• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Modern Measurement Tools Are Tricking Audiophiles Into Trusting Bad Data, Warns Veteran Speaker Designer

Status
Not open for further replies.
Another question which imho remains quite obscure is that of loudspeakers colorations, and seems to be totally overlooked while other aspects like frequency response, directivity control or harmonic distortion measurements are considered enough to evaluate the quality of a loudspeaker. You can have and compare graphically smooth enough spiroramas, but audible diferences still exists between speakers, so that what you see on those graphs is only a part of what you hear,and maybe too much attention is paid to some ugly details on a curve that is not really tragical at all, and other aspects remains masked between curves...

Probably the fact that most sales of audio gear are now realized on line have made necessary the use of a visual way of evaluating equipments, while the physical opportunity to make an audition has become less available to customers.
 
Last edited:
Has anyone ever tested how big the difference is in spinorama measurements with a microphone distance of 1 and 3 meters (approximately 3 and 9 feet)?
1 meter is not a realistic thing. Him mentioning it is confusion in that the spec for CEA-2034 says to measure at 2 meters, but report at 1 meter. So the only question here is the difference between 2 and 3 meters. I can run this analysis when I get a chance.
 
Between mix engineer & consumer is a mastering engineer, whose job is to...
I know that, but didn't want to complicate things further. Now, as you ask for it: not only the chief of mastering, also the panel that grants final approval. They sit together in front of a pair of speakers, so it's definitely not 'exactly in the middle' and so on. So: always stay relaxed. Data is important, but you can overdo it – and above all, any speculation should always be accompanied by appropriate, or at least with a serious hint on verification procedures before being shared with the public in heated discuissions :facepalm:
 
You can have and compare graphically smooth enough spiroramas, but audible diferences still exists between speakers, so that what you see on those graphs is only a part of what you hear...
That part is huge. I always find the sound of best measuring monitors to be quite similar as far as tonality. Let's remember that the research showed high correlation between listener preference and these measurements.
 
A few inputs here from someone who never does any measurements, but receive and review a lot of them:

Regarding literature on Spherical harmonics, Multipole expansion, and similar topics, you might not find what you are looking for in standard acoustics text books, but E.G. Williams "Fourier Acoustics" is a great resource here. Also, it you have studied quantum mechanics and spectroscopy (I have, years ago, alongside acoustics and more), the same mathematics is used a lot here, and Legrende Polynomials and similar will be part of the course(s).

To get a little intuition: In standard acoustics, which here means so-called 'isentropic' acoustics, we do several linearity assumptions in the derivation towards a single wave equation, and also adiabatic conditions are assumed alongside zero viscosity assumption. [I work a lot within microacoustics, where the latter two are not valid, and instead viscous and thermal effects are to be included, and so either the standard wave equation is modified to have complex values involved, or more fundamental equations are solved instead]. We can arrive a single wave equation, typically assuming steady-state conditions so we can write in the frequency domain as the so-called Helmholtz equation

grad^2 p + k^2 p = 0

We have assumed that no sources are explicitly in the acoustic setup, but energy is instead injected via boundary conditions or couplings to other domains, such as structural mechanical excitation. The p is pressure, which is a small acoustic perturbation on top of an ambient pressure.

Now, one big advantage of this is that we have a simple algebraic connection to the the temperature variation and density variation, and we could also have solved for one of those instead, but pressure is the natural choice.

But it raises some questions about complete knowledge (which is what we are going towards for an understanding of Klippel NFS) of the sound field, as something like intensity is found via the multiplication of pressure and particle velocity (not sound speed). So, how do get to this velocity? Well, under the isentropic assumptions, the sound field can be assumed 'rotation free', and so there exists a scalar potential that can describe the velocity vector field. This potential is called the velocity potential, and the wave equation could instead be written in terms of that (not so for microacoustics and other more complicated fields).

