Electrical vs Acoustic Measurements.

[rgpit]

Please bear with me as I'm not an electrical or acoustic engineer and only know enough to be dangerous. I've been thinking about loudspeaker measurement and the objective vs subjective points of view. I'd like to present the following for discussion:

As I understand it, the electrical components of our audio systems including dacs, processors, and amplifiers deal with amplitude of an electrical signal over time. One can fairly easily measure the difference between the input vs output of a signal through a given device and determine it's performance. Loudspeakers have a completely different kind of task. They first have to convert a “two dimensional” electrical signal into a mechanical signal and cast it into a “three dimensional” image for our ears. I think there are aspects of speaker performance that can be measured accurately with test gear but there are aspects that can’t easily be measured without “seeing” the acoustic field surrounding our ears. For me this is my subjective evaluation for things like a speaker’s imaging. I can’t see how one can use a traditional acoustic sensor like a microphone to measure all that a speaker must do to cast an acceptable acoustic picture in a room.

Ron

 

[pozz]

One thing to remember is that the three dimensional part takes place in your head. It's a computed, synthesized human artefact. Your ears are just two points in space, not so different from the superimposed waves of pressure sensed at the location of the microphone. With this in mind you can use a pair to simulate human hearing or as many as you like to capture the soundfield itself.

 

[pma]

One thing to remember is that the three dimensional part takes place in your head. It's a computed, synthesized human artefact. Your ears are just two points in space, not so different from the superimposed waves of pressure sensed at the location of the microphone. With this in mind you can use a pair to simulate human hearing or as many as you like to capture the soundfield itself.

Yes, but please do not forget to mention there is one big difference between 2 microphones in space and our ears+brain located at the same place. Let's forget the head shadow effect, my point is the processing of direct and reflected sound. Microphones cannot make a distinction, though our ear-brain system can. So a recording from speakers in our living room adds room acoustics and reflections (when we replay the recording) that we do not notice during normal hearing process.

 

[Hipper]

As mentioned above, the 3D image - or phantom image - is created by your brain using information received by your hearing system.

The whole stereo listening set up depends on not only the electrical equipment, then the speakers, but the room interference after the sound has left the speakers, what your ears pick up (some say even other parts of your body like the jaw), and finally, how your brain interprets this information - psychoacoustics.

So your sentence should read (missing out 'first') - 'Loudspeakers have a completely different kind of task. They (first) have to convert a “two dimensional” electrical signal into a mechanical signal.... which in turn moves air molecules.'

It's the movement of air molecules that we use to collect the information.

Although there are measurements of speakers' individual performances (such as in an anechoic chamber, using a spinorama, or the gear Amirm is using), ultimately these can only be a guide to how they might perform in a listening room. The most important measurements for a listener come from measuring the speaker in its listening room, using something like Room EQ Wizard (REW). We must then try to understand a bit about how sound behaves in a room, and psychoacoustics. After all this we should be able to form a picture of how those speakers will behave in our particular room at our listening position.

To me it seems that the speaker/room interaction plus individual hearing/psychoacoustic properties mean that it is very difficult to find measurements that will exactly decide if a speaker will work for you. That is why we rely on home listening to make a final decision. Some will say that that decision is possible with measurements, but as for example the ones that Harman have produced, namely flat on axis and smooth off axis response, seem to exclude a bunch of speaker types such as electrostats which many people enjoy.

 

[pozz]

Yes, but please do not forget to mention there is one big difference between 2 microphones in space and our ears+brain located at the same place. Let's forget the head shadow effect, my point is the processing of direct and reflected sound. Microphones cannot make a distinction, though our ear-brain system can. So a recording from speakers in our living room adds room acoustics and reflections (when we replay the recording) that we do not notice during normal hearing process.

We said the same thing in different terms:

... the three dimensional part takes place in your head. It's a computed, synthesized human artefact ... you can use a pair [of microphones] to simulate human hearing...

Our ear-brain system doesn't make a distinction either. It just computes relative levels and timing.

 

[Juhazi]

https://audioxpress.com/news/regist...ty-of-audio-systems-2020-lecture-are-now-open

The 2020 lecture on “Sound Quality of Audio Systems” presented by Prof. Dr. Wolfgang Klippel, honorary professor of the Institute of Acoustics and Speech Communication, will offer participants a deep understanding on the latest measurement and diagnostic techniques used in telecommunication, automotive, multi-media and professional applications to design small, light and cost-effective loudspeakers. The content reflects over 30 years of fundamental research by Klippel, allowing participants to improve their own designs and optimize manufacturing processes.

 

[No. 5]

I think there are aspects of speaker performance that can be measured accurately with test gear but there are aspects that can’t easily be measured without “seeing” the acoustic field surrounding our ears. For me this is my subjective evaluation for things like a speaker’s imaging. I can’t see how one can use a traditional acoustic sensor like a microphone to measure all that a speaker must do to cast an acceptable acoustic picture in a room.

In my opinion, you are touching on an unavoidable fact of sound system design, in that more then a simple set of measurements are needed to effectively "see" what's happening. Case in point: our hearing appears to have some preferences when it comes to spectrum, amplitude, direction, and phase and delay of reflected and direct sound. A single microphone location will not make sense of all that.

However, a microphone and our eardrums both "measure" the sound pressure at a point(s) in space. Therefore, I don't believe that a microphone is unable to be used for quantifying imaging, but it's more likely something we don't yet know how to use it for that.

