[Lcabid] AV specs and equipment.

[matheist76 at westnet.com.au]

Hi Guys, 
this is a list of the specs and equipment that I have come up with. Check it out to see if there is any suggestion that you can add. 

Matt 


The AV system 

=== Base System === 
This is the system that we use no matter what. It forms the basis for our more advance setup and yet on it's own it should handle all the requirements of an AV system. 

==== Specs ==== 
These are the base requirements. The equipement we use is a separate matter. The main thing about these specs is that they be compatable to the more Advance System's Specs. 

* Have all of this worked out for when we submit our bid. I still want to keep our options open but at least have a very good idea of what works and what doesn't. 

* Saving to multiple media. We want multiple copies of the uncompressed AV. One of these must be in the camera itself the other will most likely record to a pc that's designed to do it automatically (it will record for the entire then stop. 9am to 6pm [9 hours] a simple cron job will do nicely). In this case we want the camera's media to be the copy we use to record. 

* Ease of Use. An operator should know that everything is running by just looking at the camera's screen and hearing the what is being said over the microphones through the headphones. Anything more complicated than that simply won't work. If it's possible that the camera can record audio without the operator hearing through the headphones than this system is sure to breakdown. 

* Acceptable Audio. While audio and the audio levels are important, having cd quality sound isn't. Radio quality will be okay. The thing is to be able to understand what the speaker is saying and what is being asked. Also it's important not to have any irritating sounds. 

* Video Okayed. Before the full videos are published to the net they should okayed by the speakers and the lca team. This is ensure that we don't get complaints or lose speakers their jobs. This does not include live streaming as that would only be available to the attendees at the conference. 

* Open Software. We should only use Open Source software to record and transmit the AV. This includes editing and compression. I'm looking at Gstreamer scripts along with some Gstreamer apps. As to codecs our current favorites are ogg and theora. We won't use them in favour of equally open codecs (like celt which is still in development mode but would be perfect for audio stream compression as it can do high quality audio) that are better but at this point theses are our lead contenders. 

* The edit should just be very simple. Remove beginning and ending so that it make for a tighter file and doesn't mean that people have to listen to all the setup stuff (audio check that kinda thing). The important thing is that we don't do major edits but keep the video as close to possible to the actual lecture. 

==== Equipment ==== 
* Camera's. These are simple HD camera's. They must have some sort of output port that we can use to capture the AV streams. Firewire is leader at this current point but if usb or HDMI come of age we'll think about using them. They should have removable recoding medium that we can take out and parcel up and replace. 

* Recording medium. DVD, tape and SD cards all fit the bill with DVD's being the cheapest. We need a minum of 90. One for each talk. They get put into envelopes once the recording is done. 

* Envelopes. The envelopes should be printed with a standard form. Medium #, Date, Room, Session #, simple check list and a place for the speaker to put their signature of approval. 

* Microphone inputs. We think we should be able to take the inputs directly from the room's Audio system. This would be great as it mean that we don't need as many auxiliary mics and it simplifies the setup. This then plugs directly into the camera. If we use a mixer it's kept well out of prying fingers but it's output goes directly to the camera. It's always on. The audio check should be as simple as listening in on a headphone. It it has sound then it's being recorded. No it might or might not be as I found out in Hobart. It either is or isn't. Audio is the key here. Without it there really is no point in the whole setup. 

* Software. Use Gstreamer script to yank the files off the medium compress the raw files so that we can begin the editing process with the initial compressed stream being put up onto the web. 

=== The Advance System === 
This plan is what we want to do. If feasible we should at least attempt to implement this spec. One of the really big things is that we need to start really early. We should have working, at least at Humbug, all of the software for the entire AV setup at least 3 months before. We should get the cameras at least a month before we require them and just before we do that inspect the rooms to make sure that it's all possible. We would have the rooms inspected way before that, heck even before the end of the bid, so that we know exactly what we are getting into. Any extra equipment we nee to buy extra. 

==== Specs ==== 
The specs for the advance system are pretty big. They have been worked out so far with Me, James and Russell. I have been the let's keep it simple guy, James has been the I want all of this and Russell has been the side of freedom. This should be a combination of all three. The base system is mostly my spec. 

* Basic concept. All audio to Camera. Camera & Video grabber to PC. PC to Central Video server(s) via 1Gbit eth connections. Video Server sends live stream to local Webserver and in the end after editing and compression to the main file server. 

* Copies. We need to make sure that we don't lose any of the lectures. In order to do this we need to make sure that we keep copies of everything. We need to save to multiple medium. We need to have a least one complete raw data stream saved, no compression, and the preference is to have this at the camera end. The pc will save a copy as well. This may or may not be compressed. If compressed then this will be the version that is sent to the network and to the server. It may be that it's compressed by say 2:1 then later compressed a lot further. It would be good that this is a lossly compression. 

* Stream all 5 video and maybe all slides on a website to the attendies via the local network. This is so that delegates don't have to attend to watch the lectures. I'm tossing up between whether or not to serve the streams from each lecture pc or to leave it upto the video server. I guess we'll now closer to the end of the bid. 

* The video server/server farm should be saving copies of both the slide and the AV streams at the same time. It should be a very simple thing to combine the two streams together and to edit the film. This compressed final cut should be ready for download at least within 24 hours of the lecture taking place. 


==== Equipment ==== 
* Frame Grabber. Used to grab the video feed of the projectors. We are looking at 10fps maybe a bit more. This will plug into a computer along with the feed from the camera. 

* Laptops. These need to deal with inputs from both the camera and the Frame grabber. I would like for the laptop to record both inputs and then spit them out over the network. However if this is an issue we can drop it. 

* The server. This needs to be either one very fast machine with multiple video cards or multiple machines. It will take the input of all 5 camera and 5 frame grabbers. Then send the raw or compress stream on the fly to the website as a broadcast. We might save the stream and at the same time compress that file. Again this really depends on the limitations of the machines that we have. 

* Software is begining to look like a series of gstreamer scripts. One combined solution is a program called Boxtream which uses gstreamer scripts. We may need to modify these to o what we want. Again further investigation is required. 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.humbug.org.au/pipermail/lcabid/attachments/20090619/481eeb6e/attachment.html>