Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Mic suggestions for Sound source localization using Binaural Audio. (Read 2816 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Mic suggestions for Sound source localization using Binaural Audio.

Hey, this is Manuel Jenkin. I am a Final year Electrical and Electronics student and am currently doing a project on sound source localization and separation using binaural mics and machine learning. I need some help with choosing a mic for my purpose.
Here is the thesis upon which we are doing the project on

https://www.google.co.in/url?sa=t&source=web&rct=j&url=https://tel.archives-ouvertes.fr/tel-00913965/file/thesis.pdf&ved=0ahUKEwi_x9X98ZvSAhVLgI8KHWW9C1sQFggoMAI&usg=AFQjCNHbmAIYIDCywg4I2aegseQ6iSkELQ&sig2=YvxCBg5OJd-bI85_0_owJA

The procedure we are doing is given in chapter 2 section 2.3.3 . What we are going to do is have a mic placed near the ceiling at about the center of the room and trace stationary objects on the ground using sound emitted from them. The room is of size 30ft length (about 10m) and 12ft wide (about 4m). It's ceiling is about 2.5-3metres (about 9 feet). There are chairs and desks around and we'll be using only a small portion of the room for the project. We'll be placing the mic near the ceiling at about 2-2.5 metres (about 8 feet) above the floor and we'll be tracing the objects in a confined space of about 9ft x 9ft (3m x 3m) or at max about 12ft x 15ft (4m x 5m).

The criterion for the mic for the project is that i should be able to detect sounds upto atleast 5 m (preferrably upto 8m) away from it and also have good fidelity upto 10Khz. I will be putting two of them inside a dummy head similar to a binaural mic.

They seem to be using a Sennheiser MKE2002 Mic in the thesis which seems to have been discontinued. I was seeking advice from a couple of people for alternatives which are relatively less expensive and I got quite differing options.

The first advice I got was to use omni directional microphones. He suggested me to use omni lavalier mics like the Audio Technica AT803b. Two other options he gave me are - Jecklin disk with a custom disk configuration with omni mics - or - ORTF with hypercardioid mics such as AT4053b. He also adviced me to do a diy binaural head and gave me a link to see how to do it : http://home.hccnet.nl/joop.nijenhuis/headwize/hpdbmtre.htm

Then I had some chat with another guy and he gave me slightly differing opinions. He first recommended me the DPA4060 (it is a little bit high on the pricy side for me though). But then told me that lavalier mic is not good for getting sound from an entire room, but only to pick up sounds right next to it. He then advised to use small or large condenser mics. He also showed me the polar pickup patterns : http://paperclip.rcs.ac.uk/images/0/08/Polarpatterns.jpg which was quite useful to understand what I might need.
He recommended me the Rode M5, a cardoid mic. I got a doubt that if i place a couple of them the way ears are, then won't the center of the room be lost a bit because of them being condenser (which cover only the front and sides wrt mic) placed 180* away from each other. And also that the left mic won't get right side interaural cues, and right mic wont get left interaural cues, both of which are necessary to trace the exact location of the object (not sure about that though, and not sure if the sennheiser mke2002 mics record the other side info). He replied that it wont have an issue because everything that the mic will receive will be reflected from the ears of the dummy head and nothing will be from inside the dummy head. All those that wrap through the dummy head are below 200hz which dont contain much spatial cues anyway. Also he adviced me to use a cardoid over omnidirectional because omni might pick up noise from inside the dummy head.

I am kind of confused now about what are the tagwords that I can use to describe the mic required for this purpose. Would I need a lavalier, or condenser mic or anything else? Would i need a cardoid, or omnidirectional mic or anything else? I dont know the exact type of mic used in the MKE2002. I want anything that resembles the Sennheiser MKE2002 as much as it can (I know it is hard because they are discontinued, but i want something that can perform similar under the given circumstances), because the thesis is done with them and they had relatively good results.

I need to choose a mic for the project and request my university to fund it. Can you suggest me some other relatively inexpensive mics that can do this job at about 300$ to 600$ per pair, i.e. 150$ to 300 or 400$ per mic (upto to 1000$ max, but thats quite stretching it hard)? And I need multiple options because I'm from India and it is not a guarantee that all products will be available here (the AT803b, for example is unavailable here). Mostly sennheiser, shure,neumann, sony, etc are the companies that are available here.. Of course there are places where I can find other company stuff and also I can import from singapore if needed, so any option that you guys can give me will be very helpful. Also I need ideas on how I shoud proceed with the dummy head as well. I have some ideas, but not sure. Is there any setup that will work for this purpose without a dummy head? Thank you.

