I have a question regarding the streaming option in the audio importer for WebGL builds.
So our project has a lot of audio and fundamentally one of the issues of WebGL is the high memory demand. So an obvious way of by-passing this is instead of loading all of the audio clips into memory at start simply streaming them and dumping them as they’re needed. There are a few approaches to this but I think two very simple examples show what appear to be varying methods of “streaming” the audio.
I’m wondering what does setting the audio type to streaming actually do with the audio memory wise. The concerning aspect is that it implies the audio will be “streamed” to memory, however the content is included in the data file at build wich appears to be fully loaded to memory anyways.
My other approach is to simply include the audio files in the StreamingAssets folder then use WWW to grab an audio clip of them. I’m far more certain this keeps it out of memory. However just putting them in the resource folder and having platform specific building for streaming would be more ideal because it would require less work to handle on various platforms.