-
Bluetooth Jul 25, 2024
-
Speaker measuring and reporting - Spinorama Mar 19, 2024
Speaker Measuring and Reporting Standard
Spinorama
"Measurements are better than opinions," proposed F. O. Toole and other influential voices in the audio world. Their clear opinion was that a new standard for measuring speakers and presenting data should be agreed upon: the Standard Method of Measurement for In-Home Loudspeakers, ANSI/CTA-2034-A. In short, they had enough of opinions suggesting that a speaker sounds "blue" and its bass is particularly "boxy". It was time to agree on a clear and unified methodology for measuring speakers and presenting data.
Measuring the frequency response of a speaker is simple. In an anechoic chamber, the speaker should be measured by rotating it around the horizontal axis every 10 degrees and then around the vertical axis also every 10 degrees. In total, 70 measurements are needed (if the speaker is symmetrical from left to right, then 52 measurements are sufficient). Since even 52 lines are too many to understand on a graph, the standard also describes a reasonable number of summarizing graph lines that are understandable and explain the speaker's character.
Angles are defined for rather big listening room
Measurements are typically labeled as H - horizontal and V - vertical. The zero angle points directly towards the listener and increases in increments of 10 degrees. Horizontally, the angle is usually described in both directions; for example, 10 degrees to the side is denoted as H +/-10. Vertically, measurements go from floor to ceiling, so an angle of 10 degrees upwards is denoted as V10, and an angle of 10 degrees downwards is denoted as V350, although descriptions of V +/- are also found in the standard.
In ideal world where speakers have no character whatsoever, all the lines are straight, first ones horizontal and last ones declining towards higher frequencies.
On-Axis: H0
This is a well-known and common graph of speaker frequency response. We'll cover important aspects with the next graph.
Listening Window: H0, 2x H10...H30, V10, V350
This graph illustrates the sound directly from the speaker (without reflections). Ideally it should be straight line from low to highs. If the graphs of the sound on-axis and within the listening window are close, everything is fine. If not, then the speaker has a slight acoustic issue in direct view, which can be addressed by applying some toe-in.
Next, you can consider the character of the speaker. Does the bass start low enough, or is a subwoofer needed? Does the high frequency smoothly reach at least up to 10 kHz? And the important question: how straight is the graph in the middle? Now it's worth carefully examining the dB numbers on the graph axis because the same graph can be presented with very different scaling. If the fluctuation of the graph is within a +/- 1 dB range, then it is a truly world-class speaker. If the fluctuation of the graph is within a +/- 2 dB range, then it is a well-made high-end speaker. If the fluctuation is within the range of +/- 3 dB, then it is a decent speaker worth measuring. And beyond that are the typical department store sold speakers, which are usually not measured by anyone. Of course, reality is not so straightforward; for example, is the graph humped like a camel, falling towards the bass or high frequency? If this hump is around 500-1000 Hz, then it can be corrected to be straight with a common amplifier tone control knob, and it doesn't make the speaker bad. If this speaker has a high and narrow peak, for example, at 16 kHz? Even here, life is very simple; it is not audible in real listening situations. What's audible are broad undulations, so there is no need to be disturbed by individual peaks or small notches. Be cautious with the midrange, but bass is controlled by the listening room anyway, and very high frequencies are not covered by notes. It's practically indistinguishable whether there is a slight irregularity in the harmonics or not.
Early Reflections: H0, 2x H10...H90, H180, V40...V60, V320...V340
These are the primary reflections from behind the listener, from the floor, ceiling, side walls, and rear wall. This graph is often considered the most important. Along with the direct sound, it contributes the most to the speaker's overall tonality. With this and each subsequent graph, comparing the shape to the on-axis graph is important. The reason for this is very simple: if there is a significant difference in the ups and downs of the graphs, the speaker cannot be properly adjusted with tone controls or an equalizer. (Except, if speakers are made different on purpose here and onwards!)The soundstage not only becomes chaotic but also uneven and unclear with music of varying frequency distributions. The graph should be uniform in the high-frequency direction, usually steadily declining. Attention is primarily paid to the frequency range of 300 - 10,000 Hz in this and all other directivity graphs.
Vertical Reflections: V40...V60, V320...V340
These are, naturally, reflections from the ceiling and floor. Generally, vertical reflections are not considered favorable for the listening experience. Commonly, the lower this graph is compared to others, the better the soundstage. In the case of a typical multi-way speaker, a straight line should not be expected here. Where the sound transitions from one driver to another, there are gaps in the graph due to interference along the vertical angles. This graph should rather be viewed as providing character, whether it imparts any uniqueness to the speaker or not. This is also where a well-executed three-way speaker can outperform a two-way speaker. For a two-way speaker, there is typically a wandering gap in the frequency range of 1000 - 2000 Hz. This is an important range for sound quality. A three-way speaker can draw this gap apart, lower and higher. Vertical reflections can also be considered from another angle: if the listener moves around in the room, or just stands up, how will this influence the sound tonality.
Horizontal Reflections: H0, 2x H10...H170, H180
Reflections from behind the listener, from the sides, and from behind the speaker together form the horizontal directional profile. This graph is slightly more important than vertical reflections. The proximity of the graph to the listening window graph indicates that the speaker has a wide directional profile, while the distance indicates that it has a narrow directional profile. If the holes found in the frequency graph vertically are also significantly present here, then some problems in the speaker's construction have been left unresolved. Ideally, this graph should resemble the listening window, declining in the high-frequency direction.
A typical example here is the struggle of many mid-high-mid configuration center speakers in home theater systems. As the angle changes, the midrange drivers on different edges start burying and amplifying each other's sound at certain frequencies and angles, making this graph the final nail in the coffin for these speakers.
Sound Power: all measurements
Here, the weighted average of all measurements is taken, forming the entire sound power graph emitted by the speaker into the room. As before, it is important to compare this with the listening area and its consistent nature of differences.
Predicted In-Room Sound: 12% LW, 44% ER, 44% SP
Based on the same measurements, the predicted frequency response at the listening position can be calculated. This represents how the sound spread in different directions ultimately sounds at the listening position. Please see, how significant is speaker sound in all other directions, compared to Listening Window. Typically, it should be a steadily declining line from low to high frequencies, with a decrease of about 6-10 dB across the entire frequency range.
Sound Power Directivity Index: LW-SP and Early Reflections Directivity Index: LW-ER
Directivity indexes refer to the difference between graphs. The Sound Power is subtracted from the Listening Window and the Early Reflections subtracted from the Listening Window. If the speaker radiates equally in all directions, these graphs are close to zero. The more directional the sound of the speaker, the higher the rear part of the graph rises. Rarely are these graphs perfectly straight. This is where it is most clearly visible where correcting the tonality of the entire sound range is possible and where it is difficult due to irregular differences.
Trustworthiness
Spinorama is a very effective method for assessing a speaker's tonal quality. Whether the goal is to choose a speaker with a flat profile or one that can be well-adjusted with an equalizer and subwoofer depends on budget and taste. Many manufacturers present corresponding graphs on their websites, and independent measurements are also available. Caution is warranted if no graphs are found for a speaker. Rarely is the issue that the speaker manufacturer cannot produce them for some reason. More often, the problem is that the manufacturer dislikes what they see in the graphs and prefers to sell their speakers based on brand, intriguing narratives, attractive design, or competitive pricing.
Spinorama has its limitations. When considering the angles used for measuring sound in different directions, it can be said that it is most accurate when listening to the speaker from a distance of 4 meters, with the speaker and listener's ears positioned 1.2 meters above the floor, and the room height being 3.6 meters. Following the guidelines for speaker and listener placement would require room dimensions on the order of 8 x 5 meters. However, real rooms are often smaller, and speakers are closer. If the speaker is intended for use on a desk or is tilted and positioned lower, the angles for the actual listening position should be summed differently. Whether these different angles affect the summed graphs depends on the speaker. Predicted in Room coefficients can wary also in real rooms. Open room with some windows, doorways and soft furniture can result more important Listening Window sound and less for all the rest.
Should Spinorama measurements be trusted? Yes, if not nitpicking every detail. A spike could also be a measurement error (not the speaker). There are also production variations, so for very precise graphs, the manufacturer should measure and adjust each speaker individually. This is done for very expensive speakers only. Therefore, one should not expect every measurement and product graph to be identical in every detail. Additionally, there is the rounding problem, where rounding of 1/20 octave is used up to 1/3 octave at the other extreme. For product variations and measurement errors, manufacturers tend to use larger rounding. A third of an octave is a bit too much and smooths the graph, potentially losing some information. A reasonable level is 1/6 octave, where problems are clearly visible, but there is no army of tiny spikes blurring the picture, as rounding larger than this makes the high and narrow spike (often inaudible) and the low and wide spike (often audible) appear the same on the graph. Thus, the best judge is an informed observer using their eyes and the time spent on the graphs. Nonetheless, any Spinorama reveals the speaker's constructional issues and its directivity profile. Additionally, Spinorama graphs serve as clear guidance on how to optimize the speaker's performance for oneself. Good electronics are much cheaper than a very good speaker. Therefore, with an equalizer and the right amplifier, even a more affordable speaker can be adjusted to high-end quality if the directivity profile is uniform, and there are no irreparable flaws in the speaker's sound.
