Audio Hardware Design Requirements

Microphone performance requirements

Omnidirectional MEMS microphone is recommended, it has better consistency.

  • Sensitivity: analog microphones ≥ -38dBV, digital microphones ≥ -26dBFS, ±1.5dB

  • Signal-to-noise ratio (SNR) : ≥ 60dB

  • Overall-harmonic-distortion (THD) : ≤ 1% (1kHz)

  • Acoustic overload point (AOP) : ≥ 120dB SPL

Speaker performance requirements

  • Harmonic distortion (THD) : under rated power 100Hz ~ 200Hz THD≤5%, 200Hz ~ 8kHz THD≤3%

Microphone Array Design Recommend

  • The distance between two microphones should be 3.0cm ~ 7.0 cm, preferably the microphone arrays provided in the SDK.

  • All microphone pickup holes are located in the same straight line, which is parallel to the horizontal plane.

  • The microphone orientation can be at any Angle between up and forward (towards the speaker).

  • Use the same microphone models from the same manufacturer for the array. It’s not recommended to use different microphone models in the same array.

  • It is recommended to use the same structural design for all the microphones in the same array to ensure consistency.

Receive Path Performance Requirements

  1. Consistency

    • Frequency response consistency: free field spectrum (100Hz ~ 7kHz) response fluctuation < 3dB.

    • Phase consistency: phase difference between microphones (1kHz) < 10°.

  2. Leakproofness

    • External speaker playback, the overall volume attenuation (100Hz ~ 8kHz) between blocked microphone pickup hole and unblocked microphone pickup hole > 15dB.

  3. No Abnormality in the Spectrum

    • There should be no abnormal electrical noise.

    • There should be no data loss.

  4. Spectrum Attenuation

    • There should be no significant attenuation below 7.5kHz.

  5. Frequency Aliasing

    • Play the sweep signal (0Hz ~ 20kHz), and the recording signal has no significant frequency aliasing.

Echo Path Performance Requirements

  1. Loopback mode for echo reference

    • Only supports hardware loopback for echo reference.

  2. Echo reference signal position

    • It is recommended that the echo reference signal be as close to the speaker side as possible, and should be after EQ to avoid nonlinear caused by sound effects.

  3. Reference signal gain

    • When the speaker playback at the maximum volume, the echo reference signal should not have clipping, the Recommended signal peak value is -3dB to -6dB.

  4. Latency

    • Don’t have latency.

  5. Total harmonic distortion

    • When the speaker playback at the maximum volume: 100Hz, THD≤10%; 200Hz ~ 500Hz, THD≤6%; 500Hz ~ 8kHz, THD≤3%.

  6. Leakproofness

    • Device speaker playback, the overall volume attenuation (100Hz ~ 8kHz) between blocked microphone pickup hole and unblocked microphone pickup hole > 15dB.