Columns Retired Columns & Blogs |
September 12, 2008 - 7:05pm
#1
TAS Reviewer comments on JA's Harbeth measurements
Loudspeakers Amplification | Digital Sources Analog Sources Featured | Accessories Music |
Columns Retired Columns & Blogs |
Loudspeakers Amplification Digital Sources | Analog Sources Accessories Featured | Music Columns Retired Columns | Show Reports | Features Latest News Community | Resources Subscriptions |
Thanks for the heads-up, kana813. It always puzzles me that some readers prefer to ask a reviewer for another magazine specific questions about something I have written in Stereophile. My email address is hardly a secret. :-)
I note that REG writes "I am quite sympathetic to the problem of bass measurement in situations where one has no anechoic chamber available," but dismisses my mentions of the "2pi"bass boost that results from measuring in the nearfield as "casual." However, I wrote about it in detail in the article that starts at http://www.stereophile.com/features/103.
He does say in his discussion of my speaker measurements that, in respect to assessing a speaker's LF performance, "I deal with this all the time as best I can." This looks as if he does perform speaker measurements but doesn't publish them. And as he doesn't publish them, I have no idea how deasl with the the fact that a speaker that measures flat anechoically down to a low frequency will tend to sound increasingly bass heavy the smaller the room in which it is used.
I must also point out that Robert Greene is hardly a disinterested observer, as he both writes for a competing magazine and has been a very strong public champion for Harbeth speakers, especially the predecessor to the M40.1.
Thank you, here are some photos of mine:
http://forum.stereophile.com/photopost/showphoto.php/photo/166.
http://forum.stereophile.com/photopost/showphoto.php/photo/180
and http://forum.stereophile.com/photopost/showphoto.php/photo/76.
I will be publishing a follow-up on the Harbeth M40.1, addressing Alan Shaw's comments and examining how the M40.1 measured in both Art Dudley's and my listening rooms in the November issue of Stereophile.
John Atkinson
Editor, Stereophile
As real world as it gets.
Many of us can relate to this room.
I'm sorry, but that CANNOT be REG's room.
That picture gives me cooties.
The picture is from his website.
BTW, he sits on the stool and uses the books to get his ears at just the right level.
you guys are making me feel guilty:
28.5' x15.75' x 8' room. One half is the 'free space' for the audio system proper. Brick wall and fireplace on that end. 12 PZA's in the room ("Pressure Zone Absorbers"- but you'll hear more about that in the future.-tested by the NRC to exceed the known 'rules' of acoustics.) Separate power coming off the 200A panel just on the other side of the wall at the front of the room, into a 50-100Kva (depending on how it is run) AC balancing transformer. My own separate transformer on the pole outside. Everything as balanced AC.
As for the listening room of R(E.)G, if I had been drinking something when I saw the photo.. I think I might have choked on it.
I can't afford to be judgmental on that-I've been far poorer at times.
However, even when I was 16 - I knew how to set up a stereo.
And that ain't it.
As being someone who has the ability to assemble an anechoic chamber at nearly any time I desire, to utilize it in loudspeaker design, I gotta tell you it's pretty low on my list of design priorities.
Makes great ad copy, no doubt. Lab coats, clip boards and guys with cool haircuts looking really serious about their shit that they are inspecting very closely for the ad copy 'action' shot.
But, in the end, it's got squat to do with making a real world speaker. The quasi-anechoic and the close mike measurements are more than good enough to figure out what a speaker is doing, when in the hands of an intelligent person.
You may object to how his listening room looks, but look how the M40 measures in said listening room (2nd chart):
http://www.regonaudio.com/Harbeth%20Monitor%2040.html
Very nice. Looks be damned, how can any audiophile argue with results like that.
Mine's bigger than yours.
Oh yeah? I posted this from my new iPhone 3g, while sitting in said room.
I have some clean Berber and some "dawn" for that lightswitch cover if he will pay shipping..
Could this guy by any chance be a batchelor?
REG is married. His wife Paige is the webmaster for http://www.regonaudio.com.
He does have another listening room:
http://www.regonaudio.com/Living%20Room.jpg
No wonder he is a fan of room EQ. ;-)
Kal
It is indeed a good response. But with all due respect to REG, where is the info on how it was obtained? Is it somewhere else on his site? Yes, it was taken with the Liberty Audiosuite. But was it spatially averaged? (It is way too even to be a single-point measurement.) And if it was averaged, how many individual responses were taken? In what kind of grid? Were both speakers measured at the same time or one at a time? Is it the response of both speakers or of just one? What microphone was used? Was it calibrated? And who provided the anechoic response? If it was Harbeth's own, I think that should have been indicated.
I know, "picky, picky." But we try at Stereophile to give you as much supporting information as possible so that you can judge what we do. As I say in the first of my 1998 articles on measuring loudspeakers - www.stereophile.com/features/99 - measuring is a _subjective process - you have to decide what to measure and how to measure - and the result can be relatively meaningless without all the supporting information.
John Atkinson
Editor, Stereophile
Fully agreed.
