Background noise suppression in mobile communication

[ransiluj]

hey all!!
I'm planing to do background noise suppression device for mobile communication. In here it is common problem because when you talking through phone u can't hear what other person saying because of the noise of the other side. So i need make a device to suppress that noise n detect only voice of the person. I need to way to approach this project. please help me.
Thank you!!

 

[ronv]

Do you mean background noise, like other people talking in the same room?

 

[MikeMl]

Start by studying noise-cancelling microphones and Active Noise Reduction headphones used in aircraft as shown here. Makers are Lightspeed, Bose and Peltor, amongst others. I have been flying with ANR headsets for about 20years. They make a huge difference both in the Intelligibility of radio transmissions and in reducing pilot and passenger fatigue on long flights.

 

[crutschow]

hey all!!
I'm planing to do background noise suppression device for mobile communication. In here it is common problem because when you talking through phone u can't hear what other person saying because of the noise of the other side. So i need make a device to suppress that noise n detect only voice of the person. I need to way to approach this project. please help me.
Thank you!!

If you are trying to suppress the noise from the other end, that's very difficult since the noise and the voice are in the same band of frequencies. If you can make a device to separate voice from noise, it would be worth millions and you wouldn't have to work any more. Hearing aid manufacturers have been trying to do that forever.

 

[dougy83]

If you're planning on using DSP, there are many papers on noise reduction algorithms e.g. Kalman, Wiener or blind speech separation using ICA-based approaches. Do a search and you'll find many more advanced methods described too.

It's best to process the audio before it's been encoded, compressed and sent over the network - as the noise takes up bandwidth rather than the speech and will make it difficult to get decent quality. You could have a think about implementing the DSP on the phone itself if it's possible to intercept the microphone audio. This is only useful if you can get the caller to install your software though, so it's likely not a workable solution.

 

[ransiluj]

Do you mean background noise, like other people talking in the same room?

not actually other people talking. Like noise of vehicales, sound of car horns, etc

 

[ransiluj]

If you are trying to suppress the noise from the other end, that's very difficult since the noise and the voice are in the same band of frequencies. If you can make a device to separate voice from noise, it would be worth millions and you wouldn't have to work any more. Hearing aid manufacturers have been trying to do that forever.

so cancelling noise coming out from vehilces, noise of machines etc. is it not possible ? can't we design filter to cancel those noise ?

 

[ransiluj]

If you're planning on using DSP, there are many papers on noise reduction algorithms e.g. Kalman, Wiener or blind speech separation using ICA-based approaches. Do a search and you'll find many more advanced methods described too.

It's best to process the audio before it's been encoded, compressed and sent over the network - as the noise takes up bandwidth rather than the speech and will make it difficult to get decent quality. You could have a think about implementing the DSP on the phone itself if it's possible to intercept the microphone audio. This is only useful if you can get the caller to install your software though, so it's likely not a workable solution.

um planing to make separete device like plug into phone like your handfree. do u think is it possible ? build such device ? if it is not please explain. thank you

 

[Mikebits]

When I worked for Qualcomm, we had a team of engineers working on that sort of thing, after a few years, they sorta had it working.

 

[ransiluj]

When I worked for Qualcomm, we had a team of engineers working on that sort of thing, after a few years, they sorta had it working.

that means it is possible but this is my final year project. i have only one year for this. So is that time possible ? not actually one year it is like 9 months project plus i have 5 subjects in each semester. There are two semesters. so is it possible?

 

[ronv]

As Mike mentioned in post 3. noise-cancelling microphones

I don't think it is easy.

 

[ransiluj]

As Mike mentioned in post 3. noise-cancelling microphones

I don't think it is easy.

what if i use two mics? one for voice and one for noise. so i can deduct the noise part from the voice. how about that idea? someone who did this can help me for sure.

 

[crutschow]

what if i use two mics? one for voice and one for noise. so i can deduct the noise part from the voice. how about that idea? someone who did this can help me for sure.

That would have little advantage over a noise-cancelling mic which basically does that same thing with one mic.

 

[MikeMl]

That would have little advantage over a noise-cancelling mic which basically does that same thing with one mic.

Actually, the noise-cancelling boom mics on my aircraft headsets use two electret elements wired out-of-phase in series. The mic orientation is critical, one element facing the mouth close to the lips; the other element facing away. The voice stimulates the closer element while the cabin noise stimulates both elements equally, resulting in ~20db improvement in the SNR/

 

[ransiluj]

That would have little advantage over a noise-cancelling mic which basically does that same thing with one mic.

we can reduce periodic noise like noise of a generator because it's generates same sound countinuesly. (helicopter noise also same). what if non periodic sound? do you have any idea to reduce such noise?

 

[crutschow]

Actually, the noise-cancelling boom mics on my aircraft headsets use two electret elements wired out-of-phase in series. The mic orientation is critical, one element facing the mouth close to the lips; the other element facing away. The voice stimulates the closer element while the cabin noise stimulates both elements equally, resulting in ~20db improvement in the SNR/

Some noise cancelling mics simply apply the voice/noise to the front side of the mic diaphragm and the noise alone to the back side of the diaphragm, as described here. That technique would seem to be simpler, but perhaps it isn't as effective.

 

[MikeMl]

Some noise cancelling mics simply apply the voice/noise to the front side of the mic diaphragm and the noise alone to the back side of the diaphragm, as described here. That technique would seem to be simpler, but perhaps it isn't as effective.

I have some noise-cancelling Ham and Commercial two-way communication Hand mics that use this technique, they are called "ported" mics. On the aviation boom mics, the two electret mic elements together are the size of a pencil eraser, so there are two electrets back to back.