<ZipCPU>
They say that in the Army, the waters mighty fine. It's good on cuts and bruises, and tastes like iodine.
promach has joined ##openfpga
pie__ has quit [Ping timeout: 240 seconds]
<DocScrutinizer05>
azonenberg: which filter?
<DocScrutinizer05>
seems it's pretty "easy" to filter out dirt, bacteria etc. But what about toxic chemicals
digshadow has quit [Ping timeout: 256 seconds]
promach has quit [Ping timeout: 256 seconds]
<azonenberg>
DocScrutinizer05: then iodine etc wont help either
<azonenberg>
ditto for salt
<azonenberg>
Distillation might, dependin on what's in it, but that generally isnt a field-serviceable technique
<rqou>
activated charcoal?
<azonenberg>
Generally if you're filtering drinking water it's clean rainwater from a mountain stream, lake, etc
<azonenberg>
that's relatively free of anything but germs
<azonenberg>
in SAR if we expected a lack of mostly-drinkable water we'd just bring it in from a treated source
<azonenberg>
there's a large tank on our command post truck that we can load up personal canteens etc from
<rqou>
azonenberg: on a completely different topic, how hacky would a completely unidirectional 10gbe link be?
<azonenberg>
on what PHY
<azonenberg>
Base-X?
<azonenberg>
or Base-T
<rqou>
base-x
<azonenberg>
on the send side, easy
<azonenberg>
i.e. tx only is fine
<azonenberg>
RX only is trickier as the remote end wont send anything if it thinks the link is down
<rqou>
ah ok
<rqou>
i only need tx-onlky
<azonenberg>
If your board was RX only you'd have to figure out a way to tell the far end the link is up (i.e. gadget on the far end that sends idle frames etc)
<lain>
I think on my switch I can force it to think the link is up
<azonenberg>
but if you want to just send data and discard all inbound stuff?
<rqou>
i'm thinking about how to build a cheap "<all video crap> to pc" solution cheaply
<lain>
with SFP's
<azonenberg>
i see no problem with that
<azonenberg>
But why would you want to do this?
<azonenberg>
SFPs and commodity cable are duplex
<azonenberg>
If your FPGA discards all inbound data thats fine
<rqou>
the low end artix-7 parts only have 4 high speed transceivers
<azonenberg>
... oh
<azonenberg>
LOL
<azonenberg>
I see where this is going
<azonenberg>
unidirectional xaui?
<rqou>
so i was thinking "use the tx side to do 10gbe (xaui) and the rx side to do displayport rx"
<azonenberg>
That bit might be tricky as i dont know if you can run the tx and rx quite that separate
<azonenberg>
i think they may have to share a reference clock
<rqou>
yeah i don't know if that works either
<azonenberg>
but i'd have to go back and check the man page
<azonenberg>
Doing 3.125 Gbps both ways to different endpoints is totally fine
promach has joined ##openfpga
<azonenberg>
But different speeds, idk
<rqou>
yeah i get the feeling that isn't what they're designed for :P
<azonenberg>
Yeah but...
<azonenberg>
you know what channel this is right? :p
<rqou>
right
<rqou>
:P
<rqou>
anyways, it turns out "video crap" isn't a _complete_ ripoff because kintex parts are expensive :P
<azonenberg>
Lol
<azonenberg>
if you wanna talk video chat with matthaism in ##fpga
<azonenberg>
his company does broadcast video stuff
<azonenberg>
we're talking a giant stratix with three or four sodimms of ddr4
<rqou>
nah this is just brainstorming for now
<azonenberg>
umpteen gazillion HD-SDI inputs to a bunch of 10gbe and i think maybe 40gbe outputs
<rqou>
the first "video crap" device i'm likely to build will probably be "
<rqou>
"6mhz bandwidth ADC with video clamp analog stuff"
<rqou>
e.g. a capture card that understands how to do "retro" broken-NTSC correctly
<azonenberg>
Nice
<rqou>
i bet those signals will make "broadcast video" people have a stroke :P
<azonenberg>
lol
<azonenberg>
Probably :p
<rqou>
e.g. 240p-masquerading-as-480i-with-10%-faster-speed-and-slightly-off-black-levels :P
LoveMHz has quit [Quit: ZNC 1.6.3+deb1+trusty0 - http://znc.in]
<rqou>
or how about ntsc-framerate-pal-modulation or pal-framerate-ntsc-modulation? :P
<rqou>
btw you can totally put a sega genesis into the second state if you hack the vdp straps
<rqou>
ntsc50 :P
m_w has joined ##openfpga
<rqou>
an interesting project that i unfortunately don't have much time for is to have Someone Who Isn't Me write a h.264 encoder and then get whitequark to publish it
<rqou>
anyways, i should do something productive today
<cr1901_modern>
video clamp?
<rqou>
there's some circuit you need to get the ac-coupled analog video signal into the middle of your adc range
<azonenberg>
rqou: why not vp8?
<azonenberg>
or similar
<rqou>
it's often a bit weird/special-cased because most of the circuits rely on the video signal having periodic hsync pulses
<rqou>
and they make e.g. "bottom of hsync" be 0V
<rqou>
azonenberg: h.264 is much more widely supported than vp8?
<cr1901_modern>
So if it's ac coupled then there's no DC bias, right?
<rqou>
although vp8 has the advantage of not requiring Someone Who Isn't Me to write it :P
<azonenberg>
My point exactly
<rqou>
idk exactly what the specs are for analog video, but there may still be a dc bias
<rqou>
idk what the passband is supposed to actually look like
<cr1901_modern>
I'd prob have to see what you're trying to do to understand what you mean by clamp
azonenberg_work has quit [Ping timeout: 256 seconds]
<rqou>
e.g. if your screen suddenly turns all white, that might introduce a slight dc bias
<rqou>
afaik it is called a clamp because the circuits i've seen in various appnotes use a diode + some other stuff to clamp the lowest point of the signal (hopefully hsync) to 0V
promach has quit [Ping timeout: 258 seconds]
<DocScrutinizer05>
azonenberg: you ever heard of OMAP PoP soldering warp issue (CPU chip 'PCB ring' bends up/down from thermal tension)?
* cr1901_modern
thinks... yea, white in NTSC would be constant voltage separate by the sync regions IIRC
<cr1901_modern>
that certainly would be a DC bias lol
<rqou>
you can probably test your circuits on "that episode" of pokemon (with porygon) :P :P
<rqou>
i expected that to be a link to the actual pokemon clip
<rqou>
i am disappointed :P
<cr1901_modern>
I'm not that cruel, even though YT's wonderful compression would've removed the seizure-inducing components.
<rqou>
or the compression for the avi/wmv (i don't recall which) that the recording was originally released to the internet it?
azonenberg_work has joined ##openfpga
<lain>
it's so disappointing when you finally find a copy of some song or mix or whatever, only to find that it's full of compression artifacts, and that it's the /original fucking master/