Frequently Asked Questions
Sound Particles is a groundbreaking software, and as such, many questions arise...
The creation of a Windows version will depend only on the commercial success of the software. As you can imagine, we would like to bring the software to as many platforms as possible.
Yes, we are now working to bring Sound Particles to AAX plug-in architecture (ProTools), first as a Native AAX plug-in (CPU) and later as a DSP plug-in. Other plug-in architectures may follow.
Even better - schools/students/teachers can have free access to "Sound Particles U", a special academic version of Sound Particles, with the exact same features as the commercial version.
Currently the software only supports channel-based audio (e.g. Dolby Atmos 9.1 bed, Auro-3D 11.1/13.1, NHK 22.2) and scene-based audio (Ambisonics/HOA). The main issue to support object-based audio is that there isn’t a way to export metadata to other systems. Nevertheless, we are working with Dolby, DTS, Auro and Avid, to support it on a near future, as soon there is a technical way to communicate metadata information with these systems.
No. Currently Sound Particles is a standalone application. But we have plans to create a plug-in version with most features.
Due to the complexity of the processing, all audio scenes must be rendered to create the final audio stream. Nevertheless, you don’t need to render the entire scene to be able to start listen to the audio – as soon the first slots are rendered, you may listen to them.
The render time depends on the complexity of the scene. As you can imagine, if you have thousands of particles, the software will need to render thousands of audio tracks, and that can take a while. A typical 10 second scene with a small amount of particles (<100) may take around 10 seconds to render or less. As you increase complexity, the render time usually increases in a linear way.
In theory, each particle group can have a maximum of ~ 2 000 000 000 particles, and you can create millions of groups. The problem is that the software would need a LOT of time to be able to render such audio scenes, besides memory/RAM requirements that computers don’t currently have.
For each particle, the engine will calculate its position on a sample-by-sample basis (e.g. a sample rate of 96kHz means that you calculate the position of each particle 96000 times per second). Internally, all calculations use 64-bit floating-point precision. Since the scene could have thousands of explosions happening a few inches away from the virtual microphone or a simple whisper several miles away, while rendering a normalization process is applied to the final stream, preventing any clipping and optimizing the dynamic range of the output signal.