SMIl 3.0 Манипуляции временем

SMIL 3.0 Time Manipulations

14.1 Overview and Summary of Changes for SMIL 3.0

This section is informative.

The SMIL 3.0 specification leaves the SMIL 2.1 Time Manipulations Module [SMIL21-timemanipulations] unchanged.

14.2 Introduction

This section is informative.

This module introduces new attributes for advanced manipulation of time behavior, such as controlling the speed or rate of time for an element. These time manipulations are especially suited to animation and non-linear or discrete media. Not all continuous media types will fully support time manipulations. For example, streaming MPEG 1 video will not generally support backwards play. A fallback mechanism is described for these kinds of media.

Four new attributes add support for time manipulations to SMIL timing modules, including control over the speed of an element, and support for acceleration and deceleration. The impact on overall timing and synchronization is described. A definition is provided for reasonable fallback mechanisms for media players that cannot support the time manipulations.

An accessibility requirement for control of the playback speed is related to the speed control, but may also be controlled through some other mechanism such as DOM interfaces.

14.3 Module Overview

This section is normative

This section is informative

A common application of timing supports animation. The recent integration of SMIL timing with SVG is a good example of the interest in this area. Animation in the more general sense includes the time-based manipulation of basic transforms, applied to a presentation. Some of the effects supported include motion, scaling, rotation, color manipulation, as well as a host of presentation manipulations within a style framework like CSS.

Animation is often used to model basic mechanics. Many animation use-cases are difficult or nearly impossible to describe without a simple means to control pacing and to apply simple effects that emulate common mechanical phenomena. While it is possible to build these mechanisms into the animation behaviors themselves, this requires that every animation extension duplicate this support. This makes the framework more difficult to extend and customize. In addition, a decentralized model allows any animation behavior to introduce individual syntax and semantics for these mechanisms. The inconsistencies that this would introduce make the authoring model much harder to learn, and would complicate the job of any authoring tool designer as well. Finally, an ad hoc, per-element model precludes the use of such mechanisms on structured animations (e.g. applying time manipulations to a time container of synchronized animation behaviors).