Bryson Jones
Forum Replies Created
-
None of that is binding, tested or even guaranteed to be correct at all. This was a post off the top of my head, based on random field observations, not hard testing.
I suspect that what most folks “get by with” is far less than the numbers listed above.
Hope that helps.
Research, Test, Deploy, Repeat! It’s the MAM mantra.
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
1) Is it best to have the proxies stored with a managed hosting provider?
Of course the answer is… that depends. If you have 10Mb or more you can host pretty nice proxy easily. If you have 5Mb up and can use a small proxy, you’ll probably live. If you have 2Mb up and can go really low res, and don’t have many users, and don’t have “touchy” users you can deliver. 😉
2) Do small organisations setup their own hosting systems to manage this? I can imagine large orgs like NASCAR, Sony Music perhaps have their own hosting servers but for small companies this is perhaps too difficult.
Most companies host because it’s easier. Moving proxy up to a remote server, setting up the pointers and then the expense of outside hosting seems to drive this. I personally have not seen an organization use a CDN for their proxy yet. Think about it, how many people are you serving to? 5 – 10 concurrent users, the average license for Web Client, doesn’t take much to support.
3) What kind of bandwidth do I need to manage this? Are there any requirements or system preferences? If we are looking a 1000’s of clips in low res H264, what is the performance like for remote search queries?
You don’t load 1,000’s of clips in a search. Metadata comes first with a thumbnail. It’s not about clip number in most cases. (Tip: Keep you catalogs small if people are browsing! Less to load.)
4) Maybe its possible to just connect to the static IP of where CatDV is running?
Yes. Very.
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
This should be sorted in Worker 5 (which should be out any time now.) I’m betting it’s just a bug in 4.
Any of the Squarebox crew on this?
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
Honestly,
Don’t trust this to a web forum. Call your dealer and hire a pro. This is a very complex network infrastructure for many reasons.
The main reason, just to give one instance, is that “managed hosting” for your proxies are seldom mountable to the desktop in an easy fashion and currently, other than through the Web Client, you need to mount the proxy drive to the OS in order to see them using the CatDV Pro client.
To answer what I can:
CatDV Pro client needs a mounted drive path to see the proxy.
CatDV Web Client can reshare a proxy through the CatDV Server but it is of course limited in ways such as there is no way to add an asset to the db through the Web Client.
If your shop is this big, get integration help.
And for your own sanity, phase this roll out so that you don’t have to support remote locations without a solid Server install in your main location with trained users on a proven workflow before you go to the remote sites.
Just my 2 cents… and worth exactly that.
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
Geoff, the HTML export is pretty limited. It was added on as a convenience but Rolf himself admits that it’s not really for page generation, just a reference of clips processed or something like that.
If you need to create a custom web page out of the system you generally use the Worker Node to export the images you need along with an XML and then create your html from there using outside code.
The export to HTML is more for a simple listing of clips without a lot of styling choices.
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
First, you need to check how many user fields you have enabled in the Worker’s general preferences. That’s a commonly overlooked setting. The default is 3.
If you are looking for the names, if you save a field definition set in CatDV Pro, you can enter that set’s name in the server preferences in the Worker Node and it will take the names from the Server and display them for your programming pleasure.
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
No app can do everything. As you mentioned, this can be done with Compressor. Since most larger shops use something like Compressor or some other software, just due to sheer amount of media, it may not have ever come up as a feature.
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
Hey guys, just back from my post-nab vacation. (much needed.)
In the end, the more metadata in the file and folder names the better, of course but as has been discussed the most important thing is that there is a structure at all.
Folder and filenames mostly need to be unique, that’s important and harder to do as you grow larger. A lot of DAM’s actually rename the asset to have a GUID as the name and so you don’t even have a record of the filename. (This is done to guarantee unique names.) I hate this but that shows you how seriously some people take the db. In fact, a lot of heavy database folks would shudder at the idea that we use the actual user given file name.
We are not generally in that world. I find that for most shops, a date based system works fine. Year/Month/Day and then go from there.
However, if you don’t do a lot of projects and you actually work on a project basis, (let’s say more than one a week) as opposed to say news or stock footage, you can pretty much do what you like since you’re talking about 50 project folders.
Depending on the size of your operation, you can choose your way. I find that humans are remarkable in their ability to remember things. If it’s your data. But… if it needs to be handed off to other folks for the future, do yourself a favor and add some metadata to your filenames and folder names.
Oh, have you seen Focal Point? 😉
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
It is faster, but I believe he’s saying that a great benefit is that it frees up the machine to do other things. Important if you are running Worker on a server box.
Personally, I’m more of a fan of handing proxy creation off to a transcoding program and leaving a Worker node to do actions.
Remember, if it’s transcoding, it’s not doing anything else and so a few hours of video can monopolize your Worker Node. That same box running some outside software (Episode, Compressor, etc,) could be handing off those jobs and then getting back to doing other operations. Nothing like a server query not running for 30 minutes while you’re compressing a file.
One new thing for NAB, the Worker Node can be licensed to run multiple Worker instances on the same machine, so one 8 core machine can have several Workers on it, doing different actions, which is nice.
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com
-
In my opinion, these solutions are all fine. As far as “maximizing”, after a few years in this thing it’s not really about software, it’s about people and implementation.
I agree, with 60 hours of footage, unless it’s a really dense show, with very high value footage (historically significant footage etc..) there’s little benefit to digging that deep into tagging.
Choose one, learn it well and make movies. But never confuse an editing system with a DAM, no matter how many marketing articles you read about how great they are. NLE’s will come and go, but databases are forever… ask the government. 😉
Want proof? Drop 100,000 clips into any editor and report back how they handle it. (Hell, drop 200 hours of footage into Premiere or FCP and watch what they do!)
bryson
bryson “at” hidefcowboy.com
hidefcowboy.com