The 3D Tune-In Toolkit is a standard C++ library for audio spatialization and simulation of hearing loss and hearing aids.
The toolkit offers great flexibility by enabling the applications to use multiple configurations according to their needs. The library is currently being implemented and is planned to be partially released as open source for non-commercial use in spring 2017, full release at the beginning of 2018. The 3D Tune-In Toolkit supports two different setups: a simple pair of headphones using binaural sound and a set of speakers using Ambisonic encoding and decoding.
3D Tune-In Toolkit - binaural
Able to spatialize multiple anechoic sources efficiently by performing a convolution with a selected Head Related Impulse Response (HRIR), which is interpolated from those of a Head Related Transfer Function (HRTF) selected by the user. It can also add an extra shadow in the contralateral ear for sources very close to the listener head, according to a sound propagation model. The library can adapt the ITD (Interaural Time Differences) to a custom head circumference and darken sources placed far away from the listener.
Application for testing binaural spatialization features. The 3D Tune-In toolkit is able to spatialize multiple moving sources using a HRTF selected by the user.
To include the library in an application, developers can configure it with the desired HRTF, select the listener head circumference, declare the sources and start spatializing feeding it with as many audio streams as they want. Sources and Listener locations can be updated in run time and the library will smoothly adapt its process accordingly by means of interpolations.
In addition to the anechoic spatialization, the 3D Tune-In toolkit generates the corresponding reverberation, convolving anechoic sources with the room impulse response. It uses a novel approach, working with a low-order Ambisonic encoding of all sources, generating the reverberation of all sources together but keeping certain location dependent characteristics. This approach, together with an efficient convolution algorithm in the frequency domain, allows us to compute large reverberating scenes, with virtually unlimited number of sources, maintaining high spatial accuracy for the direct sound (spatialised using direct-HRTF convolution).
Application to test hearing loss and hearing aid simulation
The 3D Tune-In toolkit also implements a simulator of hearing loss and of a generalised hearing aid. The former simulates different types of hearing impairment. The latter, offers the possibility of simulating hearing aid devices, including multiband amplification and directional microphones.
3D Tune-In Toolkit - loudspeakers
For a setup based on loudspeakers, the 3D Tune-In Toolkit is able to manage multiple sources, generating multiple channels to produce spatial sound by means of a number of loudspeakers located around the listener.
Application for testing the spatialization based on loudspeakers
This spatialization is based on an Ambisonic encoding, and includes both anechoic and reverberation paths, processed independently.
The reverberation is generated using a similar approach to that previously described for the binaural setup (based on virtual loudspeakers). In this case; including the library in an application, simply requires developers to configure it, setting the exact location of loudspeakers, declare the sources and manage the eight channel output produced by the library using the appropriate audio interface.
3D Tune-In Toolkit - summary of features
- Binaural spatialization.
- Head Related Transfer Function (HRTF) selected by the user.
- Custom user head circumference.
- Specific simulation for near field sources- Smooth adaptation for moving sources.
- Reverberation using the room impulse response.
- Hearing loss simulation of different types of hearing impairment.
- Spatialization based on loudspeaker.
- Generalized hearing aid simulation.
Binaural spatialisation demo.
This is a brief demo of the capabilities of the 3D Tune-In Toolkit in terms of binaural spatialisation, including environmental acoustic and distance simulations. The first audio is the original anechoic recording:
The second one is the actual audio spatialised using the 3D Tune-In Toolkit: