Exciting dev build, but be careful

Wow! Huge thanks to the Dev Team! After our extremely stressful 2015 season, we took the last nearly 7 months off from most things Christmas lights related. I've popped in here to DIYC on occasion and here I stop by today and discover all the work the Dev Team has been doing. I just installed build 192 and played our most effect intensive sequence including lip sync and graphics at 3 minutes long and it played via the preview flawlessly - not one pause, delay, hiccup - perfect!

I have a few questions:
-Where can I find the build notes at? The release notes that were included with this dev build appear to only go to 3.2u2.
-Can someone explain how the "layers" work.
-Were any changes done to the show scheduler? We have not gone to FPP yet (need a bridge etc.) so would be nice to not have to use xlights as our player.

Can't believe this will be our 9th year of having a synchronized Christmas display and have used Vixen as our sequencer this entire time!
 
There aren't really any notes except for those in the merge notes, and those are probably pretty cryptic for a non-dev. Seeing that dev builds aren't releases, we haven't yet made release notes. The features aren't necessarily done yet, we'll be compiling the release notes closer to release time. Alternately, you can do the same thing we do to generate the notes, go to the bug ticket system and do a search for all tickets that have been completed since the 3.2u2 release.
The layer system is still evolving somewhat. The best we have so far for documentation is my post earlier in this thread.
The scheduler has not been touched except for some cosmetic form sizing issues. And there are no plans for further development on the scheduler in the near future. The reality is that most people who want more than the scheduler can handle are already moving to the FPP platform for show playback. However, the performance gains you see in the editor also applies to the scheduler. So aside from the raw memory usage required for scheduling a long show, it will definately play more efficiently and smoothly than last year. Having seen your profile, I'd recommend investing the $40 in a pi and going the FPP route. It would make things a lot smoother for you. At this point in the year, you've got plenty of time to get that configured and get thru the learning curve. (Which is pretty low to begin with)


--Jon Chuchla--

Sent from my iPhone using Tapatalk
 
Thank you for the quick reply John. I must have missed the layers post, but I'll find it. On the FPP, we'll have to see if bridges are available. Last year there were none to be had. We are still running a lot of legacy controllers which means a lot of universes and the Pi couldn't run it. But we'll do some research and see what we come up with as FPP would be the best scenario.
 
Last I checked, the pi can run half a million channels, and supports pretty much any hardware you can throw at it. Definitely more than vixen supports. I haven't met a show it couldn't handle.

If you don't find that layer post earlier in this thread let me know. I may be getting my threads confused.


--Jon Chuchla--

Sent from my iPhone using Tapatalk
 
jchuchla said:
Last I checked, the pi can run half a million channels, and supports pretty much any hardware you can throw at it. Definitely more than vixen supports. I haven't met a show it couldn't handle.

If you don't find that layer post earlier in this thread let me know. I may be getting my threads confused.


--Jon Chuchla--

Sent from my iPhone using Tapatalk
Click to expand...

Pcmom's son and linux expert here!!

Our issue with falcon pi last year was one that was unique and I believe caused by the linux kernel.

As you may or may not remember we use renards that run at 19200 baud, meaning 64 channels per universe to maintain refresh rate. We had 9 universes last year. 9 usb to serials on the pi crashed the USB driver in the kernel. Which led to the FPP not working for us last year. I will be testing the new v1.8 fpp this week to see if the issue is resolved. If the issue is not resolved I will do what I can to get the issue resolved with the raspberry pi foundation/kernel maintainers as well as the FPP guys. Last year I simply ran out of time to even try to debug FPP any more, was up against a hard deadline to make it work.

Let me know if you have any other questions.
 
scoutcamper said:
As you may or may not remember we use renards that run at 19200 baud, meaning 64 channels per universe to maintain refresh rate. We had 9 universes last year. 9 usb to serials on the pi crashed the USB driver in the kernel. Which led to the FPP not working for us last year. I will be testing the new v1.8 fpp this week to see if the issue is resolved.
Click to expand...

The FPP v1.8 SD image is Wheezy with the newer Jessie kernel and firmware files so that may have solved your issue. We upgraded the kernel and firmware to support the newer Pi models but kept the stable wheezy image for v1.x. When we release v2.0, it will be a Jessie based image.
 
Your case specific issues are outside the scope of this thread. And we can certainly discuss it further but it should be done in a separate thread. But for everyone's benefit, I'll make these comments here:
We're certainly not going to be enhancing the scheduler for the purpose of making poorly designed setups work better. And we're not going to be testing or designing for profiles with more than one or two serial type devices. You should really be looking into correcting the design inefficiencies in your show design. 19,200 is a slow baud rate. You can go much faster without taxing any hardware. With Renard, faster serial means more channels per data bus. More channels per data bus means less USB dongles. You should be able to get hundreds of Renard channels per serial bus.

