What role should listening play in speaker design

[Jaxx1138]

Much, if not all (comsol), of the acoustic (and more) properties can be predicted beforehand. Even for advanced hobbyists there is Akabak/ABEC to predict much of the directivity behaviour before building anything. Then it is a matter of selecting the correct drive units that operate well within their intended passband, etc etc..

Or perhaps I'm simplifying things.

This is generally true. But the measurements/simulations need to be extensive. On axis by itself is an extremely low resolution presentation that gives some data. But having the total sound power/distortion/ decay/compression and so on will tell you what you need to fix before you build anything. Starting with metrics on individual drive units. Far to many audiophiles will use a simple on axis response to determine the way a loudspeaker sounds and worse try to EQ to flat in a room. But I think we are moving away from that kind of oversimplification in the industry. More so, especially for consumers because of the internet, sites like this one and new publications about sound reproduction. Audio reproduction is a deep rabbit hole if you allow it to be.

 

[Jaxx1138]

So what is the answer? That we believe his ears and preferences???

Fact is that we need a standard. We are lucky that the preferred on-axis response is flat. There is zero excuse in producing a speaker with serious deviations from that. Certainly not when it is one designer's opinion.

Standards for loudspeakers have been a bit haphazard. Many times it is due to opinions of people in the industry without a strong background in the physics of acoustics and/or electrical and mechanical engineering. We all know that box specs are almost useless in discerning whether a loudspeaker is good or bad. I believe much of this stems from the fact that this is an industry where science overlaps a bit with the art. Where far too many people will conflate the two (it happens in ASR) as if they are somehow the same thing. Music production is mostly creative. Sound reproduction is all science. Psychologically getting any large group of humans to come to a consensus is fairly difficult. Even in science it requires a mountain of evidence for many of the outliers to admit that the confidence level is high enough to determine a useful standard. If you attend an audio shows you will be hard pressed to get manufacturers to admit the fact that sound reproduction is predominantly objective, (Don't want to alienate the subjectivist and lose sales). But based on my experience working in the industry, it is clearly objective. But you correct, if we are ever going to get to true fidelity in almost all loudspeakers, it will require a rigorous set of standards that determine where the metrics need to be to reproduce recordings accurately. As usual these are my conclusions, with all my bias, limited perceptions, and the sliver of available knowledge that exist in our universe.

 

[Newman]

...if we are ever going to get to true fidelity in almost all loudspeakers, it will require a rigorous set of standards that determine where the metrics need to be to reproduce recordings accurately....

We already have it. Books are out in public, documenting the science-based KPIs.

I am not sure what you are claiming does not exist? Are you seeking legislation? "A speaker shall not be offered to sale unless the manufacturer demonstrates its compliance with the following performance standards." Is that it?

 

[Keith_W]

@Newman just FYI the guy you are replying to is a Harman researcher who spent many years working with Toole and Olive. I think his photo is somewhere in the latest edition of Toole's book.

 

[Jaxx1138]

We already have it. Books are out in public, documenting the science-based KPIs.

I am not sure what you are claiming does not exist? Are you seeking legislation? "A speaker shall not be offered to sale unless the manufacturer demonstrates its compliance with the following performance standards." Is that it?

Standards well defined by a body of people like CEA , IEC , AES, IBU, or ITU that sets a standard set of parameters that defines a set of characteristics that determines a range or window of metrics that certify a loudspeaker performs within a range of acceptability for a particular application. Like a recording studio monitor or a consumer loudspeaker for a residential playback. I do not know of a standard that would fit within that description, that is widely accepted. There is research indicating that we need this and showing why. But I do not know of any standard that the recording industry adheres too that certifies a set of loudspeaker standards.

 

[Jaxx1138]

We already have it. Books are out in public, documenting the science-based KPIs.

I am not sure what you are claiming does not exist? Are you seeking legislation? "A speaker shall not be offered to sale unless the manufacturer demonstrates its compliance with the following performance standards." Is that it?

I stand corrected, I had not seen these or was not aware that anyone was adhering to this. I only knew of the THX standards (But not everyone uses those)
ITU-R BS.1116
EBU Tech 3276

I know most studios that I have seen are not following these due to the loudspeakers they use.

 

[Newman]

Yes, I was going to mention BS1116 once I knew the slant you were taking. I did not know of EBU3276, thanks.

 

[AcousticTheory]

I stand corrected, I had not seen these or was not aware that anyone was adhering to this. I only knew of the THX standards (But not everyone uses those)
ITU-R BS.1116
EBU Tech 3276

I know most studios that I have seen are not following these due to the loudspeakers they use.

These two publications are not 'standards' adopted by a panel, they are "recommendations" (per the preamble of the ITU-R document) or "tech notes". They don't have the force of a standard. I mention this just to forestall any representation that these documents represent a 'standard' adopted by an industry standards-making body for defining anything. Some requirements are proposed, but that is all they are, a proposal. An example of a 'standard' is IEC 60268, with sections -5 and -21 concerning measurement methods for loudspeakers.

 

[Head_Unit]

is a dabble into the circle of confusion, saying he doesn't really know what the source is, etc.

is the "truth" really the truth with measurements.

Fact is that we need a standard. We are lucky that the preferred on-axis response is flat. There is zero excuse in producing a speaker with serious deviations from that.

For me the circle of confusion comes about from conflating accuracy in reproduction with "what makes a lot of recordings sound good" as they are not the same.
- Accuracy in transduction would translate the input electrical impulse into a flat response on and off axis. How far off axis and how fast and smoothly it should curtail at the extremes could be debated (see next point) but all the responses should be flat. Period. THE END.
- However this might not sound good on many recordings, which are made in studios on monitors with their own myriad flaws and particular acoustics and then played back in rooms with their own particular acoustics. Recordings are mixed via such chains. Thus you get the idea of "house curves" where if the measured response at the listening position is flat into the high treble it will sound peculiar, and so forth. Or the case of Heart's 1985 self-titled album which when I would visit my best friend sounded great on her boom box but on a really good system, ugh, especially in like the upper bass.
- Plus any exaggeration from a flat response will bring out some sort of detail from various recordings that is not noted on a speaker without that particular exaggeration, just like bumping up an equalizer. Therefore some designers will make their own "house curve" of frequency response.
- And sometimes designers will accept flaws in frequency response in order to gain some other characteristic they believe important such as lower distortion or time response or directivity.
In the end I don't think you'd find may loudspeaker engineers who don't listen to their prototypes once they feel they are decent. And what do we listen with? Commercial recordings, made with all the flaws noted above. It would be really hard to 100% prioritize great measurements if a speaker made a lot of recordings sound not so good because we would be afraid nobody would buy it.

 

[Head_Unit]

I think the word "truth" needs to be defined here.

Already been done
https://www.discogs.com/master/68400-Jeff-Beck-Truth

 

[MattHooper]

Music production is mostly creative. Sound reproduction is all science.

If I may….

I would agree that sound reproduction can be all science/engineering.

And I certainly see the case for why, especially an engineer or scientist, would want it to be all science/engineering.
And that’s great. These are the people who are generally making advances in sound quality.

That said…. sound reproduction certainly admits of some creativity if that is what the listener wants to indulge in. So here I’m not first referencing the people designing and producing audio gear, but from the consumer end.

There’s a significant portion of audiophiles who approach sound reproduction, in terms of putting together their systems or even DIY, as a way of arriving at a sort of “bespoke” sonic presentation to please themselves.
They might be playing with various colorations with certain horn systems or SE tube amplifiers, and all that stuff… and there’s a certain amount of creativity possible in terms of the subjective “ playing around with sound” for such audiophiles.

I myself come from a career in Post Production film and TV sound design, in which I am manipulating sound all the time.
That tendency spills over somewhat into how I approach my own two channel audio system. I like to play around sometimes with certain colorations (eg my tube amps), and play around with speaker positioning, fiddling with room acoustics and all that to achieve specific types of sonic presentations.

I think there’s a bit of mild creativity in that, some user input. Of course everything being done is potentially measurable - and in principle of course amenable to being done more rigorously and scientifically.
But I’m not doing room measurements for every change I make to my system or my speakers or the acoustics…nor are many other audiophiles.

It’s also the case that, as you know, even among audio equipment designers, some are hardnose engineers going for neutrality, but some are not, and are essentially creatively playing with colorations in their products.
So again, there could be room for creativity and placing a personal stamp on the character of gear, especially a loudspeaker, in that sense as well.

So one may have the view that sound reproduction OUGHT to be all science and engineering. But it doesn’t seem to be the case that, in reality, everybody’s approaching it that way.

And personally, I see an engineer achieving whatever he wants, or a consumer achieving whatever they want, to be legitimate as well.
(While understanding that even when people are seeking bespoke colorations, a scientific and engineering knowledge based approach would be the most reliable way to do that as well).

Cheers.

 

[Raj21p]

Measurements get you most of the way there, especially for identifying obvious response and dispersion issues, but listening still matters because humans don’t experience sound the same way microphones do in a room. The best designers tend to use measurements as the foundation and listening as the final validation, not one instead of the other.

 

[sonitus mirus]

The way I see it, listening is not a good primary validation method for speaker design, but it is still useful as a final sanity check and as a way to discover what measurements you may have missed.

 

[Duke]

I find listening to be a valuable part of the design process in at least three areas:

1. Listening is pretty good at revealing whether there is something wrong that needs to be fixed, but is not particularly good at telling me exactly WHAT that something is. So listening can give me direction as to WHERE TO LOOK, but then the actual diagnostic information is usually found in measurements.

2. A/B listening to two different-yet-arguably-both-viable versions of a crossover will tell me which sounds better with greater reliability than eyeballing the curves does.

3. Evaluation of the effects on spatial quality of different aspects of the design is better done by ear than by measurements, and ime better done in stereo than in mono.