Sean Meighan
Software => Bugs in xLights/Nutcracker => Topic started by: bpducman on June 14, 2015, 07:38:28 PM
-
Hi Guys,
I know you guys are busting your ass trying to get all of this smoothed out BUT.
Opening an existing sequence that ran great in v3, some effects don't appear in v4. The timing is there and the effect is partially there and it seems to be associated with fades some how. I can go select that spot and remove and or add vertical fades and the effect appears to come back, or tell the effect to be normal and not effect 1 or 2. When I compare running the sequence in v3 and then in v4 there is no comparison v3 has it beat. I OH so want to use v4 but my brain is not as quick as some and I feel I am running out of time. If you would like me to send you anything just let me know
Thanks for all of you hard work and I know when this is finished it will be tremendous.
Brad
-
Things have changed so you might need to make some adjustments especially if you had effects relying on speed or fit to time. What effect are you trying to use? There's isn't any effect called a fade.
-
Hi Gil,
I understand the possible need to fine tune things. I will get you screen shots tonight after work.
I use a lot of vertical fade and horizontal fade when added on to things like color wash and for the most part it seems that if I have this VF or HF in the very beginning of the color wash effect the effect does not show up when opened in V4.
Sorry my terminology is self centered
Brad
-
When I compare running the sequence in v3 and then in v4 there is no comparison v3 has it beat.
Can you explain what you meant by has it beat? Just curious on the differences you see.
-
First I don't want anyone to get the wrong idea. I think V4 is absolutely where this software needs to go, and I am speaking from a user stand point not a programmer standpoint.
The software is still in Beta and I keep telling myself that.
Beat is the wrong word it is not a contest. And the only reason for the comparison is because when I run this particular sequence in version 3 it is to me complete. It is what I envisioned while was using the tools to make the effects I had in my mind. Then when I run the sequence in v4 things are missing and these are probably just Beta tweaks. But once the v4 is complete I would hope that you could not see any difference in the sequence if run in v3 or in v4.
Generally in my (opinion) any software ie: cad, photo, word processor. The new version should be backwards compatible and allow you to open the previous version file and have it look just as it did in the prior version and I would bet that Sean, Gil, Dan and all of the programmers are striving to accomplish this.
There are so many ++++ with V4, Timing marks flexibility, importing many forms of sequence data especially LOR and Super Star files, Visualization of models and whole house. The list keeps getting longer.
I, Me, Brad, had just got pretty good at using some of the tools available in V3 and I have not spent the 1000 hours in V4 like I had in V3 using the tools. So I am in the learning curve and it becomes frustrating when I see some of the info I did in V3 not show up V4.
Because of my "work load" I am still on the fence for the next 3 weeks before I have to comment to V4 or stick with V3 for this year. I will not have the time to fix the little things in V3 sequences and create 9 new sequences for this year. Along with dealing with the things life tends to throw at us.
Hope this helps.
Mr. Gilrock I will now open the same sequence in V3 and V4 so I can maybe give you the information you need to point you in the correct direction.
Brad
-
The sequences are not going to 100% backwards compatible. I think we are done with the v3 conversion code and if it's not perfect then so be it. I'll help show you how to achieve something you want to do but I'm not going back and adding a bunch of code to try to do a perfect conversion. In some cases the old code was broken or we just didn't like the design.
-
Got it your the man.
Here is a print screen. Look at the 21 sec area. I have no effect showing. If I press the vertical fade under the color wash then some of the effect comes back on. Just FYI.
Brad
-
OCD Strikes again.
-
Did you recently convert that v3 sequence or was it converted awhile back? I ask because we did have some conversion bugs several weeks ago that would cause major problems. Also I'm wondering why I see OFF effects on your timeline. If you attach your original v3 sequence and your latest xlights_rgbeffects.xml I can look at it tonight.
-
Here is the original xml and the one I converted tonight. The original sequence is to large to attach
Hope this helps
Brad
-
I need your xlights_rgbeffects.xml for the model definitions you use.
-
Here is the one from last year and the one from this year always a work in progress
-
Your right there is some strange-ness going on with that sequence. I'll try to figure out what's going on but I see where an effect background looks fine then you click on it and the background disappears. What's probably happening is there are new settings that are undefined and when you click on them they just to a default value. I'll let you know what I find.
-
I haven't played with those fades before but it appears to me that vertical fade is broken in the colorwash effect if you use the Effect1 blending mode. We will need to look into that but it appears that it works OK if you change the blending mode to Normal which adds alpha support to the color blending.
-
Thanks for looking at it. I used the color wash with fades allot. You are looking at 1 of 14 sequences that have that same association.
