I recently bought a digital sound level meter and decided to measure the response of my (5,000 euro) Zingali 3 speakers and my Sennheiser 600 headphones. The non-flat responses astonished me.This is some of the measurements I got:
Zingali Sennh
hz db
40 51 unavailable (below 40db anyway)
50 62 "
80 62 48
125 65 54
250 62 60
500 65 63
1k 71 67
2k 70 69
4k 72 72
6.3k 74 76
10k 72 66
12.5k 66 62
16k 60 53
20k 58 52
These are about +- 10db for the speakers and even more for the headphones(which should be flatter by accounts)
I used a well known test disc on a Teac transport thru a Pass bosoz pre and diy gainclone both of which replaced production Naim and other amplification in my system.
Are the manufacturers measurements b********* or could my digital SPL meter be that far off . I did use c weighting by the way.
What gives?
Zingali Sennh
hz db
40 51 unavailable (below 40db anyway)
50 62 "
80 62 48
125 65 54
250 62 60
500 65 63
1k 71 67
2k 70 69
4k 72 72
6.3k 74 76
10k 72 66
12.5k 66 62
16k 60 53
20k 58 52
These are about +- 10db for the speakers and even more for the headphones(which should be flatter by accounts)
I used a well known test disc on a Teac transport thru a Pass bosoz pre and diy gainclone both of which replaced production Naim and other amplification in my system.
Are the manufacturers measurements b********* or could my digital SPL meter be that far off . I did use c weighting by the way.
What gives?
Yes I have...
Hello, protos
There may be several reasons why your results look odd to you. First of all, you have used weighting in your measurements. In practice this alone will guarantee many dBs of difference from purely flat response. Apparently C-filter is a bandpass filter with -3dB frequencies at 31.5 Hz and 8 kHz. Try not to use weighting, if possible.
Did you measure your speaker's response indoors? If so, you are guaranteed to get some room modes in your measurement results. Manufacturers response curves are typically anechoic responses. You could try measuring outdoors by putting the speaker and dB meter on solid ground. Just find a place where there are no disturbing surfaces / objects nearby.
I've never tried it but I've read somewhere that measuring headphones is very difficult. The phones don't really work properly unless you put them against your head. Professionals use e.g. artificial heads to make the measurements, and even then the results are sensitive to the exact positioning of the phones. Good luck!
To answer your question: yes I have measured my speakers and the results match very well with the curves published by the original designer. It helps to own a MLS based analyzer designed for this purpose...
Pekka
Hello, protos
There may be several reasons why your results look odd to you. First of all, you have used weighting in your measurements. In practice this alone will guarantee many dBs of difference from purely flat response. Apparently C-filter is a bandpass filter with -3dB frequencies at 31.5 Hz and 8 kHz. Try not to use weighting, if possible.
Did you measure your speaker's response indoors? If so, you are guaranteed to get some room modes in your measurement results. Manufacturers response curves are typically anechoic responses. You could try measuring outdoors by putting the speaker and dB meter on solid ground. Just find a place where there are no disturbing surfaces / objects nearby.
I've never tried it but I've read somewhere that measuring headphones is very difficult. The phones don't really work properly unless you put them against your head. Professionals use e.g. artificial heads to make the measurements, and even then the results are sensitive to the exact positioning of the phones. Good luck!
To answer your question: yes I have measured my speakers and the results match very well with the curves published by the original designer. It helps to own a MLS based analyzer designed for this purpose...
Pekka
Yes, room effects are HUGE. Remember that the next time you're agonizing about which type of copper to use for a 3 cm wire run. Room effects are particularly noticeably in the discrete standing wave range.
Diffraction can cause sharp spikes/nulls. Position averaging helps.
Diffraction can cause sharp spikes/nulls. Position averaging helps.
Well I understand that in room measurements will not be as accurate particularly in the bass region (which should be mainly reinforced as I understand it and not less) and I can understand that the bass might measure a bit differently in headphones that are semi-closed on the head but I still can't undestand the big difference in the higher frequencies.The room (or the head) shouldn't account for that margin of difference .
