John Atkinson, Alan Shaw debate from 2008

[kimmosto]

^Justification:

Fig.1 electrical impedance (solid) and phase (dashed).

Fig.2 cumulative spectral-decay plot calculated from output of an accelerometer fastened to name plate on rear panel (MLS driving voltage to speaker, 7.55V; measurement bandwidth, 2kHz)

Fig.3 anechoic response on tweeter axis at 50" without grille (black), averaged across 30° horizontal window and corrected for microphone response, with nearfield responses of midrange unit (green), woofer (blue), and port (red), plotted in the ratios of the square roots of their radiating areas below 800Hz, 1kHz, and 900Hz, respectively, with the complex sum of the individual responses (black) also plotted below 300Hz.

Fig.4 lateral response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90–5° off axis, reference response, differences in response 5–90° off axis.

Fig.5 vertical response family at 50", normalized to response on tweeter axis, from back to front: differences in response 45–5° above axis, reference response, differences in response 5–45° below axis.

Fig.6 step response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Fig.7 cumulative spectral-decay plot on tweeter axis at 50" (0.15ms risetime).

Fig 1. This is okay.

Fig.2 What is scale in Pa or SPL at listening window and significance i.e. effect to sound pressure? Why measured from name plate or any other subjectively selected point if final effect to sound is not known?

Fig.3 This is time windowed, plotted from 300 Hz, and averaged within listening window. Listening window average is okay, but not alone i.e. no actual far field responses within -30...+30 deg in horizontal plane is visible (other than tweeter's axis in CSD).

Fig.4 This is normalized hiding shape of actual far field responses on...off-axis. Normalization should be off to estimate changes in absolute sound balance to different off-axis angles. This applies also to design phase.

Fig.5 Also this is normalized hiding shape of actual far field responses on...off-axis.

Fig.6 This is okay. Should be towards designed listening axis if not tweeter's axis.

Fig.7 This is also okay.

Possible in-room response. Quite okay, but not applicable to listening room of the readers.

Missing:

1. Actual far field responses at 2 m in anechoic. The truth. Using scaled near field has caused numerous misunderstandings because readers are not able to estimate baffle loss. In addition, many speakers don't have box so estimation of baffle loss is impossible even with simple diffraction simulators. Simple complex summing is okay for vents in front panel, but not so much anymore for rear/side vents. Square slot close to floor tune with the floor.

2. Off-axis far field responses at 2 m in anechoic to 0-180 deg or -170...+180 deg in hor & ver planes. This would allow at least power and DI responses, but also other CTA-2034-A.

3. Non-linearities.

4. Compression e.g. with typical music spectrum.

I really don't care about consistency as long as information is limited and not absolute which is impossible to scale to reality. I'm also aware that decent measurement environment is expensive. This might be one of the reasons why quality of loudspeakers is still low, though modern measurement and design tools and methods would allow progress compared to 70s'. Maybe also speaker designers live in the past with their consistencies.

 

[Boden]

@ head unit: MLSSA was available as of 1987/88 for approx $ 10.000,--. In 1994 the legendary IMP, later IMP/M schematic was published by our great Bill Waslo (Bwaslo here); his IMP/M software was about 50$. A Panasonic WM61 was all one needed. Before that, around 1990, a gentleman whose name escaped me also published a diy MLS measurement system in Speaker Builder magazine.
We may discuss various measuring methodological artifacts/mistakes, which point Kimmosto correctly raises, but hardware wise there is no excuse for not being able to measure since the late 80ties.
I.m.o. Atkinson has never done proper scaling of his NF bass measurements, invariably leading to 5dB bass humps around 80-100 Hz, which should not be there.







The sh

 

[Remlab]

That's a very strong statement, would you care to justify it? Because as a loudspeaker engineer I could perhaps agree with simple, but I vehemently disagree with unprofessional and pathetic, I think that is not only unkind but incorrect, particularly considering the context of the audience for that publication.

As for "No better than almost any of us could do at home" ok that might be true TODAY but it certainly was not true in the past which Atkinson keeps consistent to. I sang praises and hallelujah when someone (MLSSA) came out with a simpler system to measure magnitude and phase data of frequency and impedance and then double hallelujah out came software (LEAP, RIP Chris Strahm) could do a voltage divider crunching simulation that actually matched what I could measure on $40k of overpriced B&K boxes. How MUCH testing, a print publication certainly can't devote as much space or time as hobbyist websites. I'm just really glad Stereophile hasn't gone the way of Sound & Vision which measures pretty much nothing any more and thus has become pretty useless.

I'd also point out thank god Stereophilemeasures stuff at all, being I think the only remaining publication in the western hemisphere that does. (Websites *may* come up and replace that however that's a more recent phenomenon, and I believe in significant part because Stereophile kept the measurement flame burning).

The real value of Atkinson's measurements is that he he has Quasi-anechoically measured more speakers than anyone on the planet(I'm sure it must be well over 1000 by now, with models ranging in price from $200 to $250,000) and we have access to almost all of them. There is no better education for diy enthusiasts than that. In fact, his loudspeaker measurements are what got me into this hobby in the first place, and I'm sure that many feel the exact same way, whether they want to admit it publicly or not.
Remember this one? There are a gazillion of them at our disposal.
Jamo Reference R 907 loudspeaker Measurements | Stereophile.com

 

[kimmosto]

Limited measurement set is unfortunate especially with dipoles with electro-dynamic transducers. Directivity as normalized by axial, to half space only and near field for woofers and mids is really poor, hiding essential features of sound balance.
Big anechoic lab at university of Helsinki Finland has been available since 1969 also for local magazine(s) and manufacturers. One of my first diy designs was measured there about 30 years ago so hopefully some understatement about Stereophile's output might be understandable. But I trust that diy designers are more aware nowadays and able to outperform Stereophile at home both philosophically and technically.

 

[raymondj]

Good point about Stereophile being used as a reference, probably by us all.

Maybe the Low frequency discrepancy could be investigated to its root cause brownstone, high rise apartment or wooden house builds and their associated resonances or dimensions ?

We have done Kii3 and Harbeth, others have discussed DeVore over the years.
A jamo R 907 clone always piqued my interest, but is it too big for my room.

So what about diy aluminium and Magico clones. Why aren't they a few hundred quid? Modern drivers in an all metal box.
Difficult to get past the missus here in UK ( No wood finish) and if she ever found out how much you had spent on them, well! 🙂

Shipping of finished goods across the Atlantic seems to work badly both ways. Go figure that one.

Shipping costs and additional paperwork probably even worse for us here in Blighty in the coming months, as all of our drivers imported ones like SEAS, ScanSpeak, Audax, SB And Satori, wavecor, Founteks, Tangbands etc are in all likely to become even more expensive at least in the short term if not forever.

Question can you legally smelt bauxite in the back garden? Will the neighbours notice the fumes. MDF is causing me to cough quite a bit these days.

Only Joking and Merry Christmas to all, stay safe and warm.