Brad
-
Dan already checked in a fix for the vertical fade. He was more familiar with the colorwash effect so I let him know about it. The only issue I see left is although the effect works properly its not displaying the background on the grid when you use the horz or vert fade. I believe it's because it looks at the first row and if you use a fade that first row is all black. There is some new logic displaying that effect background that I wasn't familiar with so I'll see what Dan says about it.
-
You guys are Great
Thanks
Brad
-
Thanks Gil and Dan. It is working great. I am off the fence.
Brad
-
Thanks for finding the problem. I initially thought the problem was something to do with the effect settings changing for speed and fit to time but nope it was a real issue that needed to be fixed.
-
Just wanted to let you both know that I just spent about an hour adjusting timing and changing some of the effects in that sequence and had I done the same thing and I am not sure I could have done the same thing in V3, it would have taken 3 or 4 times the amount of time.
Thanks so much to all of you programmers you are making me look good
Brad
-
Hey Gil,
When I bring in Brian's 12 barr CCR Deck the halls using your import drop down some of his information is not getting into XL. Well I should say it is getting in it is just not operating. If you don't know the sequence you can view it from his site or look at my videos last year. It is the swirling effect around 32 sec is when it starts.
I have not found the missing link but am looking. This maybe one of those effects that super star does really well and we are just not there yet.
Let me know if you want to look at anything.
Brad
-
Well take a look at what effects are ending up on the grid at that point in the song. It would have to probably be a Galaxy effect which maps to Superstars spiral effect. Make sure an effect is even making it in and that it's length is correct. Maybe play with the effect's settings and figure out that some part of the conversion isn't right.
-
Its me again,
Here is a screen shot showing the sup file and the fseq file. The 31-32 sec time frame notice the 2 red galaxy effects. This happens thru out.
You are the man with the brains here.
What do you think?
Brad
-
Mr Gilrock,
I can see U are very busy these days. Any chance you could maybe see why NC interpreted the swirls created in Super star the way it did?
Brad
-
Have you given me anything? I am busy so if you did I forgot. I can't tell anything from a picture. Have you made any effort to click on the effect it created and see if you can adjust it to do what it should. Then you can tell me what got interpreted wrong. Can you go inside the .sup file and find the flowy for that effect and paste the XML here?
-
ok I will but I just noticed that on the other 12 bar ccr's that I have imported in not all of the twelfth bar info made it. I am using the standard 50 pix 150 rgb from lor.
-
Well isolate what's not making it in. Both SUP and xLights XML files are viewable in a text editor. It's pretty easy to open them up and match up the effect times to see what they are being converted to.
-
Mr Gilrock,
When I open up a text editor it looks like something I would print out and rap around my head and go screaming down the street!
I have simplified this down to just the ccr tree and star. If you would view please and at about 41 sec. bar 12 should show from 1/3 down to down. You will see it or you will not see it. In some arrears of the sequence that 12th bar does show up. Maybe Dan should look at it.
What text editor do you use?
Thanks
Brad
-
I played this SUP file along with the video and it was almost identical. You posted the Silent Night file, but I think you mean to post the Deck the Halls sequence.
-
in the beginning of this I was talking about deck the halls. and I will get back to that
If you would though play silent night again in xl and when you get to 41 sec and the bars start coming down pause the screen and count the blue bars. Now go and play it in super star and count the bars.
and let me know if I am loosing my mind
Brad
-
The blue bars effect at :41 is a Morph in the NC import of SS. The X value is set to 4%,95% The x,y values for morph are percentages, not fixed numbers.
For my 24 leg tree, I made up a spreadsheet of the % value needed for each leg so I can start a morph on any leg I need to.
If you want the effect to appear on all your legs, just change the Morph to x1a,x2a=0 and x1b,x2b=100.
A tree with 100 legs would have the first and last 3 legs empty if using 4,95 as the x values.
-
I haven't had time to investigate but I do have one comment that may be relevant. In SuperStar you are allowed to position multiple effects on the same layer during the same frame. We don't allow that so we place the effects that appear first in the sup file on a layer and then if we find another effect we create another layer. It's possible that in superstar the second effect is drawing above the first effect and we are drawing them in the reverse rendering order which may make things look different. Sometimes I've manually fixed my sequences by clicking the effects and moving them to change the order of which effect is on top. It might work better if when we create new layers that we always insert them above instead of below. I'm just not sure. It's difficult when their program follows different rules than ours.
-
Actually 4 % and 95% sounds correct for a 12 ribbon tree. I attempt to calculate a percentage that lands in the middle of the ribbons' percentage range because I had boundary issues with the original method I was using.
-
Here is the direct comparison side by side. Looks good to me up to the first minute.
http://tiny.cc/3m31zx (http://tiny.cc/3m31zx)
P.S. I did not render before playing this so the effects converted to images did not show up on the preview.
