Home |
Search |
Today's Posts |
#1
|
|||
|
|||
simplest way to measure a mics frequency response?
Lacking the necessary funding for brand new condenser mics, I am going to be
constructing one of my own soon. I understand the basics of construction (circuit, housing, grounding, etc) but I am not sure how to measure the frequency response once completed. What is the simplest way to do it? I don't really have much experience doing something like this by ear, unfortunately. Also, regarding the housing of the mic, I understand that if done improperly, it can effect the lower response of the mic so any suggestions in this area would be helpful also. Thank you, Chase |
#2
|
|||
|
|||
On Fri, 29 Oct 2004 07:23:00 GMT, "Chase Sunderlin"
wrote: Lacking the necessary funding for brand new condenser mics, I am going to be constructing one of my own soon. I understand the basics of construction (circuit, housing, grounding, etc) but I am not sure how to measure the frequency response once completed. What is the simplest way to do it? I don't really have much experience doing something like this by ear, unfortunately. Also, regarding the housing of the mic, I understand that if done improperly, it can effect the lower response of the mic so any suggestions in this area would be helpful also. Thank you, Chase A good reliable way is to set up an external electrode fairly close to the diaphragm. Put a high voltage on this electrode - this will distort the diaphragm slightly away from its rest position. Now remove the voltage cleanly and sharply. Read the resulting pulse from the microphone into a PC through a sound card, and perform an FFT on it. Compare this to an FFT of the edge of the voltage pulse itself, and you have the frequency response of the capsule. Most DAW programmes have an FFT (frequency analysis in Audition) that will do this for you. d Pearce Consulting http://www.pearce.uk.com |
#3
|
|||
|
|||
On Fri, 29 Oct 2004 07:30:51 GMT, in rec.audio.pro you wrote:
On Fri, 29 Oct 2004 07:23:00 GMT, "Chase Sunderlin" wrote: Lacking the necessary funding for brand new condenser mics, I am going to be constructing one of my own soon. I understand the basics of construction (circuit, housing, grounding, etc) but I am not sure how to measure the frequency response once completed. What is the simplest way to do it? I don't really have much experience doing something like this by ear, unfortunately. Also, regarding the housing of the mic, I understand that if done improperly, it can effect the lower response of the mic so any suggestions in this area would be helpful also. Thank you, Chase A good reliable way is to set up an external electrode fairly close to the diaphragm. Put a high voltage on this electrode - this will distort the diaphragm slightly away from its rest position. Now remove the voltage cleanly and sharply. Read the resulting pulse from the microphone into a PC through a sound card, and perform an FFT on it. Compare this to an FFT of the edge of the voltage pulse itself, and you have the frequency response of the capsule. Most DAW programmes have an FFT (frequency analysis in Audition) that will do this for you. d Pearce Consulting http://www.pearce.uk.com Hi Don How about the acoustic pulse from an piezoelectric gas lighter thingummy? martin Serious error. All shortcuts have disappeared. Screen. Mind. Both are blank. |
#4
|
|||
|
|||
On Fri, 29 Oct 2004 13:18:47 +0200, martin griffith
wrote: On Fri, 29 Oct 2004 07:30:51 GMT, in rec.audio.pro you wrote: On Fri, 29 Oct 2004 07:23:00 GMT, "Chase Sunderlin" wrote: Lacking the necessary funding for brand new condenser mics, I am going to be constructing one of my own soon. I understand the basics of construction (circuit, housing, grounding, etc) but I am not sure how to measure the frequency response once completed. What is the simplest way to do it? I don't really have much experience doing something like this by ear, unfortunately. Also, regarding the housing of the mic, I understand that if done improperly, it can effect the lower response of the mic so any suggestions in this area would be helpful also. Thank you, Chase A good reliable way is to set up an external electrode fairly close to the diaphragm. Put a high voltage on this electrode - this will distort the diaphragm slightly away from its rest position. Now remove the voltage cleanly and sharply. Read the resulting pulse from the microphone into a PC through a sound card, and perform an FFT on it. Compare this to an FFT of the edge of the voltage pulse itself, and you have the frequency response of the capsule. Most DAW programmes have an FFT (frequency analysis in Audition) that will do this for you. d Pearce Consulting http://www.pearce.uk.com Hi Don How about the acoustic pulse from an piezoelectric gas lighter thingummy? martin The purpose of the technique is to remove the uncertainty of the acoustic path and just test the microphone. You know exactly what you are doing with the diaphragm when you pull it electrically. Of course when you let go, you are recovering the way the microphone diaphragm moves exactly. If you fancy trying an impulse response from an acoustic source, an electric spark is a good way to do it, but it needs to be a much louder splat than you get from a gas lighter. d Pearce Consulting http://www.pearce.uk.com |
#5
|
|||
|
|||
On Fri, 29 Oct 2004 11:25:23 GMT, in rec.audio.pro you wrote:
