How to block-by-block upload a large file to S3 in Ruby, having previously encoded it in Base64?

Question

ruby

You need to upload a large media file (more than available RAM) to S3 storage, having previously encoded it in Base64.
Help optimize your Ruby code so that reading a file, Base64 encoding, and then uploading to S3 is done block by block.

blob = StringIO.new( Base64.encode64( IO.binread('big_file.mp4')), 'rb' )
      ...
      @client.put_object(
        bucket: @s3_bucket.name,
        key: target_name,
        body: blob,
        acl: 'public-read'
      )

The Base64.encode64 documentation does not provide a way to process a file stream block by block. Are there any tricks?