I am not sure how /dev/sdb4 came about, or /dev/md1. Did you mean /dev/sdb1 as in your original post? I will assume such...
Look at the contents of /proc/mdstat. The file will show every /dev/md* array, and the partitions that make them up. On your system, I would expect to see something showing /dev/md0 as being active raid1, and composed of sdb1 and sda1. If you don't see any md* devices listed, or there is no /proc/mdstat file, maybe your system hasn't loaded the appropriate module.
You will not be able to directly mount a partition that was once part of an array. You must remove the bad partition from the array, then mount the array again. You can't remove the good partition and expect to mount it separately from the array. When you get the error unknown filesytem type "linux raid member", that actually is what the partition manager is saying about the partition signature. This partition is only accessible through the software raid manager.
If I go back to your original post, and my original answer, once you break the bad /dev/sda1 partition away from the array, you should still be able to mount /dev/md0 with only the one partition in it. It will be mounted in degraded mode, but it should be there. Fortunately, you can do that since you used mirroring.
I have never used the gui disk utility tool, so I don't know how well it runs. Most gui tools though don't fully implement all disk tools: that's why I have only used the command line tools.
For more details, please look at the following page: http://tldp.org/HOWTO/Software-RAID-HOWTO.html