On Fri, 29 Oct 2004 13:18:47 +0200, martin griffith wrote: On Fri, 29 Oct 2004 07:30:51 GMT, in rec.audio.pro you wrote: On Fri, 29 Oct 2004 07:23:00 GMT, "Chase Sunderlin" wrote: snip Hi Don How about the acoustic pulse from an piezoelectric gas lighter thingummy? martin The purpose of the technique is to remove the uncertainty of the acoustic path and just test the microphone. You know exactly what you are doing with the diaphragm when you pull it electrically. Of course when you let go, you are recovering the way the microphone diaphragm moves exactly. If you fancy trying an impulse response from an acoustic source, an electric spark is a good way to do it, but it needs to be a much louder splat than you get from a gas lighter. d Pearce Consulting http://www.pearce.uk.com Ok Sloppy thinking by me. Should have thought out the basic principles before posting! martin Serious error. All shortcuts have disappeared. Screen. Mind. Both are blank. |
#7
|
|||
|
|||
Define "cheap"? Sorry, I don't think $200 or more for a mic is inexpensive. Electret condenser elements, if carefully selected, can be had for about $5, placed into a capsule and wired up. The reason I was asking about testing was because I would then need to determine where eq might need to be applied. Simply doing this by ear would not have been the way to do it for me. Chase Phil Brown Maybe $200 for a microphone isn't cheap[ to you,but there are a lot of microphones out thete in the under $100 range that have known characteristics. Might be a lot easier to modify one of those. Richard H. Kuschel "I canna change the law of physics."-----Scotty |
#8
|
|||
|
|||
On Fri, 29 Oct 2004 15:22:37 GMT, "Chase Sunderlin"
wrote: "Phil Brown" wrote in message ... Lacking the necessary funding for brand new condenser mics, I am going to be constructing one of my own soon. Huh? Mics are so cheap today I don't understand how making one yourself could possibly ne cheaper. Define "cheap"? Sorry, I don't think $200 or more for a mic is inexpensive. Electret condenser elements, if carefully selected, can be had for about $5, placed into a capsule and wired up. The reason I was asking about testing was because I would then need to determine where eq might need to be applied. Most here will suggest you construct the mic for flat response (or for whatever response you want it to have) rather than using EQ. If you want to test its response, you can probably compare it with a 'standard' - there's been a $40 Behringer model that's been described as having a flat enough response for many testing purposes. I recall the late Stephen Paul discussing mic acoustics on a webforum a few years back, regrettably he closed the forum and removed the content due to some controversy/misunderstanding, and more recently, I saw mic acoustics discussed on a webforum moderated by Harvey Gerst. There's probably one or more books' worth of info for "comparing" two mics (frequency response sweep from high quality loudspeaker, acoustically dead room (or the outdoors doesn't reflect much), bla bla bla), but this is at least a start. Perhaps that free frequency-response measuring program Arny has mentioned (Rightmark?) would be usable here. Someone suggested using a gas lighter to generate an acoustic impulse and then do FFT to get the response, and someone else gave some reasons it wouldn't work (IIRC the gas lighter's spark isn't loud enough). I would think the biggest problem would be the mechanical hammer that hits the piezoelectric element probably makes more sound than the spark itself. Simply doing this by ear would not have been the way to do it for me. You may want to look at the "Mitey Mike": http://www.audioxpress.com/bksprods/kits/kd-4.htm I saw the original Mitey Mike construction article in Audio Amateur or perhaps Speaker Builder (both of which are incorporated into the current AudioXpress magazine) many years ago. You might want to order the issue(s) describing it and read up on it. I recall that the mic element is set into the end of a fairly long (about 1 foot) tube that's just slightly larger in inside diameter than the element, to get a flat reponse and prevent the frequency-response effects of nearby objects (such as the preamp box at the other end of the tube). Also there's a DIY mic mailing list on yahoogroups.com. I recall that it's mostly for the "stealth tapers" crowd, but it may have some useful ideas. Chase Phil Brown ----- http://mindspring.com/~benbradley |
#9
|
|||
|
|||
Chase Sunderlin wrote:
Lacking the necessary funding for brand new condenser mics, I am going to be constructing one of my own soon. I understand the basics of construction (circuit, housing, grounding, etc) but I am not sure how to measure the frequency response once completed. What is the simplest way to do it? I don't really have much experience doing something like this by ear, unfortunately. Also, regarding the housing of the mic, I understand that if done improperly, it can effect the lower response of the mic so any suggestions in this area would be helpful also. You want a reference mic ? Buy Behringer's one. They've done all the hard work for you. They don't cost much. Far less than it'll cost you to make one ( assuming you could which I doubt in practice ). I admire your enthusiasm. You clearly haven't a idea about the technical issues involved though. Graham |
#10
|
|||
|
|||
A good reliable way is to set up an external electrode fairly close to
the diaphragm. Put a high voltage on this electrode - this will distort the diaphragm slightly away from its rest position. Now remove the voltage cleanly and sharply. Read the resulting pulse from the microphone into a PC through a sound card, and perform an FFT on it. Compare this to an FFT of the edge of the voltage pulse itself, and you have the frequency response of the capsule. Most DAW programmes have an FFT (frequency analysis in Audition) that will do this for you. That's just what I was going to say. Really. |
#11
|
|||
|
|||
|
#12
|
|||
|
|||
