Splitting one channel into two using 2 into 1 stereo cable giving me ZERO lags ? help !

Now that I'm using two condenser microphone inputs to a single channel , I'm getting zero lags ! ? This shouldn't happen since I'm placing the two mics at quite a distance for audio localization . But still getting the same result and 89 degrees or 90 degrees angle each time ! Can some one share a better code of sound localization using generalized cross correlation on two condenser mics connected to your laptop? Please help! Since it is the only hurdle in my project. Even a correct algorithm or way to calculate angle from GCC information will help! Thanks!

[Walter Roberson]

Could you confirm that you are merging two analog signals into one and that combined analog signal is what is being input?

What I have seen of sound localization usually involves multiple channels.

[Ali Soban]

Yes probably you have got me Walter! Actually I'm splitting a single mic input into two and then connecting the two mics into that 1 to 2 split stereo cable! Is it right what I'm doing for sound localization or I need to perform it in a different way? Any suggestions? Thanks for a quick reply ..!

[Ali Soban]

This is the cable I'm using ! Kindly see the attachment and please guide me on this issue! How to correct it and calculate accurate phase shifts using cross correlation! Thanks !

Regards,

Ali

[Walter Roberson]

Is it a stereo microphone? And you are splitting the two channels and then recombining them again? Or is it a mono microphone that you are splitting into two copies and taking those as two channels? If it is the mono microphone possibility then I would expect the signals to be identical except for possible internal noise.

[Ali Soban]

Would it give me different outputs if I use a stereo microphone? Yes microphone is stereo and I'm using the stereo splitter as shown in the attachment . But I'm using an external USB sound card which is probably not stereo!

[Walter Roberson]

USB sound cards are usually stereo. Sometimes they have additional channels as well.

I do not understand whether your intention is to combine the two channels into a single channel, or if you want the two channels separate? If you do want the two channels separate, then I do not understand why you are using a splitter cable (unless it just happens to be the case that the stereo mic breaks out the two channels instead of combining them on one cable) ??

[Ali Soban]

I want to acquire audio simultaneously from 2 audio channels with same start time! How may I do that without using any external multi-channel sound card or mixer etc. Kindly guide me because I'm stuck and unable to calculate the phase shift in sound localization! Thanks

[Walter Roberson]

Use two different mono microphones that are a distance apart from each other, and use that 2 to 1 cable to merge them into different channels on the same connector, and plug the connector into you stereo mic adapter on your sound card.

[Ali Soban]

Well thanks alot Walter. How do I acquire the sound on 2 different channels in MATLAB . I mean which command/function to use that best serves the operation of recording audio on two separate channels in the same array or may be different?

[Walter Roberson]

audiorecorder() grabs all of the channels of the audio device by default. The different channels appear as different columns of the returned data.

[Ali Soban]

But are you sure that it would be a synchronous recording this way ?

Answer 1

[Gabriele Bunkheila]

Hi Ali,

A number of things could be happeninig here - low-end or built-in laptop sound cards most often have mono mic or line inputs (despite the drivers possibly suggesting otherwise). Are your acquired signals exactly identical or simply well-correlated in time? On the other hand, your cable may be a single stereo to 2x stereo instead of a stereo splitter. I suggest you check all those possible issues in isolation.

Regards,

Gabriele.

[Ali Soban]

Hi Gabriele ,

Kindly see the stereo cable image attachment in my above comment so that you possibly be getting a fair estimate to where I'm actually stuck now. No the signals are not purely identical but they maximum correlation coefficient occurs at zero phase shift i.e.

[c lag] = xcorr(L,R)

[maxC I] = max(c)

lag = lag(I)

This lag = 0 always. I don't know the reason since the signals aren't purely identical!