That is only part of it… but the very act of using a mic (any mic) to capture the sound of a speaker to listen to is VERY limiting. Huge amounts of information are lost. All the stuff that tells the difference between a good speaker & a great speaker. And almost for certain the tiny bits of information that EnABL brings out -- espcially since much of that is related to imaging.
As to measuring -- your UMM only covers the broad strokes. You give it far too much credit. When was the last time you had it calibrated?
dave
It is less than 2 yrs old (18 mo old actually), same cal it came with. In industry and research labs we require cal once a year so by those standards I am overdue. But how many DIY people send their mics, DMM's, dial calipers, etc in for yearly cal? How old is your EMM cal? What I go by is the consistency measurement of the same driver response in the same cabinet at same amp setting and distance. The TC9FD in a Nautaloss at 0.5m and 1.4v. It comes out the same then still good.
You say I give it too much credit and I say you don't give it enough credit and discount its broad strokes way too much when it measures a crazy 10dB plateau from 600Hz to 6kHz on the CHN70, which was later confirmed by several independent measurements. The fine brush strokes of calibration accuracy you refer to is probably more along lines of 1 to 2 dB out above 10kHz. All the unsavory responses measured below 10kHz in the 5dB to 10dB variations, are unfortunately, real.
You say I give it too much credit and I say you don't give it enough credit and discount its broad strokes way too much when it measures a crazy 10dB plateau from 600Hz to 6kHz on the CHN70, which was later confirmed by several independent measurements. The fine brush strokes of calibration accuracy you refer to is probably more along lines of 1 to 2 dB out above 10kHz. All the unsavory responses measured below 10kHz in the 5dB to 10dB variations, are unfortunately, real.
Last edited:
My cals are at least that old, on all 4 mics… but i use them for what i use them for (mostly XO development) and i do not post graphs or make any claims about their accuracy, i certainly don't trust them from 8 k up. Mostly they sit in their cases in my office.
dave
dave
another suggestion (maybe for round 4 or 5 ?)
Why not making a single audio file with all the excerpts together ? Smooth one to begin and Loudest at the end so it doesnt ''strain'' the ears of the testees until the end
Why not making a single audio file with all the excerpts together ? Smooth one to begin and Loudest at the end so it doesnt ''strain'' the ears of the testees until the end
I have only a minor critique of X's measurements:
Zoom are a cheapo brand ( back from my teens when I borrowed a zoom pedalboard and gave it back quickly)
Also (take this with a pinch)...measurements taken at angles aren't flat. (citing something I Weems about measures taken at 90* to eliminate reflection from behind the mic, and the necessary work to regain a true reading)
That being said, the necessary EQ may be performed autonomously inside the Zoom.
FWIW kudos for the huge efforts in documenting your findings and sharing. I doubt the naysayers would have any grounds were you using a cal'd mic and one of the well known softwares instead of the zoom.
(I can 'borrow' a B & K pro noise mic....but I haven't been bothered to make the 200V phantom supply it uses 🙂)
Zoom are a cheapo brand ( back from my teens when I borrowed a zoom pedalboard and gave it back quickly)
Also (take this with a pinch)...measurements taken at angles aren't flat. (citing something I Weems about measures taken at 90* to eliminate reflection from behind the mic, and the necessary work to regain a true reading)
That being said, the necessary EQ may be performed autonomously inside the Zoom.
FWIW kudos for the huge efforts in documenting your findings and sharing. I doubt the naysayers would have any grounds were you using a cal'd mic and one of the well known softwares instead of the zoom.
(I can 'borrow' a B & K pro noise mic....but I haven't been bothered to make the 200V phantom supply it uses 🙂)
Zoom is for sound clips and measurement is with calibrated UMM 6 mic. Do we normally calibrate our music recording mics?
It could be argued for the test you are doing that yes we would.
The frequency response of the speaker is now also convolved with the frequency response of the mic, not something that can be ignored if not flat and not compensated for (as we do using the cal file for our measure mic)
dave
A FR for the H4n (couldn't quickly find H4) from Zoom H4n Review
Looks like 6 & 8 dB peaks at 4 & 7.5 kHz.
dave
Looks like 6 & 8 dB peaks at 4 & 7.5 kHz.
dave
Planet10,
Have X claimed anywhere any single driver measurement to be 100% correct, he just share his work at a diy sector and it is stuff as drivers plot performance under his conditions including build plans.
If a guy with a ruler that is out of spec measure a object to be 68 meter long it is no problem as long as we know it is off then we measure it up to a reference ruler and math can do the rest to get the exactly correct length of the object that before was 68 meter.
Comparison thread is scientific diy ok because it was under same condition therefor the difference is real. Because we talk UMM-6 forget the sound clips talk is about measured plots from 1 x objective and 2 x subjective threads. Does it then matter 1-2dB could be off somewhere for UMM-6 i don't think because it same for all. Back to ruler example from above REW or lot of Jeff Bagby spreadsheets can manipulate measurements and calibration files, so should you need later to have the plots not to be accurate only the comparison drivers data between then its possible have the old measurements calibrated to a pro reference limited by the build in bandwith for a UMM-6.
Have X claimed anywhere any single driver measurement to be 100% correct, he just share his work at a diy sector and it is stuff as drivers plot performance under his conditions including build plans.
If a guy with a ruler that is out of spec measure a object to be 68 meter long it is no problem as long as we know it is off then we measure it up to a reference ruler and math can do the rest to get the exactly correct length of the object that before was 68 meter.
