Speaker specification and the meaning in some terms

Ok, I try (but more knowledgeable members will post):

1. I have never understood: the 'frequency range' is supposed to be the range within which the loudspeaker is able to emit some musically audible sounds, but the next one is much more useful.

2. Usually, the really meaningful one is the low frequencies' minus 3 dB point: it means the frequency at which the sound pressure of the woofer is 3 dB lower than the average rest of its response. Since 3dB correspond to halving the perceived power, it is a meaningful indication of the limit of lower bass one can expect from an enclosure. It doesn't mean, though, that the emission of the woofer decreases vertically after that point, but only that that point is 3dB lower than the average rest of its emission, and that, likely, its response will decrease of another 3dB at each octave. Consider that 40 Hz is approximately the frequency of the empty lower string of a double bass, an E. Bear in mind, so, that if a given low frequency in a loudspeaker is indicated at -6dB, it means that the frequency that is at -3dB (hence at about half the perceived intensity than the average, full unit) is a full octave higher than that. I can't see how the same criteria applied to the high frequencies can be of any help, since any good tweeter today reaches far upper than most audiophiles' ears (given the average age of a typical audiophile). Than mine, for sure.

3. Harmonics are a bizarre thing. In music (and in physics) it means pitches that are generated by the same source emitting a fundamental tone, by virtue of a law by which a mass tends to vibrate in articulated sub-parts, and these parts have a dimension - and give birth to corresponding pitches - directly related to fractions of the whole: so, say, the 40 Hz E of the double bass, vibrating also in two halves, produces an 80Hz E, then a 120Hz B, then a 160Hz upper E again, and so on, following an order (1/1= F, 1/2=2xF/1 octave higher, 1/3=3xF/1 Fifth higher, 1/4=4xF/two octaves higher and so forth) that is called 'harmonic'. How this relates to a speaker's performances can only be more or less guessed by imagining that the mass of the cone, if not perfectly rigid – and behaving as a piston – can give birth to 'harmonics'.

4. Because a loudspeaker – and we can assume that a cable connecting a speaker to an amp is 'seen' by the amp as part of the loudspeaker's load – hasn't a steady and regular resistance, but one that is function of the frequency (and is called impedance), and is therefore different at various frequencies. But I am not sure that I have always seen cables' resistance indicated as impedance...

Hope it helps. Now, please, the true technicals may come in.

Frequency range of a speaker is the bandwidth it is able to present.   Absolute lowest and highest from the drive units and construction.

Frequency response gives an idea of how it'll sound generally .   If there is a + 3 db in 80 Hz expect a more chesty sound to male vocals and the speaker being happier in more free space,   If there is a drop of - 6db at 60 Hz yet a measure of 0 db at 60 Hz from a reflex port you would expect to hear it chuffing along .  Frequency response is best shown as a curved line .  Ultimately you want something that's as flat with +- db as little as possible.

+-3db from reference axis" will give some idea of how the speaker will eventually end up being positioned relative to the hot seat and how it could react to walls ,ceilings and floors.  Having a speaker close to a wall and firing straight ahead with no toe in could be the best way to get a flat response from some designs,   Having a speaker out in free space far from back wall and firing towards your ears might be the best way to get a flat response from other designs.

Frequency response is but one aspect of a speaker's performance. It is normally measured at a single point in front of the speaker. This is called the on-axis response.

You can have a very nice looking on-axis frequency response curve and an absolutely horrendous off-axis response. If the off-axis response curve deviates too far from the on-axis response, the speakers will probably sound terrible. In a room, you are hearing the on-axis response (assuming you're sat directly in line with the speakers) along with room reflections from the off-axis sound generated by the speaker. It helps to think of the frequency response as the 'colour' of the speaker's sound. The more the on-axis and off-axis colours differ, the more difficulty your ear-brain is going to have in sorting out the result. Consequence : listening fatigue.

A more meaningful measurement is the power response, which integrates frequency response from several axes around the speaker.

Bottom line : take specifications with a healthy dose of salt and trust your ears.

Jan

Paristhea,

With all respect, there is much to understand about the technical specifications of any loudspeaker.  Any short answer to your questions would require volumes of context for understanding.  Be assured that there is very little, if any, correlation between these specifications and the sound or suitability of any loudspeaker.  What matters is that a speaker play music well, and printed specifications will not tell you that.

I respect that you have a curiosity about this.  There are many books on the subject, anthologies of AES papers, numerous web sites.  Google can help you.  But after years of study, you still must only listen to a speaker to know how it sounds when playing music.

Charlie

If a speaker has a high +- db off axis you would expect the sound to change if you were to stand up or shift away from the hot seat, if it has a lower +- db number off axis you would expect not much change in tone and sound moving around way from the hot seat.

As for harmonics, nicely described by Max, the even order ones are generally benign and usually agreeable (euphonic). Odd-order harmonics sound - well - odd.