The crux now is that there is a relationship between pressure (scalar) and velocity (vector) as

u_vec "proportional to" -grad p

So, with knowledge of the pressure, we can derive velocity, opening up for intensity and power calculations. Also, Kirchhoff-Helmholtz integral states that if pressure and velocity is known across a closed surface, we can completely calculate the pressure field outside (or inside) of the surface. One can then imagine having a very complex source, somehow gathering information in a surface (near or far to the source, does not in principle matter, although in practice there is more to say here). Conversely, if we know the pressure at some surface, we have a big part of the picture, but to get to the velocity, we need the pressure gradient(!), which is turn means that we need an extra set of pressure; an extra 'layer' of sorts. In 1D, this would mean that knowing two pressures can give us the associated velocity as

u_x "proportional to" -dp/dx "approximately equal to" -(p2-p1)/(x2-x1)

We can do differentiation in different ways, and for numerical calculations it is important to have a high enough shape function order, but for some intuition on NFS, suffice to say that measuring in 'layers' adds important information. It is of course more important when having an unknown source in an unknown room(!), and you want to split the two, such that you have the source in question described via Spherical Harmonics or other compositions, but the insight into pressure and velocity will tells us something about the standing wave aspects of the sound field and the travelling wave aspects. While simplified here, this hopefully helps a little bit in showing how you can establish enough knowledge via measurements to a point where the source can be described via mathematical source components that in turn can then be placed into other rooms, if needed. So, if you can get the data out of NFS, you can potentially import them into COMSOL and get a very accurate pressure in some room that is also in the simulation, and you could even do auralization to hear any loudspeaker in your own room. But now we are going into some new territory ;-)

---

Now, on another note, if somebody can make a good loudspeaker it does not make them a genius, although you will see that notion a lot. The loudspeaker enclosure is very heavy, so reaction forces can be ignored compared to something like a balanced armature receiver (https://audioxpress.com/article/simulation-techniques-lumped-element-modeling-of-transducers), there are no feedback paths compared to something like a hearing aid that has many feedback paths, and a device such as smart speaker can have real-time non-linear processing going on with multiple microphones involved, and so there are many technical challenges in other applications/devices that the typical loudspeaker designer never has to think about. You can thus get away with thinking that since a battery on the terminals of a driver pushes it for example outwards, positive displacement in general drives positive pressure, thus completely misunderstanding both the transductance involved and the vibroacoustic coupling. Now, the driver design can certainly be difficult, especially when a single issue persists such as a dip or a peak, and several of my projects as a consultant are towards figuring out why such a problem is present, and to see if a solution can be found. But with properly designed drivers, hobbyist can sometimes make loudspeakers that put many companies to shame. All this to say that in order to gauge if someone is an expert/genius in a field, you will need to be at very high level within that field yourself. It is of course fine to be impressed and admire somebody, but there oftentimes is not a high level of theoretical knowledge involved (or needed) to achieve the particular end goal, but it is more down to experience and a trial-and-error/empirical approach.
 
I can run this analysis when I get a chance.
When you do, can you include a third version with (approximate) infinite distance (like 20/50m) just to see the difference?
That would be cool.
 
That part is huge. I always find the sound of best measuring monitors to be quite similar as far as tonality. Let's remember that the research showed high correlation between listener preference and these measurements.
Yes but that "listener preference" always sounds to me as a marketing argument that multinational companies owners of the big names require to rationally put some money on a project, but is far from giving the last word... :cool:
 
Yes but that "listener preference" always sounds to me as a marketing argument that multinational companies owners of the big names require to rationally put some money on a project, but is far from giving the last word... :cool:
The research for this dates back to Dr. Toole and his team at NRC. No commercial intent was involved in that, nor was a company. I suggest not promoting these conspiracy theories. If there is another "last word," let see the research on that. Not just claim that it must exist....
 
The research for this dates back to Dr. Toole and his team at NRC. No commercial intent was involved in that, nor was a company. I suggest not promoting these conspiracy theories. If there is another "last word," let see the research on that. Not just claim that it must exist....
Call that Research if you want... When Electrovoice or similar use DSP to obtain funky smiley frequency response curves from their PA speakers this is of course based on marketing Research and investigating consumer taste...;)
 
That speaker has no port on the back.

From Genelec's own website for the 8361A...
1760350559109.png



Rear view...

1760350749376.png
 
Call that Research if you want... When Electrovoice or similar use DSP to obtain funky smiley frequency response curves from their PA speakers this is of course based on marketing Research and investigating consumer taste...;)
You want to cut out the nonsense?
 
We have vastly moved the needle with respect to importance of measurements in the entire spectrum of audio products. The effect varies in category but it is 100% there or we wouldn't be have the retorts that started this thread. The impact is so big that folks are feeling threatened, trying to throw darts at measurements and creating talking points.
That is exactly what is happening here and that's why I love this forum (and appreciate people taking their time to do testing). This is also why we need independent measurements because it seems like audio industry is full of snake oil / marketing BS
 
It's that "slit", above and below the concentric, that you can see in the photo:
I believe that is the forward radiating slot for the main output from the internal bass driver. The bass reflex is the rear port, see photo that I added to my previous post after you saw it.

cheers
 
What non sense? Room or headphones curves, eq presets, etc... All based on consumer taste?

