Contents:

1. The SPFR Technology – the use of Sound Power Frequency Response.

2. Equalizing loudspeakers. The new approach to use the high resolution of Time Domain Analysis.

3. Equalizing headphones. The challenge to get an exceptional accuracy performance from headphones.

4. Comments and discussion with Dr. Floyd Toole about his paper "The Measurement and Calibration of Sound Reproducing Systems". The paper and start of discussion is available here.

 


 

The SPFR Technology - Sound Power Frequency Response

It is all about the worst element of the great music’s path to you – a loudspeaker. You would expect loudspeakers to play back music as it exists in a digital or analog form. In reality, this hardly is the case. If you would look on a precisely measured frequency response curve of your speakers, you could see something like below.

A Typical Scenario

As we can see, frequency response is far away from being linear. This means that music played back is altered, some frequencys are boosted while others are almost not audible. If you’re a professional trying to fix the curve of your monitors or a PA with never reaching that great sound, there is a solution.

APL is the Solution

With a comprehensive range of APL products you can obtain a precice curve of your speakers with ease and create a suitable digital equalizer to achieve a perfectly balanced, linear frequency response for your speakers. This way you can achieve the most optimal sound that your speakers can provide.

1. Measure the Real Frequency Curve

2. Create a Precice Equalizer

3. Enjoy Great Sound

The Empirical Approach

The goal of electrically transmitting sound without distortions from performer to listener is as old as the universe of electro-acoustics. Freedom from any linear distortion is a necessity for attaining this goal. From a superficial academic standpoint, it all might seem quite easy: measure frequency response, create an inverse correction filter and the job is done. There have been several attempts to perform corrections this way but without any satisfactory results. According to authors of these attempts and their marketing support teams, the results are tremendous. However, the expert world of professional opinion takes a more impassive and skeptical view on these solutions, for a reason.

The problem is that technical evaluation instruments of sound systems both receive and evaluate sound in a different way than the human ear. As paradoxical as it might sound, they “perceive” more “problems” than our own sense of hearing does. The cause of this is the physical interference of sound waves at the point of sound pressure measurement. The interference occurs only when two (or more) signals (sound waves) come into touch - directly or reflected.

Therefore a sound system should be measured and evaluated the way our ears do. For this we have to eliminate any effects caused by interference in our analysis by measuring the sound power of a speaker instead of a single point sound pressure. This because  sound power frequency response measurement is free of any (even theoretical) effects of interference and we can acquire reliable information (curves) about the interferences of even up to 8 speakers in cases where just one speaker is placed in a corner.

If the loudspeaker has no directivity (sound wave radiation is the same in all directions, an ideal loudspeaker) the loudspeaker’s SPFR and AFR are the same. The positive effects of this approach has been acknowledged by other “players” in the field of sound system evaluation by introducing quiet and masked averaging (especially pointed to averaging in the power domain) in their measurement systems. This is the foundation of sound power measurement.

I would like to be so bold and review the approach to achieving undistorted sound transmission. Here is the prevailing principle. The microphone stands in front of a performer in a room (studio, open air) which converts the sound pressure therein to a proportional electrical signal, regardless of frequency. Afterwards, the transmission path (preamp, radio channel, etc.) ends in speakers in a listening room. The path should transmit the signal the same way regardless of frequency. The loudspeaker should proportionally convert the electrical signal to sound pressure – again, regardless of frequency. We verified that the loudspeaker complies with these requirements in an anechoic chamber on its „acoustic axis” and, therefore, we expect success in attaining undistorted sound transmission. I need not express how futile and naive these expectations are, in fact. This can be achieved with a speaker that has ideally even directivity in all directions, but such a speaker has yet to be designed and built in real life.

To receive an undistorted sound image at the listening space, the loudspeaker should emit sound power with the same or proportional sound-spectral components and temporal characteristics as the musician emits at the place of performance. The loudspeaker must emit the same sound power as the original performer does.

In fact, this is not such a complex task. When working in a studio with good acoustics, microphones are placed at a distance from performers and thus they pick up reflections throughout the studio. This way the sound waves emitted by the performer in almost any direction are enriched in the microphone after some reflections. We know how beautiful such recordings can sound (sonically quite rich). This principle has been at work for decades for evaluating the sound power emitted by a source under tests in a reverberation chamber (as opposed to an anechoic chamber). The next step is to emit the respective sound power in the listening space, with the loudspeaker having an equalized SPFR.

The validity of this approach has been proven empirically with great success for 12 years, including demonstrations at the AES exhibitions in May and October 2007. A recording of an accordion duo was played through a corrected path ending at a pair of Radiotehnika S90 loudspeakers, renowned for their sound in the former Soviet Union. This recording was compared with the live performance of the artists. The comparison was made by cutting the recorded sound in such a way that it played for 2 bars and then were silent for 2 bars. This 2-bar silence was filled by the live performers.

The accuracy of the method can be used in studios with the best quality monitors, but at the same time the possibilities of correction are so great that you can make speakers from buckets. We even tried that…

How does the sound power correction method work in real life? Measurement of the emitted sound power occurs in many (about one hundred or two hundred) points in space, arranged on an imaginary spherical surface (or its segment in the most important direction of radiation) around the speaker in which information is collected. In simpler terms: a measurer with a microphone in his hand draws an imaginary vertical lattice for about a minute (1 min=180 points). A specially developed program fixes the value of the sound pressure at separate points and later calculates sound (acoustic) power frequency characteristics (SPFR) where the factors of interference are eliminated. On the basis of these characteristics the correction curve is then synthesized. It is created to mirror the radiated sound power frequency response curve; one can now follow this curve at a level of precision unavailable in traditional equalizers. The fact is that a FIR, a filter with finite impulse response, is used as an equalizer in SPFR correction technology. This is nothing new for radio and communication engineering, but in sound engineering it has been used quite rarely.

But there is one more thing - phase correction. In SPFR correction technology, offered to customers as APL WORKSHOP software and the FIR-based equalizer unit APL1, correction accomplishes not only of AFR (amplitude frequency response) correction but also as PFR (phase frequency response) as well, by correcting the minimum phase segment of the PFR of the loudspeaker system. And this happens, we might say, automatically. The basis of this is that, if a problem (irregularity of AFR) was caused by a minimum phase system (which  happens in most cases of electrical circuits and filters having one signal path, from input to output), then by creating a minimum phase corrector, the problem is solved perfectly – both in amplitude and phase. The correction filter-equalizer applied in the SPFR correction system from APL is a minimum phase.

Go to top of page