I'm not even sure whether this is a historian job or a philologist's (or both), but I've always wondered how on earth can someone figure out a language of which there's no speaker or at least a registry of its workings in a know language.
Any info on this process? I guess they try to spot some common used words or glyph (like a King's name) and work from there, but even that seems a very hard do grasp concept to me.
Also, when these people decipher these languages do they also learn how they sound out loud or they just can understand their meaning?
It's a job for a historical linguist, but deciphering ancient languages isn't that easy, because you always have some leads - living languages. Ancient languages are usually related to living languages. Another way how to decipher them are bilingual texts - the famous Rosetta stone, which was written in hieroglyphs and Ancient Greek.
Of course, deciphering the hieroglyphs was a difficult cryptographic work for Champolion, but he eventually succeeded.
Another example is the cuneiform script - it was desiphered through bilingual texts from ancient Persia. Then it was used to decipher Akkadian, which is an ancient Semitic language, and there are many livig Semitic languages. And then, Sumero-Akkadan dictionaries were found. Sumrian is an isolated language without any living relatives, but because of the dictionaries it was possible to reconstruct it.
As for the phonology, for Sumerian we don't precisely know how it sounded, but we have a good idea. With Akkadian it's somewhat better, because it's related to other Semitic languages, so we can reconstruct roots and find out how they subsequently developed, because linguistic changes have their rules.