The irreparable flaws referenced multiple times are: frequency gaps caused by crossovers that distribute sound to different loudspeakers and when the directional profile of the speaker varies greatly at different frequencies (wavy). Irreparable flaws (but not visible from the Spinorama) also include timing errors of the loudspeakers, problems arising from the characteristics of loudspeaker membranes and motors. Rectifiable flaws include: straightening out the overall frequency response curve and adding bass by using a subwoofer. The latter cannot even be considered a flaw, as many speakers are intended to be used with a subwoofer, so it cannot be seen as a disadvantage but rather simply as a specific feature of the speaker.
Sample Graphs (ANSI/CTA-2034-A, pg. 39)
The standard provides graphs as examples for manufacturers to present data. We can use them as examples for analysis.
Let's assume we are using this speaker with a standard amplifier, where the bass and treble knobs have an adjustment range of +/- 10 dB with a midpoint at 700 Hz. We'll move from low to high frequencies.
Let's also look at the grid of the graph; each line interval corresponds to 1 dB, and there are no concerns about scale or rounding. There are plenty of spikes, and the speaker is presented according to the measurement data.
In the adjustment range (-10 dB), the available bass starts at 50 Hz. However, it cannot be obtained with the bass knob because increasing it would also raise the bass in the 150 Hz range. The actual available bass is around 65 Hz. The rapid drop in bass below 50 Hz indicates that it's a ported or passive radiator speaker, and below this drop, bass cannot be obtained with any formula, although there is some additional resonance around 30 Hz. Both the port and the resonance raise questions about the purity of the bass section; is there perhaps a lingering echo? The bass starting from around 65 Hz is mostly suitable for listening to music, but for Hi-Fi quality, a subwoofer would be needed. Then the question arises: why invest in this 65 Hz, if instead, one could choose a smaller but higher-quality speaker where the bass drops out around 80 Hz. Continuing with the bass section, everything from 140 Hz to 1 kHz is excellent. Visually, the graph may seem wavy, but looking at the numbers, it's within +/- 1 dB, which is a very good result. With this, the bass section is long over, and we're in the midrange.
Moving upward from the midrange, the speaker clearly has a problem around 2.5 kHz. The graph drops by about 4-5 dB, and this is audible. This is a sensitive range for hearing, important for many string instrument sounds, and a significant part of the spatial sound comes from this range. Unfortunately, this dip cannot be corrected with a standard treble knob because the effect on the high-frequency knob is minimal here, and turning it up would affect even higher frequencies. However, nothing prevents correcting this part with an equalizer—the gaps between the graphs are uniform, and a 4 dB drop does not indicate a fundamental problem. The next small dip is around 6.5 kHz but only on-axis. It's likely due to some acoustic effect. This cannot be filled with an equalizer because it would create a peak in the room sound, but toe-in of the speakers may help because this dip is not present in the listening window or other graphs.
Moving higher, the next dip in the sound image for this speaker is around 9.5 kHz. Again, this can be corrected with an equalizer because it's consistent across all graphs. The high frequency starts to drop around 14 kHz. It's slightly too early for main speakers, but fortunately, it's quite a consistent and correctable line.
In conclusion, if one had to choose this speaker from among others for purchase, they wouldn't choose it. Although the important lower frequency part is nice, this speaker cannot be trusted for lower bass. A subwoofer would still be needed, and there are several unnecessary fluctuations at high frequencies. Fortunately, all these fluctuations can be corrected with an equalizer. If this speaker is already owned and the bass and directivity profile are pleasing, there's no need to give up on it. Without an equalizer, it may be cozy but still not a high-end speaker.
Returning to the initial topics of blue and boxy sound, one could also try to describe this speaker in words. The bass is punchy, and male vocals sound lively. However, cinematic effects fall short. Guitars and violins don't ring as they should. With some songs, consonants may hiss. Sparkling sound crystals in icy air are not heard from this speaker. Rather, it has a warm sound, but may be dull for some tracks.
Of course, such a description is absurd. Look at the graphs instead of reading incompetent reviews!
This analysis is not comprehensive and it does not need to be. It is just meant to make initial correlation between graph lines and sound, to raise more questions one can analyze further to resolve the sound system.
All TiFi Speakers have spinorama measurements, look for the link at the end of product description. Shortcuts: Tifi Elf Spinorama, TiFi Poseidon Spinorama, TiFi Helios Spinorama, TiFi Argos Spinorama, TiFi Hestia M Spinorama