In RAID 5, users combine up to three or more hard drives to make up a large volume using all of the drives except one which is used for parity bits. It's easy to think of it that way since the parity bits are actually not stored in a single drive, but distributed across all of the drives. However, parity storage will take up the capacity equivalent to one of the drive in the whole array. Simply means, if you have four 500GB hard drives, your total volume size in a RAID 5 array would amount up to 1500GB, with 500GB used for parity bits.
Parity is much like an error correction bit used to provide added data integrity. Even if you have one of the drives failed in the entire array, the RAID volume would still continue to run as if nothing has happened since the parity bits would kick in with enough information to rebuild your data. Thus, RAID 5 makes very efficient use of your hard drive's capacity as it requires less capacity to rebuild your data, if you compare it to mirrored volumes in RAID 1 arrays.
However, RAID 5's optimal array setup is limited only to about four hard drives. Theoretically, you can use more drives in a RAID 5 setup, but more drives would mean lower survivability in terms of disaster recovery. This is because in a RAID 5 setup, you cannot afford to have more than one hard drive failing you at any one time.
This is where RAID 6 steps into the picture. RAID 6 is very much like a RAID 5 except that it now uses two parity blocks instead of one. Similar to RAID 5, the two parity blocks are distributed across all the drives. This makes a RAID 6 array more robust in terms of disaster recovery as the array can withstand up to two drive failures at any one time. Thus, in a RAID 6 setup, users can often afford to use more than four hard drives to create their volume since the chances of multiple drive failure would be significantly lowered.