Author Topic: ISO: high channel count LSP musical sequence  (Read 12104 times)

Offline nutz4lights

  • Full Member
  • ***
  • Join Date: Dec 2014
  • Location: Melbourne, FL
  • Posts: 191
  • Kudos: 3
Re: ISO: high channel count LSP musical sequence
« Reply #60 on: January 28, 2015, 08:55:44 PM »
You mean ping?
-Louie
Videos @ Cape Sable Lights

Offline arw01

  • Hero Member
  • *****
  • Join Date: Oct 2013
  • Location:
  • Posts: 836
  • Kudos: 0
Re: ISO: high channel count LSP musical sequence
« Reply #61 on: January 28, 2015, 11:03:38 PM »
nope the tracert, it also shows an average time.  it should just be a couple of hops, but currious what it says to your various different controllers on the unicast.

Offline nutz4lights

  • Full Member
  • ***
  • Join Date: Dec 2014
  • Location: Melbourne, FL
  • Posts: 191
  • Kudos: 3
Re: ISO: high channel count LSP musical sequence
« Reply #62 on: January 29, 2015, 08:11:33 PM »
nope the tracert, it also shows an average time.  it should just be a couple of hops, but currious what it says to your various different controllers on the unicast.
So I ran the tracert tests tonight on 8 out of the 10 controllers (two are still not hooked up... I'm working on getting them setup in the off-season storage location)... The tracert results to all five of the J1SYS p12s/d controllers was < 1msec across the board... same with the single SanDevices e6804 controller... the two Advatek controllers (one PixLite4 and one PixLite16) came in at around 4-9msec though...

See anything wrong with those values?

-Louie

Offline arw01

  • Hero Member
  • *****
  • Join Date: Oct 2013
  • Location:
  • Posts: 836
  • Kudos: 0
Re: ISO: high channel count LSP musical sequence
« Reply #63 on: January 29, 2015, 11:18:26 PM »
No those should be good, did the tracert come back pretty quick, no long pauses like it might have been dropping some packets?

The 6-9ms controller would be off maybe  by theory off 1/2 a frame that would never be detected by the eye.

I've not run your big sequence on my full 32K license computer.  On my 8K license I saw no lag with the sequencer playing it and when I turned on output no issues, but it only put out the first 8K on the wire.

hope to get to the full 32K this weekend.

Alan

Offline nutz4lights

  • Full Member
  • ***
  • Join Date: Dec 2014
  • Location: Melbourne, FL
  • Posts: 191
  • Kudos: 3
Re: ISO: high channel count LSP musical sequence
« Reply #64 on: January 30, 2015, 04:37:01 AM »
Forgot to mention this but that was from a temporary laptop (dual core) to the controllers.  I didn't know if it was possible to do that test from the FPP to the controllers (not a Linux guy).  The show computer(s) are in the house and I didn't have time to connect them back to the show network.  Let me know if that is important.

There was definitely a pause from the time I hit enter to the time it showed the reading... maybe several seconds.

Louie

Offline zwiller

  • Falcon Beta Team
  • **
  • Join Date: Mar 2013
  • Location: Sandusky, OH
  • Posts: 938
  • Kudos: 12
Re: ISO: high channel count LSP musical sequence
« Reply #65 on: January 30, 2015, 12:34:18 PM »
my experience showed that, based on the recommendation of helpers on the LSP forum, the macros applied to the entire display had to be set to 10FPS to reduce the number of timing marks... that was how I knew I had no chance of doing musical sequences... because very rarely do you sequence to the whole display when doing music... you want individual banks of lights doing different things... which in LSP creates more, and more, and more timing marks...

Just wanted to respond to this (while you guys geek out).  First off, the average beat track only adds 360 timing marks to a sequence (120bpm 3:00 song) and that many marks should not have any affect on this sort thing.  While I do agree that LSP should allow you to sequence he way you do (everything discrete) I have a hard time seeing the effects in smaller elements.  Unless you setup each smart mini tree exactly the same way and draw a match of it in the visualizer you have little chance of the effect carrying through.  IE wrapping 100 nodes around a 2' mini tree and then adding a tree shape in visualizer.  In addition, although it would be cool to see a wreath color chase 360 degrees, it would be pretty hard to notice a 24" wreath in context to a large 2 story home display. 

In this video at the end I use whole house macros at the end 1:45.  http://vimeo.com/115857387  My minis and live trees are smart but programmed as dumb.  Despite of this, I think the effects carryover well.  Maybe that's what you are concerned with?  Until I tried it, I originally thought I would need for everything to be smart/discrete to pull it off... 
Sam

Last year's video: https://vimeo.com/150560653

Offline arw01

  • Hero Member
  • *****
  • Join Date: Oct 2013
  • Location:
  • Posts: 836
  • Kudos: 0
Re: ISO: high channel count LSP musical sequence
« Reply #66 on: January 30, 2015, 02:40:42 PM »
I would run it on the show computer, but tracert always has a bit of lag, but that can be because a reply does not come back, if you do a long route like to a server in another country, some hops populate fast, others take a long time.