As to C weighting my booklet advises using this to measure freq.
resp. in preference to A weighting which is used for workplaces etc. I don't have a non-weighting mode on my meter.
As to C weighting my booklet advises using this to measure freq.
resp. in preference to A weighting which is used for workplaces etc. I don't have a non-weighting mode on my meter.
THAT'S for sure!! Whenever I want to demo my subs to unsuspecting visitorsSY said:
I ask then to stand next to the front door. 
Then you should get enough info on the weighting curves in order to take these into account (i.e. calculate the "real" response of your device under test after the measurement).
I too think that headphones have to be measured on an artificial head.
Measuring your speakers: Apart from the room modes, the speakers themselves may have a slightly peaky response in the upper frequency range as it is the case with almost every horn speaker. I suggest that you don't measure each point with one single frequency but with several frequencies spread over 1/3 octave. After that you take the average.
Regards
Charles
SY said:
<grin> Nice to see I'm not the only cynical one here 🙂
Circlotron said:
THAT'S for sure!! Whenever I want to demo my subs to unsuspecting visitors
A friend brought his sub over once to try out and said the best place to put it was the point in the room where the bass sounded strongest, only problem was that that point was upstairs in the toilet 🙄
Tony.
Room effects can be this large. Check out my measurements of my proac clones:
http://www.diyaudio.com/forums/showthread.php?threadid=18020&highlight=active+proac
Then check out the review I linked there. Regardless of how you feel about my measurement technique, you'll see that Stereophile noted the same effects I did despite anechoic flat measurement. These include non-subtle effects in the 5-20kHz region.
http://www.diyaudio.com/forums/showthread.php?threadid=18020&highlight=active+proac
Then check out the review I linked there. Regardless of how you feel about my measurement technique, you'll see that Stereophile noted the same effects I did despite anechoic flat measurement. These include non-subtle effects in the 5-20kHz region.
C-weighting
Protos,
see e.g. http://www.norsonic.com/web_pages/correlation.html. It'll give you an idea of the effects of C-weighting.
If you measured with point frequencies you can get room effects even at higher frequencies, not only in bass. Constructive or desctructive interference may happen between the sound coming from the speaker and the echoes from floor, ceiling and walls. This means that you may get either emphasis or attenuation, depending on the wavelength (i.e. frequency), room dimensions and speaker placement.
If your test CD contains narrow band noise or "wobbler" tones the room effects can be masked quite well - except in low frequencies (I've read).
Regards,
Pekka
Protos,
see e.g. http://www.norsonic.com/web_pages/correlation.html. It'll give you an idea of the effects of C-weighting.
If you measured with point frequencies you can get room effects even at higher frequencies, not only in bass. Constructive or desctructive interference may happen between the sound coming from the speaker and the echoes from floor, ceiling and walls. This means that you may get either emphasis or attenuation, depending on the wavelength (i.e. frequency), room dimensions and speaker placement.
If your test CD contains narrow band noise or "wobbler" tones the room effects can be masked quite well - except in low frequencies (I've read).
Regards,
Pekka
Found the C- weighting measurements here
http://www.safetyline.wa.gov.au/institute/level2/course18/lecture54/l54_09.asp
This would suggest that the speakers and the phones are peakier at around 6khz than supposed although the response does flatten a lot over 8khz and a touch under 63hz.
It looks like weighting is about -3db per octave from 63hz and lower andthen from 4khz and higher.
http://www.safetyline.wa.gov.au/institute/level2/course18/lecture54/l54_09.asp
This would suggest that the speakers and the phones are peakier at around 6khz than supposed although the response does flatten a lot over 8khz and a touch under 63hz.
It looks like weighting is about -3db per octave from 63hz and lower andthen from 4khz and higher.
This is the conversion factors to transform dB(A) to linear dB in octave bands. So if you measure 20 dB(A) at 125 HZ, you get 36.1 dB(lin).
