Enabling a 2nd universe to Artnet creates jitter

[nmaddix]

I'm having the following problem... I have universe 1 configured for Artnet output. Works fine. I have 72 LEDs connected and they go blinky exactly as I want them to.

As soon as I tick the Output box for the same artnet network on another universe, the problem occurs.

The pattern of blinking LEDs goes from orderly and exactly as designed to jittery and stuttering every so often.

E.g. instead of lights sequencing 123412341234 they go 123343412234 with the time delay between them sometimes varying.

Works fine in other software with multiple universes. Any ideas?

QLC v4.12.2 on Max OSX

[GGGss]

Wifi or hardwired?
Are you unicasting? or broadcasting? -> See options inside the Artnet options.
I suggest unicast because broadcasting can overswarm the network.
And wifi access points have filters built-in that omit broadcasts to protect available bandwidth.

[edit] and are you sending the 2nd universe to another Artnet-node-universe combo? [/edit]

[nmaddix]

Hardwired. Not sure about unicast/broadcast. The only options I see are Full and Partial. It works as described on Full and when I select Partial it's a much bigger mess, with lights coming on that aren't even in the program as well as severe timing issues.

[nmaddix]

[edit] and are you sending the 2nd universe to another Artnet-node-universe combo? [/edit]

The 2nd universe has nothing configured except ticking the Output box for the same Artnet network IP as the 1st. I don't even have anything set up to use that universe yet.

[GGGss]

In the output Artnet options:
Multicast = x.x.x.255 as the destination address (risk of broadcast swarm control inside you switch / router)
Unicast = x.x.x.y <-- use this one where y is the Artnet interface you want to talk to.

Enabling the 2nd universe ... make sure you set it up all right Artnet-node-universe combination.
If 2 QLC+ universes were to send signals to the same A-n-u combo, who will get priority then?
Even if nothing is set up in the 2nd QLC+ universe - there are keep-alive-packets transmitted.

And remember that Artnet universe 0 is Artnet universe 1 on devices.
[edit] (there were cases where this was true - not sure anymore now) [/edit]

[nmaddix]

Ah yes, I had tried specifying the IP address specifically, thus making it unicast. Same result.

My LED controller can control up to 4096 LEDs, thus needing several universes sent via the same Artnet network. Is there another way to set this up to avoid collision and priority issues?

Maybe I don't fully grasp your meaning behind "Artnet-node-universe combination".

[GGGss]

On the controller, you can drive 4096 LED's,
Somehow they have to be divided into universes - otherwise, you couldn't address them.

So my assumption would be that 8 universes are used. 4096/512
but then the groups of channels vs physical leds would be broken since only 170 leds can fit inside 1 universe. (512/3 = 170,67)
making channel 507 the start address of the last led inside one universe.

This would explain why you complain about RGB panels not doing what it's supposed to do.

Now the flicker and jitter thing:
In QLC+ you have to instruct where it has to send its info to.
Artnet-node-universe combo is the combination of Artnet-node (led controller in your case IP-address) and universe (which Artnet universe inside the controller).
In your case: Leds 1-170 are universe 0 (or 1 depending on how Artnet is interpreted) and leds 171-340 universe 1 (or 2) and so on.
In QLC+ you have to setup Artnet output to send to the IP of your controller + you have to tell it to send info to universe 0 to 8 (1 to 9).

Sending 2 QLC+ universes to the same controller with the same universe will cause problems, because of priority.

[nmaddix]

I feel I am understanding this just fine. It's quite logical. Only the last line is confusing:

Sending 2 QLC+ universes to the same controller with the same universe will cause problems, because of priority.

What do you mean by same universe? It's set to use different universes. This is a standard H802RA LED Artnet box. It works in other software just fine without jumping through any hoops during setup. Literally plug and play, 2 32x8 panels in series, no jitter, both working fine.

[GGGss]

That is because other software scans the Artnet line for devices and does a quick patch.
But with your panel setup to work, you will have to tweak this other software too.

BTW thanks for letting me know the controller you work with. This controller will only send to 680 pixels (2040 channels) and not to the 4096 channels in DMX.
Spec's and sample setup: https://f00.psgsm.net/p/870451/H802RA_user_manual.pdf

[nmaddix]

I don't understand what you're trying to tell me. Can you think of any way to get rid of the jitter in QLC+? If not, I have to move on and spend my time on some other software.

[GGGss]

Thanks for appreciating my time with you.
When 2 devices are talking at the same time - who will be heard?

As said before: you have to change the parameters of Universe 2 so that it will send to Universe 2 in Artnet !!
Thank you

[nmaddix]

Thanks for appreciating my time with you.
When 2 devices are talking at the same time - who will be heard?

As said before: you have to change the parameters of Universe 2 so that it will send to Universe 2 in Artnet !!
Thank you

I do appreciate you trying to help. I think there might be a language barrier since on my end it feels more like you're telling me that I don't know what I'm doing. I'm sorry if I'm not understanding your suggestions but I'm doing my best and we appear to be going in circles.

