An interview with Floyd E. Toole,
Harman International.


1. Please tell us about when and how you came to see the importance of ABX tests in audio?

Before I joined Harman International, I was a research scientist, at the National Research Council in Canada. In the 1970's I realized that there was a need for some serious research into the psychoacoustics of loudspeakers and rooms, and I examined the literature that existed at the time. There were plenty of measurement techniques, to generate technical data - graphs and numbers - but there were no reliable rules for interpreting them in ways that related to human perceptions. At the same time, there were people saying that "we all hear differently", and that there can be no rules for what sounds good. I simply did not believe that, and decided to test it. Looking into the methods of experimental psychology, it was clear that blind,and double blind testing was necessary, and that it was also necessary to control other factors, such as loudness, if we were to have any hope of examining the true opinions of listeners. When we did listening tests with even very simple experimental controls, we found that most people, most of the time, liked and disliked the same loudspeakers. There were exceptions, and when we examined why, we found that hearing performance was a main factor. People with close to normal hearing thresholds, and who had some experience in critical listening, all agreed very closely in their preferences. So, it is true that we do not all agree on what is good, but those persons with normal hearing, and some skill in making judgments about sound quality, show remarkably similar opinions. When people differ from this group they differ in random ways, so it is not possible to rely on such listeners for guidance or advice. All of this is described in great detail in:

1. F.E. Toole, “Listening Tests, Turning Opinion Into Fact”, J. Audio Eng. Soc., vol. 30, pp. 431-445 (1982 June).

2 . F.E. Toole, “Subjective Measurements of Loudspeaker Sound Quality and Listener Performance”, J. Audio Eng. Soc., vol 33, pp. 2-32 (1985 January/February)

 

A more recent description can be found in a chapter in the Focal Press book, "Loudspeaker and Headphone Handbook" edited by John Borwick, a new edition (the 3rd) will soon be released. ABX testing is just one of many techniques for evaluating sound quality. It is very useful for settling "is there an audible difference" kinds of tests - e.g. wires, CD players, amplifiers, perceptual encoders. The results of such tests are, ideally, yes or no. For loudspeakers, the differences are clearly audible, and the question is more one of preference and why there is a preference, so we use multiple comparison techniques, which give listeners a better "context" within which to form what is a very complicated opinion.

2. Loudspeakers cannot show such a good frequency response, distortion figures or step response as amplifiers, do you believe they will ever come close? And will it be possible to make loudspeaker boxes without a boxy sound? What quality is most in need of improvement?

There are two fundamental properties of a loudspeaker - frequency domain behavior (amplitude and phase response - both of which define the time domain response) and directivity. With the excellent transducers that we can now make (e.g. the new Infinity C.M.M.D. ceramic metal matrix diaphragms that have no diphragm resonances within the audible band) and digital equalization, it is possible to make loudspeakers that are remarkably close to the ideal performance. Directivity is more complicated, being influenced by the shape and size of the enclosures, and the number and sizes of drivers. But, even there, things are now getting very much better than in the past. Time-domain behavior (you call it step response), is not a serious factor, as humans are very insensitive to phase shift. Only very large group delays - larger than occur in normal loudspeaker designs - are audible, and even then normal listening room acoustics make the differences even more difficult to hear. Right now, some loudspeaker designs come close enough to the ideal that the opinions of listeners are more influenced by the variations in recordings and listening rooms. Those are the factors most in need of improvement. Many recording studios use monitor loudspeakers that are seriously bad, and all rooms control the bass response that we hear. All rooms are different.

The boxy sound of an enclosure is caused by acoustical and mechanical resonances. Good engineering can reduce these to inaudible levels. I remember, in a blind test of several years ago, listeners complaining of a "boxy" coloration from a certain product. It was audible to me as well, and moderately annoying. When the curtain was opened, the listeners saw that the boxy sound was coming from a large full-range panel loudspeaker. The resonance was in the panel - there was no box. It turns out to be difficult to design panel loudspeakers that do not have resonances. There is no "magic" about panel speakers, and there is no need for a box to sound boxy.

