Reproduction classes¶

As introduced in the overview, the reproduction modules link between the rendering modules and the physical hardware. VA provides different classes each being useful for different combinations of renderer and hardware setups.

In the following, an overview of all reproduction classes will be given. Then, the individual reproduction classes are discussed in detail. This starts with general reproduction classes which are not restricted to a specific hardware type, followed by classes designed for headphone-based and loudspeaker-based reproduction.

Reproduction class overview¶

General reproduction¶

Talkthrough¶

This class simply forwards the output stream of a renderer directly to the connected hardware channels without modifying it. Typical use-cases are plain headphone playback or channel-based renderings for loudspeaker setups, e.g. VBAP.

LowFrequencyMixer¶

This class is designed for the reproduction of low frequency content (e.g. using a subwoofer array). For this purpose, the channels from the rendering output are mixed using the sample based average. The result is forwarded to every attached hardware channel. The user can specifiy which renderer channels are used for the mixing process in the VACore.ini:

Headphone-based reproduction¶

The following reproduction classes are designed for headphone reproduction. Thus, all have a two-channel output stream.

Headphones¶

This class takes a two-channel input stream and applies an inverse Headphone-to-ear transfer function (HpTF) using FIR convolution to equalize the headphones. The data must be provided using an audio file which is specified in the VACore.ini:

BinauralMixdown¶

Attention: This reproduction module is currently under construction and does not work yet

The idea behind binaural downmixing is taking a channel-based input signal rendered for a loudspeaker array and reduce this to a two-channel binaural output signal which can be played via headphones. For this purpose, the loudspeaker array is placed in a virtual scene (free-field). Each loudspeaker can be interpreted as a virtual source with position and orientation defined in the hardware configuration. The real-world pose of the receiver is used to specify position and orientation of the listener within this array. Typically, it is assumed that the listener resides within the center of the array and only rotates its head. Then, the binaural output signal is created convolving each loudspeaker signal with the respective HRTF and mixing all signals. Playing this via headphones, the result should sound similar as standing within the actual loudspeaker array.

For this class the user has to specify two hardware output groups in the VACore.ini: The Output parameter refers to the real hardware used for the actual playback (typically headphones). The VirtualOutput refers to the "hardware setup" describing the virtual loudspeaker array.

BinauralAmbisonicsMixdown¶

This class is similar to BinauralMixdown but does an ambisonics decoding before doing the actual downmix. Thus, it requires an ambisonics encoded input signal and has a binaural two-channel output stream. In addition, to the VACore.ini parameters of the BinauralMixdown class, there are parameters for the Ambisonics decoder:

Loudspeaker-based reproduction¶

NCTC¶

This class uses a cross-talk cancellation (CTC) implementation to reproduce a binaural two-channel input with an arbitrary number of loudspeakers. For a dynamic CTC the real-world pose of the listener, i.e. position and orientation within the array, can be tracked. The following parameters can be specified in the VACore.ini:

HOA¶

This class takes an ambisonics encoded signal as input and decodes it for the connected loudspeaker array. The user can set a truncation order for the ambisonics decoding and select between different decoder algorithms in the VACore.ini: