<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 17/05/2011 8:24 PM, Tim Bowden wrote:
<blockquote cite="mid:1305635064.3485.80.camel@fangorn" type="cite">
<pre wrap="">On Tue, 2011-05-17 at 19:50 +0800, James Bromberger wrote:
</pre>
<blockquote type="cite">
I suspect USB3 may be a requirement; I see PCI-E cards for $22
and $17
on deal extreme. But that does then rule out laptops and rule in
SFF
desktops (most nettops don't have PCI/PCI-E).
</blockquote>
<pre wrap="">
If it can be passed through a USB3 port on the shuttle fast enough (ie,
as fast as the shuttle HDMI input gets the stream), then there is no
need for a camera PC at all. </pre>
</blockquote>
The main thing that I am yet to see is the performance of USB3. And
I don't mean the quoted bandwidth, but the load placed upon the CPU.
The main reason why firewire wins over USB2 is not speed; its the
load that USB puts on the CPU (by way of interrupts) that firewire
doesn't. Firewire has rock solid timing, so no data is ever delayed
or buffered; USB is subject to CPU scheduling.<br>
<br>
<blockquote cite="mid:1305635064.3485.80.camel@fangorn" type="cite">
<pre wrap="">The shuttle device would replace that.
The only PC in the whole setup would be the mixing box. Downside is
we're limited on the maximum distance of cameras from the mixing box "as
the cable runs" to 25m. Probably a bit short as the max distance from
the mixing box, but then why can't the mixer sit at a second low powered
box (even in another room/annex?) and remote login to the full powered
headless mixing box that could sit anywhere suitable for the cameras?
That way the maximum distance between any two cameras could be up to 50m
(theoretically).</pre>
</blockquote>
So I tried using dvswitch over VNC; that's a no go- the main video
window, which issue Xv to some such to render, is not shown over a
VNC link.<br>
<br>
I would still like to have each cam op on a headset, and that
requires the local PC to be working. We dint try mumble last week as
we were in a very small space so it didn't make sense. We should try
this somewhere bigger - more spread out. So I dont think we get rid
of the camera capture boxes; the ones we have now, although large
and in need of quieter fans, are doing the job.<br>
<br>
I just re-watched our video, and the flickering of the P-in-P
happens from both cameras. I suspect it was the box we were using
being underpowered. But we have a much more powerful box in AV 2 now
up and running. <b>Jason: </b>what's the size of the fan thats
really noisy in AV2 there with you?<br>
<br>
Having another machine outside that's doing stuff - well, definitely
an option for the OGG encoding; that's what we did at UCC. But if we
get a quad core CPU into the AMD mobo AV2 box, then we may be
sorted??<br>
<br>
James<br>
<br>
<div class="moz-signature">-- <br>
<em>Mobile:</em> +61 422 166 708, <em>Email:</em>
james_AT_rcpt.to<br>
PLUG President 2011: <a href="http//www.plug.org.au">http//www.plug.org.au</a><br>
Perth.pm Organiser 2011: <a href="http://perth.pm.org">http://perth.pm.org</a></div>
</body>
</html>