Triggers, video textures and model accessibility

Hi everyone. Thought I’d go about this one a bit differently than with the previous one and actually ask for advice before the possible problems arise.

We are launching this online trade fair on October 15th: 3D scene

The plan is to focus on the speed of the model; render everything as low as possible and keep the model size manageable even for the low-end hardware and mobile. So far I think it’s working smoothly. Every booth will have a trigger that will direct the visitor to a smaller model with only that single booth - but this time with better render quality and a full content blast. This way the exhibitors can market their own “room” through their own channels and bring visitors straight to their content.

The problem is I’d still like this main scene to have as much content and awesome stuff as possible. So the questions are:

  1. Extensions. Do they make the model heavier in any way? Should I just keep these in the smaller models and leave them out of this main one? There is going to be around 200 of them.
  2. Videos. If I were to put a video texture control extension on every one of these, so that they would only start playing 1 at a time, what kind of effect would this have? Does the visitor still have to load all of the 40 videos upon entry or how does it work?

Thanks in advance for any feedback.

P.S. I have a huge wish. Would it be possible to give us a possibility to add additional triggers for a single extension? That would be very good for my mental health @jan @wojtek

  1. Extensions are quite lightweight. This photo exhibition: https://bps.shapespark.com/exposicaofinal2/ has the largest number of HTML Label extensions that we have seen so far.
  2. Video textures are created when a scene is loaded. If videos are paused the browsers won’t be continuously streaming content of these textures, but it is possible that the browsers will stream some initial buffer for each of these paused textures. I’m not sure what will be performance and download speed impact of having 40 paused video textures, we haven’t yet seen such scene.

Extension configuration is stored in a cover.json file in your scene directory. You can access a nicely formatted version of this file by going to: http://localhost:5000/scenes/YOUR-SCENE-NAME/cover.json?pretty
If your extensions configuration is very repeating, you could edit and copy-paste parts of this file. Just make sure to make backup copies in case your changes break the file.

You can also use a validator like this one: https://jsonformatter.curiousconcept.com to make sure the file has valid format after your edits.

1 Like

Thanks for the answers and suggestion. Cool, I haven’t dared touch the json’s yet. Have to give it a try.

We are now running tests and the sky has yet to fall even though the scene has 30 or so 20mb videos running with no texture controls yet in place. Very promising.

My third question is about the lights. I think you wrote somewhere that not using probes makes scenes a lot lighter. How about the lights themselves: is it better to use spots, points or area types if we once again were to concentrate solely on speed?

Lights are baked into lightmaps, so their number and type do not affect performance and the download size of the scene in any way. Lightmap textures size is important for the download speed and memory requirements of the scene, it is best to use one or at most two lightmaps. Lightmap texture size depends on the lightmap resolution and the ‘Max lightmaps’ setting, but not on the number of lights.

Lightprobes have some memory and runtime cost, but are not super expensive. If you have a couple of them, you shouldn’t notice performance difference.

1 Like