I see what you mean, but a few thoughts popped into my head as I read through:
- I don't think anyone has questioned preference. People enjoy the sounds their systems make or try to improve their enjoyment of them. That's great. Whether they are seeking the "absolute sound" (recreation of a live event or the master tape or however they define it) or simply something subjectively pleasing to them, that is fine. I don't worry about that. Heck, the latter people might be happier!
- Art is art, but that isn't (in my mind) what is being discussed. You do the things in the studio outlined in your second to last paragraph and produce a piece of art. And of course we can choose to enjoy it or not. And some will yearn for a system that reproduces it as you heard it in the studio, which I assume is how you wish it could be heard, to enjoy your art to the fullest. This involves the science being discussed in this thread.
- In the beginning of the high fidelity era, my perception is that the goal was the accurate reproduction of a live, acoustical event- real musicians playing instruments in a real space, primarily classical. This has morphed, perhaps based on practicality (or anatech may argue based on reality) to an accurate reproduction of what is on the master tape, even in the classical realm. The current reality that much music produced today never existed in a "real space," such that there can't be any reference, except to what the artist/engineer/mixer heard in the studio/mixing room.
- Because there isn't any reference to what we think we should be hearing, we have seen a shift toward what some have called "radical subjectivism," i.e. if it sounds good it is good, accuracy be damned. Again, fine. If that is the way you want to enjoy the music, go for it. But there remain people trying to achieve "accuracy" (however that is defined), hence the study of the sciences of transducers, acoustics, psychoacoustics, perception, etc. And don't we think these pursuits will benefit all listeners?
- There is much that isn't known in correlating measurements to perception, hence the deep dive that people have gone to in this thread and why many of us followed with keen interest.
- I think the big stumbling block was thinking people were discussing production and differing perception rather than reproduction.
- I would still rather listen to the Hot 5s and 7s, Charlie Parker, Coltrane, or EMI classical recordings of great artists after they moved away from their golden age on a table radio than a perfect recording on a million dollar system in a perfect room of Amazonian nose whistling.
- The OPs initial question was a reasonable one, and the answers very interesting.
Bill
I think that original goal you cite is not only still apt (as evidenced by many of the contributors to this forum), but also defines the extent of high fidelity loudspeaker engineering (notably where accuracy can be defined). Without an acoustical event in real space as you refer to it, I believe we are into the realm of designing "pleasant" sounding loudspeakers instead, which is not the same pursuit (since accuracy is not properly definable), and which often produces different results from those pursuing the original goal.
I believe too that the engineering goal will also be limited to sensation, where we will also likely find better correlations with perceptions too - even just by making better use of the measurements we have available now, for example. But the influence of factors such as life's prior experience and emotional states on perception imparts such a degree of non-linearity, that the accurate modelling of an individual's perception remains unlikely. Moreover fully overcoming this limitation appears somewhat implausible, even with a computer the size of a planet.
I think your perception of this thread's content is significantly different from any other contributor here. In particular, I am not aware of any contribution that has attempted to characterize the creation or enjoyment of music (by applying scientific method or otherwise). My perception is instead that this thread concerns what happens in between the creation and enjoyment of music: The engineering chain from microphone to ear, wherein scientific method is entirely applicable.
I think this is one difference between viewpoints. I have no problem with preferences people may have. I design both tube and solid state equipment, update and repair same. I apprenticed with tube era technicians and my formal training has been mostly solid state. However no matter the technology, the lower the distortion products and noise are, the better something sounds. I have never once had a client disagree with that. Mind you, I am looking at measurements most of you may not be familiar with as my equipment is current.
One fact remains. Electromechanical devices have seen the greatest advancement in the last few decades. This coincides with advances in measurement technology. I used T/S parameters in enclosure design as soon as I learned of them, and my products were instantly better. Today's speakers are a great deal better on average than ones made in the early 1980's, same with phono cartridges. We didn't get here "by ear", I can assure you. Every top speaker manufacturer is invested in measuring their product to make them better. They also audition them. I think you need to do both and have always said this.
The negative aspect is that measurement equipment is expensive. So is attaining the knowledge required, both in time and courses. What I see is the subjective type enthusiast wanting to be involved without all the training, equipment and experience. That's fine, but there are then limits and you actually do need more to progress. Early in time (1970's and before) we knew very little compared to today. You could hear problems we frankly were not able to easily measure or even figure out. So today we have much of the electronic signal path figured out. We are still figuring out the electromechanical devices. But today with electronic devices, if you hear something, we can measure and quantify it, period. We will get there with speakers.
But never confuse preference with accuracy. Accurate does not mean it sounds bad at all. I have heard so called accurate systems costing a great deal that did not sound very good. But done right, a properly engineered system always sounds better than one designed by ear. The designed systems are also (not surprisingly) more reliable.
As for art .. yes, to lay down good tracks is an art. To mix it well is an art. Never said it wasn't (I was just in the studio last night as it happens). But once it has been laid down as a master, it is a product and that is the reference. If you deviate from that on purpose, the result is not accurate - sorry. It may be what you prefer, and that's fine, but it is not what the engineer and producer intended. Reproducing a signal is not art, it is a science. All we can do is reproduce the signal the best we can.
The art in speaker design (as with anything) is the compromises to balance competing requirements to attain the best result. No shock there, all engineering is this balancing act. That's the same with the listening room, costs and what your partner may find acceptable. Some day we may get to a point where compromises are down to how speakers look, all speakers reproducing accurately. Until then, all you can do is avoid what you don't like and balance the rest.
One fact remains. Electromechanical devices have seen the greatest advancement in the last few decades. This coincides with advances in measurement technology. I used T/S parameters in enclosure design as soon as I learned of them, and my products were instantly better. Today's speakers are a great deal better on average than ones made in the early 1980's, same with phono cartridges. We didn't get here "by ear", I can assure you. Every top speaker manufacturer is invested in measuring their product to make them better. They also audition them. I think you need to do both and have always said this.
The negative aspect is that measurement equipment is expensive. So is attaining the knowledge required, both in time and courses. What I see is the subjective type enthusiast wanting to be involved without all the training, equipment and experience. That's fine, but there are then limits and you actually do need more to progress. Early in time (1970's and before) we knew very little compared to today. You could hear problems we frankly were not able to easily measure or even figure out. So today we have much of the electronic signal path figured out. We are still figuring out the electromechanical devices. But today with electronic devices, if you hear something, we can measure and quantify it, period. We will get there with speakers.
But never confuse preference with accuracy. Accurate does not mean it sounds bad at all. I have heard so called accurate systems costing a great deal that did not sound very good. But done right, a properly engineered system always sounds better than one designed by ear. The designed systems are also (not surprisingly) more reliable.
As for art .. yes, to lay down good tracks is an art. To mix it well is an art. Never said it wasn't (I was just in the studio last night as it happens). But once it has been laid down as a master, it is a product and that is the reference. If you deviate from that on purpose, the result is not accurate - sorry. It may be what you prefer, and that's fine, but it is not what the engineer and producer intended. Reproducing a signal is not art, it is a science. All we can do is reproduce the signal the best we can.
The art in speaker design (as with anything) is the compromises to balance competing requirements to attain the best result. No shock there, all engineering is this balancing act. That's the same with the listening room, costs and what your partner may find acceptable. Some day we may get to a point where compromises are down to how speakers look, all speakers reproducing accurately. Until then, all you can do is avoid what you don't like and balance the rest.
Actually, in DBLTs, electrostats don't do as well as some well designed electromagnetic drivers ... and some using Unique Fibrous Material (UFM aka paper) do a LOT better than Beryllium and other fancy diaphragm materials
Poorer 'waterfalls' for da ELS units point to some of the reasons why.
Last edited:
Please do. I've actually got clear rankings of fancy measurements and how they relate to 'sound' but there's too much noise in this thread to consider posting here.
I'll stick around a bit more to get to grips with a couple of fancy measurements I haven't had experience of.
If my single remaining brain cell hasn't gone short circuit from all the above supa dupa stuff, am I to deduce there are NO BISPECTRA MEASUREMENTS OF SPEAKERS OR UNITS on record?
So let’s examine the accuracy and reality. No matter the performance wether it acoustic classical/chamber recorded from a soundstage or multi tracked and produced in a studio, the packaged product is ‘produced/mixed/mastered in a studio environment after the fact…..and done while monitoring through a playback system and then verified from the master….tape and vinyl back in the day (sometimes today still) or digital file. Given the above, true accuracy to that recording would be taking a trip to the original mastering lab with the original 2 channel master under your arm to take a listen in the hopes that nothing has changed…..the environment, playback system, etc are all as they were. That’s ‘accurate’. There’s really NO OTHER way to verify the recording, performance or the embellishments placed upon it. And after listening, going home and then having to rely on our very poor aural memory, compare that event at home……provided you somehow managed to smuggle the master out under your coat.
Now let’s consider the true standard of accuracy ( I don’t think there’s a debate here, is there?) for a recorded piece of music……and apply the premise to this discussion……how would speaker measurements resulting in development of drivers and systems get us to the above standard of accuracy?…..is there something within those measurements that will allow our speakers to recreate the acoustic 3D space the master was created in?……will our choice of drivers to build the ultimate two way in any way approximate say a set of 3 or 4 way ATC monitors?….or PMC?, JBL, Genelec, and so on and so on.
I like to consider myself a pretty rational guy…..so applying the above logic, I personally don’t ever strive for accuracy….it’s not a reality…..or better put…anywhere near practical. I say better to return to the subjectivity of reasonable folks who objectively create a listening environment.
Last edited:
Please do. I am all earsPlease do. I've actually got clear rankings of fancy measurements and how they relate to 'sound' but there's too much noise in this thread to consider posting here.
But they wouldn't set it to ~60dB SPL especially to judge sound quality. They would want to set it to the original level .. and this is independent of background noise levels. Remember, they are trying to get the physical 'size' of the sound sources correct.
Hello mayhem13,
As unreasonable as that is for most people to do, I have.
For one, studio mains are not accurate, nor are the nearfield monitors. We all know this. The acoustic environment is not a listening room environment - you also know that. Each control room is set up differently, uses different electronics and speaker systems. Different engineers may even use headphones occasionally. But they have a practised ear and have a good idea how a mix will sound on their systems. For years the common speakers were NS-10 or NS-10M for nearfield monitors, many had JBL 412's or similar for mains. They did this so engineers could move between studios and have a similar sound for mixing. The awfultones are another matter again! lol!
However, whatever route they used to get there, the produced mix is the product.
Now, speakers. The only thing a good manufacturer can do is to produce a system with a flat response, low distortion and good transient response. That is the only target they can aim for. Likewise, producers of electronic reproduction equipment have the same targets. Now it comes down to the room and listening position. That is completely out of the control of everyone except the person setting that space up. So defects there are not the concern of manufacturers. You can purposely bounce sound all over the place (omni directional or rear high drivers). Those didn't do the trick, mostly because they were second guessing the space and placement. Early speaker (and pretty cheap, low quality) drivers didn't help either.
So what are we left with? All you can do is provide a speaker with known, agreed upon characteristics. In fact, this seems to work better and the way to do this is (thankfully) by measuring performance and working to correct impairments. Under this method, speakers have improved greatly over the last few decades. You can't argue with success.
Do we know everything that matters right now? Probably not. We have enough problems that we can't solve everything so we are down to trying to balance the design for the best result. Some people are more sensitive to certain problems, so their opinion will support the speakers that bug them the least - for what they can afford to spend.
So going to the original question. "Do measurements of drivers really matter for sound"? Yes, they certainly do. Our measurements have steadily improved over time and I expect this to continue. Speakers have improved in concert with better understanding born from better measurements.
How can you argue with that last statement?
As unreasonable as that is for most people to do, I have.
For one, studio mains are not accurate, nor are the nearfield monitors. We all know this. The acoustic environment is not a listening room environment - you also know that. Each control room is set up differently, uses different electronics and speaker systems. Different engineers may even use headphones occasionally. But they have a practised ear and have a good idea how a mix will sound on their systems. For years the common speakers were NS-10 or NS-10M for nearfield monitors, many had JBL 412's or similar for mains. They did this so engineers could move between studios and have a similar sound for mixing. The awfultones are another matter again! lol!
However, whatever route they used to get there, the produced mix is the product.
Now, speakers. The only thing a good manufacturer can do is to produce a system with a flat response, low distortion and good transient response. That is the only target they can aim for. Likewise, producers of electronic reproduction equipment have the same targets. Now it comes down to the room and listening position. That is completely out of the control of everyone except the person setting that space up. So defects there are not the concern of manufacturers. You can purposely bounce sound all over the place (omni directional or rear high drivers). Those didn't do the trick, mostly because they were second guessing the space and placement. Early speaker (and pretty cheap, low quality) drivers didn't help either.
So what are we left with? All you can do is provide a speaker with known, agreed upon characteristics. In fact, this seems to work better and the way to do this is (thankfully) by measuring performance and working to correct impairments. Under this method, speakers have improved greatly over the last few decades. You can't argue with success.
Do we know everything that matters right now? Probably not. We have enough problems that we can't solve everything so we are down to trying to balance the design for the best result. Some people are more sensitive to certain problems, so their opinion will support the speakers that bug them the least - for what they can afford to spend.
So going to the original question. "Do measurements of drivers really matter for sound"? Yes, they certainly do. Our measurements have steadily improved over time and I expect this to continue. Speakers have improved in concert with better understanding born from better measurements.
How can you argue with that last statement?
AFAIK, we don't have absolute detectors of loudness in our ears (nor brightness in the eyes). Our hearing adapts to new conditions within a few hours, it happens subconsciously, and not only in humans. You will hear how blood streams along your veins, how your stomach works, etc. You may even 'hear' how your brain works, pushing more blood to the active zones, and other fantasies. https://journals.physiology.org/doi/full/10.1152/physrev.00011.2022, for example
60 dBSPL would subjectively sound as loud as 80 - or even louder. the sound would be 'big'.
I accept all your points but I'm describing OBSERVED PHENOMENA; that some recording engineers can set a playback level within 1dB of their original recording environment, even in the presence of different background noise.
I can do this with my Soundfield mike recordings and also those using a single omni measurement mike. Several pages of why it's easier to do this with a B&K 4134 or Calrec Soundfield deleted
Another person on my DBLT panel that can do this party trick was 'Bat Ears Faulkner'. He prefered simple mike arrangements too but not as simple as my single Soundfield
I think your theory has to explain the OBSERVED PHENOMENA. Of course if the recording is completely artificial eg multi-miked, then no 'original recording environment spl' exists and this is moot.
If you could wait a bit Mike. I'm still trying to understand your LTI 'measurement' I need to do a lot more listening to the files you've provided and try to correlate them to our "distortion audibility tests" of da previous Millenium
I totally agree. I have not met such recording engineers. The set of people with whom I worked in anechoic chambers is rather small (~5) and is not statistically representative. I well may be drawing conclusions based on insufficient and/or cherry-picked evidence.I accept all your points but I'm describing OBSERVED PHENOMENA; that some recording engineers can set a playback level within 1dB of their original recording environment, even in the presence of different background noise.
I can do this with my Soundfield mike recordings and also those using a single omni measurement mike. Several pages of why it's easier to do this with a B&K 4134 or Calrec Soundfield deleted
Another person on my DBLT panel that can do this party trick was 'Bat Ears Faulkner'. He prefered simple mike arrangements too but not as simple as my single Soundfield
I think your theory has to explain the OBSERVED PHENOMENA. Of course if the recording is completely artificial eg multi-miked, then no 'original recording environment spl' exists and this is moot.
I’ve agreed on all of the above points earlier and elsewhere…..measurements are a great tool for research and design no doubt.
You could make the case for modern improvements is speakers to a point…..but I think it’s important to recognize the great designs of the 70s and 80s during the hifi boom that were for the most part derived of good old collective voicing…..lots of these were done before even TS Parameters were a thing……and folks had slide rules in their pockets to make calculations based on the available applied science at the time. There were some pretty amazing systems designed this way and are held in high or holy regard til this day.
In that lies my mantra……I’m often disgusted with the attempts to dehumanize nearly everything these days through tech and the examples of this are all around us in the subjective beauty of what was….and the cold, ugly and sterile appearance of the concrete and steel jungles of today. And this is only going to get worse with AI generated content and design. People in the creative arts fields are losing their jobs daily….print media artists and designers, photographers, marketing , publishing……all the first wave of the AI genocide. The writers are next…..producers are currently working and experimenting with AI generated content for TV. Full on AI musicians are right around the corner for the studio environment……we’ve already got great drummers and keyboards with huge MIDI banks…..won’t be long before they can create their own content to a requested key and tempo……..and be ‘humanized’. They won’t be late to show up, they won’t be drunk or high…..they won’t have an attitude and they won’t cost nearly as much……the temptation won’t be necessary…..it will be essential to meet budget once available.
And sighted listening?……using measurements to reverse align a subjective experience with an objective factor……I advise those using this methodology to proceed with caution……science taught us early on that this leads to classic conditioned behavior…VERY evident and on display with today’s YouTube kingdom of audio reviewers that use this method to analyze speaker systems………one fella makes up all sorts of nonsensical explanation for what he hears based on what his klippel shows him on screen…….i call it pre programmed prose……the ugliest spoken word art there is. I’d rather listen to Pavlov’s dog barking.
Last edited:
As I spoke of at the start of this thread, early academic studies appear simply to have been ignored for no good reason that I can fathom. However, there are several publications of loudspeaker analyses using the Wigner Distribution, and a handful of further publications that describe non-linear windows to best model our auditory capabilities. Several have been published by the AES and provide good tutorials with lots of 3D displays too.If my single remaining brain cell hasn't gone short circuit from all the above supa dupa stuff, am I to deduce there are NO BISPECTRA MEASUREMENTS OF SPEAKERS OR UNITS on record?
I am not aware of any commercial loudspeaker manufacturer who has published bispectral results, nor of any loudspeaker specific-studies. Yet you can even find mention of the information to be gained in the work that enabled Ambisonics and the development of the Soundfield microphone in which you were involved. I have only the two speculative reasons I supplied previously to explain why loudspeaker and related engineering still makes no use of these tools.
In other domains where "psychacoustics" are involved, or where the like aim is to separate information (such as in intercepting submarines, incoming missiles and the like), it is not treated as "supa dupa stuff", but part of the standard analysis toolbox - and for very good reason. Instead we routinely discard information that we already have in our measurements.
Purely as an example (and without any undue criticism), MLS measurements do very well indeed in measuring a linear frequency response with good noise immunity, but do so by dispersing the non-linearities in the noise. We all know that loudspeakers can sound different, but we often fail in identifying the cause from measurements. It should not be such a mystery: We have access to the information we need, just our measurements and analyses often obscure it.
Please post them! Far from confusion, there is unification to be had given the maximal information content that physics permits.Please do. I've actually got clear rankings of fancy measurements and how they relate to 'sound' but there's too much noise in this thread to consider posting here.
I'll stick around a bit more to get to grips with a couple of fancy measurements I haven't had experience of.