[Discuss] BLU meeting live-stream experiment

[Tom Metro]

John Abreau wrote:
> The only real question was whether the live-stream would replace the
> old camera, or merely augment it. Based on the results we got this
> month, it looks like it will be best to use both. The webcam does a
> great job of live-streaming the meeting, but the old camera produces
> a much higher quality video.

There are a lot of inexpensive ($100 to $200) point-and-shoot cameras on
he market that claim to do 1080p video. (Some sub-$100 cameras can be
found on Amazon.) Not only should such a camera offer better quality
than a typical web cam, but you also typically get a zoom lens so you
can more easily frame what you want to record.

Many YouTubers use cameras like this, and later graduate to DSLRs. I was
curious to see if there were any that could not only record the video
but stream it as well over USB. There are cameras that can, but
apparently they are quite rare, and it is hard to identify them based on
specifications.

Even third party reviews aren't of much help, as they usually don't
address the live streaming need. You can find lots of reviews for
cameras useful for recording YouTube videos, but they all assume a
record-and-upload process.

(The reviews, however, do point out that most of the low cost
point-and-shoot cameras fall short when it comes to capturing audio, and
a mic jack is unlikely to exist until you move up to a DSLR. But as long
as the live video doesn't have much latency, you can just feed audio
separately into Hangouts.)

Despite the difficulties of finding the right camera that supports this
functionality, I still think this is probably a better than a web cam.


Jerry Feldman wrote:
> As for sound, there are microphones at the podium we can use.

If you have the budget, a wireless lapel mic would probably be best for
the speaker.

Some of the MIT rooms have ceiling mics to pick up the audience, which
could be used for the Q&A portion, but might be simpler to just use an
omni or 180 degree pattern USB mic for that, like
https://youtu.be/fANP9Q6OKmQ

For Boston.pm we often deal with remote speakers, so micing the speaker
isn't our responsibility, but micing the room is, and for that using the
laptop's built-in mic isn't ideal, so we're considering an external mic.


John Abreau wrote:
> I could use help figuring out the delay. At one point early in
> Brian's talk, I stepped aside and pulled up the youtube feed on my
> tablet to verify that everything was working, and the live stream was
> delayed significantly from what Brian was saying at the time.

In my experience this is typical.

When using Hangouts, there are two ways of accessing the stream: 1. As a
contributing participant in the Hangout. For these users the delay is
kept to a minimum. (If you were having a video call with someone, a long
delay would make the product unusable.) 2. As a view-only observer
through YouTube or the Hangouts UI. For these users you'll see a delay.

My guess is that for users in group #2, regardless of which UI they are
using, Google is passing the stream from the Hangouts servers to the
YouTube servers in order to allow you to scale up to hundreds or
thousands of viewers, and propagation delay is added in the process.

Also keep in mind that YouTube lets you pause the live stream. When you
unpause it resumes from where you left off. (I'm not sure if that is
limited to a maximum window, beyond which it jumps forward in the
stream.) But this is another way in which you can end up with a
significant delay between what you see and what is live.

In any case, do you see the delay as problematic? A 30 second delay
should be tolerable, even during a Q&A session, if the questions are
submitted by IRC or the Q&A sidebar in Hangouts.

 -Tom

-- 
Tom Metro
The Perl Shop, Newton, MA, USA
"Predictable On-demand Perl Consulting."
http://www.theperlshop.com/