He physically spliced together Axl making various sounds to create the words. It really is Axl's voice. He did the same thing when he reconstructed Silkworms out of the shitty live performance. He used high quality audio of Axl speaking to splice different sounds together and create studio quality vocals.
It's a speaker issue as far as I know. It sounds very good on some systems and not so good on my phone. My phone typically plays things very clearly, though.
I've always thought that too, but I accepted that he must have just been on a good run because not only were the vocals on point, his whole stage presence and attitude were back. So who knows