[volunteers] November Presentation Report

[Rick Moen]

Scott wrote:

> Thus, if we were to edit the recordings and just post the "juicy"
> parts, we might have an opposite effect where people love the clip and are
> motivated to want more?

Meta-comment:  It makes me a little concerned when conversation is
almost solely about end-results and there is almost nothing about
_process_, i.e., almost everything about what might be cool to do and
very little about how to do it.  It's my experience that starting out to
do things without considering process leads to sadness.

Rather than fuss about that, though, I'll talk process.


Digital video cameras are great and getting better all the time.  A bit
over $300 will get you a decent one bundled with a 16GB micro-SDHC flash
card.  Typically, video can be written in AVCHD or MP4 (MPEG-4), so
let's talk about the latter because it's a lot smaller.  There's
always a built-in condenser microphone and a microphone jack for
the sound component.

Let's say the default capture is going to be HD (1080i) at 30fps, and
I'm estimating that you'd fit 90 minutes on that 16GB card.  (If you
have more 16GB micro-SDHC cards at ~$100 a pop in very round figures, or
if you had a 32GB card, you'd be able to record more.

OK, then!  Wait, the sound is absolutely vital, and I doubt the mic on
the camera is going to do the trick, so you're talking either a cabled
microphone at some distance from the camea, or some sort of cordless
lapel mic.  Add that to the purchase.

You're now set for sound - sort of.  Probably you are _not_ going to
have a second microphone that gets walked around the audience to the
people interacting with the speaker.  I mean, you could, if you wanted
to.  If you do that, you have to somehow manage the switching back and
forth between mics.  I'm no expert, but think that's usually done on a
small mixer board.  If you have only one mic, then someone, preferably
not the person running the camera (but call him/her the sound tech),
needs to wrangle both the speaker and interacting members of the
audience.  The audience members' input can be handled either of two ways:

1.  Sound tech keeps reminding the speaker to please repeat the question
or comment from the audience, just before responding.  Experience shows
that many speakers slack off on that, and some just never try much, so
the sound tech has to keep gently persuading.

2.  Sound tech coaxes each member of the audience wanting to have a
question or comment to come up to the microphone first - or brings it
to the audience member.  Sound tech becomes the sound gopher.

Almost ready to go!  One last thing, camera technique.  The last guy
who recorded speakers at SVLUG, John Regan, would plant his camera in
the middle of the audience, and focus it on the speaker.  He didn't
remember to go over with the speaker how far from side to side his
picture 'window' extended, so the speaker occasionally wandered out of
rrame, and John whipped the camera sideways on its mount to follow.
Ziiipp!  But wait, there's more:  Then,  audience members were
recognised for questions/comments, and John swivelled the camer
180 degrees to focus on the (inaudible because non-miked) audience
member, then whipped the camera back to the front when he thought the
speaker was about to speak again.  Given frequent back-and-forth
between speaker and commenter, John pivoted that camera around like
crazy.

I hear you needed Dramamine or closed eyes, to enjoy the footage.
On the bright side, we got a comprehensive example of what _not_ to do.

Warned by that experience, you start out with a rational plan to
keep the camera trained on the front, and the speaker is aboard
with this plan and knows where to stand.  You have a rational plan to
capture audio with one or two mics.

Roll (non-)film.  Two hours of talk, you have the MP4 with a
comprehensible and complete sound track.  Good job!

Now, mind you, the file sizes for 2 hours of video don't have to be
quite that big.  Big, but not _that_ big.  That's a question of video
quality, which can be expressed in bitrates.

File sizes for 2 hours of video could very reasonably be cut to maybe
3-4 GB for a compromise bit rate (4Mbps), not HD (which is 24Mbps, I
think), but who needs HD for a technical lecture?  Maybe if you're
lucky, the bitrate captured by the camera can be dialed down _on_
the camera.  If not, you'll need to transcode to a lower bitrate
on a beefy computer with a big hard drive, using something like
ffmpeg, prior to uploading.

And, of course, you might also want to do some video editing anyway.
Get to know Cinelerra!

Ok, you're going to upload it to www.svlug.org, right?  No, you most
certainly are not.  Blowing our monthly Linode bandwidth quota on video
is not in the cards.  Where can you put a couple of hours of video?  No,
not YouTube unless you want to cut it up into a very large number of,
whatever it is, 10 minute segments(?).

Ustream might do it.  There's a gratis service, where viewers are
obliged to watch advertising as well as the provided material, and there
is the paid service (_you_ pay; SVLUG doesn't have a treasury), which
is commercial-free.

And we then put links on the front page of www.svlug.org to the Ustream
links.

That will all work fine.  All it needs is someone willing to do it.

Not complaining, but I'm pretty sure this subject has been covered
before, here.