50.2 Description

The Asynchronous Sample Rate Converter (ASRC) converts the sample rate of an incoming audio frame without affecting quality. It supports input and output sampling rates up to 192 kHz.

The ASRC must be used when the rate of an input (or output) audio stream is different from the rate of the audio system processing the audio stream (for example, a 44.1 Ksps audio stream must be processed in a 48 Ksps environment).

The ASRC must be also used when the received (or transmitted) audio stream and the audio system both work at the same nominal rate, but are driven by different clock sources (for example, two different crystal oscillators with the same nominal clock frequency, differing by a few ppm).

The ASRC is made of 4 independent digital signal processing (DSP) modules. Each DSP can be configured with specific input and output sampling frequencies and is associated to one of the DMA channels.

Each DSP can process mono or stereo audio streams.

Data can be provided to the DSPs via up to four input channels and data can be read from the DSPs via up to four output channels. A channel comprises a holding register that can be accessed via software or via a channel of the central DMA. The ASRC generates the trigger events for the central DMA.

Up to 8 channels are available.

Depending on the channel index and its configuration, a channel manages from one up to eight audio frames.

The ASRC can be configured to process up to four stereo audio streams in a fully independent manner. Input/output sampling frequencies differ for each DSP.

The ASRC supports TDM audio stream. The ASRC is capable to concatenate/split audio streams while resampling frequencies. The ASRC can be configured to resample and merge several audio input sources of the same or of different sampling frequencies carried by several DMA channels into a single output audio stream sampled at a unique frequency and carried on a single DMA channel (TDM audio stream creation from multiple sources). The ASRC can also split and resample audio streams (for example, TDM stream) carried on a single DMA channel to multiple audio streams, each having the same or different resampling frequencies.

Note: There is no anti-aliasing filter. There is no limitation when up-sampling an audio stream but when down-sampling, the input signal frequency bandwidth upper boundary must be lower than half the resampling frequency.