At http://www.acourate.com/images/REGHarbeth.jpg you can see a picture of a one-point measurement of the Harbeths taken at REGs listening chair. Including the unsmoothed frequency response. Testsignal was a logsweep from 20 - 20050 Hz of duration 60 sec. This is a non-EQ'd result!
The inter-channel cross-correlation coefficient (acc. to IACC inter-aural coherence [Blauert/Braasch]) has an excellent value of 0.632 over 1.4 sec duration. This means that the speakers match well and the room does not introduce much uncorrelated noise.
My conclusion: I have seen much worse measurement results.
BTW the pictures showing his rooms also show both speakers and the listening chair. Think about arrangement and distance, about direct sound and reflections!
A nice environment does not necessarily mean good sound.
Uli Brueggemann
www.acourate.com
Welcome, Uli!
Great first post.
The interesting story to me is, how is it that you have REG's measurements laying around? I'm guessing that he uses your software and somehow came to share that with you, rather than be seen posting here.
Simple explanation: once I got a chance to visit Robert and to take some measurements.
Uli
Uli-
You made a trip from Germany to CA just to measure REG's
system?
Dan
The guys from TAS never pass up an opportunity to take pot shots at Stereophile and it's writers. Thankfully, JA and Company have chosen to take the high road which is a credit to their character. Even when MF felt the need to respond to a criticism of one of his reviews (I believe it was on the laser turntable) he did it in a very respectful letter to the editor in TAS. Just one of the many reasons that while I subscribe to both my preference is for Stereophile.
Competition is good for everybody. A bit of good natured ribbing and legitimate
discourse keeps everybody motivated to be better at what they do. TAS, however,
is acting like the Chihuahua chasing the bulldog. I think they probably spend a
little too much time in the industry (advertisers)/ magazine writers echo chamber.
TAS's target audience comes across to me to be the advertisers.
So, are we saying that if you measure this speaker sitting in that room, about two feet from the speaker, at tweeter level, you get flat broaband response?
I am dubious.
Maybe we can also measure the new Sunfire speakers as REG listens while holding them at arms length!
Very interesting. I assume the time window used for the FFT used to derive the frequency response from the calculated impulse response was the 399ms indicated in the graph? I use Fuzzmeasure for my in-room response which also uses a log chirp, BTW, because it runs on my Mac PowerBook.
Me too.
Thanks very much Uli.
I will certainly check out your website.
John Atkinson
Editor, Stereophile
No, the frequency chart shows 32769 bins as the pulse response typically used with Acourate is of length 65536 samples. This means an overall pulse length time of 1.486 seconds as the samplerate has been 44100 Hz. Greater pulse lengths are allowed too.
John, you are of course welcome for a test of Acourate. All I need is two pulse responses in pcm- or dbl-format (pulse length >=65536, about 6000 samples before main peak) and your favourite wav-track. Acourate can also process offline wav-files. So no additional hardware is required.
Uli
www.acourate.com
REG is still at, here's his latest post:
"Just as a point of clarification, people who are interested in this
subject of bass in rooms(and I suppose we all are) might like to
read Atkinson's explanation of why he measures speakers as he does.
The basic drift as of his 1999 article at least, if I may paraphrase
as impartially as I can, is that, even though he knows that the
close miked woofer method is going to produce a considerably larger
amount of measured bass than the anechoic-measured amount, he
feels that this larger measured amount is representative of what one
hears in actual roms of domestic size.
There is of course "room gain" in the extreme bottom, as I am sure
you all know. (We have often talked about the advantage of slow roll
out of the extreme bottom and you can find a discussion of room gain
and subwoofers in my discussion of the Audio Physic Minos in TAS.)
But it is odd to my mind to see the big bulges
up,usually around 60-70 Hz, in his close miked measurements and then
look at what are often extremely smooth and flat in room RTA bass
measurements--which, unless the DSP people are all way off bass,
will actually give a correct amount of bass(for a flat -in- the -
bass target curve). Room gain in my sense of the word would bring
up the RTA at the listening position, typically, not just the
measrement close to the speaker. And in any case, it is the bass
quantity at the listening position that would seem to me to matter
most. In short, the speakers that would measure flat for
his "anechoic" (close miked woofer curves) would apparently be down
in level in the bass in his room in the sense of the RTA at his
listening position. In more detail:
Since dB are additive in the obvious sense, to get his close miked
measurement flat one would typically need to have depressed anechoic
response at those frequencies--a dip to counteract the bulge that
arises in the measurement when the speaker is anechoic flat. And in
turn , this would lead to a dip in the in-room response (RTA) in his
room--since typically that is quite flat in his room when the
speaker is anechoic flat.
As far as I am aware, few people feel that speakers with that
behavior sound anything but quite bass shy.(Room gain is a much
bigger issue in the true bottom octave, in my experience, unless one
has a very very small room.)
Anyway, just for your information
here is a link to his article(the bass discussion begins on page 6)
www.stereophile.com/features/103
REG"
M40.1 measurement done by an REG forum member:
http://forum.stereophile.com/photopost/showphoto.php/photo/1776/size/big/password//sort/1/cat/500