Chrisb,
I agree with you 100%. If you can hear it, it IS measurable, but we may not know how. I don't know how to measure "detail", but I hear the difference between a DX3 and an FE207E, even with the Lowther hump EQ'd out. I have no clue how to measure EnABL. Either or both could be beyond my equipment.
I do remain steadfast in my belief that these things CAN be measured. After all our ears are just measurement instruments.
Bob
I agree with you 100%. If you can hear it, it IS measurable, but we may not know how. I don't know how to measure "detail", but I hear the difference between a DX3 and an FE207E, even with the Lowther hump EQ'd out. I have no clue how to measure EnABL. Either or both could be beyond my equipment.
I do remain steadfast in my belief that these things CAN be measured. After all our ears are just measurement instruments.
Bob
I reckon that most of us would agree on that score; I know Mark, Dave & I certainly do. If you can hear it, then somewhere along the line, some clever bloke in a white coat is going to figure out how to measure it or explain why it has x effect. Actually, I suspect the technology already exists; it's figuring out how to apply it that's the issue. Going with a gut feeling, I suspect a set of LDV comparisons might show something, but that's just a guess. Until then though, & probably in conjunction with it, a set of blind trials with a reasonable sample group would at least be a useful step forward in helping to get a handle on general trends.
Last edited:
I'd tend to think of the ear as a microphone(perhaps compromised at best), "just" turning compressions/rarefactions of air into electrical signals; the real work is done inside the brain - and guess how long away we are from understanding how that works
What is LDV?
Laservibrometry might have something to say... maybe this one(http://www.microflown.com/index.htm) -- i read in some thread that some Japanese reserchers had used this last to detect HOMS
dave
Same thing I think (http://en.wikipedia.org/wiki/Laser_Doppler_Vibrometer)... otherwise, velocimetry (as in wind tunnel measurement).
Hi guys,
Excellent posts, good reading. There is technology available that can measure the performance of components and complete drivers to a more advanced level. I had the pleasure of spending some time with Dr. Wolfgang Klippel here is Asia. I’m now working with Matsubara San, former lead designer at Fostex. We are using measurement techniques, derived mostly from aerospace, to give us extended insights into the complex behaviour of drivers. Sadly, much of this technology is under very tight wraps for commercial reasons.
I’m regularly cooking up new designs, making samples and measuring their emittance characteristics. But then I hit a wall. There’s little substantial applied research that tells driver makers how human’s will react to their driver developments. A classic example is my recent work on paper cones, the EL70 being the first born. The only way I could gain any predictive insight into the effectiveness of my work was to pass samples to lead Diyers and use the feedback to adjust the final production.
Many makers rely on past traditions in both design and manufacture, a worthy bank of knowledge but not directly related to extensive research into the human perception and their behavioural responses to transducer technology. It partly explains why most driver products have changed little over the years.
What we need is a kind benevolent fella (Bill Gates?) to put a large sum of cash into the hands of a suitably qualified research institution, gather up lots a listeners (blind trials), work on analysing the direct relationships between humans and transducers. Oh bliss, another dream of mine!
I agree with Bob Brines and Scott, things can be measured, people also. We need to measure both and work out the causal-effect relationships.
Cheers
Mark.
Excellent posts, good reading. There is technology available that can measure the performance of components and complete drivers to a more advanced level. I had the pleasure of spending some time with Dr. Wolfgang Klippel here is Asia. I’m now working with Matsubara San, former lead designer at Fostex. We are using measurement techniques, derived mostly from aerospace, to give us extended insights into the complex behaviour of drivers. Sadly, much of this technology is under very tight wraps for commercial reasons.
I’m regularly cooking up new designs, making samples and measuring their emittance characteristics. But then I hit a wall. There’s little substantial applied research that tells driver makers how human’s will react to their driver developments. A classic example is my recent work on paper cones, the EL70 being the first born. The only way I could gain any predictive insight into the effectiveness of my work was to pass samples to lead Diyers and use the feedback to adjust the final production.
Many makers rely on past traditions in both design and manufacture, a worthy bank of knowledge but not directly related to extensive research into the human perception and their behavioural responses to transducer technology. It partly explains why most driver products have changed little over the years.
What we need is a kind benevolent fella (Bill Gates?) to put a large sum of cash into the hands of a suitably qualified research institution, gather up lots a listeners (blind trials), work on analysing the direct relationships between humans and transducers. Oh bliss, another dream of mine!
I agree with Bob Brines and Scott, things can be measured, people also. We need to measure both and work out the causal-effect relationships.
Cheers
Mark.
now that's a very interesting cross pollination
well, I just got this e-mail from the Nigerian lottery .....
As someone who both works in aerospace and depends on good human performance modeling, I fear that this is an unattainable goal. The best we can do is quantify the variance in human perception and design to a least common denominator that captures suitable numbers for efficiencies of scale... otherwise, Bill Gates might be the only one that can afford the drivers! Then again, if the majority of perceptual variance can be accounted for in physical characteristics (head shape, ear folds, basic hearing test, etc.) rather than within the translation of the aural response in the 'inner' meat servo, than a fitting may be all that is required to tailor the technology much like canal phones are fitted.
As someone who's been involved in aerospace and human performance modelling, you much surprise me with your comments.
Blind trailing as a form of research to establish outcomes is well understood and implemented on various degrees of scale.
Agreed, larger sample sizes will generate more reliable data, (hence my Bill Gates reference). That said, small scale research can still be considered indicative provided a accredited methodology is employed and the sample selected is sufficiently representative.
The blind trail would be based on attitudinal analysis. Trailing of this type is done by a huge number of enterprises (big and small) as part of their product/service marketing efforts.
Blind trailing as a form of research to establish outcomes is well understood and implemented on various degrees of scale.
Agreed, larger sample sizes will generate more reliable data, (hence my Bill Gates reference). That said, small scale research can still be considered indicative provided a accredited methodology is employed and the sample selected is sufficiently representative.
The blind trail would be based on attitudinal analysis. Trailing of this type is done by a huge number of enterprises (big and small) as part of their product/service marketing efforts.
A start of the groundwork for some of this stuff has been done. Floyd Toole's new book covers a lot of interesting material. Required reading.
dave
dave
I am not implying that such methods cannot produce a statistically valid model of human performance, only that said model would necessarily include natural variance in human response/perception. This variance, if greater than the design/manufacturing 'precision' available, would mean that a specific design would only be for a very small population (thus losing economies of scale). I have no idea how close the existing models are, and maybe there is great improvement to be made; but every model of human performance I've seen has significant variance.
- Status
- Not open for further replies.