Comparisons of theoretical limits for source coding with practical compression algorithms