#apertus IRC Channel Logs

off to bed now ... have a good one everyone!

Bertl: I hope this little sleep is healthy for you... Do you need help?

regarding help, there is always stuff to do and you can certainly help out if you like

I meant more in regards to your health and that. I do expect to become a contributor, but not too soon, considering the photogrammetry stuff I need to rush before renovations start, and then be better off producing results in time for 36c3. I got a genesys 2 devkit at home, waiting for me to turn it into an axiom beta mod. I'm kinda commited therefore, and I assume at least some of the work needed for my

will be useful for/make it's way into the mainline fpga software.

what are your plans to connect the Axiom Beta with the Genesys 2?

Hardware wise I'm mostly interested in getting it wired up to the FMC connector with "only" 4 layers from the chinese vendor, likely using post-wall shielded paired-up stripline for the 10GHz signal, to allow longer passive (twinax) DAC cables (optics result in a lot of delay, and DAC might have sufficiently low delay to run TCP against full TCP offloading in the PCIe NIC, all without needing to buffer

so you might want to go with the Axiom Alpha design then

The hardware is not nearly as field-handling-optimized as the Axiom Beta, but my focus is more to have something light I can strap on a copter or maybe a heaxapod, and offload full sensor bandwidth to a computer I can easily outfit with a PCIe SATA HBA for enough HDD bandwidth, or NVMe, GPU(s), all without having to write custom drivers that work with an FPGA.

Bertl: yeah, as said, I'll drop most of thw hardware design and try to re-use peripheal electronics to save time. It's not made easier by wanting to experiment with some exotic optical approaches/designs (some playing with light field imaging is planned, as well as astronomical (speckle) imaging. Unfortunately there seem to be no specs describing what happens when you carefully cool the sensor down with

coolant. The astronomical stuff is likely limited by how close I can get to shot noise limited sensing (there appear to be physical limits to resolution enhancement attainable in post-processing to fix atmoshperic turbulence, and they seem to be based on shot noise limitations (so even an avalanche photodiode array won't help)

Enough of the wishful thinking though. I need to write emails, and get to prepare software/search for implementations of sturcture from motion multi view stereo pipeline components that include rolling shutter correction. I'd prefer not to have to re-write that, and I kinda fear I would.

Yeah. One last thing: do you know if anyone thought about cooling the CMV12000 to below the datasheet while in operation? Or maybe a way to search irc logs without throwing a spider at it (which I'd prefer to not do, in aprreciation of the server load that'd likely cause if the pages are dynamic)?

but if done properly, I do not see a big problem with it

Bertl: thanks. From your experience, does the manufacturer seem likely to offer support/guidance for these magic register values, that have to be changed to adjust for the signficantly different analogue behaviour?

back now ...