Tubes amps tend to have higher levels of second harmonics than transistor amps. Many like this added colour and one switching amp manufacturer (NuPrime) purposely augments second harmonics in their designs, to simulate 'tube' sound.

(thanks Charlie, you said it better than I could)

Jan

Jan-Erik Nordoen posted:

Bottom line : take specifications with a healthy dose of salt and trust your ears.

Jan

+1

I have just bought a pair of Focal 1008be. The Specs state it has a lower frequency of 46hz, my Living Voice state 35Hz. In my room the Focals plumb as deep, with better control and more tunefulness.

M

Massimo - great response. To your answer on harmonics... harmonics are crucially important as they define the timbre of a sound and is made up by a precise combination of the fundamental and its harmonics and the amplitude ratio of the harmonics with respect to each other  - thanks to good old Fourier.

Therefore if the ratio of harmonics in terms of amplitude is markedly different through non linearities or other distortions then the timbres reproduced will change - and if you were to draw the continuous sound pressure wave representing the sound- its shape would vary.

Now of course our brains don't help (or should I say do help) as they tend to fill in the gaps and correct errors.. so if we hear a guitar play on a poor system - it might be quite inaccurate and have a poor representation to a real guitar - but our brains say that is a guitar... but we feel the sound is quite synthetic or not real.

When I was an undergraduate a developed some software to replay sound built up from harmonic composition - and it was quite interesting to see what effect changing harmonics had... although my model only worked on a single fundamental at a time - however  in real life most sounds are made up of multiple fundamentals and their harmonics and interact with each other producing intermodulation sounds (sums and differences).

Also I was reading some AES research that showed that our brains can often superimpose missing fundamentals such as from bass sounds through speakers / systems that just can't reproduce that part of the fundamental signal effectively. This effect is taken advantage of by some mass consumer audio electronics and smart devices with very small speakers

Simon

Jan-Erik Nordoen posted:

As for harmonics, nicely described by Max, the even order ones are generally benign and usually agreeable (euphonic). Odd-order harmonics sound - well - odd.

Tubes amps tend to have higher levels of second harmonics than transistor amps. Many like this added colour and one switching amp manufacturer (NuPrime) purposely augments second harmonics in their designs, to simulate 'tube' sound.

Hi Jan,

you are (sensibly) speaking of an amplifier: that's why I found it a little unusual to read about harmonic distortion in a loudspeaker... 

Anyway, the distinction between even- and odd- order harmonics in music doesn't make much sense: the even order ones – let's say from 2nd to 16th – are essentially octaves of the fundamental and octaves of its major third and perfect fifth; yet, the first perfect fifth (the really deciding event of all the harmonic 'horizon') is the third one, hence an odd order one; and the first major third – another essential component of the tonal structure of natural sound – is the fifth harmonic, hence again an odd order one.

From harmonic 1 to 16, the tonally congruent ones are the large majority (with tonally congruent ones I mean all multiples (=octaves) of the three harmonics building what we call the perfect chord, i. e. the 4th, the 5th and the 6th – which, it could be of some interest for some, are elaborated by the same small group of cortical receptors, as if they were a single phenomenon); from then onward, since each new subdivision generates its own series of harmonics, mathematical ratios - and so perceptual consonance or dissonance - inside the series become more complex, and tonally congruent pitches become more and more separated from each other. But each is equally vital to the timbral and tonal constitution of sound.

So it's understandable that since the difference between a musical instrument and a technical instrument is that the first creates music from the null, while the second tries to reproduce an already generated (hence harmonically complex) sound, every epiphenomenon created by the microphone, the tube, the cone adds something that shouldn't be there. As long as it is a simple, very low in energy octave of a fundamental it can pass almost unnoticed, but harmonic complexities have a right to be only in music. Electronics must be discreet - if you allow me the pun.

Best

Max

This reminds me of something. Years ago, 'my' organ got restored. A colleague organist tried to get a 'sesquialter' (certain sound for the non-organists) added on the organ. Not possible, no funds and the organ case was already full.

During the first service after restoration I prepared for one of the songs a quite complex accompaniment. Lefthanded, I played the base chord. Base the pedals. Righthanded, I played on another clavier the overtones / harmonics to mimick a 'sesquialter'.

My colleague organist was very confused.

Great practical fun.

More ontopic: I like the comment that many speakers can produce tones which are much lower than their specification, but I've noticed that it also depends on the other sound material which gets played simultaniously on the speakers. If the frequency range is quite full, the lower tones get worse.

Well, my observation which is of couses my perception. Could be recording as well.

Maybe time for an Hicap on the supernait.

The harmonic distortion, in the OP's question 3,  is a specification which attempts to characterize (ever so simplistically) a loudspeaker's departure from linearity.  Most listeners, I think, would prefer a speaker which is highly linear (other things being equal).  All speakers exhibit some non-linearity, often increasingly so at higher volume.  Low order harmonics, as Max eloquently describes, are components of musical instruments and may tolerable.  When noticeable, they alert the ear/brain of the listener that HiFi system is present, and can distract from musical enjoyment.  Worse, the presence of non-linearity in the speaker also causes intermodulation between multiple voices and instruments playing simultaneously.  This distortion creates "noise" which fills in the silence that should exist between notes in time, frequency and space (for stereo).  Naim's DR power supplies are famous for the improvement in music when these intermodulations are minimized.

