Earl,
The link I posted contained some real measurements. Svante measured several which have a +3dB to +4 peaks.
This thread in the topic I posted:
http://www.diyaudio.com/forums/showthread.php?postid=1457797#post1457797
Contains several calibration files which range from +2.5dB to my one with two 6dB peaks. This calibration file was not created against a reference microphone, but by doing a delta from several measurements of a Seas 27TBFC/G and 27TDFC tweeters. As I said, it was the only form of calibration I could do given I was not prepared to spend hundreds shipping overseas and back for calibration. I never said it was as good as a proper calibration and admit some error is no doubt included. I think the process I followed was reasonable and better than not using any calibration at all.
David.
The link I posted contained some real measurements. Svante measured several which have a +3dB to +4 peaks.
This thread in the topic I posted:
http://www.diyaudio.com/forums/showthread.php?postid=1457797#post1457797
Contains several calibration files which range from +2.5dB to my one with two 6dB peaks. This calibration file was not created against a reference microphone, but by doing a delta from several measurements of a Seas 27TBFC/G and 27TDFC tweeters. As I said, it was the only form of calibration I could do given I was not prepared to spend hundreds shipping overseas and back for calibration. I never said it was as good as a proper calibration and admit some error is no doubt included. I think the process I followed was reasonable and better than not using any calibration at all.
David.
Acoustic Measurement Seminar
Keith Larson will be giving an acoustic measurement seminar at the Lone Star Audiofest, if any of you are interested. Admission is free. The Lone Star Audiofest is May 2, 3 and 4 in Dallas. The measurement seminar is Saturday at 1:00pm in the conference room. See link below for more info:
Keith Larson will be giving an acoustic measurement seminar at the Lone Star Audiofest, if any of you are interested. Admission is free. The Lone Star Audiofest is May 2, 3 and 4 in Dallas. The measurement seminar is Saturday at 1:00pm in the conference room. See link below for more info:
Dear Dr Geddes,
Interestingly measured response of my primitive ribbon microphone (-3.8dBu sine sweep) shows a kind of symmetrical impulse response. It has rising top end (slit/gap/transformer issue?). S-transform (Stockwell et al.) shows both room reflections and rapid decay = very low storage energy.
http://unizet.net/pub/S-TransformRibbon.png
Pictures:
http://unizet.net/pub/
Jack,
Interestingly measured response of my primitive ribbon microphone (-3.8dBu sine sweep) shows a kind of symmetrical impulse response. It has rising top end (slit/gap/transformer issue?). S-transform (Stockwell et al.) shows both room reflections and rapid decay = very low storage energy.
http://unizet.net/pub/S-TransformRibbon.png
Pictures:
http://unizet.net/pub/
Jack,
Thanks for the input guys - the info about a good impulse/big enough room was very helpful. . I think I'm SOL trying for a decent impulse in this small room. It's 15-1/2 X 11-1/2 feet and the speakers are 80 inches tall. 😀
I can't seem to move enough stuff around to get a good, clean first reflection in the impulse, which probably explains why the CSD looks so weird all the time. I'd try and take them outside, (I've got a coupla thousand acres all around me) but they're 4-way active and it'd be a mess to drag all the amps, crossover, preamp etc outside. Fortunately, I should be able to enjoy them anyway. 😀
I can't seem to move enough stuff around to get a good, clean first reflection in the impulse, which probably explains why the CSD looks so weird all the time. I'd try and take them outside, (I've got a coupla thousand acres all around me) but they're 4-way active and it'd be a mess to drag all the amps, crossover, preamp etc outside. Fortunately, I should be able to enjoy them anyway. 😀
AJ said:
Just make a long speaker cable(s) - can even use an extension cord (put the male prongs at the speaker end so somebody on the other end can't accidentally plug into 120VAC!) Leave the equipment inside. Yeah, probably more running back and forth, in and out of the house but better than moving everything . . . If your equipment room has a window, it's not so bad.
-- mark
Mark - I thought about that too and did some measuring. If I just moved them to the driveway maybe 25 feet from the house, the only meaningful first reflection (other than from the driveway) would be whatever bounces off the house. I could get about 25 feet on each side of the driveway; the driveway has two rows of 60 foot high cedars on each side that wouldn't be very reflective as they drape all the way to the lawn. The driveway itself ends at a gravel road and across from that is pretty much infinitely clear for half a mile. If I had the backs of the speakers facing the house (they're dipoles) then all I would be seeing would be the ~50 foot reflection (25 to the house, 25 back to the speaker) + the distance to the mic, which should allow a decent IR if nothing else.
So, maybe ~40 feet of speaker cable would do it? Is that too much length? I like the extension cord idea - I'm using Home Depot for speaker cables right now anyway. 🙂
So, maybe ~40 feet of speaker cable would do it? Is that too much length? I like the extension cord idea - I'm using Home Depot for speaker cables right now anyway. 🙂
AJ said:
Don't get obsessed with distance and LF capability, its a loosing proposition. Get a "good" clean 15 ms. and only go down to about 200 Hz. Then do a near field measurement and blend the two together. They will have enough overlap that this is not difficult. That way you can get 20 Hz - 20 kHz without much problem.
AJ said:
Should work great.
Cables are horribly overrated. (please, folks, no speaker cable flame wars intended; we're talking measurements here)
Assuming an ~8-ohm speaker, any decent 50' 12-14 ga outdoor extension cord should be more than adequate. My 2 cents. I'll bet Earl agrees, but I hope he'll chime in to confirm.
