Some databases can store blobs in the filesystem, but managing it by records in the tables.
Typically, the closely coupled recordset structure is a very alien structure from the storage of files, particularly if you have large ones and many of them.
My experience is that it makes the db maintenance costs and performance costs too high. Backups will be slooow, the db starts to become buggy, it will be more vulnerable for crashes, and it will perform very badly, compared to a dual "files in a directory managed by records in the db" structure.
If you want a good performing application, then there is no free lunch, you need to develop it. It is not a big overhead, essentially, instead the blobs, you will have filenames.
There is also the option to serve these files directory by your webserver, without involving the backend (which is tuned probably for complex tasks with small strings, and not for the primitive, but super-optimized http serving of large files).