Patent Application
20240390793
The present application claims priority from United Kingdom Patent Application No. GB2307782.9, filed May 24, 2023, the disclosure of which is hereby incorporated herein by reference.
This invention relates to computer-implemented methods of generating sound effects for video games.
In video game development, there is often a need to produce sound effects associated with the large number of objects that will appear in the video game. For example, there might be large number of weapons available for the player to use during gameplay. A weapon such as a sword might require one sound effect representing the sword being swung, another to be played when the sword hits a wooden surface, yet another for hitting stone, and so on. Each of these interactions will also be associated with a different animation (e.g. one animation representing the swinging of the sword and another representing the sword striking a solid object). Often, there will be many objects belonging to the same category (e.g. several different swords) each requiring a unique set of animations and sound effects representing the different interactions that can be performed with that object. However, there will generally be some similarity between the sound effects and animations that are developed for objects belonging to the same category: returning again to the swords example, it would be expected that an iron sword and a steel sword would sound similar to one another when striking the same enemy or object. One conventional approach to generating sound effects for a set of objects belonging to the same category has been to start with one basic set of sound effects and produce unique sounds for the various objects by modifying this basic set. For example, starting with a sound effect representing an iron sword striking stone, a convincing sound effect for a steel sword undergoing the same interaction could be produced by small adjustment of the initial sound effect. However, this approach is in practice very time consuming, requiring the developer to manually generate a modified sound for each object, and involves a large amount of trial and error since the modifications applied will not always produce a convincing result. There is therefore a need for a faster way of developing sound effects for objects in video games.
The invention provides a computer-implemented method of generating a sound effect for a video game, the method comprising:
In this method, a bespoke second sound effect is generated for the first animation by modifying the first sound effect file (which could be, for example, a stock sound effect not associated with any specific object in the video game) based on constraints derived from the first animation. As noted above, the first sound effect file comprises one or more sound effect features. The sound effect features represent elements of the sound effect and typically will be layers of the sound effect file. An example of a suitable file format for the first sound effect file is the Wave (.wav) format. This speeds up the development of sounds effects to be associated with animations during the development of the video game since it overcomes the need for the sound designer to create new sound effects manually.
Based on the received first animation, constraints to be applied to the sound effect feature(s) are determined. The constraints define the characteristics that the adjusted sound effect features may have and are chosen so as to make the resulting sound effect features suitable for association with the first animation. One simply example of a constraint that could be determined based on the first animation is the duration of the sound effect feature(s): for example, a constraint could be applied whereby the duration of the sound effect feature(s) may be no greater than the duration of the first animation. In another example, if it were determined that an object represented in the animation is made of steel, the constraints could be chosen such that the adjusted sound effect features give the impression of the sound of steel (e.g. a suitable pitch range and range of possible durations). The animation may be received in the form of an animation file-for example a .anim file. This animation file would typically be produced using a graphics design tool as part of the process of developing the video game. The animation file may contain data defining the visual content of the animation, e.g. a sequence of frames.
As noted above, the sound effect feature(s) may advantageously be layers of the sound effect file. Each layer could represent a basic sound, and several layers can be combined in the sound effect files to create rich, complex sounds. For example, in a sound effect representing a weapon striking an enemy, one layer could provide the basic sound of the impact and another layer could provide a bass sound coinciding with the impact to provide depth to the sound effect.
It will be appreciated that different constraints may be determined for each sound effect feature. For example, for a sound effect feature (e.g. a layer) representing the basic sound of an impact as just described, a constraint could be placed on the pitch of this layer based on a material identified in the animation. For the bass layer, the pitch may be unchanged but the duration or volume could be altered to vary the impression of depth created by this layer.
In preferred embodiments, adjusting the sound effect features comprises, for each sound effect feature (e.g. each layer), varying one or more sound parameters of the sound effect feature, wherein the one or more sound parameters preferably comprise one or more of:
By “sound effect parameter” we may any property of the sound effect feature (e.g. a layer) that can be varied. Such properties include the examples listed above. By “temporal position” we mean the span of time over which the sound effect feature is played in the sound effect. The time at which the sound effect feature is played could be delayed by moving the sound effect feature to a later temporal position, for example. The “temporal separation” is the relative separation in time between two sound effect features—e.g. in the case of a steel sword striking an object, the sound effect could include a sound effect feature representing the sound of the impact and then, at a slightly later point in time, the sound of the steel blade ringing.
Adjusting the sound effect features may comprise randomising the one or more sound parameters of each sound effect feature.
In some preferred embodiments, the randomisation of the sound parameters is performed a plurality of times so as to generate a plurality of versions of each sound parameter; and wherein producing the adjusted sound effect feature satisfying the determined constraints comprises selecting, from the plurality of generated versions of each sound parameter, one version of each sound parameter that satisfies the determined constraints. In other words, the sound parameters of each sound effect parameter are randomised multiple times, resulting in several versions of the first sound effect (each version of the first sound effect being constituted by a respective set of the generated versions of the sounds parameters), and the version that meets the determined constraints is selected.
In other preferred embodiments, the randomisation is controlled such that each randomised sound effect parameter falls within the determined constraints. In other words, sound parameters will be randomised within the bounds of the constraints—e.g. if one of the constraints is a pitch range, the pitch of the sound effect feature to which this constraint is applied may be adjusted to a random value within that range.
As noted above, the determined constraints are preferably limits on the one or more sound parameters of the sound effect feature. For example, each constraint could comprise an upper limit, a lower limit, an allowed range, or a specified value for the respective sound parameter of the sound effect feature.
Preferably the determining, based on the received animation, one or more constraints to be applied to the sound effect features, is based on animation parameters of the animation, the one or more animation parameters preferably comprising one or more of:
The animation parameters may be determined based on the content of an animation file of the kind described above-for example based on the visual content of the animation (e.g. the frames of an animation file) or other data associated with the animation file, for example metadata describing features of objects in the animation (e.g. the material of the objects).
The animation parameters may advantageously be identified using a computer vision algorithm. Such techniques are particularly suitable for identifying properties such as shapes and materials and therefore provide a way of taking these variables into account for the generation of the second sound effect. Other visual analysis techniques (e.g. pattern recognition) could alternatively be used. A computer vision algorithm could be applied to an animation file of the kind described above.
The first animation is preferably received in the form of a file comprising data representing at least some of the animation parameters. Some of the parameters listed above, such as duration, can conveniently be identified in such a file: for example, the number of frames in the animation could be determined from the file, from which the duration of the animation could be inferred.
Examples of methods in accordance with embodiments of the invention will now be described with reference to the accompanying drawings, in which:
The first animation is of a first sword 100 swinging. The first animation has four frames: frame 101 at a time T1, frame 102 at time T2, frame 103 at time T3, and frame 104 at time T4. The second animation is of a second sword 200 swinging and has three frames: frame 201 at time T1, frame 202 at time T2, and frame 203 at time T3. The first animation therefore has a temporal duration of 4 frames and the second animation has a temporal duration of 3 frames.
The second sword 200 is smaller than the first sword 100, so the swing of the second sword 200 is represented in the animation as being faster (and therefore having a shorter temporal duration) than that of the larger, slower first sword 100.
In development of the video game, the sound developer will need to produce sound effects for the two animations representing the sounds of the swords 100, 200 cutting through air. The two sound effects will be similar but, because the swords have different sizes, shapes, weights and other properties, not identical.
In a first step S101, a first sound effect file is received. This sound effect file will typically be either (i) a stock sound effect file representing the general category of object for which a sound effect is being generated (e.g. a generic sound effect representing a sword swinging through air) or (ii) a sound effect that has been developed for an animation of one object in a category in which sounds effects for other similar objects now need to be developed. The sound effect file comprises a plurality of sound effect features. The sound effect features will typically be layers of the first sound effect file, since often the sound effect will be formed of a plurality of layers that may each represent a basic sound sample and which are combined in the sound effect file to create a complex, bespoke sound effect. The sound effect features are variable, by which it is meant mean that parameters of each sound effect feature (e.g. layer) may be varied: for example, the volume, pitch, duration, and temporal position can each be controlled. The sound effect file will typically be a .wav file.
In step S102, a first animation is received. The first animation is an animation for which a sound effect is to be generated. The animation could be received a variety of formats, one example of which is a set of image files representing the frames of the animation (as shown in
In step S103, based on the received first animation, there are determined a set of constraints to be applied to the sound effect features. Typically the constraints will be numerical limits to be placed on the sound parameters of each sound effect feature in the first sound effect. For example, a constraint applied to one sound effect feature (e.g. layer) could be upper and/or lower limit on the volume of the sound effect feature, or a limitation to be applied to the pitch range of the sound effect feature.
As noted above, the constraints are determined based on the first animation. The principle here is that a set of constraints defining an appropriate adjustment of the first sound effect can be determined based on the first animation: one very simple example of this is identifying the duration of the animation (e.g. from the number of frames) and determining from this information a constraint specifying that the duration of the sound effect features should be the same as that of the animation. In another example, it could be determined from the first animation that an object in the first animation is made of a material such as steel, and that a sound effect representing an action such as the object striking a hard surface should have a pitch that is sufficiently high and a duration of sufficient length to convincingly imitate the ‘ring’ of the steel when struck. In the example shown in
Determining the constraints could be performed based on visual processing of the first animation: for example by applying computer vision to the frames 101, 102, 103, 104 of the first animation to identify animation parameters such as material and shape of the object. If the animation is received in the form of a file containing an animated 3D model, the animation parameters could be derived from data in this file—for example, data in the file might specify the dimensions of the object and the texture in which it is to be rendered, from which its material could be inferred.
Once the constraints have been determined, the sound effect features of the first sound effect file are adjusted so as to produce a corresponding set of adjusted sound effect features satisfying the determined constraints. In the
In an alternative approach to adjusting the sound effect features, the sound parameters of each sound effect feature could be randomized a plurality of times and then from the plurality of randomized sound parameters, one version that satisfies the determined constraints would be selected.
Once the adjustment of the sound effect features has been completed, in step S105, a second sound effect file comprising the adjusted sound effect features is generated. This second sound effect file represents a customised sound effect for the first animation that was received in step S101 and its sound effect features satisfy the constraints that were determined, based on the first animation, in step S103. This second sound file may be output and associated with the first animation.
| Number | Date | Country | Kind |
|---|---|---|---|
| GB2307782.9 | May 2023 | GB | national |