16 56,7
31 39,4
62 26,2
125 16,1
250 8,6
500 3,2
1000 0
2000 -1,2
4000 -1
8000 -1,1
F
16 56,7
31 39,4
62 26,2
125 16,1
250 8,6
500 3,2
1000 0
2000 -1,2
4000 -1
8000 -1,1
F
Protos
You didn't say what meter you are using, but assuming that it is the Radio Shack one, then expect some fairly serious inaccuracies from the meter. I fooled around for ages with mine trying to figure out why my speakers had such a peak at 6.3 kHz and then rolling off steadily from there. Eventually, having swapped amps, CD players and speakers I realised that it was the meter.
I have since modified my meter to give linear readings, plus put the mic capsule at the end of a tube (all as described by Eric Wallin at http://www.gti.net/wallin/audio/audio.html ) This has helped, but there still seems to be a peak at 6.3kHz and not much response past about 12 kHz.
To help reduce room response try measuring nearfield.
Mick
You didn't say what meter you are using, but assuming that it is the Radio Shack one, then expect some fairly serious inaccuracies from the meter. I fooled around for ages with mine trying to figure out why my speakers had such a peak at 6.3 kHz and then rolling off steadily from there. Eventually, having swapped amps, CD players and speakers I realised that it was the meter.
I have since modified my meter to give linear readings, plus put the mic capsule at the end of a tube (all as described by Eric Wallin at http://www.gti.net/wallin/audio/audio.html ) This has helped, but there still seems to be a peak at 6.3kHz and not much response past about 12 kHz.
To help reduce room response try measuring nearfield.
Mick
Kanga...have you looked into the Rives Test CD2 (for use with the rat shack meter)
there's a quote from their website. I bought one of the disks a couple months ago...seemed easier than using the correction tables that are floating around the internet.
take care>>>>>
there's a quote from their website. I bought one of the disks a couple months ago...seemed easier than using the correction tables that are floating around the internet.
take care>>>>>
Interesting RS SPL expereince
I had an interesting experience with a RS SPL.
I used one of the old analogue ones to ruff in some White Instruments EQ's before my friend could come over with his B&K gear to do a proper job.
When I told what him what I had done he smiled and opened up his equipment case and showed me something...
His Bruel and Kjaer microphone calibrator fit the microphone capsule on my $40 meter perfectly. This cheap little gadget turned out to be within a few dB of the laboratory gear.
Of course there where lots of operational differences, (The B&K had 1/3 octave input filters!) but for a ruff analysis, my little toy meter did OK.
My EQ settings were very close to what John finally came up with. In the end I don’t think the difference was his fancy equipment, I’m quite sure it was his experience and feel for what the room should sound like that made the biggest difference.
One of the things you get of course when you plunk down megabucks on laboratory gear is repeatable results that trace back to calibrated standard references. A week or so latter I watched that little cheap meter I had exploded into pieces when I dropped it on the floor.
I had an interesting experience with a RS SPL.
I used one of the old analogue ones to ruff in some White Instruments EQ's before my friend could come over with his B&K gear to do a proper job.
When I told what him what I had done he smiled and opened up his equipment case and showed me something...
His Bruel and Kjaer microphone calibrator fit the microphone capsule on my $40 meter perfectly. This cheap little gadget turned out to be within a few dB of the laboratory gear.
Of course there where lots of operational differences, (The B&K had 1/3 octave input filters!) but for a ruff analysis, my little toy meter did OK.
My EQ settings were very close to what John finally came up with. In the end I don’t think the difference was his fancy equipment, I’m quite sure it was his experience and feel for what the room should sound like that made the biggest difference.
One of the things you get of course when you plunk down megabucks on laboratory gear is repeatable results that trace back to calibrated standard references. A week or so latter I watched that little cheap meter I had exploded into pieces when I dropped it on the floor.
- Status
- Not open for further replies.