Perhaps it's a case where the right kind of data is needed. Consider how imaging is affected by loudness, phase, and delay. And then add to that the fact that it's quite common for two "matching" loudspeakers to not quite be matched with respect to frequency response and phase. The more similar the sound of the two sources is, the better the imaging. Informally experimenting with this myself, and I have observed spacial differences with a fraction of decibel, and fraction of a millisecond, mismatches between stereo channels. Now both speakers in a stereo pair may both measure similarly well, but are they similar enough to not have a negative effect on imaging? Would you even know if only one of the speakers was measured? I can't say that's the answer, but it seems logical to me that it would help in the right direction.

 

[NTK]

I was very fascinated by the several recent threads related to the topic of 3D sound perception. So I did some internet searches for more information. There is a significant body of research in this area, and here are links to two papers I found particularly helpful to me. They gave some very good background information.

The first one is "Fundamentals of Binaural Technology" by Møller.
https://vbn.aau.dk/files/54564101/Moller_1992.pdf
Below is the very first paragraph (after the abstract).

The idea behind the binaural recording technique is as follows: The input to the hearing consists of two signals: sound pressures at each of the eardrums. If these are recorded in the ears of a listener and reproduced exactly as they were, then the complete auditive experience is assumed to be reproduced, including timbre and spatial aspects. The term binaural recording refers to the fact that the two inputs to the hearing are reproduced correctly.

Thus, the basic premise is that if one can accurately reproduce the sound pressure at both ear drums, the entire listening experience, including spatial, can be reproduced. (Note: Other hearing pathways such as bone conduction is ignored.) Møller then went onto describing how to determine the head related transfer function (HRTF).

The second paper "Systems for Virtual Sound Imaging" by Nelson et al gave a brief summary of a model of how humans detect sound source location. It then went on to discuss the strategies for binaural reproduction. Inevitably, when a person starts looking into the research on "3D" sound reproduction, he/she will quickly run into statements about the fundamental limitations of stereo sound reproduction.
https://eprints.soton.ac.uk/337590/1/IUCS2011_fullpaper_Phil_comp.pdf

 

[hardisj]

I made this post in last month on my Facebook group page about plans for my measurements and it fits right in with the post above...


One of the things I want to do for my speaker reviews is to record the stereo sound like you are actually *in* my room or car, hearing it first hand. I have binaural mics but they aren't good for higher SPL recordings and the setup to make sure I get repeatable measurements is a pain. Repeatable is key. So this is what I‘ve settled on:

Zoom H3-VR

https://amzn.to/309YcZO

I will place it at the seated position and use its built-in binaural processing to record the sound system. The audio will be published to a link and most likely my YouTube page (albeit, with high compression). The idea is that, over time, I will have amassed enough of these “virtual system recordings” that people can compare characteristics of the speakers. It will not replace a speaker in your room but it should give a reasonable baseline to draw some conclusions about the performance; especially in regards to interaction with the room when it comes to speaker directivity (I.e., a highly directional waveguide vs an omnidirectional speaker). Additionally, it’ll be an easy way to discuss things to listen for in songs as well as compare differences in DSP settings (think: room correction vs without).

And, of course, all efforts will be made to match the signal as best I can do each recording is apples to apples. Details will follow once I get to the point they need to be. I’m not there yet.

The listener will have to listen back with headphones to get the desired effect. But ... it's pretty dang awesome.

Here’s an example. The testing was done in my home theater. (Just for the record, I'm using a JBL Pro 15" midrange and a 2" compression driver on a huge freaking horn for my left/center/right.)

https://www.dropbox.com/s/gb2trjj2xo50tug/Test 2.flac?dl=0

 

[No. 5]

@hardisj if you haven't already done so, you may be interested to read up on what David Griesinger has written on the topic of binaural recording. I had emailed him recently for his thoughts on using them for room acoustic investigation and loudspeaker comparisons, his reply was to the effect of "that is why he uses it."

 

[RayDunzl]

Were you to record what is presented to the eardrum...

Wouldn't you have to back some of that out, in order to play it from speakers, and have the ear again impinge upon the waves on their way to the tympanic membrane?

Measurement Apparatus and Modelling Techniques of Ear Canal Acoustics - https://www.researchgate.net/profil...delling-Techniques-of-Ear-Canal-Acoustics.pdf

(download button at right)

 

[hardisj]

I’ve been using binaural mics since about 2010 for RTA purposes in my car. It’s way more revealing in that environment than joke because HRTF is a big factor in the car since you’re sitting off-center and therefore the left speakers are around 10-20 degrees off-axis whereas the right side speakers are as much as 40-50 degrees off-axis. It’s very interesting how standard EQ operating procedure - matching left and right side response with an average of multiple single-mic measurements- in this regard doesn’t necessarily work like it does in a home where the listener is symmetrically centered between the two sides.

 

[hardisj]

Were you to record what is presented to the eardrum...

Wouldn't you have to back some of that out, in order to play it from speakers, and have the ear again impinge upon the waves on their way to the tympanic membrane?

Measurement Apparatus and Modelling Techniques of Ear Canal Acoustics - https://www.researchgate.net/profil...delling-Techniques-of-Ear-Canal-Acoustics.pdf

Generally speaking, if you use binaural mics you also need to use headphones to listen to the playback.