AES Journal Forum

The author raises a wide range of problems in the field of loudspeaker usage, but is quite skimpy in proposing any real solutions.

He points to a „widely accepted” steadystate amplitude response measurement that is being used just as a „matter of faith”, and which „now has penetrated consumer audio” simply as „an enticing marketing story”. He mentions the use of time windowed measurements as a next stage of development of steadystate amplitude response measurement. At the same time, he points out that „the simple measurements therefore cannot be definitive” and „there are reasons to exercise great caution in the application of equalization based on conventional in-room measurements.”

The author only touches on some solutions very briefly and has not shown the source or basis for such conclusions:

„The on-axis curve by itself is insufficient data. Full 360 data, appropriately processed, is important information.”

„...spatially-averaged measurement to reveal the underlying curve.”

Let’s look at the article, reading more closely.

aplaudio.com/downloads/Reading_Dr_Toole.pdf

“A response to Raimonds Skuruls’ comments on my recent paper

 I thank Mr. Skuruls for his response to my paper.  Dialog is needed on this topic, and that, in part, was a motivation to write it.  I am accused of “being skimpy in proposing any real solutions”.  I think a cautious approach is necessary in a peer-reviewed paper that is proposing changes to long-established internationally-standardized measurements. In fact, several dimensions to the needed solutions are, I think, unambiguously identified in the paper, and others that need further research are noted. Even in 30 pages (!) something has to be left out.  Most of the detailed explanations are in the references, especially in my 550-page book, and the numerous references therein. There is a lot of relevant data relating technical measurements to perceptions.  We are close to being able to assert in unambiguous terms what needs to be done, but my “research scientist” caution inclined me to postpone that until some ongoing investigations are complete.

Mr. Skuruls is in the business of selling his devices and services to customers in the audio industry.  As such he can choose to ignore the widely-used recommendations from the ITU, ISO and SMPTE that employ calibration methods based on steady-state in-room measurements and minimum-phase equalization. For decades past, and right now, all of these recommendations are being employed in calibrating movie sound dubbing stages and cinemas worldwide, and in setting up listening venues and facilities for broadcasting as well as university research. The widespread assumption is that they are all that is needed to ensure both good and consistent reproduced sound quality.

 I am currently engaged in committees looking to update some of these recommended practices, employing improved measurement methods and applying objectives guided by psychoacoustic relationships learned from disciplined, double-blind, listening tests. All of these documents exist because of a belief that the quality of reproduced sound in recording and broadcast studios, dubbing stages, cinemas, and homes should be fundamentally good and similar.  Mr. Skuruls appears to disagree, saying: “But it is a well-known fact that the “good” work of a recording engineer (producer) sounds good as a piece of art through any speaker that is used. Of course it is not the same “sound”, but it is the same work of art.”   So, he asserts that the “circle of confusion” in Figure 1 of my paper is irrelevant − that it is sufficient to recognize the “melody”, the “rhythm” and “lyrics” of a song, and that the bandwidth, spectrum, linear- and non-linear distortions and sound levels do not alter the “art”. I beg to differ.

 At several points in his comments it is asserted that my focus is on consumer products, that “professional” loudspeakers are inherently superior. I admire his faith, but numerous measurements in Section 2.4 and Chapter 18 of my book show that professional loudspeakers are as susceptible to design inadequacies, as are consumer loudspeakers. The best consumer loudspeakers and the best professional loudspeakers, as they measure and sound, are almost indistinguishable. The flawed ones exhibit an infinite variety of misbehavior, not all of which are capable of “correction” after the fact. Identifying and correcting the flaws is made greatly more difficult if the only data come from measurements in reflective spaces.

Mr. Skuruls claims that recording/mixing engineers are unbiased – “as close as it comes to being objective in the absence of a blind test”. The listening ability of recording/mixing engineers was elaborately tested early in my career and published as “Subjective Measurements of Loudspeaker Sound Quality and Listener Performance”, J. Audio Eng. Soc., vol. 33, 1985 (30 years ago!). In those tests, professional recording engineers and producers were mixed in with audiophiles as subjects in double-blind subjective evaluations of loudspeakers intended for use as broadcast/recording monitors. As a result of noting that several of the professionals were unable to repeat their subjective ratings in subsequent randomized presentations, a problem was revealed that is now widely acknowledged, but rarely discussed. Hearing loss is an occupational hazard in the audio business, and, as a result, the opinions offered by people so afflicted are less reliable, and may exhibit more bias than those from people with more normal hearing. The topic is also discussed in sections 17.4 and 19.1.2 in my book.  Professional and “amateur” listeners with relatively normal hearing exhibited closely similar preferences in sound quality. Some of the highest scoring loudspeakers in those tests were consumer products. A couple of the recording engineers commented at the end of the tests that they had never heard such good sound before. They had rejected some of their previously favored monitors, even after some repeat tests using their own master tapes. Needless to say, the highly rated loudspeakers exhibited the least-flawed anechoic measurements, which was discussed in the paper, and others that followed.

 Section 17.5 in my book shows how both professional and amateur listeners can be biased by visual information about the products they are listening to – we are all human; blind testing is important.  Subjective opinions are the basis for evaluating any sound reproducing system, and controlling the variables, including who is doing the listening, is essential.  See Figure 17.6 in my book, and Olive, S. (2003). “Difference in Performance and Preference of Trained versus Untrained Listeners in Loudspeaker Tests: A Case Study”, J. Audio Eng. Soc., 51, pp. 806-825.  It shows that experience is a significant factor in terms of the consistency and range of sound quality ratings, but that, in the end, the relative ratings of the products are essentially identical for all listeners – that is unless one’s “microphones” (ears) are damaged.

Mr. Skuruls is focused on performing useful in-situ measurements – not a bad thing – but this inclines him to downplay the usefulness of comprehensive anechoic data on loudspeakers. That is fine, but the reality is that anechoic data are required by the designers of the loudspeakers (if not, why not?) and as such should be available. These data can be elaborated to the level of the “spinorama” shown in Figure 20 in the paper. The frequency resolution is not limited by the measurement venue (time windowing), and is sufficient to reveal audible resonances over the entire audible frequency range: see Section 19.2.1 in my book and Toole, F.E. and Olive, S.E. (1988). “The modification of timbre by resonances: perception and measurement”, J. Audio Eng. Soc., 36, pp. 122-142.  Such high resolution is “difficult” in non-minimum-phase rooms, so such problems may go unnoticed.  The finding that the spectral “bump” is a more reliable indicator of audibility than time-domain ringing is of importance.

The calculation of a sound power estimate is a natural byproduct of doing full anechoic orbits of frequency-response measurements, and is more accurate than a reverberation chamber estimate for devices that are strongly directional, like loudspeakers (at the NRCC I had access to both kinds of chambers).  The frequency resolution of the sound power calculation is that of the basic measurements – in this case 1/20-octave – which helps in identifying the presence of resonances – see Figure 21 in the paper.

I could go on, but all that I would say is either already in the literature, or awaits the results of the psychoacoustic investigations needed to fill in the few blanks left in the Journal paper that provoked these remarks. It is hoped that a new generation of peer-reviewed sound quality recommendation documents will emerge at some point. Then Mr. Skuruls may see the “real solutions” he desires – I was not being deliberately skimpy.  In the meantime he has created his own solution, described on his website.

Floyd Toole, Nov. 15, 2015