Regarding your suggestion of changing the parameters of universe 2 so that it will send to universe 2, as far as I understand and feel like I have described in great detail, that is exactly what I've done.

Attached is a screenshot of the artnet configuration.

Note that I've got some other LEDs using Universe 1, so the LED panel I'm referring to is using Universe 5 and 6 (4 and 5).

If I've missed the mark could you kindly explain to me specifically where to look and what setting to change.

[sandinak]

Moving discussion from Reddit . Reading the comments above I believe GGGss has the right of it .. but to expand a little:

• Given "The issue specifically is that when I enable a 2nd universe on the same ethernet adaptor, I start to get jittery output on the first universe, even without sending any data to the second universe." This means you're sending 2 U to the same ArtNet Target U.
• So assuming you have QLC U 5 sending scene values and U 6 sending nothing; and both pointing at ArtNet U 1, the way ArtNet works when things are configured this way .. is that the target U will get both the scene values from U 5 and 0's from U 6 at approx 44Hz ( DMX Speeds ) And so get the "jitter" you're experiencing.
• this got me here https://youtu.be/SdmkEUZQ8zo?t=755 .. out of focus i know .. but if you look at the bottom step second from the left when the flag appears .. you can see a flashing ( it's in rows 3/4 of that matrix ) where exactly this happened to me because of a misconfiguration.

So QLC:

• Generally ( and by Default ) : ArtNet U = QLC_U -1 ( so for Universe 1 in QLC the ArtNet universe is 0 ) when you configure things.
• ArtNet U doesn't have to track (QLC_U -1).. just has to be unique when pairing them up. So since you're using QLC U 1, 5 and 6 .. You could still use ArtNet 0, 1 and 2.
• *generally* does a good job of handling sending the same fixture two different DMX data concurrently .. but especially with matrix it can give a poor result so potentially you could have that issue

A quick thing here would be to send a copy of your showfile here .. and i can look at it and help get you straight. however as noted:

• select the ArtNet output by source IP
• click on the wrench/screwdriver to open the ArtNEt config
• for the output for the first U .. the QLC Universe is 1 .. ip is broadcast or (recommended) unicast and ArtNet Universe is 0
• for the output on the second U .. the QLC Universe is 2 .. ip is same as above and ArtNET Universe is 1
• you can use Full or Partial if the target device supports it.

the important thing here is to make sure your ArtNet outputs and inputs are matched 1 to 1. Happy to help .. reach out any time.

[nmaddix]

Thanks for the details. I really feel like GGGss explained it the same way and frustratingly, I really feel like I've done exactly what is suggested. You can see in the screenshot I've posted in a previous message that all my universes are mapped to different Artnet universes. I don't see how anything could be expected to work otherwise, in fact.

I've also attached my project file in case that helps. Am I being obtuse and missing something here?

By the way, I'm patching in Artnet universe 5 and 6 (4 and 5) instead of 2 because I've plugged some LED panels into the 2nd output connector on my H802RA box. The second output connector is universe 5 (4). The panels work fine, apart from the jittery timing, so I know they are connected correctly and mapped to the right universe.

[sandinak]

Looked at the file.. here's the config I see:

Code: Select all

   <Universe Name="Universe 1 " ID="0">
    <Input Plugin="MIDI" Line="1"/>
    <Output Plugin="ArtNet" Line="4">
     <PluginParameters outputIP="192.168.6.11"/>
    </Output>
   </Universe>
   <Universe Name="Universe 2" ID="1"/>
   <Universe Name="Universe 3" ID="2"/>
   <Universe Name="Universe 4" ID="3"/>
   <Universe Name="Universe 5" ID="4">
    <Output Plugin="ArtNet" Line="4">
     <PluginParameters outputIP="192.168.6.11"/>
    </Output>
   </Universe>

That's what yours looks like now .. which maps U1 -> AU0 and U5 -> AU4 ... And to be clear about your intent here:

• U1 -> AU0 - the 8 - 36 Channel LED Bars which are connected to H802RA Output Channel 1 .. I know this because it's going to the H802RA IP
• U5 -> AU4 - first Matrix is connected to the same H802RA Output Channel 2 .. I know this because it's going to the H802RA IP
• U6 -> AU5 - second matrix is connected to the same H802RA Output Channel 3 .. tho not configured in the file you sent me
• you have configured the H802RA like this:
  • Channel 1 - AU 0
  • Channel 2 - AU 4
  • Channel 3 - AU 5 ( maybe? )
  • Channel 4 - Unused

Found the config guide for the H802RA here: http://www.myneopixel.com/u_file/1906/f ... 2c7ba9.pdf
So some thoughts:

• Instead of creating extra Universes to move the target ArtNET out .. you might instead just create U0, 1 and 2 .. and then remap the output Universe via the ArtNet control panel so it should look something like:

  Code: Select all

  <PluginParameters outputIP="192.168.6.11" outputUni="4"/>

• if you only have 1 H802RA .. which given the IP address mapping that appears to be the case, why are you jumping universes from 1 -> 4? Looking at the config guide for the device it appears you can only set the offset for the first U/Channel .. which by default is 0 .. and then the rest would be 1,2 and 3
• given the H802RA has 4 channels .. and you're using AU 0 on channel 1 .. going to universe 4 would be outside the default configuration of unit .. eg you would have to change channel 2 distinctly to AU 4.. and I don't see how you can do that from the config guide . Similar to question above .. why the jump?
• I have seen strange behavior from devices like this when you unicast transmit data that is outside their configured ranges .. eg if this H802RA is using the default config .. then AU4 is outside it's range of 0-3. It *should* not care .. but some devices don't behave as you might think.

Over all my thinking here is to simplify as much as you can the configuration relative to the LEDs so that you're using default setup/config whereever possible. I have found .. especially with third party devices they work best when you're using them with defaults. I'd suggest:

• Move the matrixes back to QLC U1 and 2
• Use the default config of the H802RA as designed
• test.

hope this helps a little more .. I didn't see anything else in your file that was odd or mis-configured from QLCs POV.

[GGGss]

[edit] ^^ Sandinak already did some digging while I was investigating ... here is what I also found as being wrong or questionable[/edit]

Observing your screenshot ... you know you will have to use offsets in the led-studio program?
You can only use 4 universes with the controller - so sending U5 & U6 to the controller would need for connector 2 to have an offset in led-studio.
The same goes for connector 3.

Now your project file.
All but 2 RGBMatrix functions are wrong. You really need the exact layout inside RGBMatrix as being 8x12px. If you want to extend to 2 panels, then you need to adjust your panel setup + functions accordingly. I see RGBMatrix functions addressing only 6x12px and another 8x32px... this possibly leads to an unwanted outcome.

I suggest you create a new project and precisely watch parameters, setup, and functions. And please - keep it as simple as possible so that most faulty bits aren't in the same equation. Debugging something from a distance is difficult enough.

So 1 panel to start with, on connector 1, no offsets, no transferring Universes ... clean 1 to 1 situation.
After positive results, you can start extending.
(tip: to make a tidy end of universe conversations, put 1 dimmer on channel 512 and set it to 0 - now the Dmx-chain is terminated by a sturdy '0'-value and 512 value-packets are sent per Universe).

Good luck and report back.

[nmaddix]

My god guys, this is insanely complicated. I promise you I did not change the H802RA configuration one bit. It's being used factory stock. I plugged my panels into the 2nd port so that I wouldn't have to tie them to my existing LEDs and then played around until I found that the second port starts at Universe 5 (4). As far as I know that is how the unit works. Obviously documentation on it is poor.

I'll look at this more closely when I recover some brain cells from this project. I need a rest from all this to look at it with fresh eyes.

What's especially frustrating is that this works flawlessly with all other lighting software I've tried, in the exact same configuration.

Thanks both for your feedback so far!

[nmaddix]

I will do a new project from scratch to simplify as much as possible. But one thing I want to address now to clarify.

All but 2 RGBMatrix functions are wrong. You really need the exact layout inside RGBMatrix as being 8x12px. If you want to extend to 2 panels, then you need to adjust your panel setup + functions accordingly. I see RGBMatrix functions addressing only 6x12px and another 8x32px... this possibly leads to an unwanted outcome.

I don't see what's wrong with my matrices. I literally have physically in front of me 8 sets of 12 LEDs wired in series and attached to port 1 of the H802RA. Separately I have 2 32x8 panels wired in series and attached to port 2.

And again, the programming works exactly as expected except for the jitter. QLC addresses every light exactly as it should. It worked great until I added the 32x8 panes to port 2 and enabled the universe to be able to send data to them.

Back to the drawing board...

[nmaddix]

Ok quick update, I made a new very basic project to simplify. Uploading the project file and videos of the 2 LEDs mapped in action with and without jitter.

Jitter happens as soon as I enable universe 2.

Hoping that sheds some light on things (pun intended).

Correction, the video files are too big to upload to the forum so here are dropbox links:
https://www.dropbox.com/s/lapvrbiow6gr0 ... r.mov?dl=0
https://www.dropbox.com/s/platnlfipyrtq ... r.mov?dl=0

[GGGss]

oke ... hard nut to crack...

In your new project-file you are sending now to U4 on your interface TWICE... we told you not to.
And again - why U4; it should read U0 - we cannot predict what the interface is going to do when you send info outside of the scope.
The scope being U0-3 for DMX = headers 0-3 (or 1-4 depending on labeling)

The jitter really looks like a stay-alife packet coming in. A network sniffer could confirm that; be you don't need that because QLC+ is sending both universes to the same output Universe.
Please change that