By Ryan Mihelich, Chief Audio Engineer and Chris Kyriakakis, Chief Audio Scientist
With few exceptions, product specifications have become a numbers game where bigger is better. Audio products fall squarely within the big numbers game. After all, who wouldn’t want more amplifier power, more audio channels, and louder playback?
One way to boost the numbers is to measure performance in creative ways. While the numbers you see may be technically correct, they often come without describing how the measurements were taken. For example, a small speaker measured at a distance of 10 cm can produce the same loudness as a bigger one measured at 1 m distance. But distance is almost never mentioned in specification sheets, and without that added context, any loudness measurement is effectively meaningless.
When we designed Cell Alpha, we focused on delivering high quality performance that is immediately obvious to the listener rather than producing large numbers for specs that are meaningless or misleading. That’s why, when we launched in May 2021, we intentionally did not release frequency ranges or loudness measurements. But we understand that the public has come to expect numbers with their hardware. So in this article, we’re sharing our perspective when it comes to powered loudspeaker performance and how we used these insights in the design of Cell Alpha.
Sound, Frequency, and Frequency Response
Let’s begin with the basics: what is sound and how is it characterized? Sound is generated by rhythmically pushing the invisible particles in air to create increases in pressure followed by decreases in pressure. As these patterns of high and low pressure arrive at our ears, we hear the sounds they represent. When the pressure fluctuations oscillate at a fast rate, they produce high frequency sounds (treble) and when the rate is slow, they produce low frequency sounds (bass).
The range of frequencies that we can hear is generally considered to be from 20 Hz (oscillations per second) to 20,000 Hz (20 kHz). Practical products limit this range to something less. A high quality loudspeaker (like Cell Alpha) may have a range from 30 Hz to 20 kHz while a mobile device ranges from 300 Hz to 10 kHz.
There is very little standardization in the industry on how to actually measure frequency response.
Frequency relates directly to the wavelength of sound being emitted, the distance from one pressure peak to the next peak, in the wave. The lowest audible frequency, 20 Hz, represents a wavelength of 56ʹ in air, whereas 20 kHz is about 3/4″. This dramatic range of frequencies and wavelengths is what causes many of the considerations that apply to loudspeaker designs.
A good sound system should reproduce a wide frequency range with a smooth frequency response (no pronounced peaks or dips) over that range. The wide range ensures that all (or most of) the frequencies in the original content are being reproduced, while the smoothness ensures that they are all reproduced without changing the octave-to-octave balance. Anything less leads to audible changes that cause coloration of the original sound.
From the discussion above, it is natural to ask: “Can we judge loudspeaker quality by using a microphone to measure the frequency response?” While there has been important research in this area that shows a correlation between a good (and properly measured) response and sound quality, there is very little standardization in the industry on how to actually measure frequency response. The desire to show the flattest and smoothest graph often leads to questionable measurement methods that provide little (and sometimes misleading) information on how the speaker will actually sound in a real room.
Directivity: Where Are You Listening?
A loudspeaker in a room radiates sound in various directions. The sound arriving at the listener is a combination of direct and reflected sound that a microphone weighs very differently from the way we perceive it. A frequency response measured at a single location provides little insight on the performance of the speaker. A more important consideration is where (and how) the sound output is “aimed.”
At low frequencies the wavelengths are so long and the drivers so small that all practical loudspeakers radiate sound uniformly into a spherical output shape. While low frequencies are omnidirectional, loudspeakers become increasingly directional as frequency goes up. Above a certain frequency (that depends on the size of the driver) the loudspeaker starts to “beam,” directing more energy out along its front axis. Thus, the output can be seen as omnidirectional at low frequencies, gradually becoming a directional “laser beam” as frequency goes up.
This change in radiation pattern from low-to-high frequencies is a problem for loudspeaker performance. The lower (omnidirectional) frequencies interact much more with the room before arriving at the listener while the high frequencies arrive through a mostly direct path. This imbalance causes audible sound quality degradation that changes as the listener moves around.
An ideal loudspeaker would be able to beam the widest possible range of frequencies to the listener with the same radiation pattern. But this is, of course, not possible because multiple drivers are used, each with its own beam shape that changes with frequency. The combined radiation pattern suffers from uneven directivity at the frequencies where one driver hands off duties to the other (the crossover points) because in that frequency range both drivers are active at the same time. A well-designed speaker should, therefore, be judged by how wide of a frequency range it can cover with a nearly constant radiation pattern. This was a critical consideration that guided the design of the Cell Alpha.
We first implemented controlled directivity in the Cell by using a custom-designed horn that provides a carefully-shaped path for the wave radiating from the loudspeaker to follow, guiding the wavefront and controlling the off-axis dispersion. This reduces the interaction with the other drivers in the crossover region and produces a smoother combined beam pattern.
An ideal loudspeaker would be able to beam the widest possible range of frequencies to the listener with the same radiation pattern.
In most traditional loudspeakers, the drivers are spaced vertically on the front baffle. This can cause variations in the time of arrival of high and low frequencies because they are radiating from a different physical location. While it is possible to time-align the drivers via a proper crossover network design, this only typically addresses the issue in the horizontal plane and not the vertical. In the Cell, we placed the tweeters and midrange drivers coaxially on a centrally co-located common axis. This enables proper time alignment that results in a much smoother directivity in the crossover region and over a wide frequency range (from about 600 Hz to 20 kHz).
