Fran said: Hello, Roll20 VTT team member here! Thanks so much for your in-depth response! (And thanks to everyone here!!) To #1: Alpha channel - We DID look into this - but still worry about big performance hits. Alpha channel checking would probably have to be an option on a per-asset basis - don't quote me on that - and used judiciously. We felt that it was a bit much to come roaring out of the gate with a performance concern. Regardless, its something we can look into if others are interested. To #2: Noted! Are you - the GM - Clicking on doors and windows during map set up, or is it difficult for players to click on them during game time? To #3: You can adjust your Campaign Settings (Game Details Page > Settings Dropdown > Game Settings > Game Default Settings > Foreground Tab.) One thing that quickly emerged in our research was that we needed to be able to adjust defaults, as we worried that this new feature would add too much prep time! To #4: Specifically for elevation , Ill point you here: <a href="https://app.roll20.net/forum/post/12296098/add-multi-level-slash-elevation-support-to-the-foreground-layer#post-12297566" rel="nofollow">https://app.roll20.net/forum/post/12296098/add-multi-level-slash-elevation-support-to-the-foreground-layer#post-12297566</a> . To #5: We've noted peoples comments about the layer order thus far and have seen what we found through research - which is that no one order appears to make everyone happy. The team chose to prioritize the following: 1. Keeping "Token" up top, as the "Main" layer. 2. Grouping Token, GM and Map, as it has the strongest shared mental model. 3. Grouping "Foreground" and "Light" together, as they appear "locked" for free users. I hope this sheds light on our thought process. :) Hey there! I would love it as a "use at your own risk" type option, though I absolutely understand performance concerns. Having made a lot of maps, I know that a lot of times there are weird edge cases with shapes and bounding boxes of overhead layers that will otherwise cause oddities with the bounding box version of the test. I have struggled with this both as a DM and as a player. In my Bazaar test scene, in order to click on doors to use or adjust them as a DM, I have to physically move the overhead layer assets out of the way. Link to a video of that. As a player, I ran into this issue on Wednesday, as my DM had just added an overhead layer to the dungeon we were exploring (top level cave walls, beams, sconces, etc that we could walk under), but these obscured and paritally hid doors that we encountered, and I and other players kept clicking and closing/opening doors we forgot were there accidentally, as we couldn't see the icons. Oh excellent, I missed that! Glad you've already got that handled. Already commented on/upvoted :) I continue to disagree with that grouping, especially the idea of Map as a "mental model" not being grouped with the map-related layers (Foreground, Lighting), but I seem unlikely to change the design team's opinion. My only comment is that it seems odd that you're designing the UI experience for the users of the actual paid Plus/Pro features in a way that makes the most sense to free users (as the stuff they can't use gets grouped together). I know that you do have to design for both, it just is a bit backwards from common UI design principles. Thank you very much for responding to my feedback though! keithcurtis said: KatValkyrie said: The
'bounding box' test used for obsuring assets can be very clunky if the
asset is not generally rectangular. I ran into this problem with
overhead assets authored on the diagonal, and assets that have odd
protrusions or cut outs, and thus a lot of space where you can walk into
the bounding box and not the asset. As someone that works in 3D art, I
might suggest the possibility of looking into an alpha-test based
solution (ie. checking the transparency values (alpha chanel) of the
pixels) for a more accurate bounding box. For
non-rectangular items (alpha channel has been determined to be a
potential performance inhibitor) you can slice up irregular shapes into
chunks and upload them separately. Re-assemble them on the foreground
layer, and group them. The group will follow whatever behavior you set
for it, with any member of the group triggering the behavior. Thus an
L-shaped roof could be assembled from two pieces and grouped. Either leg
of the group members will trigger the behavior, but not the negative
space formed by the L. It's a lot easier to show than to tell, so let me
know if a diagram would be of service. Oh I'm fully aware that further slicing up and grouping overhead assets is an option, but as someone that made and released 150+ maps for a specific module I was running just for fun, the onus really shouldn't be on map makers to have to slice assets that should cover one building, up into many smaller pieces specifically for weird issues with Roll20's specific use case. It's confusing for less technically savvy end map users/DMs to set up and understand, and I've found from comments people have left on my work that people already have issues utilizing overhead pieces sometimes without extremely specific guidance, and "XYZ goes here" type diagrams, extrapolating that to say 3x the number of pieces to help solve issues like that is not a reasonable ask. Many other mapmakers I've seen also don't split their overhead tiles up if there are some smaller individual sections on a map, and instead export them as an image the full size of the base map. This was something a DM friend of mine encountered, when she was experimenting with the Foreground layer herself, and so many others will probably run into this issue also.