adding-sound-to-a-sim.md

Checklist for Adding Sound to a PhET Simulation

This checklist is intended to guide developers through the process of adding sound to a simulation. It is written at a fairly high level, and assumes certain knowledge of how the tambo library works and of PhET's development processes. Before using this for the first time, it is highly recommended that a developer reads through the Users Guide.

• Create a sound design document and work with the sound designer(s) to get to the point where enough has been worked out that it becomes desirable to start prototyping. For an example, please see the Molecules and Light Sound Design document.
• Create an issue for adding sound to the simulation, use the title "Implement sound design". Here is an example GitHub issue, though it doesn't use the suggested name.
• Figure out if it's okay to turn sound on for the simulation. This will hinge upon whether there will need to be releases made off of master before the sound implementation is complete. Generally you'll want to turn sound on, but in cases where that isn't possible, skip the two sub-steps below, and the query parameters supportsSound and supportsEnhancedSound can be used to turn sound on for testing during development.
  • Turn on sound by adding the soundSupported: true key-value pair to the phet sub-object in the package.json file for this simulation. See https://github.com/phetsims/friction/blob/master/package.json.
  • If the sound design includes enhanced sound, turn this on by adding the enhancedSoundSupported: true key-value pair to the phet sub-object in the package.json file for this simulation. The main effect that this is has is to add the 'Enhanced Sound' item to the PhET menu. See https://github.com/phetsims/friction/blob/master/package.json.
  • Run grunt update to get the updated configuration into the HTML file.
• Explicitly turn off sound for any common UI components that now produce sounds that are not needed, if there are any. This is done by looking at the options and setting any sound players to Playable.NO_SOUND. Search through the code base for example usages.
• Add any behavior for any common-UI sound generation that is different from the default behavior. This is done by creating a Playable (often a SoundClip instance) and passing it in as an option for the sound player.
• Decide whether or not to have a separate "sound view" or to use the existing ScreenView files. Both approaches have been used, and both are legit, and it probably depends on how much sound generation is needed and whether adding it all to the ScreenView file(s) is likely to make the file too large and/or difficult to maintain. As of this writing, WavesScreenSoundView is an example of a separate class where the sounds are hooked to the model, and FrictionScreenView is an example of where sound was interwoven with the visual view code. Note that in some cases it is necessary to know what the user was doing that triggered the need for sound generation, and in these cases the sound generation will need to be inside the view element so that the code has access to all the needed information. An example where this often comes up is if different behavior is needed when a value was changed via keyboard interaction versus mouse or touch interaction.
• Decide whether a "Sound Options" dialog is needed to allow designers to compare different sound design ideas in context. This has been found to be very useful when iterating on a sound design. There is more information on this in the User Guide, and [an example in the Tambo demo] (https://github.com/phetsims/tambo/blob/master/js/demo/SoundOptionsDialogContent.js).
• Add sim-specific sound generation. See previously sonified simulations for examples on how to do this, but the general idea is to create sound generators and hook them up to the model and/or view elements that they are meant to sonify, all based on the sound design document.
• Iterate on the sound design. Regular meetings and good note taking (generally in GitHub issues) have been found to be quite helpful for this. This is basically like the implementation process for all other portions of the sim, and involves publishing dev versions, getting feedback, refining the sim, rinse, and repeat. One observation: It seems to be more difficult for people to imagine how they will like a sound in context than a visual design element, so be prepared for a lot of iteration, and use the 'Options' dialog if and when it can help.
• Once all of the sounds have been finalized, do a "mix" step where the primary sound designer goes through the sim and sets the volume level for all sim-specific sounds.
• Once the sound design is thought to be complete, or nearly complete, publish a dev version and create a checklist of all of the sound designers and other stakeholders in the original sound design GitHub issue and ask them to either approve of the implementation or log their objections. Example: phetsims/molecules-and-light#233 (comment)
• Finalize the names of all sound files that have been added and get uncompressed versions of them and add them to the assets directory. Make sure the names match, e.g. bonk.mp3 in the sounds directory and bonk.wav in the assets directory.
• If an options dialog was added or if content was added to an existing options dialog to support comparitive evaluation of sounds, remove said content
• If there were individuals who were only involved in the sound design and thus not already on the team or other lists in the credits, add them using the soundDesign key.
• Once the sound design is approved, mark the issue as ready-for-testing and have it tested with the other aspects of the simulation through the publication process.