-- Mark
Tubamark said:
Of course - the cables arn't going to effect the measurements one bit - as long as they have a low enough resistance, which 12' - 14' gauage will.
gedlee said:
I agree on the speaker cables but you need to verify how long your mic and preamp cables can be without impacting the measurement. I find with my mic (Liberty Instruments Panasonic electret) and preamp, using a long (30') inexpensive patch cord between the mic preamp and mic results in some rolloff of the high frequencies. So I use the long cable between the mic preamp and PC with a relatively short cable between the mic and preamp with no problems.
Regards,
Dennis
ALso, it is possible to simply exclude the wire resistance even if it is high. Just measure as input the voltage AT the speaker and in a two channel measurement the wire loses will be excluded.
Yes, always locate the mic preamp as close as feasible to the mic, and run the higher level signla from the preamp to the PC.
Yes, always locate the mic preamp as close as feasible to the mic, and run the higher level signla from the preamp to the PC.
Paul W said:
Earl,
Here is my ECM 8000...calibrated by Kim G.
For a $40 mic, it's darned good from 20-20k. There is some unit-unit variation so this one isn't necessarily the same as other examples.
Paul
Just to follow-up, I just got my Behringer ECM8000 back from KimG and the frequency response is much worse than Paul's. Same basic curve, but +6.6db @ 11khz -- similar to what Dave reported. If someone is really interested, I can post a graph...
Regards,
John
Calibration results
... here you can find some detailed informations about the calibration of microphones. The result of 200 calibrations is shown:
http://diylautsprecher.de/index.php?option=com_content&task=view&id=138&Itemid=35
Regards
Heinrich
PS: ARTA 1.4 is comming on sunday
... here you can find some detailed informations about the calibration of microphones. The result of 200 calibrations is shown:
http://diylautsprecher.de/index.php?option=com_content&task=view&id=138&Itemid=35
Regards
Heinrich
PS: ARTA 1.4 is comming on sunday
nullspace said:
I think Dr. Geddes has said it all in very straightforward language. The most important thing are repeatable impulse responses that are distinct from room reflections. If you are measuring a giant loudspeaker in a small room, well, you're just going to have buy some long wires for the drivers, and measure the thing out on the driveway. The only wire that needs to be short is the coax connection between the microphone and the preamp, which typically sits at the base of the microphone stand.
Do not hold the microphone; use a mike stand. This is probably really obvious, but your body and the nearby PC reflects sound just fine, so stay well outside the flight path of the sound.
In terms of general observations, if moving the mike a small distance (a few inches) at a speaker-to-mike distance of 2~3 meters makes a big difference to the measurement, you have a problem. Assuming that room reflections are not the cause (see above), you might have issues with too many drivers scattered over too big of an area - Dahlquist DQ10 or line-radiator enthusiasts, take note.
P.S. Make sure the card table that is holding your computer (and other miscellaneous junk) is strong enough to hold it. It hasn't happened to me, but I've seen some expensive gear at Tektronix land on the concrete and it wasn't pretty.
Another brief note - it is possible to have RFI get into the microphone, cable, or mike preamp. I have had TV sync buzz (60 Hz and 15.75 kHz square waves) get into my system when I was in a high RFI area, so I've gotten into the habit of monitoring the output of the mike preamp with an inexpensive headphone amplifier, just to see if everything is free of electrical noise.
You should hear a very hifi mono version of reality when you audition what comes out of the measurement microphone - no AM radio stations, no TV sync buzz. RFI will definitely degrade your measurements - amusingly, Stereophile for years published measurements with visible artifacts from TV sync buzz at 15.75 kHz. Go back to an early-Nineties issue and see it for yourself.
You should hear a very hifi mono version of reality when you audition what comes out of the measurement microphone - no AM radio stations, no TV sync buzz. RFI will definitely degrade your measurements - amusingly, Stereophile for years published measurements with visible artifacts from TV sync buzz at 15.75 kHz. Go back to an early-Nineties issue and see it for yourself.
I've always wondered why no one has tried to subtract room reflections from measured data so that we can measure down to however low we like in any reasonably sized room.
nullspace said:
I agree that +/- 1dB FINAL measurement error margin is fine. This must include the input/output loop of the soundcard. If we get an excellent calibration file from a third party for our mic, we have to be sure of low/high pass aberrations in the rest of our chain, and integrate them to the file too.
For simulation work, the phase turn points at the edges of the measured spectrum can break the correct predicted summation at Fcross even if it is located smack centre. A check with Hilbert transformation can attest to that.
To conclude: Absolute precision on a single arbitrary axis for an already synthesized system is saying nothing much. But excellent wide phase data is good to have for trusty design work, and to have absolute calibration helps there.
P.S. Very good thread idea.
soongsc said:
I have tried this, many years back. I wrote a cople of papers on it. Its not that easy. LAUD uses the same techniques and gets very good results. But I find that near field for LF and far field for HF with a blend for the two measurements works just fine.
gedlee said:
Is there any incompatibility issue about our auditory system time of integration in LF and the integration time of pinking with an RTA when evaluating best positioning of subwoofers for instance?
There was some suggestion from Charles Hansen that this thread is its spin off and we would like your knowledge contribution on that.
- Status
- Not open for further replies.