How a mirror works
RAID 1 writes every block to two drives, so each one holds a complete copy. The array is as big as one drive. Two 4 TB drives give 4 TB usable from 8 TB raw, which is 50% of the raw space.
You can add a third copy. A three way mirror of 4 TB drives still gives 4 TB from 12 TB raw, and it survives 2 failures. Reads can come from either drive, so a good controller or file system can spread them over both. Writes go to both drives and finish when the slower one does, so a mirror writes at about the speed of a single drive.
A rebuild is a straight copy
There's no parity math. When a drive fails, the survivor keeps serving data, and a replacement gets a block by block copy. For a 4 TB drive that copy takes at best 7 hours at the modeled hard drive speed. A 16 TB drive takes at best 28 hours.
The copy reads the whole surviving drive while it holds the only copy of your data. At a spec rate of 1 in 10^14 bits, the chance of at least one unrecoverable read error in a 4 TB read is at most about 27%, and at 1 in 10^15 bits it's at most about 3%. For a 16 TB read the ceilings are 72% and 12%. Those are ceilings too, but one unreadable sector still means a block the copy can't reach.
What a mirror doesn't protect
A mirror copies everything, mistakes included. Delete a file and it vanishes from both drives. Ransomware that encrypts the volume encrypts both copies. A power surge, a theft or a fire takes both drives together. RAID 1 keeps the machine running through a dead drive. A backup is a separate copy in a separate place.
Where RAID 1 fits
Boot drives, small servers and two bay NAS units. It's the simplest redundant layout, with no parity to calculate and nothing to learn. When you need more space than one drive holds, the options are RAID 5 or RAID 6 for capacity, or RAID 10, which stripes across mirrored pairs. Eight 8 TB drives give 32 TB in RAID 10 and 48 TB in RAID 6.