Harmonic distortion is used as a specification because is is convenient to measure (using sine tones or sweeps) and easy to compare.  Ideally it should be plotted as a function of frequency at many different output levels - all to often it is quoted at individual frequency/level points.  However, it is not just harmonics we hear when listening to music (unless we listen to sine tones - heaven forbid).  As non-linear distortion increases, a speaker most likely will sound less convincingly like real music.  Comparing two speakers, one can reasonably conclude that the lower distortion speaker will sound more like music, all other things equal.  Of course, other things are rarely equal so again this specification may or may not be useful as a predictor of good music reproduction.  One must then resort to listening.

Charlie

Mr Underhill posted:

Jan-Erik Nordoen posted:

Bottom line : take specifications with a healthy dose of salt and trust your ears.

Jan

+1

I have just bought a pair of Focal 1008be. The Specs state it has a lower frequency of 46hz, my Living Voice state 35Hz. In my room the Focals plumb as deep, with better control and more tunefulness.

M

It is not related to the main question, but I like to ask if you can provide a more detailed description of the Focal 1008BE sound please, as it is one of the speakers i am considering, along with the B&W 805D3.  Also what is Living Voice, is it an instrument?

Ardbeg10y posted:

I like the comment that many speakers can produce tones which are much lower than their specification, but I've noticed that it also depends on the other sound material which gets played simultaniously on the speakers. If the frequency range is quite full, the lower tones get worse.

 

Other than what's already written , this is I think to do with the speakers sensitivity. Most speakers put down on paper being either an 8 ohm or a 4 ohm design - but some frequencies like for instance 85 Hz might drop down to 1 ohm on some speakers. So your amp will be working much harder keeping music signal in this frequency in relative perspective to other frequencies.

Thanks Max, I appreciate the explanation. You're correct, the electronics should be discrete, but unfortunately, they rarely are completely so. To illustrate, below is an image taken from Zaph Audio's website showing harmonic distortion for a Peerless 830884 8-inch mid-woofer. The worst offender is F2, but as it is the second harmonic, it should be subjectively benign. I built a pair of speakers using these drivers and if it's any indication, I'm still having great fun with them even after eight years. The crossover took a year to get right though. Zaph tests speaker drivers using a sweep, rather than single tones, as the former is more revealing than the latter. Which points out once again that unless we know how a manufacturer did the harmonic distortion tests, the results are difficult to interpret.

The Scanspeak 22W8851T00 (below) would be an interesting choice for the next build, but there are far more parameters to consider than frequency response and harmonic distortion...

Jan

Phase angles ?  I'm a bit cloudy on what they are.

TOBYJUG posted:

Phase angles ?  I'm a bit cloudy on what they are.

They are the phase relationship between the voltage (potential difference across the terminals) and the current flowing through the speaker at a given frequency.  As has been pointed out speaker have an impedance (not resistance) that varies with frequency.  Clearly the frequency affects the rate of change of voltage and current, and when this is put across something with inductance and capacitance, this is what causes the complication.

However for sine waves of a single frequency it can be explained relatively simply...

-ve phase angle is more intuitively obvious:  The voltage from the amplifier increases, but the load is mostly inductive, so the current is delayed because energy is being stored in the magnetic field in the inductor and only builds more slowly than the voltage.  When the voltage starts to fall, the energy from the magnetic field starts to be released in the form of additional current.  In other words, the current is delayed behind the voltage.

+ve phase angle is the reverse:  The amplifier tries to increase the voltage, but the speaker is hungry for current, storing the energy in a capacitor.  As the capacitor fills with charge the voltage across the speaker rises.  The current leads ahead of the potential difference across the speaker terminals.

Although minimum impedance of a speaker (not it's nominal impedance or sensitivity) is often the largest factor determining how difficult a load a speaker is,  This phase angle effect can make one speaker a much more difficult load than another even if they have the same minimum impedance.

Jan-Erik Nordoen posted:

As for harmonics, nicely described by Max, the even order ones are generally benign and usually agreeable (euphonic). Odd-order harmonics sound - well - odd.

Tubes amps tend to have higher levels of second harmonics than transistor amps. Many like this added colour and one switching amp manufacturer (NuPrime) purposely augments second harmonics in their designs, to simulate 'tube' sound.

This may be true Jean-Erik but we must not forget that a loudspeaker impedance vary greatly below and above it's nominal impedance and the output of a valve amp will rise and fall in an attempt to maintain constant current to the speaker.

More than the level of second harmonics, this could explain why there is a difference in sound between ss. and valve amps depending on the linearity of the impedance curve for a given loudspeaker.