LDL (yeah I know): anyone else having the maximum light radius capped very very low?

My maximum light radius for tokens appears to be set to about 16 feet when using LDL all of a sudden, which is basically the end of my ability to use LDL. I know it won't be fixed but I'm really not looking forward to the massive CPU usage of UDL--has anyone else run into this problem and found a workaround?

Hi Alex! I haven't personally had a problem with CPU usage on UDL, though I know some have reported it. Have you tested? I'm not sure exactly what the issue you are having is, though. Do you mean that no matter what value you input, the light radius is capped at 16'? Or that no value above 16 is accepted? Have you tested in a brand new game to verify that some setting isn't getting in the way of the unexpected behavior?

Hey Keith! Thanks for the response. Re the Long-term solution of switching to UDL: I have and still do periodically. UDL seems to do fine for very simple setups when it's just me, but once I start having multiple light sources and walls and multiple tokens with vision (a party of 6 with 2 or 3 pets is just brutal!), it starts grinding to a halt for me as GM. The last time I gave it a shot, a couple of months ago, it also ground to a halt for a couple of players and then... somehow crashed itself and no other part of the VTT and showed them the entire map as though all DL had been disabled? It's kind of funny, and if I can't get LDL to keep working I'll head more to UDL when I really need it, but so far it has also been nothing but pain. Of course, it's not like LDL has been perfect, avoiding self-closing polygons is also a pain. This particular thing: Here, what's happening is that there's a brand-new artificial cap on the light a token can emit, beyond whatever settings we can control. I haven't done a ton of testing yet so if there are edges where this behavior starts and ends I haven't found them, but that's also kind of what I'm asking about. Here's one example, where I want to fill a map with dim light and the 180' of dim light is capped: You can kind of see the circles here; the settings dialog is open for the orange circle you can see, which is on the token layer. It should be filling the map with dim light, but it is... much smaller than that. You can kind of see another circle underneath that dialog where I was testing a copy of said orange token on the DL layer, with bright lighting--it behaves the same, although you can tell something  weird is going on. It's kind of interesting in an academic sense; the light radius expands and keeps its feathered edges up until it hits that cap, at which point the feathering stops and you get a hard edge of light as though the light is still being emitted properly by one system , but a second part of the DL system is capping it or overriding it or covering it up or something.

The only thing I can think of that would give results like that on LDL is if the page grid was set for something other than 5 feet per cell.

keithcurtis said: The only thing I can think of that would give results like that on LDL is if the page grid was set for something other than 5 feet per cell. Yeah, that's not it, and it's definitely a cap, I can grow or shrink the light source up to that 16-foot-ish cap. Although you do raise a good point! Changing the measurement style has no effect, and changing the grid scale has no impact either. I'll keep an eye on this but if you haven't seen a workaround, I suspect one hasn't been found yet. Thank you for following along and chiming in!

I did load up an LDL page, and wasn't able to duplicate that cap; I was able to illuminate the whole map. Did you test on a brand new game?

keithcurtis said: I did load up an LDL page, and wasn't able to duplicate that cap; I was able to illuminate the whole map. Did you test on a brand new game? Oh that's kind of freaky. Thanks! I haven't tested completely clean-slate yet, so that is a very promising avenue of inquiry!

I still use LDL and played around a bit in my campaign. I couldn't duplicate the problem, but I did notice the following: For Infravision, I use the following settings: 60/-60. What I saw was that the range of vision was falling about 15 ft. short. So, in order to give a PC infravision out to 60', I now have to use the settings 75/-75. For a regular lit torch, I use the settings 25/12. This still worked as expected.

Thanks for the additional help. That's some interesting results! I did a quick scan through some of my extant rooms, every room I tested had the same problem. Made a new room from scratch, same deal. Tested with and without character sheets, so there's not a strange bug in that  interaction that I'm missing, I'm pretty sure. And in a way, that's kind of reassuring! That means it's not a setting that's breaking the room itself, at least. Miiiight be hardware--there was a recent Roll20 WebGL update to limit crazy CPU usage, right?--and I'm running Linux so maybe there's some interaction between some combo of weird Linux Nvidia drivers, Chrome hardware acceleration, and/or WebGL that other people aren't running into? Hmmm. Hmmmmmmm. This is a complete tangent, but as far as infravision/darkvision ranges go, I've usually done 90/-10 for my Pathfinder 60' of darkvision (basically the same solution?), so I don't think the range mismatch is a new thing, exactly! The darkness gradient LDL generates just starts at #000 (pitch black) and increases linearly with distance, so it stays way too dark at the edges. At least, that was my experience. (If I cranked up the brightness in either hardware or software, I could see out to the full range, but that wrecked everything else of course.) And different monitors/physical lighting situations/etc can change how much of that contrast we can perceive. But if that is actually a real change, I am very  curious!

There has been some anecdotal reporting of issues with Linux lately, but more about lag than DL, IIRC.

Vision seems to work as expected in Firefox, so that's another helpful boundary there. Shame Firefox is so much less performant!

Don't have more time to debug this tonight, BUT, the one other quick thing I could do fixed the problem. Disabling hardware acceleration in Chrome fixes the issue. IIRC there are a bunch of detailed settings you can dig into and as I get some time I'll work through them, but hardware acceleration is a VERY clear on/off switch, broken/not-broken. 

What does that do to your performance?

keithcurtis said: What does that do to your performance? Will need to test! It was pretty bad on an older machine, but this one's pretty beefy, so it might not be awful. I reserve Chrome for a few resource-intensive things, Roll20 and video streaming mostly, so disabling hardware acceleration wouldn't impact very much (but it might hit the things that really care). A couple other data points: a friend of mine on Chrome/Windows 10 just reproduced the error, so it's not targeting Linux specifically . I also discovered that enabling Vulkan also  seems to fix the problem. Vulkan support on Chrome for Linux is experimental and unstable. OH. Look at me, making assumptions. The real headliner here is that it's not just LDL. It hits BOTH dynamic lighting systems! Crud crud crud. But this does mean it's something Roll20 can look at.