If a slow scan time is making more frequent incremental backups a poor option, you could look at the optimized scan method in v.10 (formerly known as data classification in v.9). This should make the scan phase a snap.
You could also setup semi-frequent local shadow copies of the file server data volumes to protect files changed during the day. These take very little space since they are copy-on-write snapshots. End users can access the contents of these shadow copies themselves via the "Previous Versions" tab of objects on the file share.