[Caspian Brand]

I have used FreeNAS, and am in the process of migrating to OpenFiler.

Reason: FreeNAS doesn’t support User Level or Folder Level Security, which is very important when trying to share projects.

While AFP may be simple and somewhat powerful, I personally don’t care to bind myself to it. NFS and SMB can be faster, though they can be a bit more trouble to setup the first time, not that you’ll necessarily reach that speed comparison depending on how powerful your storage back end is and what codecs you’re using and how many clients are hitting it.

OpenFiler also supports SAN/NAS, which I think is a nicer approach for those attempting to roll their own on a shoestring budget. You have the options of both platforms in one solution. Plus though OpenFiler is Open Source and “free” you can pay for support and get it if you want it, which to me would be important to have an extra tech resource for such a system.

It may be a little more time and headache to setup the first time, but you’d be amazed what you can get out of a NAS built on older PIII and P4 hardware with a decent RAID controller card. All hail Linux!

Just my 2 cents.

[Caspian Brand]

There are also quite a number of educational facilities in all corners of the country where one may find a number of iMacs for teaching FCP and other tools, where Ethernet is the only network connection option. These also make great inexpensive systems for off-line editing.

Studio Network Solutions’ product EVO provides a multitude of connections, including Fibre Channel for those needing Uncompressed HD, as well as an optional 10GbE uplink port, for capable switches, providing the bandwidth required for multiple systems to stream compressed HD.

From my MacBook Pro 15″ I have handled 3 x streams of PRORES HQ over a single GbE connection to an EVO, ~70MB/sec. Works quite well, of course you’ll need at least 4GB of RAM, but that goes without saying for any edit system in my opinion.

[Caspian Brand]

QuadLogic in Bristol, England are also very knowledgeable, techs are on par with SANTech (I know and have worked with them both). I know they have built XSANs running multiple streams of 10bit RGB 1080, which has a very intensive data rate on par with Uncompressed 2K.

https://www.quadlogic.co.uk/

[Caspian Brand]

I’ve set up a lot of video RAIDs in my day, and with RAID 5 I do not like to create a Logical Volume larger than 12 physical disks, and then only when a Hot Spare is configured. The magic number of disks for most video (for the exception of some Uncompressed HD formats) is 6 – 8 drive members. If you have a 16 Bay array I would recommend two 8 disk RAID 5s. If you have a 12 Bay array I would recommend two 6 disk RAID 5s. As Bob mentioned, RAID 5 is the best for video. If you want more physical drives and one larger volume, then you can use your server to create a RAID 0 stripe across multiple RAID 5 logical volumes (also known as RAID 5+0 or RAID 50). This approach keeps your ratio of redundant drives reasonable, while increasing the total number of disks in play for performance. The downside is what can happen if your Server looses it’s RAID 0 striping information.

RAID 10 is a good solution for a Nearline Archive, which you may want to consider for your second enclosure.
Or if you have a Quad port HBA, you can RAID 50 across both controllers using four RAID 5 volumes.

In short, your safest bet is multiple RAID 5 volumes for video, and possibly a separate RAID 10 array for Nearline. Make sure your server is maxed out on RAM if your clients are funneling through network shares to the storage.

Depending on how much realtime work you end up doing you may outgrow the approach of a Server with Direct Attached storage. At such time you may need to opt for a high performance NAS or SAN.

[Caspian Brand]

Hi Phil,

I would recommend looking into a product called CatDV, it’s like Final Cut Server, on steroids.
It can be a little more expensive than Final Cut Server, but they have the ability to publish HTML pages of proxy files which can be annotated and edited remotely. This way you can mark which parts of the clips you want to spend time transferring high-rez versions of the files over the internet or other methods with. CatDV allows you specify what size and codec to use for your proxies, and everything is handled over XML, so you can edit with your proxies, then replace the footage with your on-line HD content when you’re ready.

https://www.squarebox.co.uk/index.html

Another option, that isn’t server based, but is good for transferring and synchronizing files (sans workflow automation) is Dropbox. I’ve been using this as a very simple tool to collaborate on a music recording project where I update session files and add new audio tracks for our producer/mixer in California, as well as notes on recording takes, associated session pictures, etc.

https://www.getdropbox.com/screencast

The downside here is the maximum account size is only 100GB.

[Caspian Brand]

I am having a similar problem.

I am shooting DV in 16:9. I use the DV Anamorphic Preset when capturing into FCP.
FCP Displays the video properly when editing.
When I export a QT Reference or Directly to Compressor, even with the Anamorphic settings checked, QT will not display the output file properly. I am even exporting H.264 .mov and .m4v files.

My intent is to post these to the web, so I don’t need to letterbox.

The only workaround I’ve found is to adjust the frame size settings in Compressor to 853 x 480.
The trouble with Compressor is that it takes forever until the end of time to deinterlace. I’ve attempted to use Handbrake for this (which is much faster and looks good) however, it won’t recognize the modified presentation setting of 853 x 480 from an exported QT file.

What do I do?

Frustratedly Yours,