How are footage files written to the NAS

Solution

• Footage files are written to disk using an OS blocking write call
• Symphony writes frames in chunks (multiple frames at a time) where the chunking size is optimized for the stream/hardware.
• The frame sizes vary depending on CODEC, settings, resolution, and whether we are writing difference frames or key frames
• Frame size can vary widely
  • ~5KB/frame (H.264 P or B frames at standard definition)
  • ~500KB/frame (MJPEG HD camera)
• Symphony uses NTFS
• To tweak performance you can set the following variables in Server Manual Configuration Editor:
  • Type: Global
  • ID: <empty string>
  • Section: Main
  • Key: FootageBufferBytes
  • Value: 65536

NOTE: Value is 65536 by default. We recommend a power of two, ideally matching or exceeding your disk's allocation unit size.

• Type: Global
• ID: <empty string>
• Section: Main
• Key: FootageBufferAutoGrow
• Value: 1

NOTE: Value is 1 by default. If it is anything other than 0, Symphony automatically increases the buffer you specified above to ensure it holds at least two key frames.

• Footage files are created with the filename as the UTC time at the start of the video
• Symphony rolls over to a new footage file when it becomes 3.75GB or 6 hours has elapsed (whichever comes first)
• To force the footage buffer to start at a size other than 64K set the FootageBufferBytes variable
• To stop the buffer from automatically increasing set FootageBufferAutoGrow to 0