That depends on many factors, including how much buffer pool memory is allocated, for example. If there is lots of buffer pool space allocated, there is not much synchronous physical I/O going on.
In theory it does help to have indexes on a seperate disk drive or disk array than the tables, so having different tablespaces for them makes that possible if you set it up that way.
You do not need to have each table in it's own tablespace. However, tablespaces have their own page size, bufferpool, extents, and prefetch size, which could differ among the tables you have.
Backup and recovery of related objects at the tablespace level is also a consideration.