-
I will add that while not all the effects come over 100%, it's pretty cool thing to see that Gil's code is able to bring in so much that it's more than good enough to use in a sequence. With some minor tweaking you can get it to look just like the original and like everyone else who has that sequence (BORING) :)
Interesting to see the acceleration differences where the morph in the NC sequence actually gets ahead of the SS sequence. Not an issue in my book.
In the beginning the overlapping balls didn't overlap in the NC sequence as they did in the SS, but that can be fixed in NC with the layering settings. All simple stuff to fix.
The big takeaway here is that you should be finding the good parts of these sequences and saving the good effects as presets and then use them in other sequences.
-
Yeah I pretty much had to create my own acceleration algorithm. I really didn't try to match superstar. And that reminds me after Dan got the node level viewing working I could see that my negative accelerations aren't working properly. I need to take another look at them and figure out how to fix it.
-
If you guys want to ask Brian for his acceleration calculation I'll make them match. :)
-
I can do that
Brad
-
I have made the request I shall let you know
-
By the way this is my accceleration adjustment code. The ratio is how far along the effect is in its duration. Then it runs into this routine to alter that percentage. So I'm basically taking the duration percentage to the power of the acceleration. I divided by 3 because it seemed too strong. What it ends up doing is slowing down the start of the effect. So you input 20% and you get out about 7.5% but by the time you hit 100% you are even because 1 to the power of anything is 1.
double RgbEffects::calcAccel(double ratio, double accel)
{
accel /= 3.0;
if( accel == 0 ) return 1.0;
else if( accel > 0 ) return std::pow(ratio, accel);
else return (1 - std::pow(1 - ratio, -accel));
}
-
I think I like it better when its magic 8)
-
One things that's pretty cool is drop down a morph and then expand so you can see the nodes and then adjust acceleration and watch what happens on the nodes.
-
Yea the whole thing is pretty cool. But I think you need to give me a little more info how do you drop down so you can see nodes I see where you can make acceleration adjustments.
-
Double-click the row heading for a model and it will reveal the strands. Double click a strand and it will reveal the nodes for that strand.
-
When I open up a text editor it looks like something I would print out and rap around my head and go screaming down the street!
What text editor do you use?
Thanks
Brad
We all pretty much use Notepad ++ for editing XML files like this. It will automatically organize and parse the files, making them much easier to read. You will probably just "shout and walk around the room."
https://notepad-plus-plus.org/ (https://notepad-plus-plus.org/)
-
I have Notepad++ but I also bought a license for UltraEdit which I really like.
-
I will give it a try
8)
-
Hey Gil this is my communication with Brian. maybe I did not ask correctly. but here it is.
me
I have 2 request if you could. Some how I have destroyed my sup file for wizards in winter I got from you. And I am trying to run your Silent Knight in Nutcracker and having an issue with the 12th ccr bar and I have been told that it is probably an acceleration problem. Is there any chance I can get your acceleration calculation. I really don't want to alter your sequence it is very special to all that have viewed it down here.
Brian
There should be no problem with the acceleration in the morphs. It exports as commands the same as the non-accelerated morphs. And accelerated morphs are not isolated to the 12th CCR, so I don't think that is the problem. Can you describe the problem you are having?
And when you say you are running the sequence in "Nutcracker", exactly what do you mean? Are you merging it with nutcracker effects?
me
So last year I was creating effects in nutcracker and then c&p into LOR very slow process file sizes were approaching a gig.
This year NC has developed a import file that sees the 12 bar ccr and brings it right in. When you run your Silent Knight @ about 41 sec in you have blue bars coming down from the red top. I see all of the bars being made right up to the 12th bar and it does not appear. Now the 12 bar does show up in other areas of the sequence. They tell me that they interpreted that effect as a morph and the timing has something to do with the 12 bar showing up. They do have a way to adjust the timing, start and stop points.
But Yes I am merging it in or I would like to anyway.
Brian
The biggest new feature in LOR S4 is the ability to export in the new "intensity data" format. And it includes the "Pixel Editor" which is really LOR's enhanced version of nutcracker. The "Pixel Editor" allows creation of the nutcracker effects at any time you want and then you can export them to the sequence editor in the new "intensity data" format and the exported files are smaller.
Superstar exports in the new "intensity data" format as well. And the sequence editor will automatically play the exported superstar sequence alongside the exported "Pixel Editor" sequence. So you don't have to do copy/paste anymore.
As for the blue bars that come down at about 41 seconds into "Silent Night," the blue effect that comes down is identical in all 12 of the CCRs. I don't know why the first 11 would work and the 12th one would not.
The rest of communication was just other things.
I know this did not help.
Brad
-
I'm sure Brian will be at the expo. Might be something to ask in person.
-
You'll have a better chance of getting the formula to Coca-Cola IMHO.
