I have worked further on xirmat and ermat to use run-length coding. Now, the output of xirmat is compressed and then hard-coded into ermat’s souce code. Ermat then uncompresses the data, one byte at a time.

The compression is highly effective, and reduces the 113 KiB file to a 3 KiB one, which can be hard-coded into a header file.

I still need to solve the problem on how to accurately display the right image on right time.