<peeps[win]>
that is to say, i'd love to see the "slow" version of the code and prove it can be optimized
<peeps[win]>
without python LUT generation
snaked has joined #openscad
<teepee>
uh, oh, qr code challenge :)
LordOfBikes has quit [Ping timeout: 240 seconds]
snaked has quit [Quit: Leaving]
snaked has joined #openscad
LordOfBikes has joined #openscad
mmu_man has quit [Ping timeout: 245 seconds]
<InPhase>
peeps[win]: The python qrcodegen module I usually use is 900 lines and about half of those are comments. And it doesn't seem to import numpy or really any heavy-lifting libraries. It's basically self-contained.
<othx>
guso78 linked to "QR Code Door hanger by gigl" on thingiverse => 1 IRC mentions
<guso78>
But IT also works having everything White and use the light cavities as black. The of course the qr Code needs to BE inversed.
fling has quit [Ping timeout: 246 seconds]
fling has joined #openscad
qeed__ has quit [Quit: qeed__]
qeed has joined #openscad
mmu_man has quit [Ping timeout: 245 seconds]
mmu_man has joined #openscad
<Scopeuk>
Hmm the "light cavities" triggered a though, I wonder if just fiddling with the geometry might work (back lighting aside), should be possible to set one up where "white" is a flat top surface and black is a slightly sunken surface setup to scatter the light and cause a much lower response
<Scopeuk>
A bit like how an anechoic chamber works
snaked has joined #openscad
<J23k33>
I thought about this too - or as a light screen - but the QR code has unconnected voids so you can't print only the black or only the white. But print the whites .5 lower put a drop white paint on and use a cardboard edge to squeegee over .. Let it sit a bit and wipe on a wet tissue (or with ipa) to clean residue .. or sanding
<Scopeuk>
I also wonder how a smooth/glass's build plate for light and then bridge for recessed dark might work.
<Scopeuk>
The though experiment is interesting but I know I'm very lazy
extor has quit [Quit: ZNC 1.8.2+deb2build5 - https://znc.in]
qeed has quit [Ping timeout: 245 seconds]
qeed has joined #openscad
<joseph__>
teepee: I have been working on eliminating errors from VertexArray::addAttributePointers. From my investigation it still uses CPU-side vertex arrays. And that code seems to get called when doing F6 shader rendering, so I believe it needs to actually be replaced with something modern instead of just turned off entirely by a macro.
<teepee>
joseph__: ah, it's possible, there's still some ancient code in some places. replacing those code path would be a good step forward
<guso78>
J23, qr Codes use one of 8 different patterns which get xored to the result , which ensures that there are Not too big voids or too big faces in local areas. You dont need to worry about too "lonesome" Locations.