Ben Bradley wrote: Someone suggested using a gas lighter to generate an acoustic impulse and then do FFT to get the response, and someone else gave some reasons it wouldn't work (IIRC the gas lighter's spark isn't loud enough). I would think the biggest problem would be the mechanical hammer that hits the piezoelectric element probably makes more sound than the spark itself. There is a low frequency rolloff from a spark of about 12 db per octave and the corner frequency is in inverse proportion to the gap length. Even at a half inch, I've measured that corner to be a couple of kHz. There isn't enough low frequency energy in a spark gap to get anywhere near ambient noise much less above it. An exploding wire might be a good idea. I built a high voltage, high energy supply for this purpose from a kit a couple of years ago but when I came to realize just how lethal the thing is, I decided to let it gather dust rather than turn it on. Impulsive methods are problematic for other reasons too, the main one being signal to noise ratio. Their advantage is mainly for measuring halls and spaces because they are omnidirectional. When all the energy is in a pulse, the total isn't very much at all. Swept sin methods are much better because you are spreading your energy out over time at relatively high power at all frequencies and can get _much_ better SNR in your measurement. To make this work, you need to calibrate your sound source with a calibration mic and do impulse response division of all measurements by the one measured by the calibration mic. The Beringer mic is supposed to be flat enough but since it contains a Panasonic WM-61A capsules I just don't believe that. From the factory they have a 12 kHz presence peak that is plus or minus about 2 kHz and plus or minus about 6 dB by my measurement of about 50 of them. I really don't believe Behringer selects for matched characteristincs, the standard deviation is just too broad, and they just don't come flat. A much better choice for a calibration mic that is useful also for recording is the Earthworks TC30K. The actually do the capsule qualification and selection and ask you to pay for it in the price. With a matched pair you get the calibration curves and they _are_ flat. But then thats clearly outside your budget. It's gonna be hard to do what you want on the cheap. (Maybe I ought to sell measurement mics that don't pretend to be flat but are cheap like the Behringer and come with their calibrated impulse response on a floppy. :-) If you want to take this on, Chase, I've done a lot of work with all this (and still am) and would be happy to advise via email. Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#13
|
|||
|
|||
On Sat, 30 Oct 2004 02:32:51 -0700, Bob Cain
wrote: Ben Bradley wrote: Someone suggested using a gas lighter to generate an acoustic impulse and then do FFT to get the response, and someone else gave some reasons it wouldn't work (IIRC the gas lighter's spark isn't loud enough). I would think the biggest problem would be the mechanical hammer that hits the piezoelectric element probably makes more sound than the spark itself. There is a low frequency rolloff from a spark of about 12 db per octave and the corner frequency is in inverse proportion to the gap length. Even at a half inch, I've measured that corner to be a couple of kHz. There isn't enough low frequency energy in a spark gap to get anywhere near ambient noise much less above it. Aha, that explains why a step response would be a better idea. I saw the description in this thread of electrostatic excitation of a step response, but I can't think of a way to do that acoustically. ... you want on the cheap. (Maybe I ought to sell measurement mics that don't pretend to be flat but are cheap like the Behringer and come with their calibrated impulse response on a floppy. :-) CD-R's are cheaper, not that the price of the media matters much, but more importantly, newer computers don't have floppy drives anymore. Also, you can put more data on a CD-R. But that does sound like an interesting business idea. Are there services that will make such a measurement on a mic the customer sends in? You could do that as well. If you want to take this on, Chase, I've done a lot of work with all this (and still am) and would be happy to advise via email. Bob ----- http://mindspring.com/~benbradley |
#14
|
|||
|
|||
Ben Bradley wrote: Aha, that explains why a step response would be a better idea. I saw the description in this thread of electrostatic excitation of a step response, but I can't think of a way to do that acoustically. Yes it is very difficult to make atmospheric pressure suddenly jump to a new value and stay there. :-) I'm not sure what the value of electrostatic excitation is. Might tell you information about diaphragm tension and port accuracy if you have something standard to compare it with but it sure isn't going to tell you much about how it transduces sound. you want on the cheap. (Maybe I ought to sell measurement mics that don't pretend to be flat but are cheap like the Behringer and come with their calibrated impulse response on a floppy. :-) CD-R's are cheaper, not that the price of the media matters much, but more importantly, newer computers don't have floppy drives anymore. Also, you can put more data on a CD-R. My age is showing. :-) But that does sound like an interesting business idea. Are there services that will make such a measurement on a mic the customer sends in? You could do that as well. I have thought about it. With the general availaibility of convolution, having the impulse responses and the inverses at several orientations could be useful. Not sure how many people would know how to use the results though and I doubt they'd pay just for a graph. Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
common mode rejection vs. crosstalk | Pro Audio | |||
Topic Police | Pro Audio |