Hi all,
I've been pondering this one for a while. I couldn't think of a sensible explanation by myself, so thought I'd see if anyone around here has any clues.
My aim is simple: I'd like to be able to measure microphones. I'm a live sound engineer, and use a variety of mics for my work. I'd like to be able to measure them to ensure they're working properly, performance isn't degrading, and also to satisfy my own curiosity.
My test speaker isn't anything special. It appears to be a Vifa TG9, but with a black cone. The cone still has the same texture of the usual white version, so I expect it was done for a custom order. It's in a sealed box and good for about 150Hz-15kHz. REW's calibration holds up okay for slow rolloffs, so I'm happy collecting data down to about 50Hz, so long as there's no external noise to mess things up.
Moving on towards the mystery:
When I put a microphone in front of the speaker and run sweeps, there are plenty of reflections clearly visible. The room is small, and doesn't have much absorption. In order to eliminate some of the reflections, I made an absorbing tube to put over the cone of the test speaker, so I could put the mic at the other end and reduce the amount of sound thrown sideways. NB: most stage mics are pretty insensitive to sound arriving from the rear, so sound going past the mic and bouncing back is a smaller concern.
The construction of the tube is very simple. There's a layer of steel speaker grille (the most "open" I could find) which is curved into a 5" diameter tube, and then I added an inch thickness of recycled wool underlay to the inside. That matched nicely to the 3" cone of the test speaker. I then wrapped the outside in another inch of wool. In essence, then, there's 2" of wool and a thin steel mesh enclosing the mic and speaker. Open at the mic end, closed at the speaker end.
The impulse response showed improvement, with a lot of nearby HF reflections disappearing almost entirely. An excellent result!
What surprised me, however, was the dramatic change in the frequency response:
Unfortunately, there was a power cut while REW was open, and my knackered old laptop shut down instantly. I've recovered the with-cylinder measurement (green) as a text file which I'd saved for future use, and compared it to a previous measurement (blue) of the speaker.
The response was pretty flat, and now there's a ~30dB swing from 150Hz to 3kHz. Same mic, same interface, same speaker. I had expected no change in the >10kHz range, where the speaker's beamwidth is narrow, and for the absorption to tame some reflection-related "fuzz" in the range where the absorption would be happening. ie, no effect below 500Hz-ish, but better above that.
So, any ideas about why the response shape has changed so much? Shall I just add it to my speaker calibration file and call it good?
Thanks in advance. This one really has me stumped.
Chris
I've been pondering this one for a while. I couldn't think of a sensible explanation by myself, so thought I'd see if anyone around here has any clues.
My aim is simple: I'd like to be able to measure microphones. I'm a live sound engineer, and use a variety of mics for my work. I'd like to be able to measure them to ensure they're working properly, performance isn't degrading, and also to satisfy my own curiosity.
My test speaker isn't anything special. It appears to be a Vifa TG9, but with a black cone. The cone still has the same texture of the usual white version, so I expect it was done for a custom order. It's in a sealed box and good for about 150Hz-15kHz. REW's calibration holds up okay for slow rolloffs, so I'm happy collecting data down to about 50Hz, so long as there's no external noise to mess things up.
Moving on towards the mystery:
When I put a microphone in front of the speaker and run sweeps, there are plenty of reflections clearly visible. The room is small, and doesn't have much absorption. In order to eliminate some of the reflections, I made an absorbing tube to put over the cone of the test speaker, so I could put the mic at the other end and reduce the amount of sound thrown sideways. NB: most stage mics are pretty insensitive to sound arriving from the rear, so sound going past the mic and bouncing back is a smaller concern.
The construction of the tube is very simple. There's a layer of steel speaker grille (the most "open" I could find) which is curved into a 5" diameter tube, and then I added an inch thickness of recycled wool underlay to the inside. That matched nicely to the 3" cone of the test speaker. I then wrapped the outside in another inch of wool. In essence, then, there's 2" of wool and a thin steel mesh enclosing the mic and speaker. Open at the mic end, closed at the speaker end.
The impulse response showed improvement, with a lot of nearby HF reflections disappearing almost entirely. An excellent result!
What surprised me, however, was the dramatic change in the frequency response:
Unfortunately, there was a power cut while REW was open, and my knackered old laptop shut down instantly. I've recovered the with-cylinder measurement (green) as a text file which I'd saved for future use, and compared it to a previous measurement (blue) of the speaker.
The response was pretty flat, and now there's a ~30dB swing from 150Hz to 3kHz. Same mic, same interface, same speaker. I had expected no change in the >10kHz range, where the speaker's beamwidth is narrow, and for the absorption to tame some reflection-related "fuzz" in the range where the absorption would be happening. ie, no effect below 500Hz-ish, but better above that.
So, any ideas about why the response shape has changed so much? Shall I just add it to my speaker calibration file and call it good?
Thanks in advance. This one really has me stumped.
Chris
