How to normalize the bam files from different samples

[ykdang]

I am new to bioinformatics. So recently I got several RNA seq data. I was able to blast them with tophat, generate bam files and load them onto IGV. However, since those data is not normalized. I could not compare them. I knew that by using cufflinks and cummeRbond I could compare RPKM of different samples. But we are eager to see it at genome wide. I heard that people could convert bam into Bed file, which can directly load as graph in IGV and UCSC. But how could I normalize these bam files so that they are comparable? \

Hopefully someone can be answer this naive question, Thanks!