So here are the folder with all the files used : https://drive.google.com/file/d/0B1ZKYtpB0vu2dkM2eDRyX1hranc/view?usp=sharing

 

EOF doesn't give me any warning except for the fact that i haven't created FHP. (using 22/05/2017 hotfix)

 

The toolkit doesn't throw any error. (v2.8.2.0-273b0ba2)

 

Yet whatever path i choose to play ingame, Rocksmith crash right after tuning and i have no idea what could be the cause.

 

I regenerated the wem.

Deleted showlights just in case.

With or without DDC, i still get the same crash.

 

So if anyone has an idea, i'm all ears!

 

Just working on wwise 2016 integration, I could look at it after, but nothing I could promise <____>

The two CDLC that I updated today also crash after tuning when I build them with the latest version of the toolkit. I haven't had time to look into it more closely, but the culprit are the showlights (yet again). Using the showlights file from the previous, working version of the CDLC, they don't crash.

 

Try using an older revision of the toolkit and see if that works.

At one point i think i will ask if i can create the showlight myself if it's that much of a problem to have it automated, this is starting to become quite a nightmare to have this thing just randomly starting to crash everything...

So yeah, used the v2.8.2.0-3c7d4efa of the toolkit and everything works fine...

@@firekorn, not a big deal we could go back in time with statically empty show lights and it's likely to make more mistakes by hands than with algorithm I all we need is figure out bug and write test so we won't get regression in future ez as abc but who cares anyway

 

And thnx for the safe point so we could trace it faster with simple diff on those show lights anyway showlights just love to ctd ewerything if it'not initialized before 10 seconds may be the case with dad and all it should generate off reference track which is full arrangement but even then it should be validated for initial values venue fog color an stage lights color.

It's simple just like that

 

 

Gimme ideas for showlights and maybe we'll put edit form for them in blocklike style so you'll see your initial colors for lights and song's sections or phrases with key note +color suggestion based on rs default scheme for the lights, right? But I that requires more research than just fixing current generate function imo

@@Alex360 It's just that showlight creation has been a bit inconsistent, no offense to you guys who make the toolkit i know you tried to fix an issue but the fact it resulted in another bug is a bit enraging sometimes for the creator (especially when we have no clue why things go down).

 

Also while i got you, would it be possible to have a new stable release (maybe using the v2.8.2.0-3c7d4efa since i don't think it has any major flaws as far as i can tell) so that the stable release of the toolkit can handle abridged xml from EOF.

 

As for the showlight idea, it's something that i had in the back of my mind for a bit of time but since i don't have the venue mode activated, i'm not bothered with whatever the result is from the algorithm but i do think it would be cool if any charter could have control over it if he wants too. But that's not a priority in any way over the basics of the toolkit :D

I did some testing with my two crashing CDLC.

 

In the first showlights file, there were no notes from the 42-59 range. I tried adding one 42 at the beginning and the crash was gone.

 

As for the second showlights file, there were two notes from the 42-59 range near the end. 43 and 45. At first I thought they came too late in the song or something, but I tried adding the note 55 (which I happened to choose randomly) between them, and that got rid of the crash.

 

Then I ran search in the showlights XML files from the on-disc songs, imported RS1 songs and DLC and the RS2014 DLC I own. While there are 1207 hits for the note 42, there are zero hits for the notes 43-47.

 

Editing the showlights file and just changing the note 45 to 55 fixes the crash.

Showlights is a bit of a black box and has been hit miss depending on the song.  We devs have been trouble shooting this as iminashi has done for a long time.  Thanks again @@iminashi for the help/sugestion.  I will revise next beta so that showlights excludes 43-47 range and make sure it starts with 42.

 

Footnote:  The issue with @@firekorn CDLC is because notes are appearing before 10 seconds which current showlight generate did not handle/accommodate.  Will fix this in next beta.

I'll need to do some more testing and fiddling with the showlights, but here is my uneducated guess on why it crashes.

 

The total number of showlights and how early they need to be set doesn't seem to matter. In official content the first showlight note is almost always at 10.000, which is where the beatmap begins in most official content, but the game doesn't crash if there are no showlight notes at the beginning. And the minimum number of showlight notes you need seems to be 2.

 

There needs to be at least one note from the fog color set (24-35) and one note from the spotlights/effects set (42, 48-59). The game loads them into tables or lists or whatever and the programmers hadn't prepared for the case where they have zero entries, causing the game to crash.

 

I have tested the following showlights files and they didn't crash the game for me (the notes are set at a time that is very near to the end of the audio): 

<showlights count="2">
  <showlight time="63.455" note="32" />
  <showlight time="66.646" note="55" />
</showlights>


<showlights count="2">
  <showlight time="66.646" note="24" />
  <showlight time="66.646" note="42" />
</showlights>


<showlights count="2">
  <showlight time="66.646" note="35" />
  <showlight time="66.646" note="59" />
</showlights>

@@cozy1 That's an interesting information, the time 10.000 correspond to what specific times in seconds/milliseconds?

Time is in seconds for showlights.

@@iminashi

Have you seen ODLC where two showlights notes appear at the same time as in your example?