Published technical research from the NRC is considerably more credible than vendor market research. Watch your tone here.
 
Ecco alcuni suggerimenti da qualcuno che non effettua mai misurazioni, ma ne riceve e ne esamina molte:

Per quanto riguarda la letteratura sulle armoniche sferiche, l'espansione multipolare e argomenti simili, potresti non trovare quello che cerchi nei testi standard di acustica, ma "Fourier Acoustics" di EG Williams è un'ottima risorsa. Inoltre, se hai studiato meccanica quantistica e spettroscopia (io l'ho fatto anni fa, oltre ad acustica e altro), la stessa matematica è ampiamente utilizzata qui, e i polinomi di Legrende e argomenti simili saranno parte del corso.

Per avere un po' di intuizione: in acustica standard, che qui significa la cosiddetta acustica "isentropica", facciamo diverse ipotesi di linearità nella derivazione verso una singola equazione d'onda, e anche condizioni adiabatiche sono assunte insieme all'ipotesi di viscosità nulla. [Lavoro molto nell'ambito della microacustica, dove le ultime due non sono valide, e invece devono essere inclusi gli effetti viscosi e termici, e quindi o l'equazione d'onda standard viene modificata per includere valori complessi, o vengono risolte equazioni più fondamentali]. Possiamo arrivare a una singola equazione d'onda, tipicamente assumendo condizioni di stato stazionario, quindi possiamo scrivere nel dominio della frequenza come la cosiddetta equazione di Helmholtz.

grad^2 p + k^2 p = 0

Abbiamo ipotizzato che nessuna sorgente sia esplicitamente presente nell'impostazione acustica, ma che l'energia venga invece iniettata tramite condizioni al contorno o accoppiamenti ad altri domini, come l'eccitazione meccanica strutturale. La p è la pressione, che è una piccola perturbazione acustica che si aggiunge alla pressione ambientale.

Ora, uno dei grandi vantaggi di questo è che abbiamo una semplice connessione algebrica con la variazione di temperatura e la variazione di densità, e avremmo anche potuto risolvere una di queste, ma la pressione è la scelta naturale.

Ma solleva alcuni interrogativi sulla conoscenza completa (che è ciò a cui miriamo per comprendere la teoria della non-frequenza di Klippel) del campo sonoro, poiché qualcosa come l'intensità si trova moltiplicando la pressione per la velocità delle particelle (non la velocità del suono). Quindi, come si ottiene questa velocità? Ebbene, secondo le ipotesi isentropiche, il campo sonoro può essere considerato "privo di rotazione", e quindi esiste un potenziale scalare che può descrivere il campo vettoriale della velocità. Questo potenziale è chiamato potenziale di velocità, e l'equazione d'onda potrebbe invece essere scritta in termini di questo (non è così per la microacustica e altri campi più complessi).

Il punto cruciale ora è che esiste una relazione tra pressione (scalare) e velocità (vettore) come

u_vec "proporzionale a" -grad p

Quindi, conoscendo la pressione, possiamo ricavare la velocità, aprendo la strada a calcoli di intensità e potenza. Inoltre, l'integrale di Kirchhoff-Helmholtz afferma che se pressione e velocità sono note su una superficie chiusa, possiamo calcolare completamente il campo di pressione all'esterno (o all'interno) della superficie. Si può quindi immaginare di avere una sorgente molto complessa, che in qualche modo raccoglie informazioni su una superficie (vicino o lontano dalla sorgente, in linea di principio non ha importanza, anche se in pratica c'è altro da dire). Al contrario, se conosciamo la pressione su una superficie, abbiamo una parte importante del quadro, ma per arrivare alla velocità, abbiamo bisogno del gradiente di pressione (!), il che a sua volta significa che abbiamo bisogno di un ulteriore insieme di pressioni; una sorta di "strato" aggiuntivo. In 1D, questo significherebbe che conoscere due pressioni può darci la velocità associata come

u_x "proporzionale a" -dp/dx "approssimativamente uguale a" -(p2-p1)/(x2-x1)

