Refactored the TTS subsystem to more properly handle message garbling. Added a volume preference for TTS. (#75559)

TTS subsystem refactor.
---------

Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
Co-authored-by: Iamgoofball <iamgoofball@gmail.com>
Co-authored-by: Kyle Spier-Swenson <kyleshome@gmail.com>

code/modules/assembly/voice.dm

@@ -36,9 +36,9 @@
 /obj/item/assembly/voice/Hear(message, atom/movable/speaker, message_language, raw_message, radio_freq, list/spans, list/message_mods = list(), message_range)
 	. = ..()
 	if(message_mods[WHISPER_MODE]) //Too quiet lad
-		return
+		return FALSE
 	if(speaker == src)
-		return
+		return FALSE
 
 	// raw_message can contain multiple spaces between words etc which are not seen in chat due to HTML rendering
 	// this means if the teller records a message with e.g. double spaces or tabs, other people will not be able to trigger the sensor since they don't know how to perform the same combination
@@ -49,6 +49,7 @@
 	else
 		if(check_activation(speaker, raw_message))
 			send_pulse()
+	return TRUE
 
 /obj/item/assembly/voice/proc/record_speech(atom/movable/speaker, raw_message, datum/language/message_language)
 	switch(mode)