A software laser framework with GUI, for up to 4 lasers live actions with ethedreams DACs. Think creative like Laser "battles", planetarium, sharing available lasers in demoparties for competition, ...
- A "tracer" per etherdream/laser that take its given point list, correct geometry, recompute in laser controller coordinates, send it to its controller and report its status to the "manager".
- A "manager" that talk to all tracers (which client number point lists to draw, new geometry correction,...), handle IOs (webui functions, OSC commands,...) and plugins.
- A web GUI in html, css, and vanilla js. This UI can be used in a tablet, computer, whatever. LJ does not come with an html server but absolutely can . This GUI has a (currently slow) simulator, but one can also use an etherdream/laser emulator (see nannou simulator below) to work without physical lasers !!
- A network available database (redis). "Plugins" can send directly their pointlists to redis. Each "tracer" is instructed to get one of the avalaible pointlist in redis.
Important : for best performance LJ is meant to run in a dedicated computer especially with multiple lasers and highly multitasked load : if you watch video, use live webcam face recognition, webui simulator,... and run LJ on the same computer, well you need a bunch of cores. If you don't, spread the load : you can use webui on a tablet, the livecam on a phone, run pointlists generators on another computer,...
- Web User Interface in your browser : open www/index.html. Javascript is needed. By default it connect to localhost. If you want to control remotely, you need edit webui choice : python3 configure.py
- A compiled version for os x and linux of nannou.org etherdream+laser emulator is included. For more informations, like license see https://github.com/nannou-org/ether-dream
LJ accept up to 4 groups = virtual "scenes" of 4 "pointlists" each (= one pointlist per laser), so up to 16 pointlists can be sent to redis at anytime from anywhere in the network. The idea behind this is to easily share actual lasers. Imagine in demo party :
Erica needs 4 lasers, that's the 4 pointlists of "scene" 0.
Paula and Jennifer use only 2 lasers each, so they can share "scene" 1.
And so on..
The server/network/webUI idea allows to spread cpu intensive tasks on different cpu cores and especially give tracers enough cpu to feed etherdreams DACs smoothly. Of course all this can happen in one computer if *you have enough cpu/computers/network ressources*.
It's obviously overkill for one laser in a garage, but for several laserS games events, laserS art, laserS competition, laserS planetarium,... LJ will handle the complexity. Content providers like artists, demomakers,... just need create plugin in whatever langage, send the points to redis.
To change current scene used by lasers/tracers use the command : /scene/scenenumber/start 1
Registering the plugin in LJ.conf is absolutely not mandatory.
Needs at least : an etherdream DAC connected to an ILDA laser, RJ 45 IP network (gigabits only !! wifi and wired 100 mpbs doesn't work well with several lasers). Seriously : if you experience frame dropsyou need to upgrade your network and use a dedicated computer to run seperately main program from plugins, youtube,...
LJ is tested with Firefox, supports Linux and OS X. Windows is unkown but welcome, if someone want to jump in.
LJ is in dev : versions in this repository will always be core functionnal : accept and draw pointlists. New features can be not fully implemented, wait for the next commit. Any feedback is welcome at any time.
The configure script to enter your etherdreams IPs,... is launched during install. You can use anytime, i.e if your network/etherdream setup change : python3 configure.py
- Webserver ?
A webserver is useful if you want to use the webUI remotely. Bring Your Own Webserver.
1/ Say the laser server computer (running LJ) IP is 192.138.1.13, the client computer is 192.168.1.52, First remember to check on the server computer, if the redis server is listening to the right IP : edit /etc/redis/redis.conf
The server approach is based on redis, so you can write and run your laser client software in any redis capable programming langage (50+ : https://redis.io/clients). An external program that just send pointlists to redis is a "client". If you want some interactions from the webUI, like text status area support, crash detection, autostart,... it's a "plugin" and some default code is needed. LJ and plugins signaling is over OSC.
- Generate at least one pointlist array (say a square) with *enough points*, one point is likely to fail for buffering reason.
- Feed your point list array in string format to redis server. i.e use "/pl/0/1" redis key to feed scene 0, laser 1. See /pl/ command in command reference below how to send your pointlist to i.e /pl/0/1 redis key.
First open custom1.py and learn how to declare different objects. custom1.py is a 2D shape example in 3D rotation (red/green anaglyph rendering) that use 2 layers : one for left eye and one for right eye. custom1 is a copy of square.py
Dest1 = lj.DestObject('1', 1, True, 1 , 0, 1) # Dest1 will send layer 1 points to scene 0, laser 1
3/ Different layers to one laser ?
You should consider adding all your points to one layer, but same as 1/ it's a destination problem, just add another destination with the same scene/laser for this layer
Dest1 = lj.DestObject('1', 1, True, 1 , 0, 0) # Dest1 will also send layer 1 points to scene 0, laser 0
4/ I want to animate/modify anything on the fly : I'm doing a game and suddenly my hero change color.
It's a drawn object problem : there is two kind of drawn ojects :
- "Fixed" objects : you generate points in 2D from 0,0 top left and 500,500 is bottom right. Say Hero is a Fixed Object, you can directly change value of
- "Relative" Object : is a kind of laser sprite : your points in 'objectname.xy' has to be around 0,0,0. The other properties let you describe the actual position (xpos, ypos), resize,..
i.e for a character "PNC" vanishing in one point declared as a "Relative" Object, you can decrease resize parameter : PNC.resize
DrawDests() will take care of all your declared drawn elements/"objects" and Destinations to generate a point list per scene/laser sent to redis. In client point of view a "pointlist" is the sum of all its declared "layers".
Nannou visualiser kind of emulate an etherdream and display in a window what a real laser draws.
2 compiled nannou visualisers are included, one for Linux, one for macOS. It's pretty old versions but much more compatible with "old" processors/computer, as you may want to repurpose an old computer to run LJ.
To use this visualiser as one of LJ's lasers, in LJ.conf edit one of line ip = someipaddress with the IP of the computer running the visualiser. Relaunch LJ. One visualiser per computer.
Nannou visualiser emulation is better and better but one can find a few known non-working situations. See it's github repository (https://github.com/nannou-org/ether-dream/tree/master/dac-emulator) for more recent versions.
- Because wifi will always finally sucks for many reasons, our computers (LJ server and plugins) are *gigabits wired* with IP 192.168.1.10 and after. Don't trust end user gear marketing on wifi. We have a big gigabits switch for the *laser only lan*. We provide Internet through wifi on a different network like 192.168.2.x if really needed.
The need for a dedicated computer to act as "laser server" usually depends on how many lasers you want to control and your main computer load. If you seen flickering with small point lists, try the dedicated computer idea and/or stop process interfering like redis monitoring,...
For nice lines and angles all user pointlists are automatically resampled by tracers. There is 2 cases defined for resampling strategy : short and long distance between 2 points.
short distance : has one step : (1.0, number repetition at destination position). 1.0 = 100% = end point of the line between the 2 points. For example : (1.0, 8) means the end point will be repeated 8 times.
long distance has 3 steps : (0.25, 3), (0.75, 3), (1.0, 10) : means an extra point at 25% is created and repeated 3 times, another at 75% repeated also 3 times and destination point is repeated 10 times.
For glitching experience you can change resampling strategy live with "resampler" command. See command reference. [short distance, long distance] = [(1.0, 8),(0.25, 3), (0.75, 3), (1.0, 10)]
LJ is compatible with TLL and analog modulation. Each point color is an int value, wich is simply the hex color in decimal. Example : white = #fffff = 65535.
This program suppose that the ether dream is configured in a certain way especially for its IP address. We write an autoplay.txt file inside an SD Card within the ether dream DAC, with the following lines you can adjust i.e for pps or fps. Yes, there is a builtin DHCP client in the ether dream DAC but if you run multiple lasers, having a fixed dedicated network makes you focus on laser stuff.
See links section for great etherdream managing tools.
About hardware setup, especially if you have several lasers : ILDA cables are insanely expensive. For each DAC, buy a very small ILDA cable and RJ 45 cable, all DAC goes to a local switch and only one long cable to your
You may also consider the Power Over Ethernet 'POE' option. Buy a POE splitter and connect everything to the ether dream fixed near your laser. You can have then a simple and very long network cable and use a Power Over Ethernet injector or switch close to the driving computer. Beware some vendors use 24V POE Injector : POE injectors and splitters must match.
- Sitter will display all etherdreams available on the network and their state (playing, idle,...). python sitter.py or use the compiled version (for macOS). May need tkinter :
Every point must be : (x,y,color). Color is the hex color like #FFFFFF in decimal. For new laserist : if point A is green and next point B is red, line AB will be red.