A further innovation in the Cell design was to use three of these coaxial horn elements placed at 120 degrees spacing in what we call the Triphone. An advanced beam forming algorithm was developed to selectively combine the outputs of the three horns and create sound beams that have a constant radiation pattern down to 300 Hz. This is a significantly wider frequency range of directivity control than traditional loudspeakers and overcomes the audible degradations that occur at the crossover frequency regions of traditional loudspeakers.
It also provides another unique advantage: the Triphone can project multiple beams in different directions simultaneously. For example, the song vocals can potentially be directed to the listener while the guitars and keyboards to the left and right sides of the Cell. This creates a much more spacious soundstage compared to traditional single-box wireless loudspeakers where all the content is summed to mono and sent out in one direction.
Amplifier Power: What’s Watt
An audio amplifier is the device that converts the audio signal from your source into a higher power signal that can cause electro-mechanical motion to occur in a loudspeaker motor. Historically, audio systems have been made up of several components, speakers, amplifiers, receivers, equalizers, etc.) often sold and marketed by different companies. These components all needed to work together, with no knowledge of what the other components were. An industry blossomed around this where people could custom tailor their audio systems by buying new components and listening to potential sonic differences between them. Specifications for each component were important for users to know whether components were compatible or well suited for each other.
The importance of amplifier power boiled to the surface and became the single most advertised numbers for audio systems. The power wars began and manufacturers sought the highest numbers possible. These numbers are nuanced and require careful interpretation to be useful and comparable. Often neglected in these specs is an understanding of distortion, duration, and the load impedance. All that came to matter was a rating of BIG Watts.
More power does not make a better audio system, better sound makes a better audio system.
Fast forward to today and things are much different. The era of component based audio systems has been relegated to the HiFi world. Mainstream audio systems are integrated devices containing loudspeakers, amplifiers and a processor in a single package. Plug in your source device and you are listening to music in moments. In this world, the engineers have carefully designed each component in the system to work optimally together. There is no need for the customer to be concerned about interoperability of speakers and amplifiers when they are already integrated. The amplifier power numbers of old are technically meaningless and irrelevant to the end consumer.
More power does not make a better audio system, better sound makes a better audio system. Well designed, low power systems can outperform poorly optimized high power systems. Listeners judge the audio system by the sound it produces, not the irrelevant, technically incorrect (there is no such thing as Watts RMS that some manufacturers list), and misleading numbers that manufacturers often put on spec sheets.
What truly matters in defining the performance of a loudspeaker is its efficiency: the power that turns into sound, not the wasted power that turns into heat inside the device. Cell Alpha is designed with efficiency in mind. The speakers and amplifiers have been optimized to deliver beautiful sound at reference sound levels while drawing minimal power. This allows the Cell to achieve a very high level of performance in a very small package and with minimal heating, all without fans. You won’t have to worry about burning your hand on Cell Alpha.
How Loud is Loud
The loudness of sound is specified using sound pressure level, or SPL. It is the level of acoustic output, measured in decibels, or dB SPL. It is a relative number that is referenced to the softest sound we can hear (called the threshold of hearing). Quiet sounds such as whispering measure around 20–30 dB SPL, normal face-to-face conversation is around 60 dB, a truck 10 feet away can be 90 dB, and a rock concert near the stage can easily exceed 100 dB.
Placing the mic closer would give a higher number!
The sound pressure level depends greatly on the distance from the source. If it is not specified, then the number is meaningless. For example, that same truck at 100 ft away would be a much more polite 70 dB.
In a powered loudspeaker, the SPL rating usually refers to the maximum acoustic output that the speaker can produce before the output is limited (due to insufficient power or distortion). But, this number cannot be listed in isolation without specifying the distance at which it was measured. Placing the mic closer would give a higher number! A proper SPL rating should also describe the test signal type (e.g. noise or sine tones), and signal duration. For example, a loudspeaker can produce a very high output for a short period of time, but that is not a realistic number for long term listening. This is called peak SPL and is the largest number that will be found on a spec sheet, often (but not always) measured at 1 m from the speaker.
An SPL rating based on a longer term average is much more useful to inform the consumer as to how loud the speaker can be expected to sound during continuous playback.
As is the case with power ratings, there is no industry standardization for measuring SPL. Manufacturers often choose to measure the SPL of their products with short bursts of sound at close distance thus boosting their numbers and making it impossible for the consumer to compare.
When we built Cell Alpha, we focused on creating a high fidelity listening experience. We know that our product can compete with any other speaker out there in terms of stats, but we also know that most of the stats out there are not actually useful to any end consumer.
So while the Cell reproduces a wider frequency range than speakers of similar size, it does so in a way that makes it difficult to compare to other powered speakers. With advanced directivity control and an extremely efficient electroacoustic design that doesn’t waste power into heat, it’s unlike any other speaker out there. You might have to hear it to believe it, but once you do, you’ll understand. This is no ordinary speaker — it’s Triphonic.
Ryan Mihelich is Chief Audio Engineer for Syng and has previously served in a similar role for Apple and Harman International. Ryan has been designing loudspeakers and audio systems for the consumer, automotive and professional markets for more than 25 years.
Chris Kyriakakis is Chief Audio Scientist for Syng and Professor of Audio Signal Processing at the University of Southern California. He has published more than 100 technical papers and a textbook on Immersive Audio. Chris is a Senior Member of the Institute of Electrical and Electronics Engineers (IEEE) and a recipient of the World Technology Award in Media.