Re: [Debconf-video] Report from last week-end's single-user single-day DebConf Video sprint
On Tue, Jun 27, 2017 at 12:51:40AM +0200, Nicolas Dandrimont wrote:
> Hey all,
>
> I've spent an evening and a day working on video stuff last week-end, using the
> opportunity of a pair of lent cameras to do some work. I've put some pictures
> up at https://krkr.eu/dcvideo/
>
> Here's what I can report:
>
> Camera testing
> --------------
>
> I had both the Sony PXW-X70 and the JVC GY-HM200E set up. Both cameras are
> largely similar in functionality and feature set, and they work just fine with
> the vocto setup.
>
> I recorded to YouTube : https://youtu.be/BleLh5hXICY; at the beginning of the
> video, the JVC is on the left, the Sony on the right; I switched them out at 12
> minutes : most of the video has Sony on the left and JVC on the right.
>
> The differences I could notice were:
>
> - Low light behavior: using dawn and the blinds in the room allowed me to test
> for crappy light conditions
>
> (reminder: at those timecodes, Left = Sony, Right = JVC)
> test 1 : https://youtu.be/BleLh5hXICY?t=9282
> test 2 : https://youtu.be/BleLh5hXICY?t=10293
>
> feel free to stroll through that video for more tests.
>
> What I noticed:
>
> The JVC high gain for low light setting is manual and fiddly; Sony has proper
> automatic gain control. The JVC's auto white balance is slow, the Sony seems to
> do the right thing pretty fast even in very bad conditions (playing with
> blinds, light switches, ...). The JVC also was very grainy when fully zoomed in
> and in low light (see the very beginning of the video).
In addition, the JVC's autofocus seems pretty bad, often requiring
several seconds to be done; the Sony seems faster in that regard.
Its grainyness is due to the larger sensor on the Sony; that's perfectly
expected behaviour (and is why my friend suggested the Sony in the first
place).
I notice that the JVC seems to be able to zoom in further; is that a
correct assessment? Given that it has a smaller sensor, this is also to
be expected; but given that the Sony's sensor can do 4K and we only need
720p, we can probably get away with so-called "digital" zoom without
that much quality loss (there will be some optical artifacts, but not
really something you'd call showstoppers)
Did you run a test on how far away you can get a close-up of someone?
> - Size: the Sony is substantially smaller than the JVC and we should be able to
> fit two in a case the size of the ones we've been using for the dvswitch
> laptops.
That's good, yes :-)
> - SDI connectivity: the Sony does SDI-6g (up to 4k) while the JVC does not
> (AFAICT, pegging the SDI to 1080p). Not very useful for us, but who knows
> what we'll do in the future...
Right, but I doubt we'll ever want to do more than 1080p (if even that),
so it shouldn't matter at all.
> - Audio input setup: I only had batteries for the tests (no AC adapter) and I
> noticed the Sony would lose audio input settings when rebooting. As it
> defaulted to shoe mic it's not a showstopper.
It is a bit annoying, and might be confusing. There is a slight risk of
people switching off the camera and thereby losing settings, resulting
in the recording not being the way it should be.
I agree it's not a showstopper, but it is something we need to make sure
we've got covered for volunteers; we should tell them never ever ever to
switch off the camera, or have a cheat-sheet of "post-boot operations"
with every camera.
Probably best to do the latter, actually.
> Also the silly stuff like phone remote control works...
Heh. Cute :-)
> IMO the Sony is superior to the JVC, mostly thanks to its smaller size and
> better low light behavior. Happy to hear opinions.
>
> Streaming setup
> ---------------
>
> I had a very productive chat with markvandenborre, member of the FOSDEM team.
So am I ;-)
> They use nginx with the RTMP module to get the feeds out of the rooms into HLS,
> then a bunch of nginx caching frontends to back video.fosdem.org. I think
> that's a setup that we will be able to easily replicate.
For streaming, this makes sense.
FOSDEM also uses that setup for recording, with many woes. It makes some
things easier, and in that light it does make sense for FOSDEM given the
huge huge (*huge*) (no really, *HUGE*) amount of rooms they need to
record, but I would recommend against doing the same for DebConf. You'll
be sorry afterwards.
[...]
--
Could you people please use IRC like normal people?!?
-- Amaya Rodrigo Sastre, trying to quiet down the buzz in the DebConf 2008
Hacklab
Reply to: