5

When you listen to a recording of your own voice it seems alien to you, because it does not include the sounds that were conducted through your skull to the ear.

Is it possible to recreate your "inner voice" by using a recording of the "outer voice"? Maybe by filtering/increase/decrease some frequencies of the original and combining both (as you still here your voice through air), maybe even adding a small delay if the propagation velocity is different? I would imagine manifacturers of bone-conducting headphones do the same in reverse to counter the distortion from the bone, but I cannot find any sources or even what the best keywords are (I'm not even sure this is the right stackexchange). Any information on how sound is affected by traveling through bone and how the human ear perceives it would be helpful.

1
  • Sounds like a fun experiment :)
    – musicin3d
    Aug 14, 2019 at 4:31

2 Answers 2

1

I am quite sure it's possible... It would make a beautiful art resource - and after that, lose its magic as an audio plugin like any other hehe.

Yes, I think this is one of the right StackExchanges to ask :). The other would be the Signal Processing StackExchange.

This could be achieved either by modelling the sound transmission inside the subject's head or by recording with a (maybe not standard) dummy head like the ones used for binaural recording. The effectiveness of this modelling (or of the recording with a dummy head) will be impacted by the similarity of the model/dummy head in relation to the head of the subject that will hear this processed/recorded audio - in the same way binaural recordings made with dummy heads work well for the people whose hrtf (head-related transfer function) is similar to the one of the dummy head used in the recording, while these don't work very well for subjects with heads that are too different to the dummy head.

In terms of modelling, this task looks analogous to Dereverberation (removing the reverberation of a sound). Techniques range form physical modelling to training neural networks.

Some keywords may be dereverberation; hrtf (head-related transfer function); deconvolution

Keep us posted ;)

0

Unfortunately not. the "inner voice" to which you refer is the sound of your voice as perceived by the processing done by your brain, consequently it cannot be recorded.

Additionally, everyone has a different bone-structure and consequently even if it was possible to capture this sound and create some sort of 'impulse response', it would be different for every single person as no two skulls or brains are the same.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.