Chunking algorithms are often used by storage solutions in order to factorize and deduplicate data. Such algorithms make the assumption that the consecutive versions of a file share a lot of similarities. Unfortunately, file formats often use compression algorithms and minor changes have the potential to completely reorganize the internal layout of a file. In consequence, chunking algorithms become less efficient in factorizing data. In this paper, we evaluate content-defined chunking with file formats that use data compression. We show how content-defined chunking algorithms can take the file format into account. Finally, we demonstrate that adding file format knowledge to a popular chunking algorithm significantly improves its performance.