Jump to content

How Researchers Go About Making the Subjective Experience Objectively Meaningful


Recommended Posts

This white paper is a reference in a number of research papers presented at the recent Audio Engineering Society's headphone conference in Denmark, covered by Tyll Hertsens of Inner Fidelity. The approach is not restricted to headphones. The most interesting part, imo, is on "sensory measurement": "Tests where people use their senses to provide assessments are called sensory measurements. There are two types of sensory tests, the objective/perceptual measurement and the subjective/affective measurement."


Link to paper: http://assets.madebydelta.com/docs/share/Akustik/TECH_Document_Perceptual_characteristics_of_audio_UK.pdf



The "Sound Wheel", presently a work in progress, is central:




Here's the conclusion of the paper, which summarizes better than I can:


The technical data on their own do not provide information on the sound character of the audio

products. By supplementing the technical data with the results of listening tests, it is possible to get an

objective characteristic of the perceived sound. However, it requires that the listening tests are carried

out as structured blind tests with trained listeners in a listening room that does not colour the sound.

For such tests, among other things, a lexicon of well-defined attributes that the listeners are trained in

must be used. The sound wheel is such a collection of attributes, of which the timbral attributes have

been qualified in a number of tests. The aim of the sound wheel is to create a common language to

describe systems for audio reproduction.

An objective characteristic of the perceived sound can be used in product development, but it can also

be used as sensory product information, which the consumer will benefit from. On the basis of the

results of objective listening tests, the customer's preferences can be modelled and thus predicted.


Note that blind testing by a trained panel is part of the process.


If you take the time to read the paper, I'm interested in any responses.



Link to comment
Share on other sites

A messy subject with temptations to grab one of many hundreds of handles to hold onto. I don't know how workable it is, but their early results are encouraging and the goal admirable. It is something Harman has approached in a similar way and also a different way as they are more intent on describing measurable elements to match with preference.


As they more or less state, they are trying to develop a common language about speakers with more precision and one that allows you to get a better idea of how a speaker sounds before you hear it via description. I think of how that works among friends I have known a long time and spent lots of time with. It is a shared experience that is the basis of comparative descriptions. The actual language is influenced by our own areas of expertise, influence from publications and applied to common things we have together. It works pretty well for us, but our own group experience and language are fully beneficial only among our group and in much more limited ways otherwise.


More concrete examples of what I have in mind. A friend has a new speaker. One of the things he says is the way it handles voices is almost Quad like. I owned Quads for many years, a mutual friend had some all three of us have spent hours hearing Quads together. That is a pretty clear useful description. My friend then goes on to contrast other aspects not like Quads and perhaps mentions our old buddy so and so's monitor speakers for bass. Etc. etc. In a few minutes with a bit more elaboration from our collaborative experience I have a reasonable idea of what I will hear and even if it suits me. I might hear, "you know how you liked what speaker X did with the imaging, well this is much like that". Or I know you never liked how LS3/5a's imaged super precisely, but this speaker mixes that with a panel's open midrange and detail.


It appears to me this group is trying to develop that language that would be equally useful among all audiophiles. I just don't know if it will work extremely well without hearing some of their reference designs or a long experience of many speakers of different types.


Also, I don't like the color wheel. Too complex and the color itself could 'color' for lack of a better term perception at some level. Not sure right off of a better idea. I might try and develop a 3D plot like a waterfall plot with appropriate labels on the axis. Fewer terms, but then thinking is 3D might be bad too.


Still, all in all an interesting project. The big bugaboo is even the stuff communicated among friends, it is not blind and can be effected by many things. So blinding is needed yet among friends it would mean your experience then has no label to access the experience. I once used room correction to make a 2 way sealed box speaker sound as close as possible to some panel speakers. They were in the room with the tall slim box speaker just outside the panels. It managed to make a couple people think they were listening to panels until I said otherwise. Suddenly they noticed differences. I did the same thing once with box on left and panel on right. Close your eyes and you could hear it wasn't quite balanced. Open your eyes and it fooled one person into thinking they were listening to the panels. Blinding is essential to move forward with such a matter.

And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. 

Link to comment
Share on other sites

I agree with Denis that the colour wheel presented here is too complicated and the use of colour doesn't help either.

I have seen such wheels before and some were in my view much more effective designs...


But I think that if this works we'll finally have made user and critic reports/reviews useful; as it stands audio reviews are nearly always worthless for they are guided by taste and not observation-lead.

I have not read the article yet but I have worries regarding the methodology, which in my view should make the exclusive use of live unamplified music compelling in auditions (this of course implies that the listener will have attended live concerts).

Listening to pink-noise might help too, but this would require that the listener/critic be trained with a transparent system for the task to be effective.

Speakers would have to be setup in a way as to provide the most balanced in-room response possible (and not follow the user's taste in "soundstaging") and this of course requires measurements, but all rooms add their own colorations which makes it difficult to judge.

And what about partnering electronics?

Too many obstacles, in my view, to make this possible at home.



"Science draws the wave, poetry fills it with water" Teixeira de Pascoaes

Link to comment
Share on other sites

Harman has been doing this stuff for a while but apparently they use multiple seating spots (even though we all know there's only one that works) and I'm not sure they take much care at optimally positioning the speakers in the room.


They've even designed a listener training software:


Harman How to Listen



"Science draws the wave, poetry fills it with water" Teixeira de Pascoaes

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Create New...