Audio beamforming for smartphones: hearing aid
Vložit
- čas přidán 23. 08. 2024
- A working acoustic beamformer prototype in tablet form. Design description, demonstration, and how a multiple beam revision could be used.
Smartphone manufacturers have largely run out of ideas for hardware features but beamforming has not yet been done. Beamforming could also be useful for handsfree phone calls in cars.
I loved the outro the most
Very interesting and informative. Thank you for the in-depth explanation!
Thank you Kevin.
If you want a fun graphical math puzzle and to better visualize, pick N microphones and a target, then steer the beam in 2D to locations where you get an amplification of N, N-1, N-2, ... 1 for a particular frequency. Afterward, sweep the frequency to see how sharpness increases with frequency.
By the way, I have found that scilab.org offers a compelling alternative to Matlab: I think it is better in some ways and it is also free.
I rarely write comments but I just wanted to say thank you for that video!
I have to write a couple of pages about acoustically holography and beamforming for the university and even tough I read a couple of hours in books that video just manged to explain something I did not understand until...
Thx from Germany!
Thanks for the thx, I appreciate it. Also of possible interest is underwater beamforming, which is the natural follow on to basic sonar. It can see where there is no light or where water is muddy. I worked on a prototype back in 1986. Hope you get a good mark on your paper :)
Hi, any chance to contact you regarding the potential business inquiry on the beamformer related stuff?
Beamforming has long been public domain (not patentable) and my ideas regarding how to use it are now public domain, which was my intent. I thank you for broaching the idea of business inquiry but my days are full. An email or phone number in the comments is asking for trouble. If you have a question, I'll try to answer it.
@@htpac Thanks for reply, limited days are not an issue. Will really appreciate if you can drop me an email for details. See my info at czcams.com/channels/IfFIwgL60X6P79xkUUJHxg.htmlabout
Is it possible to carry out the same project with an ST NUCLEO board?
Thanks for watching.
Sorry, but my blunt answer (opinion) is no. Consider the parallel thoughput going on, the data rates required, and the computations you would need to do for the filtering/decimation.
I have reverence for STM32 and noted long ago that the ST upper management mostly have engineering backgrounds, which is why it is a good company.
If you can't find a way to beamform with an STM32 then the natural thought progression is to a fancy DSP and now it becomes an overly complicated programming problem on an expensive chip, when my design is trivial on an ASIC/FPGA.
I knew a Chinese engineer who would pull out the epithet “cut the feet to fit the shoes” when we were trying to make existing hardware fit a problem it was never meant for :)
@@htpac Thanks so much for the super response. I have the obligation to use analog microphones to carry out beamforming. I could kindly ask you about the material to better understand the theory. Assuming we can digitize the output of analog mems with A / D and read the data with STM32, what remains is to program the microcontroller to introduce delays, right? I should fill arrays with the digital values of the microphones output and work on the sum by translating the cells. I thought so. I'm doing it wrong? I would really appreciate it. Good day