Noise reduction from speech in real time

I would like to briefly describe my issue:

I participate in the development of equipment for excursions.
We have a mobile device with 2 (or more) microphones.
The guide speaks into the first microphone (it is voice with noise).
The second microphone is located at a distance of 20-30 cm from the first one and gets the surrounding noise.
Please, see example of records here: (there are ‘voice.wav’ and ‘noise.wav’ files in the zip file).

My task is to get a clean voice without noise in real time mode (with not more 5 miliseconds delay), using two or more microphones.

What would be the best option to solve this issue?

Thank you so much

I’m not sure what is your setup, and what you have for recording, but you should be able to set plugin on input line on insert and that would work in real time, just like any hardware. For this you could use Izotope RX suite, dialogue denoiser, but to get it in 5ms delay…I’m not sure you will get with anything.

Thank you,
I mean delay is not more 5 milliseconds (fixed already), sorry. What Izotope RX module do you recommend? I could try it.

In Izotope RX6 suit is called Voice De-noise in earlier versions is called Dialogue Denoiser