Comparison thread is scientific diy ok because it was under same condition therefor the difference is real. Because we talk UMM-6 forget the sound clips talk is about measured plots from 1 x objective and 2 x subjective threads. Does it then matter 1-2dB could be off somewhere for UMM-6 i don't think because it same for all. Back to ruler example from above REW or lot of Jeff Bagby spreadsheets can manipulate measurements and calibration files, so should you need later to have the plots not to be accurate only the comparison drivers data between then its possible have the old measurements calibrated to a pro reference limited by the build in bandwith for a UMM-6.
Last edited:
No he hasn't but he has been using his tests to make authorative statements about the speakers he is testing.
I am just pointing out where, in the subjective test in particular, where potential errors that could skew the results are. Part of any scientific test is taking care to point out were errors could/can occur. The differences people hear in this test include a whole lot of things that are not the speaker (lets ignore information lost in the process), but the conclusions are being made about the speaker.
For instance, XRK shows an peak for the FF105 at 8k, if his H4 is similar to the H4n, then that will be close to doubled by the mic used.
That is not the mic we are talking about in the subjective thread. It was a mistake on my part that i assummed it was used.
dave
I am just pointing out where, in the subjective test in particular, where potential errors that could skew the results are. Part of any scientific test is taking care to point out were errors could/can occur. The differences people hear in this test include a whole lot of things that are not the speaker (lets ignore information lost in the process), but the conclusions are being made about the speaker.
For instance, XRK shows an peak for the FF105 at 8k, if his H4 is similar to the H4n, then that will be close to doubled by the mic used.
That is not the mic we are talking about in the subjective thread. It was a mistake on my part that i assummed it was used.
dave
same as a driver, a mic is a transducer as well (but on reverse; it tranforms acoustic energy into electrical energy)
so, yes, the frequency response of a mic is not flat. Also bandwith and amplitude is limited (mics for high SPL, drums, voice, etc..) Same as all drivers
compensation file can be used, with more or less reliability, or you can use calibrated mic which at least gives you a reliable base.
differences with individual hearing, headphones and their interaction with the ear canal add more variables. A Zoom X-Y mic should capture some flavor. I don't understand how opamps are compared and differentiated from one another online when there's more opamps/caps in the path to the listener.
That seems logic when members from all over the globe in a poll rank drivers based what they hear from the sound clips, then he look at his measurements same drivers and remember he is lot scientific and objective minded (Hope he not kill me saying that) and make conclusions.
You the last one althouh in subjective way to talk about making authorative statements about the speakers you tested and what you expect is good and bad about upcoming drivers or tech.
Think experienced member Bob Brines is or at least was skeptic to what to extract of facts from a sound clips test, hat of he joined at both rounds and made his points and voted. Hat of if you join round 3 and make a vote, make your points or not, and if not no problem think Jay will be there 🙂.
The H4 has selectable mic cal profiles to emulate classic studio mics. So it depends on which profile was used. I use the Neumann U87 (a well known classic $3k professional mic used for a lot of recording work).
Factory graph -- 5 dB peak at 8k or so:
An externally hosted image should be here but it was not working when we last tested it.
and a user measure ofhis actual mic (the LF stuff is likely the room):
dave
UMM66 for the data taken in room? Then please excuse me, I had assumed Zoom was used for all tasks.
MIC emulation...thats a sticky point. Not great to record a recording with a mic then software emulating a common (coloured) mic used for vocals (?)
A little like using an SM58 to record vocals....
Personally id use no emulation and cal to flat. Then at least what is recorded is as close as possible to what you are hearing in room, during the tests you make.
Point being no studio mic is flat, nor is it intended to be in many cases.
But yours probably should be if your not going to introduce misleading frequency response changes.
MIC emulation...thats a sticky point. Not great to record a recording with a mic then software emulating a common (coloured) mic used for vocals (?)
A little like using an SM58 to record vocals....
Personally id use no emulation and cal to flat. Then at least what is recorded is as close as possible to what you are hearing in room, during the tests you make.
Point being no studio mic is flat, nor is it intended to be in many cases.
But yours probably should be if your not going to introduce misleading frequency response changes.
Last edited:
You could always revert back to listening to the calibrated mic recorded impulse in a virtual audition... 😉
Dave, any thoughts on going back to round 2, listening to the clips, and providing your feedback? We know your views on the methodology. But I know you have a lot of experience with the Alpair 7.3 in particular. You could listen to just the A7.3 clips and give your impressions about the recorded sound versus your experience with those drivers in real life. Or anyone for that matter, that has extensive first hand experience with any of the tested drivers. Something like "I've used XYZ extensively in my own designs, and the recordings do capture these particular characteristics [...]; but the recordings fail to capture [...]"
In the end, I don't think there's any other way to do a "virtual get together". I mean it would be great if this could be done with top tier equipment, in a pro studio, by experienced recording engineers... But until someone is willing to bankroll that, we'll have to settle for what we can get.
I would assume most people understand the limitations of this test, and don't take it as the end all, be all of driver research. It's first and foremost a way to have fun, and at least learn a little about the drivers.
Xrk, clearly the next evolution of this test is to have us all over to your house. 😉
In the end, I don't think there's any other way to do a "virtual get together". I mean it would be great if this could be done with top tier equipment, in a pro studio, by experienced recording engineers... But until someone is willing to bankroll that, we'll have to settle for what we can get.
I would assume most people understand the limitations of this test, and don't take it as the end all, be all of driver research. It's first and foremost a way to have fun, and at least learn a little about the drivers.
Xrk, clearly the next evolution of this test is to have us all over to your house. 😉
- Status
- Not open for further replies.