This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Solution using Audio Front End with Acoustic Echo Cancellation

Good day,

My name is John Leon from DeepSea Developments Company (https://www.deepseadev.com/en/).

We are working in a low-size solution using Audio Front End. We have a microphone + esp32-32s + speaker (https://www.espressif.com/en/solutions/audio-solutions/esp-afe). But, the speaker is too near to microphone and need solve the issue about echo, so we are looking a chip that allows a special implementation for AEC (acoustic echo cancellation) when the user talks.

Please, could you let me know if you have a chip:

1.Low consumption

2.low latency

3.small size ( not too many pins - package)

4.AEC + Noise suppression (NS)

 

I look forward any comment.

Thanks and Best Regards,

  • Hi John,

    In order to implement AEC you would need to use a codec with miniDSP on board. For example this would be the AIC3111, AIC3120, AIC3254, AIC3262, etc. Because of this the chips are going to have at least 32 pins. Most are 5x5 VQFN.

    Best regards,

    Jeff

  • Thanks Jeff. The AIC3254 (6PAIC3254IRHBRQ1)  looks great and the price is good. But do you know if there is an example that already runs that in a circuit or a development kit? I have to be sure if the miniDSP is enough to implement our solution and more technical data about your start up.

  • Hi John,

    We have Evaluation Modules available for the 3254: https://www.ti.com/tool/TLV320AIC3254EVM-K

    This will let you prototype the device in your solution. The miniDSP is configured using PurePath Studio: https://www.ti.com/tool/AICPUREPATH_STUDIO

    Best regards,

    Jeff

  • Hi Jeff, how are you?

    I have been working on this project, reading documentation and looking for theory. We haven't bought the development kit yet because we need to be sure that the solution works.

    We need the loop: I2S_mic (read) + miniDSP or codec + I2S_Amplifier  (write) + Speaker, where at the same time the circuit capture the voice (reading the I2S mic) and playback by speaker (writing on I2S amplifier) with good sound quality (low latency, free of echo, gain without saturation, volume control, noise or feedback suppression , etc).

    If you can also provide a demonstration that show how it could work on the miniDSP software (https://www.ti.com/tool/AICPUREPATH_STUDIO) It would be helpful.

    Thank you so much,

    I look forward any comment

    John

  • Hi John,

    I apologize that PurePath isn't the most user friendly thing, but if you right click any component and press "help" it will take you to a page that explains that component in more detail. I've found that to be the best way to learn and troubleshoot. 

    Below I've attached a screenshot of a very simple process flow that you might use:

    The most important thing here is the framework block. That allows PPS to create the code for the right part in the right sampling rate. Here I have I2S in, Stereo GEQ, and I2S out. The other yellow block there is because this DSP has two cores that separate the work between the analog and digital portions. In other words the ADC side and the DAC side. Because I2S in is on the DAC side, and I2S out is on the ADC side, there needs to be that handoff block you see in yellow. If that block is missing, you'll see an error that explains where the disagreement is between which side of the DSP the components are on.

    Again, the help function will give you more details on these blocks to help you create the process flow you want.

    Best regards,

    Jeff