Possiamo effettuare la differenziazione in diversi modi e, per i calcoli numerici, è importante avere un ordine della funzione di forma sufficientemente elevato, ma per una certa intuizione su NFS, basti dire che la misurazione a "livelli" aggiunge informazioni importanti. È ovviamente più importante quando si ha una sorgente sconosciuta in una stanza sconosciuta (!), e si desidera separare le due, in modo da avere la sorgente in questione descritta tramite armoniche sferiche o altre composizioni, ma la comprensione di pressione e velocità ci dirà qualcosa sugli aspetti delle onde stazionarie del campo sonoro e sugli aspetti delle onde progressive. Sebbene semplificato, si spera che questo aiuti un po' a mostrare come si possa acquisire una conoscenza sufficiente tramite misurazioni al punto da poter descrivere la sorgente tramite componenti matematiche della sorgente che a loro volta possono essere posizionate in altre stanze, se necessario. Quindi, se si riescono a estrarre i dati da NFS, è possibile importarli in COMSOL e ottenere una pressione molto accurata in una stanza presente anche nella simulazione, e si potrebbe persino effettuare un'auralizzazione per ascoltare qualsiasi altoparlante nella propria stanza. Ma ora ci addentriamo in un territorio nuovo ;-)

---

Ora, un'altra cosa: se qualcuno riesce a realizzare un buon altoparlante, non significa che sia un genio, anche se questo concetto è ricorrente. Il cabinet dell'altoparlante è molto pesante, quindi le forze di reazione possono essere ignorate rispetto a qualcosa come un ricevitore ad armatura bilanciata ( https://audioxpress.com/article/simulation-techniques-lumped-element-modeling-of-transducers ), non ci sono percorsi di feedback rispetto a qualcosa come un apparecchio acustico che ne ha molti, e un dispositivo come uno smart speaker può avere un'elaborazione non lineare in tempo reale con più microfoni coinvolti, e quindi ci sono molte sfide tecniche in altre applicazioni/dispositivi a cui il tipico progettista di altoparlanti non deve mai pensare. Si può quindi pensare che, poiché una batteria sui terminali di un driver lo spinge, ad esempio, verso l'esterno, lo spostamento positivo in generale genera una pressione positiva, fraintendendo così completamente sia la trasduttanza coinvolta sia l'accoppiamento vibroacustico. Ora, la progettazione di un driver può certamente essere difficile, soprattutto quando persiste un singolo problema, come un calo o un picco, e molti dei miei progetti come consulente mirano a capire perché si presenta un problema del genere ea vedere se è possibile trovare una soluzione. Ma con driver progettati correttamente, gli hobbisti a volte riescono a realizzare altoparlanti che fanno impallidire molte aziende. Tutto questo per dire che per valutare se qualcuno è un esperto/genio in un campo, è necessario essere a propria volta ad altissimo livello in quel campo. Va bene essere impressionati e ammirare qualcuno, ma spesso non è necessario (o necessario) un elevato livello di conoscenza teorica per raggiungere l'obiettivo finale, ma è più una questione di esperienza e di un approccio empirico/per tentativi ed errori.

A few inputs here from someone who never does any measurements, but receive and review a lot of them:

Regarding literature on Spherical harmonics, Multipole expansion, and similar topics, you might not find what you are looking for in standard acoustics text books, but E.G. Williams "Fourier Acoustics" is a great resource here. Also, it you have studied quantum mechanics and spectroscopy (I have, years ago, alongside acoustics and more), the same mathematics is used a lot here, and Legrende Polynomials and similar will be part of the course(s).

To get a little intuition: In standard acoustics, which here means so-called 'isentropic' acoustics, we do several linearity assumptions in the derivation towards a single wave equation, and also adiabatic conditions are assumed alongside zero viscosity assumption. [I work a lot within microacoustics, where the latter two are not valid, and instead viscous and thermal effects are to be included, and so either the standard wave equation is modified to have complex values involved, or more fundamental equations are solved instead]. We can arrive a single wave equation, typically assuming steady-state conditions so we can write in the frequency domain as the so-called Helmholtz equation

grad^2 p + k^2 p = 0

We have assumed that no sources are explicitly in the acoustic setup, but energy is instead injected via boundary conditions or couplings to other domains, such as structural mechanical excitation. The p is pressure, which is a small acoustic perturbation on top of an ambient pressure.

Now, one big advantage of this is that we have a simple algebraic connection to the the temperature variation and density variation, and we could also have solved for one of those instead, but pressure is the natural choice.

But it raises some questions about complete knowledge (which is what we are going towards for an understanding of Klippel NFS) of the sound field, as something like intensity is found via the multiplication of pressure and particle velocity (not sound speed). So, how do get to this velocity? Well, under the isentropic assumptions, the sound field can be assumed 'rotation free', and so there exists a scalar potential that can describe the velocity vector field. This potential is called the velocity potential, and the wave equation could instead be written in terms of that (not so for microacoustics and other more complicated fields).

The crux now is that there is a relationship between pressure (scalar) and velocity (vector) as

u_vec "proportional to" -grad p

So, with knowledge of the pressure, we can derive velocity, opening up for intensity and power calculations. Also, Kirchhoff-Helmholtz integral states that if pressure and velocity is known across a closed surface, we can completely calculate the pressure field outside (or inside) of the surface. One can then imagine having a very complex source, somehow gathering information in a surface (near or far to the source, does not in principle matter, although in practice there is more to say here). Conversely, if we know the pressure at some surface, we have a big part of the picture, but to get to the velocity, we need the pressure gradient(!), which is turn means that we need an extra set of pressure; an extra 'layer' of sorts. In 1D, this would mean that knowing two pressures can give us the associated velocity as

u_x "proportional to" -dp/dx "approximately equal to" -(p2-p1)/(x2-x1)

We can do differentiation in different ways, and for numerical calculations it is important to have a high enough shape function order, but for some intuition on NFS, suffice to say that measuring in 'layers' adds important information. It is of course more important when having an unknown source in an unknown room(!), and you want to split the two, such that you have the source in question described via Spherical Harmonics or other compositions, but the insight into pressure and velocity will tells us something about the standing wave aspects of the sound field and the travelling wave aspects. While simplified here, this hopefully helps a little bit in showing how you can establish enough knowledge via measurements to a point where the source can be described via mathematical source components that in turn can then be placed into other rooms, if needed. So, if you can get the data out of NFS, you can potentially import them into COMSOL and get a very accurate pressure in some room that is also in the simulation, and you could even do auralization to hear any loudspeaker in your own room. But now we are going into some new territory ;-)

---

Now, on another note, if somebody can make a good loudspeaker it does not make them a genius, although you will see that notion a lot. The loudspeaker enclosure is very heavy, so reaction forces can be ignored compared to something like a balanced armature receiver (https://audioxpress.com/article/simulation-techniques-lumped-element-modeling-of-transducers), there are no feedback paths compared to something like a hearing aid that has many feedback paths, and a device such as smart speaker can have real-time non-linear processing going on with multiple microphones involved, and so there are many technical challenges in other applications/devices that the typical loudspeaker designer never has to think about. You can thus get away with thinking that since a battery on the terminals of a driver pushes it for example outwards, positive displacement in general drives positive pressure, thus completely misunderstanding both the transductance involved and the vibroacoustic coupling. Now, the driver design can certainly be difficult, especially when a single issue persists such as a dip or a peak, and several of my projects as a consultant are towards figuring out why such a problem is present, and to see if a solution can be found. But with properly designed drivers, hobbyist can sometimes make loudspeakers that put many companies to shame. All this to say that in order to gauge if someone is an expert/genius in a field, you will need to be at very high level within that field yourself. It is of course fine to be impressed and admire somebody, but there oftentimes is not a high level of theoretical knowledge involved (or needed) to achieve the particular end goal, but it is more down to experience and a trial-and-error/empirical approach.
Thanks for your difficult-to-understand but intuitive explanation. I believe that some designers, if not geniuses, certainly possess above-average knowledge. The fact remains that even the "non-genius" who is passionate about his work, such as Andrew Jones, having studied science, is able to delve deeper into the topics you've discussed in this post, something inevitably inaccessible to the "do-it-yourselfer." In this specific case, we're not talking about an assembler, but rather a designer who, having a need, participates in the creation of transducers. I'll end by thanking you again, but I don't admire the subject; it's my ears that thank that concentric circle.
 
Published technical research from the NRC is considerably more credible than vendor market research.
Well the problem with Toole's, Olive's and other's works is that they are precisely part pure technical Research and part vendors market Research, and the last one is basically what their companies pay these costs for + the prestige and credit reported by their capacity to do so...
 
Last edited:
Grazie.
Se un libro fosse bastato, avrebbero risolto il problema. Invece, ci sono stanze che suonano bene e stanze che non si riescono a far suonare bene. È un processo di perfezionamento costante.
Why don't you read the book? It will be give you perspective on what distinguishes different kinds of rooms and the actual capabilities of acoustic treatments.
 
Status
Not open for further replies.
Back
Top Bottom