Combining ESP32 and XMOS VocalFusion EVK as Alexa Client

[TeeVee]

We try to combine ESP32 (as ESP-WROVER-Kit or ESP32-DecvKitc) with the XMOS VocalFusion AVS Kit (http://www.xmos.com/products/voice/vocalfusion-avs) to build an Alexa client system.

We took MrBuddyCasino’s repository (https://github.com/MrBuddyCasino/ESP32_Alexa) for the basic Alexa client and modified the
I2S settings for the audio renderer as well as for the audio recorder.

The XMOS board has one single I2S port which is used bidirectionally for the microphone signals (to ESP32) as well as to the output codec (from ESP32). It’s role is defined fix as Master.

So we changed the I2S controller’s role for both ports to Slave with Rx for the audio_recorder and Tx for the audio_renderer.
Both BCLK and LRCk from the XMOS board were connected to both I2S ports on the ESP32 respectively. As they are now both Slaves clocking as completely done by the XMOS board and shouldn’t show any issues.

However we do not have any success neither on the Tx nor on the Rx side. We see ESP32 connecting wuirelessly and exchange of token
with AVS as well as starting of http2 transmission but we do not get any sound from Alexa.

Any ideas what to try?