Offline nutz4lights

  • Full Member
  • ***
  • Join Date: Dec 2014
  • Location: Melbourne, FL
  • Posts: 191
  • Kudos: 3
Re: ISO: high channel count LSP musical sequence
« Reply #67 on: February 01, 2015, 11:29:16 AM »
Just wanted to respond to this (while you guys geek out).  First off, the average beat track only adds 360 timing marks to a sequence (120bpm 3:00 song) and that many marks should not have any affect on this sort thing.  While I do agree that LSP should allow you to sequence he way you do (everything discrete) I have a hard time seeing the effects in smaller elements.  Unless you setup each smart mini tree exactly the same way and draw a match of it in the visualizer you have little chance of the effect carrying through.  IE wrapping 100 nodes around a 2' mini tree and then adding a tree shape in visualizer.  In addition, although it would be cool to see a wreath color chase 360 degrees, it would be pretty hard to notice a 24" wreath in context to a large 2 story home display. 
My post might get a bit long, and whatever the tones sounds like, please don't take anything I am typing as an attack on you as that is not remotely close to my intention.

I don't think you're understanding what I'm saying about sequencing.  I started this whole "Christmas lights to music" in 2005-2006.  Prior to the whole pixel and RGB thing, I used LOR and their software.  The way I did my songs was to create the grid with divisions every 0.10 seconds.  My original technique involved tapping out several beat tracks manually, then placing effects to line up with marks in those beat tracks.  I had 12 twig trees, 32 micro trees, 20 mini trees, six banks of flood lights, a 12 way divided megatree, a six bank set of net lights in my bushes, the roof, and two oak trees.... all of which had red, green, and white, so multiply by 3 (some elements were ganged, like the 32 micro trees were grouped into sets of 4).  I obviously would have the mini trees doing one thing, the megatree another, and micro trees another, the twig trees another, etc.  To do that, I would simply go down to that element group and manually put in a set of ramps to create effects, then move on to the next element group.  While that was fine for 32, 64, and eventually 192 channels that I had, there is no way I'm doing that with the 25k to 30k that I have now.

So, when I switched over entirely to pixels & RGB, and using LSP, I was interested in doing the same thing, but using the power of macros (that is the one thing LSP has that makes it so powerful, so why not use it?!).  I created a layer with all the twig trees, another with all the mini trees, another with all the micro trees, another for the megatree, another for the roof, another for all of the bushes, the small palm trees, the tall palm trees, etc. etc.  I first decided to start a blank sequence with that same 0.10 second grid that I was so familiar with in LOR software.  I then went and created a set of effects applied to the first element group using a macro and WHOA, it created a whole bunch of additional marks in the sequence!  And those marks didn't line up with the existing grid that I had created which means I now had, in a one minute sequence, 600 + 900 (15FPS is the default) = 1500 timing marks.  Well that stinks, so go back and create another blank sequence this time without the 0.1 second starting grid... create that first set of effects with a macro and I get 900 timing marks like I got before.  Ok, on to the next set of elements... create another macro and WHOA, why the heck isn't it snapping to the existing timing marks?!  There's already 900 perfectly good timing marks there, why can't I snap to those?!  And that is where the problem lies... yes you can go back in and try to drag those macros around to line things up perfectly... but you shouldn't have to... and not only that but 1 out of every 10 tries the program locks up trying to move those macros around... and if you don't line them up perfectly, the sequence doesn't run smoothly.

I ended up giving up sequencing individual element groups and decided to see what would happen if I created a macro applied to the entire display (since I can't remember if I have posted a link yet, the videos I ended up the season with are located here  https://vimeo.com/album/2723819 ).  I was successful (as seen in the videos) at getting the display to run out of LSP Scheduler with all 28,000 channels running out of LSP scheduler, but in order to do so, I had to create "whole display macros" using 10FPS (literally it had to be 10FPS, when I tried to create those same macros with 15FPS default it was very laggy).

I am running a Core i5 quad core desktop with 16GB DDR3 RAM and a top of the line (commercial) Samsung Pro SSD, 2GB video card and a fresh install of Windows 7 setup cleanly.  I have a Intel PCI-E gigabit NIC installed, and have a Netgear gigabit switch that sends the signals to the controllers.  I understand that I could run a Core i7, but other than that, everything is top of the line (maybe I need to contact IBM and have them build a supercomputer for me?!).  Anyways... that is why we are all here... the FPP should run this smoothly... but the sequencing is still up in the air...

Quote
In this video at the end I use whole house macros at the end 1:45.  http://vimeo.com/115857387  My minis and live trees are smart but programmed as dumb.  Despite of this, I think the effects carryover well.  Maybe that's what you are concerned with?  Until I tried it, I originally thought I would need for everything to be smart/discrete to pull it off...
I understand what you and arw01 are saying about having pixels in the mini trees, but using them as dumb and I don't want to do that for one reason... I want the ability to blink every light in the display and it is my understanding (could be wrong) that if you setup the element as dumb, you can't do that... I want the whole display to sparkle at points in the song...

-Louie

Offline arw01

  • Hero Member
  • *****
  • Join Date: Oct 2013
  • Location:
  • Posts: 836
  • Kudos: 0
Re: ISO: high channel count LSP musical sequence
« Reply #68 on: February 01, 2015, 12:08:38 PM »
I'm not advocating for setting them up as dumb, but using a probable work around with the extreme channel count to do some grouping of pixels in the controller itself.  E.g. group them as 2 or 3 nodes per "pixel", would drastically cut the channel count and might bring you into the realm where things are behaving as they are supposed to be doing.

I will copy your sequence over to my show computer here shortly and see if I see anything wierd with output on 32K channels, but I don't have much up still.  Might try changing your mega tree over to artnet and see if I can get some output on my "mega tree" but you have more nodes than I do.

Offline zwiller

  • Falcon Beta Team
  • **
  • Join Date: Mar 2013
  • Location: Sandusky, OH
  • Posts: 938
  • Kudos: 12
Re: ISO: high channel count LSP musical sequence
« Reply #69 on: February 01, 2015, 04:36:18 PM »
I AM advocating setting them up as dumb.  I spent 6 months testing stuff and minis don't really do much.  That said, sparkle would be cool BUT in the context of an entire display, no big deal. 

Same disclosure about "tones" here: Don't take this the wrong way...  Showed my wife the first video link and she like really liked your display.  I asked her if the show was more "sophisticated" than mine.  She said, NO WAY!  I explained to her that that the channel count was 3x that of mine ("layout is 3 times larger").  She said she can't tell.  Now, I can tell you are running 32k channels but it doesn't seem like it at all.   To me, it looks like 128 channels of incans, like Holdmans in the day.  Very elegant and stately, but not "pushing the limits of technology" looking.  When I see someone post they're using 32k channels, I am expecting to get my mind blown... 

Offline nutz4lights

  • Full Member
  • ***
  • Join Date: Dec 2014
  • Location: Melbourne, FL
  • Posts: 191
  • Kudos: 3
Re: ISO: high channel count LSP musical sequence
« Reply #70 on: February 01, 2015, 07:18:41 PM »
Same disclosure about "tones" here: Don't take this the wrong way...  Showed my wife the first video link and she like really liked your display.  I asked her if the show was more "sophisticated" than mine.  She said, NO WAY!  I explained to her that that the channel count was 3x that of mine ("layout is 3 times larger").  She said she can't tell.  Now, I can tell you are running 32k channels but it doesn't seem like it at all.   To me, it looks like 128 channels of incans, like Holdmans in the day.  Very elegant and stately, but not "pushing the limits of technology" looking.  When I see someone post they're using 32k channels, I am expecting to get my mind blown...
Its kind of hard to push the limits of technology when the setup won't let you!  What you are seeing in the display is "what works" which admittedly is not much   ;)

That being said, we all have different preferences in the way we do our displays, what effects we're after, what music we prefer...  Just so that you can understand what my display looked like at our old house when I was "in the groove" with everything working (192 channels LOR, around 75,000 incandescents and LEDs), here is a link:  https://vimeo.com/album/1493575   I typically only work with classical Christmas music and I am into the elegant, stately displays compared to the flashy, techno type displays.  Again, personal preference.

Part of the reason of getting into RGB was quite simply to reduce the light count hence reducing installation time.  I can check that box, never has the display gone up and come down so quickly!  The second reason of getting into RGB was to lighten the load on some of the elements... and give the option for multiple colors to some elements that simply couldn't support it.  I can check that box too.  So you would think that with two boxes checked, I would be happy with "dumbing" things down a bit, but I haven't been able to give up the dream of all pixels.

Folks that have been coming to the display for years get a kick out of the fact that there's only 16,000 lights on the house... I always get asked "are you over 100,000 lights yet?" from those that knew that there were 75,000 lights up back in 2010 and 2011 before we moved in 2012...  Shoot, 16,000 lights sounds so amateurish, but it sure is bright   8)

-Louie

Offline zwiller

  • Falcon Beta Team
  • **
  • Join Date: Mar 2013
  • Location: Sandusky, OH
  • Posts: 938
  • Kudos: 12
Re: ISO: high channel count LSP musical sequence
« Reply #71 on: February 02, 2015, 07:33:39 AM »
Louie, kinda freaky but both my wife and I used those exact words describing your display: "elegant and stately".  Your display is.  I get it.  My neighborhood is older and traditional and most displays are all white incans. 

People asking about your light count is boorish...  Who cares?!  That said, I did like your older vids, but I am the kinda guy wanting TSO's light rig bolted to my house...   ;D

Are you adding .1sec (100ms) timing marks now like you did with LOR?  You aren't by any chance doing 10ms?  That would be .01sec. 

 

Back to top