[LAD] wfs streaming project report

Jörn Nettingsmeier nettings at folkwang-hochschule.de
Sun Jan 18 12:00:06 UTC 2009


nescivi wrote:
> On Wednesday 14 January 2009 12:54:12 Jörn Nettingsmeier wrote:
>>
>> http://stackingdwarves.net/public_stuff/event_documentation/wfs_live_transm
>> ission_2008/WFS-Report-web.pdf
>>
> 
> wow!
> 
> the paper did not mention this, but did you have any packet losses through 
> birds? or bird losses through packets?

very rarely we did indeed have packet losses across the laser link, but
since they were so few and far between (even in bad weather), i don't
have reliable data. one possible weakness in the whole scheme is that
the UDP redundancy methods of both jacktrip and netjack will send
redundant packets right next to each other, so that if you have a burst
failure (which is common), you are screwed.

for me, the morale is: lasers can be made reliable enough if you can
tolerate the occasional single or short burst packet loss (loss rates of
about 0.0001%), the general internet cannot, unless you get end-to-end
QoS, but you can sneak past that if you have lots more bandwidth than
you are going to need. but nothing in the world short of http streaming
will protect your ass against crappy border gateways and switches that
barf on udp stream traffic.

as to bird losses, the class 3 lasers operated at 8mW, so the chance of
a bird being vaporised is, ahem, slim.

the main issue with respect to laser safety was eye damage. the minimum
safe distance to look directly into the laser was about 50m. but since
IR lasers do not trigger a lid-closing reflex (you only see a dim red
shimmer), this minimum safe distance is determined for a duration of
100s of continuously staring into the lens (for lasers in the visible
range, this duration is below 1s, iirc). so you would actually have to
hover in front of the laser for well over a minute before your retina
takes serious damage. therefore, we can safely rule out eye damage to
birds as well, unless they are very skilled flyers, very bored and very
very stupid.

as to stupid people: on the spire, the laser was mounted so that
accidentally looking into it was impossible. it was ascertained that no
members of the public would ever come near the device without a guide
who would inform them of the dangers. the khm site was off-limits to the
public, and on both sides, barriers and warning signs discouraged direct
access to the laser devices.

no photons were harmed during the creation of this event.



More information about the Linux-audio-dev mailing list