 

count="2">
time="66.646" note="35" />
time="66.646" note="59" />

 

I ask, because I have not.

 

The important aspect of your discovery, is actually what the toolkit has attempted to do in the code, i.e, makes sure that showlights begin with a FogNote followed immediately by a BeamNote.  This seems to be the key to initializing the in-game showlight generation properly (which prevents in-game hanging).

so @@iminashi is saying that all official DLC starts around the 10 seconds mark? damn i thought it was starting around like 6/7 seconds... I might have to change some stuff i do then.

so @@iminashi is saying that all official DLC starts around the 10 seconds mark? damn i thought it was starting around like 6/7 seconds... I might have to change some stuff i do then.

 

There is a little over 10 seconds of silence at the start of an ODLC and the beatmap usually starts at exactly 10 seconds. Try importing some of them into EOF and check for yourself. But in the game it doesn't take 10 seconds for the song to start. I've been wondering how that works.

@@iminashi i will investigate cause i'm intrigued by this.

@@iminashi

Have you seen ODLC where two showlights notes appear at the same time as in your example?

 

<showlights count="2">

<showlight time="66.646" note="35" />

<showlight time="66.646" note="59" />

</showlights>

 

I ask, because I have not.

 

That was actually a mistake I made as I copy pasted, but they are out there.  For example "All I Wanna Do" starts with 

  <showlight time="10.000" note="24"/>
  <showlight time="10.000" note="52"/>

@@iminashi - Regarding the wait time when playing not equaling the amount of leading silence, I've noticed that in some customs I've made that if I start the beatmap fairly far into the song (which sometimes happens if there's a piano intro that's totally untabbed) that when playing it in game it will skip to just a few seconds before the beatmap starts. I haven't tried checking to see exactly how much of the "unmapped" audio gets skipped, though, so I can't tell you anything more specific regarding that.

Showlights seem to start at 10 seconds regards of the actual start of music.  You will notice that the showlights/stage iris opens up around this time regardless of when the actual music started.  I have tried to start showlights sooner but it does not seem to work.  This may be hard coded into the game.

New toolkit beta is available with revised showlights for testing.

@@cozy1 You released a new version just as I was writing a reply.  :)

I'll need to test the new version to see if what I was about to write still stands. One thing I noticed with the code is: shouldn't rnd.Next(48, 59) be rnd.Next(48, 60) in order to get a random number between 48-59 (ditto for 24-35)? Or maybe I'm reading the wrong documentation...  :)

 

 

Anyway, I did some more testing with the showlights and here are my conclusions:

 

42, 48-59 control the spotlights/beamlights/whateverlights.

42 is "off" and 48-59 are the colors in the same order as the fog colors.

 

67 turns on the laser lights and 66 doesn't appear to do anything.

 

About 67/66:

 

If there is no note 67 in the showlights file, the laser lights seem to be on all the time. In ODLC both 67 and 66 are used in each song exactly once, with 67 always coming before 66. The one exception in the DLC that I own where neither are used is Megadeth "Peace Sells". Maybe 66 is supposed to be "turn laser lights off" but doesn't work? In ODLC it often comes at the very end.

@@iminashi. I also unable to reproduce stroboscope effect from one of Rise against songs It would be cool to get one!
I guess it's binary masks but it's ok to operate them in numerical form so not a big deal here.

@iminashi  Thanks for the details and blackboxing the note numbers and what they produce.  I will make the changes per your notes. 

 

It is interesting that the showlight file only needs two elements to work, i.e. a fognote and a beamnote.   The game seems to produce a default showlight effect when this is the case because it looks and plays fine.

 

Is there a corresponding off note for fog like 42 is for beams?

 

random .Next(min, max) is inclusive was the way I was reading it which is wrong.  You are correct the upper limit is exclusive.  Thanks

@@iminashi. I also unable to reproduce stroboscope effect from one of Rise against songs It would be cool to get one!

I guess it's binary masks but it's ok to operate them in numerical form so not a big deal here.

 

Which song is it? I'd like to see it myself.  :)

 

 

Is there a corresponding off note for fog like 42 is for beams?

 

I don't think so. There are zero hits in the XML files for notes 0-23 and 36-41.

For anyone interested, here are screenshots of the showlights:

https://drive.google.com/open?id=0B_0HDnEbZi4PTWF1bVl2SXdZVlE

 

There are two screenshots of 35. I have no clue why the first one (the glitched one) was green + purple.

In the screenshots for the fog colors, beams and laser lights were off. In the screenshots for the beams, the laser lights are on and the fog color was 25.

 

The exact names of the colors aside, in the note key in the toolkit showlight editor, 33 and 34 should be switched. 33 is a blue color and 34 is a yellow color. 

 

Also, the exact number and type of lights affected by 42,48-59 depends on the venue. My screenshots were from a small venue, but in some venues they also affect the color of a moving light/lights.

@@iminashi
Where these test run with showlights xml that contained just two elements (fog and beam)?  I am trying to reproduce results.  I am finding by T&E that the sequence of notes seems to have effect on which lights are on and off.  Is this your experience?