NOTE : I am new to the forum and not sure if this question belongs here or any other subthread. If there is any other subthread, where this question belongs, I request you to redirect me there!. Thank you.

Re: Mic suggestions for Sound source localization using Binaural Audio.

Reply #1 –
Any "professional" lavalier mic from one of the pro manufacturers (Shure, Audio Technica, Sennheiser, etc.) should work.    It will probably be an electret condenser because that type of mic can be made small and high quality.      A professional headset mic should work too.  (Not a cheap computer gaming or communication mic.)

Note that condenser mics have electronics inside so they need power.  An electret condenser may have an internal battery.  If not, it will need 48V phantom power, which will be supplied through the XLR connector on the mixer/interface.   

A regular consumer soundcard will provide 5V for a computer mic, but it's the wrong interface for any pro mic.    "Computer" mics tend to be low-quality, the preamp built-into a regular soundcard/laptop also tends to be low-quality, and pro & computer mics are not interchangeable.    Pro (stage or studio) mics are low-impedance balanced (3-wire) with an XLR connector.   (There are some good quality "studio style" USB mics, but in general you can only use one at a time.)

A cardioid mic is directional, so it's the opposite of an omnidirectional mic.    Directional mics tend to be larger and if you stick one into an artificial head you'll probably block the ports and it will no longer be directional (and you'll change the sound characteristics).    And, I'm pretty sure you need to pick-up sounds from all directions because your direction-finding needs both mics.

Quote
But then told me that lavalier mic is not good for getting sound from an entire room,
No.  A non-directional microphone doesn't "know" anything about the distance.  Cardioid and certain communication mics do behave differently with close & distant sounds.    (Cardioid mics have something called proximity effect which boosts bass at close frequencies, and close-to-the mouth communication mics are designed to cancel sounds that are not very close and from the front.)

Re: Mic suggestions for Sound source localization using Binaural Audio.

Reply #2 –
Quote from: DVDdoug on
Any "professional" lavalier mic from one of the pro manufacturers (Shure, Audio Technica, Sennheiser, etc.) should work.    It will probably be an electret condenser because that type of mic can be made small and high quality.      A professional headset mic should work too.  (Not a cheap computer gaming or communication mic.)

Note that condenser mics have electronics inside so they need power.  An electret condenser may have an internal battery.  If not, it will need 48V phantom power, which will be supplied through the XLR connector on the mixer/interface.   

A regular consumer soundcard will provide 5V for a computer mic, but it's the wrong interface for any pro mic.    "Computer" mics tend to be low-quality, the preamp built-into a regular soundcard/laptop also tends to be low-quality, and pro & computer mics are not interchangeable.    Pro (stage or studio) mics are low-impedance balanced (3-wire) with an XLR connector.   (There are some good quality "studio style" USB mics, but in general you can only use one at a time.)

A cardioid mic is directional, so it's the opposite of an omnidirectional mic.    Directional mics tend to be larger and if you stick one into an artificial head you'll probably block the ports and it will no longer be directional (and you'll change the sound characteristics).    And, I'm pretty sure you need to pick-up sounds from all directions because your direction-finding needs both mics.

Quote
But then told me that lavalier mic is not good for getting sound from an entire room,
No.  A non-directional microphone doesn't "know" anything about the distance.  Cardioid and certain communication mics do behave differently with close & distant sounds.    (Cardioid mics have something called proximity effect which boosts bass at close frequencies, and close-to-the mouth communication mics are designed to cancel sounds that are not very close and from the front.)

Hey thanks. I got some replies for the same question at reddit.

https://www.reddit.com/r/audioengineering/comments/5xmnvs/mic_suggestions_for_sound_source_localization/deji4qo/

I was recommended some omnidirectional measurement mics because it seems like other music/field oriented mics colour the spectrum for compensation for their position etc.. Measurement mics seem to be geared towards my application of spectral analysis

I am leaning towards the presonus prm1 or audix tm1 or any other measurement mic if you can recommend them

https://www.bhphotovideo.com/c/product/922336-REG/presonus_prm1_precision_reference.html

and using the focusrite scarlett 2i2 interface

I am just worried if this combination will have good enough reach/sensitivity for my application . Being able to record sounds well for volume level of a general conversation at about say 6 or 8 metres away from mic at all 360*, especially at a right angle to the mic, because of the way I position the mics in binaural head relative to sound source.

Thank you.