-
I think you mixed up two different issues. The 12th strip issue doesn't have anything to do with acceleration. Steve had commented on how the accelerations were different so I was just saying if I knew Brians internal algorithm on how he uses the acceleration value I might be able to match it. It was sort of a joke because I don't expect him to show what he's doing in his code.
If you want to describe it to him the way to say it is "xLights has created the ability to purchase your SuperStar sequences and import them into xLights and it converts all the SS effects into equivalent xLights effects that are editable."
He might not be happy about the ability we've provided because now you can create a lot of those same effects without buying his program. But on the other hand it increases the number of folks that may purchase a sequence from him and import it into xLights if they were not a LOR user to begin with. So it's hard to say whether the stuff we've added helps or hurts his bottom line. That's why he might not be very open to telling you how he uses the acceleration value.
-
Oh I am sure I mixed them up.
I am going to let this one just sit.
-
Well I figured out that 12th ribbon issue. The downside is I believe the correct fix is inside the math for morph effect on how it calculates which ribbon to use based on the percentage. The SuperStar import is doing it the way I want which is if you have a 12 ribbon tree each ribbon has about an 8% band. The problem with the way the morph math is working is that half that 8% band is lost at the top and bottom ribbon so it changes from the first to the second ribbon at 5% when it really should have been 9%. I've figured out the correct equation which will give each ribbon an equal proportion but the issue is existing sequence might have a morphs that could shift by 1 ribbon if they were at certain percentage values. Out of the 101 percentage locations 24 are impacted. The good news is the folks doing SuperStar imports will see an immediate correction to their positions. I think it's the right thing to do. If it's really a problem for someone I could write a script to do a one time translation from the old to new scheme.
-
It sounds like it is worth a shot.
Thanks
Brad
-
I already checked in the change and it fixes the SuperStar import. Now I'm trying to figure out if its worth it to develop code that detects and modifies all the morph positions. I hate having to have a big block of code that is only in there to fix up changes that happened during the beta phase but will be left in there forever.
-
I was trying to write some conversion code but it gets too complicated. The problem is I can't convert the percentages unless I know the model size. We store our effects in database lookup format so that if you reuse an effect with the same settings it takes up less space. So I would have to cross reference each morph which is possible but then I run into the fact that the same morph settings could exist on multiple models in which case it cause them to no longer be the same settings and then I would have to split it into two effect settings lookup objects. I really don't want to figure all that out. For most of the morphs that use 0 and 100% it won't affect them. For the ones in the middle it might shift the start point by on column or row.
-
I have a chart for my Megatree for morphs. I know what percentage values for X to set for any given leg of the tree. I would like to keep using those same values. Is there a bug that would require changing it up?
-
sounds like it is best just to leave it alone.
I am sure you have better thinks to work on right for now
And with LOR changing all of their data base it maybe better to wait and see.
JMHO
-
I have a chart for my Megatree for morphs. I know what percentage values for X to set for any given leg of the tree. I would like to keep using those same values. Is there a bug that would require changing it up?
Yeah the problem I found was the two ribbons on each end were only getting half of the percentage range that every other ribbon was gettng. And the halves that were lost on each end got distributed into the middle ribbons. The problem is worse for trees with less ribbons. Like for a 12 ribbon tree ribbon 1 only had 0%-4% but ribbon 2 had 5%-13%. Now that they are evenly distributed ribbon 1 will be 0%-8% and ribbon 2 will be 9% to 16%.
So yes you might need to tweak a few values on the chart. It will depend on whether your percentages were closer to the middle of the range for that ribbon or whether you had selected a value that was just barely on the edge of the range.
Dave and I talked about this yesterday and he thought we should change it and make the math right instead of finding it 2 years later and wishing we had found it sooner.
-
I have a chart for my Megatree for morphs. I know what percentage values for X to set for any given leg of the tree. I would like to keep using those same values. Is there a bug that would require changing it up?
Dave and I talked about this yesterday and he thought we should change it and make the math right instead of finding it 2 years later and wishing we had found it sooner.
I absolutely agree, it should be fixed now. We all know this is still beta and there are risks in sequencing shows until a full release.
James
-
Ok it took a couple hours but I wrote code that will convert your morph positions. Values of 0 and 100 won't be touched but any other value in between will calculate the old line it was on and then calculate a percentage to place it on that same line for the new logic. So the values will all end up in the center of the band. So like on a 12 ribbon tree column 10 was anywhere between 78% and 86% so any value in that range is converted to 79% which is close to the center of the new column 10 range of 76% to 84%. It actually calculates to 79.9% but gets converted to an integer.
-
So will that get incorporated into the next version?
Brad
-
Yeah its already checked in. If were setup to compile the code you could have it today. Actually I guess that could pose a problem. It would run the convert logic every time you open the file unless you know how to modify the version number in the build.
So I advise you do a backup before running the new version.