:space:]]*$//')
echo "$voiceKey"
fi
fi
}
get_current_rate() {
# Get current speech rate
local currentRate
currentRate=$(WINEPREFIX="$WINEPREFIX" "$wine" reg query "HKCU\\SOFTWARE\\Microsoft\\Speech\\Voices" /v "DefaultTTSRate" 2>/dev/null | grep "REG_DWORD" | sed 's/.*REG_DWORD[[:space:]]*//' | sed 's/^0x//')
if [[ -n "$currentRate" && "$currentRate" =~ ^[0-9a-fA-F]+$ ]]; then
# Convert hex to decimal
echo $((16#$currentRate))
else
echo "5" # Default rate
fi
}
set_voice() {
local voiceName="$1"
local voiceKey=""
# Find the voice key for the given voice name
local i
for i in "${!voiceList[@]}"; do
if [[ "${voiceList[i]}" == "$voiceName" ]]; then
voiceKey="${voiceKeyList[i]}"
break
fi
done
if [[ -z "$voiceKey" ]]; then
msgbox "Error: Voice '$voiceName' not found."
return 1
fi
# Check if this is an L&H voice (cannot be set as system default)
if [[ "$voiceKey" =~ ^LH_ ]]; then
msgbox "L&H voices cannot be set as the system default SAPI voice. They use a proprietary API and must be selected by individual applications that support L&H TTS."
return 1
fi
# Kill wine server to ensure registry changes take effect
WINEPREFIX="$WINEPREFIX" "$wineserver" -k 2>/dev/null
# Set the default voice using wine reg
local fullVoicePath="HKEY_LOCAL_MACHINE\\SOFTWARE\\Microsoft\\Speech\\Voices\\Tokens\\$voiceKey"
if WINEPREFIX="$WINEPREFIX" "$wine" reg add "HKCU\\SOFTWARE\\Microsoft\\Speech\\Voices" /v "DefaultTokenId" /t REG_SZ /d "$fullVoicePath" /f 2>/dev/null; then
infobox "Voice set to: $voiceName"
return 0
else
msgbox "Error: Failed to set voice."
return 1
fi
}
set_rate() {
local rate="$1"
# Validate rate (0-10)
if ! [[ "$rate" =~ ^[0-9]$|^10$ ]]; then
msgbox "Error: Rate must be between 0 and 10."
return 1
fi
# Kill wine server to ensure registry changes take effect
WINEPREFIX="$WINEPREFIX" "$wineserver" -k 2>/dev/null
# Set the speech rate using wine reg (convert to hex)
local hexRate
hexRate=$(printf "%x" "$rate")
if WINEPREFIX="$WINEPREFIX" "$wine" reg add "HKCU\\SOFTWARE\\Microsoft\\Speech\\Voices" /v "DefaultTTSRate" /t REG_DWORD /d "0x$hexRate" /f 2>/dev/null; then
infobox "Speech rate set to: $rate"
return 0
else
msgbox "Error: Failed to set speech rate."
return 1
fi
}
test_voice() {
local voiceName="$1"
# Set the voice temporarily for testing
set_voice "$voiceName" || return 1
# Create test script
mkdir -p "${WINEPREFIX}/drive_c/windows/temp"
cat << "EOF" > "${WINEPREFIX}/drive_c/windows/temp/speak.vbs"
dim speechObject
set speechObject = createObject("sapi.spvoice")
speechObject.speak "This is a test of your chosen voice. It contains multiple sentences and punctuation, and is designed to give a full representation of this voice's qualities."
EOF
# Test the voice
WINEPREFIX="$WINEPREFIX" "$wine" cscript "c:\\windows\\temp\\speak.vbs" 2>/dev/null
}
initialize_sapi() {
# Initialize SAPI if not already done
mkdir -p "${WINEPREFIX}/drive_c/windows/temp"
cat << "EOF" > "${WINEPREFIX}/drive_c/windows/temp/init.vbs"
dim speechObject
set speechObject = createObject("sapi.spvoice")
speechObject.speak ""
EOF
WINEPREFIX="$WINEPREFIX" "$wine" cscript "c:\\windows\\temp\\init.vbs" 2>/dev/null
}
# Handle command line arguments
while getopts "hr:v:" option; do
case "$option" in
h) help ;;
r)
if ! [[ "$OPTARG" =~ ^[0-9]$|^10$ ]]; then
echo "Error: Rate must be between 0 and 10."
exit 1
fi
requestedRate="$OPTARG"
;;
v) requestedVoice="$OPTARG" ;;
*) help ;;
esac
done
# Check if wine32 bottle exists
if [[ ! -d "$WINEPREFIX" ]]; then
msgbox "Error: Wine32 bottle not found at $WINEPREFIX. Please create it first."
exit 1
fi
# Get wine version if available
if [[ -r "${WINEPREFIX}/agm.conf" ]]; then
source "${WINEPREFIX}/agm.conf"
export WINE
export WINESERVER
fi
wine="${WINE:-$HOME/.local/share/audiogame-manager/wine32/bin/wine}"
wineserver="${WINESERVER:-$HOME/.local/share/audiogame-manager/wine32/bin/wineserver}"
# Check if wine executable exists
if [[ ! -x "$wine" ]]; then
msgbox "Error: Wine executable not found at $wine"
exit 1
fi
# Initialize SAPI
initialize_sapi
# Get available voices
if ! get_voices; then
msgbox "No SAPI voices found in wine32 bottle. Please install SAPI voices first."
exit 1
fi
# Handle command line options
if [[ -n "$requestedRate" ]]; then
set_rate "$requestedRate"
fi
if [[ -n "$requestedVoice" ]]; then
set_voice "$requestedVoice"
exit 0
fi
# Show interactive menu if no voice specified
currentVoice=$(get_current_voice)
currentRate=$(get_current_rate)
# Verify arrays are properly set
if [[ ${#voiceList[@]} -eq 0 ]]; then
msgbox "Error: No voices loaded."
exit 1
fi
while true; do
# Build menu with current settings
declare -a menuOptions=()
# Find current voice name for display
currentVoiceName="None"
if [[ -n "$currentVoice" ]]; then
for i in "${!voiceKeyList[@]}"; do
# Clean up the voice key for comparison
cleanKey=$(echo "${voiceKeyList[i]}" | tr -d '\r\n' | sed 's/[[:space:]]*$//')
if [[ "$cleanKey" == "$currentVoice" ]]; then
currentVoiceName="${voiceList[i]}"
break
fi
done
fi
menuOptions+=("voice" "Select Voice (Current: $currentVoiceName)")
menuOptions+=("rate" "Set Rate (Current: $currentRate)")
menuOptions+=("test" "Test Current Voice")
menuOptions+=("quit" "Exit")
choice=$(agm_menu "Set Speech" "SAPI Voice Configuration" "Choose an option:" "${menuOptions[@]}")
case "$choice" in
"voice")
if selectedVoice=$(voice_menu "${voiceList[@]}"); then
if [[ -n "$selectedVoice" ]] && set_voice "$selectedVoice"; then
currentVoice=$(get_current_voice)
fi
fi
;;
"rate")
if newRate=$(rate_menu); then
if [[ -n "$newRate" ]] && set_rate "$newRate"; then
currentRate=$(get_current_rate)
fi
fi
;;
"test")
if [[ -n "$currentVoice" && "$currentVoiceName" != "None" ]]; then
infobox "Testing voice: $currentVoiceName"
test_voice "$currentVoiceName"
else
msgbox "No voice selected. Please select a voice first."
fi
;;
"quit"|"")
break
;;
esac
done
exit 0