1. Loudspeaker Measurements and Their Relationship to Listener Preferences”, J. Audio Eng, Soc., vol. 34, pt.1 pp.227-235 (1986 April), pt. 2, pp. 323-348 (1986 May).

2. F.E. Toole and S.E. Olive, “The Modification of Timbre by Resonances: Perception and Measurement”, J. Audio Eng, Soc., vol. 36, pp. 122-142 (1988 March).

 

3. Music reproduction via a surround matrix like the Fosgate "Classical" mode is something I have found beneficial to get a more natural ambiance in the listening room when listening to two channel stereo recordings. It sets the music free from the loudspeaker in a refreshing way. I would not spend a larger amount of money on any audio reproduction equipment restricted to only two channels, regardless of if it would only be used for music listening. Many audiophiles do not seem to be of the same opinion, but claim that more than two channels have not got a place in a serious music reproduction system. At the same time, the big manufacturers, (with an interesting exception for Harman International), do not want to do anything with matrix extracted ambience, but instead offer some reverb-effects in their multichannel amplifiers, that produces a sameness coloration to everything that is played. What are your thoughts about this?


My opinions are explained in great detail in a two part article that appeared in the American "Audio" magazine, May and June 1997. It is on the www.harman.com website as a white paper "Direction and Space, the Final Frontiers". I absolutely agree that most (but not all) stereo recordings sound better when reproduced through a really good surround matrix. The Fosgate 6-axis is one, and a newer, more powerful one, is the Lexicon Logic 7 algorithm. It is more powerful because it is all digital, and the listener can adjust the parameters of the steering algorithm to suit personal tastes, or the particular method of recording. Since stereo recordings were not made with this kind of decoding in mind, some amount of artistic twiddling can be beneficial.

The addition of reverberation to the playback of stereo recordings is absolutely not comparable to what I have discussed above. I find it to be generally not acceptable, and frequently very offensive. I never do it.

The idea that two-channel stereo is somehow more "pure" than multichannel audio is misguided. In spite of some examples of very bad taste in the days of quadraphonic audio, and now, the benefits are there, and gradually recording engineers will learn how to use the new medium. It is the better solution. However, I feel that more than five channels are needed. I live with the Logic 7 system in its 7-channel mode, and I cannot now go back to 5.1. The sense of envelopment and space is much more realistic with the additional channels (two sides and two rears), and it remains so even when you turn your head or move around the room. Remember, multichannel audio is a social experience - more than one person can enjoy it.

4. I believe that many audiophiles would get more from their equipment if they would transfer some of their interest and money for audio cables into acoustics and room adaption, but since audio cables seem to be of such big interest, maybe you could share what you think are the relevant qualities when it comes to loudspeaker cables?

Cables are very profitable products, and that is the main driving force behind them. At a time when advanced technology has reduced the number of tweaks that audio enthusiasts can play with, it is natural that these products should become topics of conversation. I call the most extreme of them "audio jewellery" , in that they do nothing for the audio system except make the owner feel better or more proud. Superbly performing audio cables can be purchased for very moderate prices. Even "bad" cables, are not bad enough to be audibly worse than the truly nasty things that some rooms or poorly designed loudspeakers can do.

5. In Sweden we have got the motto; "There are no shortcuts to the perfect sound" Do you think it is at all possible to find a path to the perfect sound, and what would be the biggest obstacle before getting there?

I am not sure what "the perfect sound" might be. If it is to transport a listener to a concert hall, then we have gone about it the wrong way - we should have focused our efforts on binaural recordings and playback, or long before now moved into multichannel reproduction, using many more than 5.1 channels. If it is to reproduce in the home the exact sounds heard by recording engineers and musicians in a studio control room (the original sound for most recordings), then we should long ago have enforced some standards in control room loudspeakers and acoustics, so that that experience more closely resembles what we are able to reproduce in homes. Now, when we do listening tests, it is very difficult to find recordings of any musical genre that are not modified by colored microphones or equalization added during the mix or mastering. Did you know that many commonly used microphones have much worse frequency response at high frequencies than the tweeters in even moderately priced loudspeakers? This is embarassing for an industry as mature as audio.

 
© Interviewed by: Per Arne Almeflo
Also translated to swedish for publication in Musik & Ljudteknik.