Describe time, space and acoustic response

A shared timeline needs an explicit mapping#

Audio frames, video frames, MIDI events and animation ticks do not automatically share a clock. VAO 0.5.0 describes explicit timebases and mappings. Rational rates avoid ambiguous decimal approximations; half-open intervals include their start and exclude their end. A region from frame 0 to frame 48,000 at 48,000 frames per second is therefore one second long, without counting an extra endpoint sample.

A linear mapping relates coordinates as target = source × scale + offset. Both coordinate units matter. For a source measured in audio frames and a target measured in seconds, the scale has units of seconds per frame. Clock drift, residual uncertainty and discontinuities require their own treatment. A single offset can align an onset without explaining drift during a long session, and a mapping must not interpolate across an unmodeled dropout.

source coordinate: audio frame index at 48,000 frames/second
target coordinate: video time in seconds
scale = 1 / 48,000 seconds/frame
offset = 0.125 seconds
frame 96,000 maps to 2.125 seconds

Keep the measured alignment uncertainty and any discontinuities separately.

A position means something within a frame#

A coordinate triple is incomplete without units, axis definitions, handedness and a reference frame. A microphone pose in an instrument frame differs from its pose in a room frame. A geographic longitude and latitude pair must not be silently interpreted as local Cartesian x and y. Explicit transformations and their provenance let another client reconstruct the intended relations or identify a missing conversion.

For a measured response, preserve the source and receiver positions, orientations where relevant, instrument and room state, and the method used to establish them. An approximate position should remain approximate. Giving more decimal places to an estimated coordinate does not improve the underlying measurement. A change of coordinate frame is a transformation; it is not a new acoustic observation.

Keep the response and its domain together#

An impulse response characterizes a specified source–receiver arrangement under particular conditions. Its association with a VAO should identify the exact signal realization, channel interpretation and the scene or measurement to which it belongs. Applying that response elsewhere is a modeling choice. A player needs an explicit policy for positions outside the measured or modeled domain, such as a declared fallback, instead of suggesting that one response captures every listening position.

Declare the capabilities the object actually uses#

VAO 0.5.0 requires Core and Dynamic Delivery for every conforming object, including an embedded-only package. Additional profiles follow the represented capabilities: Multimodal Timeline for timebases and synchronized tracks, Spatial for frames and poses, Acoustics for relevant response or scene records, and Scientific for scientific records. Physical Instrument, Playable and Deterministic Runtime address further aspects when present. Profile declarations are not a promise that every client will execute every feature.

A useful inspection sequence is to read the profile declarations, identify required materialization groups, verify their availability, and then check the application’s implementation scope. A Blender importer may inspect a spatial package and load geometry without executing its acoustic behavior. A runtime capable of convolution still needs the correct responses, mappings and dependencies. The object contract and the application capability are both necessary to interpret what the user hears.