Pretty much all Renard controllers can also run Dmx hardware. By design, you can run 512 Dmx channels on a single port. You'd only need two Dmx universes to accomplish your 9x64ch Renard setup, and it will refresh much faster as well.
Even if you want to keep 9 serial busses for wiring convenience, you really need to be getting yourself a few bridges to do this for you. 9 USB anything plugged into any one machine is just asking for trouble. A design with more than one or two streaming USB devices needs something other than USB. That won't even work well on a full blown PC with multiple USB host controllers. That's probably why you need to keep your rate so slow. USB itself is your bottleneck.
There's now several products that will serve as sACN to serial bridges to one extent or another. Diyledexpress makes one, but is often not available, but most of the of these products are readily available most of the time throughout the year. Falcon, sandevices, pixlite, and even a few cheap Chinese products (that they shamelessly copied from designs born on this forum)
Bottom line is that it's now July, you've got plenty of time to correct your signal topology.
And even if you want to stick with the slow USB adapters, you can spread them out across several FPPs. The FPP itself can act somewhat like a bridge accepting sACN input and USB Renard output. The trick is keeping the USB devices per host down to one or two. I haven't run the numbers, but this may even be cost comparable to bridge devices.
And lastly, if you believe your scenario is unique, step back and ask yourself why it's so unique and why it needs to be. Usually when I find myself in a unique situation, it's because I didn't bother to step back and look to see if a better way already exists. Which ironically is exactly the reason we're not working on the vixen scheduler anymore.


--Jon Chuchla--

Sent from my iPhone using Tapatalk
 
jchuchla said:
If you don't find that layer post earlier in this thread let me know. I may be getting my threads confused.
Click to expand...

Found your post on Layers - Post #10. Good description.

CaptainMurdoch said:
The FPP v1.8 SD image is Wheezy with the newer Jessie kernel and firmware files so that may have solved your issue. We upgraded the kernel and firmware to support the newer Pi models but kept the stable wheezy image for v1.x. When we release v2.0, it will be a Jessie based image.
Click to expand...

CaptainMurdoch - good info. Our son is already troubleshooting and this appears to be the info he needs to get our FPP up and running.

Jon - on your last thread - we are aware of our limitations and planned to upgrade via bridges last year but they simply were not available. And of course we are a hybrid display with both traditional incans and Pixels and we continue to grow each year. Just getting started for 2016 after a much needed break, hitting the ground running!
 
Build 193 brings significant improvements to the preview editor when it comes to working with large matrix type displays. They're no longer clunky and hard to manipulate. This build also incorporates a number of other preview optimizations.
194 is a change in behavior for the new curtain effect. Note that this change may break curtain effects added in the past week or so because it changes how colors are defined. You may need to delete those effects and add them again. We try not to do this often, but it's only been in the dev builds a short while and the enhancements are worth it.
195 is a fix that eliminates the artificial lower limit for matrix sizes in preview. You can now draw very small matrices. This is useful for things like Coro canes where you can make a tall but very narrow matrix to represent it in preview.


--Jon Chuchla--

Sent from my iPhone using Tapatalk
 
Thank you so much for implementing layers. It replaces the last feature from Vixen2 that I really truly missed, Boolean Paste. I struggled with the inability to control the intensity of certain effects in V3 and even requested it several times. And now it's all come true! AWESOME!
 
Build 196 includes more optimizations as well as adding a new effect to Vixen, Plasma. This is patterned off of the effect of the same name in xlights. It's a native vixen effect and adds some additional flexibility and functionality.


--Jon Chuchla--

Sent from my iPhone using Tapatalk
 
Well I'm certainly a bit late out of the blocks so I will give it a spin. I managed to get 1 song sequenced so far and still haven't ordered my pixels. Just mailed Ray for a quote. Hoping to do 3 window frames and a front porch extension and maybe 2 arches.

Wish me luck!
 
Oh out of UK, might give Paul Zhang a ping too, a lot quicker on getting things done and out the door than Ray, but if you want custom, Ray normally won't mess those up as much as Paul's suppliers can.
 
Build 209 adds some new highly requested audio based effects effects to the main branch. These effects have been around a while in unofficial alpha form and are now tweaked and polished enough to join the main line effects.

This brings us down to 12 more tickets in the queue for merges before release. Some of these may miss the release cutoff. But it gives you the idea on just how much we've been working on. We're already at almost 40 builds since the last release and another 12 to go. Many of these are huge projects, others are smaller and more targeted fixes and additions. But nonetheless, this is probably over 1000 hours of coding going into this release.


--Jon Chuchla--

Sent from my iPhone using Tapatalk
 
Jon,

Awesome news. I just downloaded Dev Build 208 last night. I'll grab 209 tonight and play around with it. Can you expand on what the new "audio based effects" are?
 
The audio based effects are effects that respond to the audio in that place on the timeline. Think of effects that make a string behave like a VU meter or a matrix like a spectrum analyzer. They make the lights respond to the music.
Honestly i haven't worked with them myself in many months and it was Jeff that did the final polish on them recently. So I'm not the best one to explain the details.


--Jon Chuchla--

Sent from my iPhone using Tapatalk
 
Jon,

Sounds like some cool effects. I was actually thinking about adding some Pixel columns by my front porch this year (if time permits) and was envisioning these types of effects. I think I'll go ahead and add them into some sequences and play around with them.

Keith
 
Awesome! I have three columns on the front of my porch, previously the fire effect was the close as I could come to something like that. I'll shall play with those audio based effects tonight!
 
I put the effect on each pole separately if I did it on the group it came out dif so you might wanna play with it like you said I didnt relaly play with the setting too figure them out yet
codec101 said:
Awesome! I have three columns on the front of my porch, previously the fire effect was the close as I could come to something like that. I'll shall play with those audio based effects tonight!
Click to expand...
 
You must log in or register to reply here.
Back
Top