Really enlightening talk, thanks Dan. Very interesting and well explained how harmonic distortion can give the impression of detail. Explains why a particular brand whose headphones have very high distortion claim that they are the most detailed.
Glad you enjoyed it!
I've found that many people think of distortion as being hard-clipping in amps but it's obviously a more complex subject than that as transducer distortion occurs more gradually and tends to be dominated by lower-order distortion until it gets severe, whereas hard clipping in an amp is more abrupt and contains more high order harmonics at higher amplitude.
I thought a few analogies might be helpful to folk...
It was a good presentation, and the example of how distortion can color the sound was fascinating.
If anything, I think the presentation was detail-shy and not technical enough and left me with a lot of questions! How can a headphone create a 3D soundstage when the wavefront always hits the pinnae from the same angle? Also, apart from tuning the group delay, impulse response, and distortion, what are the advantages of non-DSP frequency response tuning? How does individual-specific HRTF computation fit in to statistical average approaches of finding a target preference curve like Harman 2018? How much do interpersonal anatomical variations stymie research on finding preference curves, especially considering the limited diversity in the subjects (e.g. differences between female and male subjects, or even interracial differences)? Finally, in a world of DSP and HRTF/otoacoustic emissions measurements/canal gain measurements, what can be gained from bypassing the pinnae in headphone design (i.e. IEMs)? what is lost?
The Harman research did note some gender related preferences. The Harman curve was derived as an aggregate of responses and so it represents a "typical" perspective, as well as both "amateur" and "pro" listeners, but it should be clear that it's kind of like the Pirate's Code, more of a guideline, really, and that there's an envelope of "reasonableness" that can surround variations to the curves...
The thing to remember about DSP is that it's used most typically for frequency response corrections. But this is limited in effect because if you have a standing wave null, you can't EQ it away. Things like 3D soundstage require your brain to receive crosstalk signals from the opposing ear, it's the absence of this that causes the "in your head" experience. I always thought this was a great little explanation:
"The pinna is the flap of skin surrounding our ears. Reflected sound off the pinna combines with the direct sound into the ear to create high
frequency comb-filtering effects (typically above 6
kHz). These effects change as a function of angle of arrival, so that each angle of arrival has a unique sound quality. Our brain uses this quality as one of the ways to localize sound at each ear individually. The effect seems most persuasive in the vertical realm, so it is reasonable to hypothesize that we localize horizontally mostly by time difference while in the vertical axis the pinna effect is used more."
So in essence what we have done with some of our AMTS work is address the vertical positioning of the image, where we've tried to engineer it so the soundstage is "to the front" rather than "15 degrees up." But side to side soundstage is handled with cross-feed or sometimes DSP, though to date I've never heard a DSP implementation that clicked for me. I know research in that is ongoing, I'm sure more on this will evolve.
Yes it was... and what Dan said was also true .. some member at ASR drew that FR chart and I am a member here
.
Dan, as of today, has my permission to use it in his presentation.
I really like his talk and honesty and matches with most of my views.
The plot was made by playing around with a graphic equalizer and matching what I heard with terminology used on various audio(phile and phool) forums.
Most terminology in the plot is widely accepted in the audio consumer world.
Yeah! I knew I saw the originator somewhere on this forum. Nice work! There are a number of acoustic engineering sources with similar graphs but I thought your use of terminology was very spot-on.
For me, the talk put the data into perspective from a designer’s point of view. It can help a lot to have an expert put the pieces together in a coherent way.
One question I had, if the membrane/diaphragm is under tension, does the quality of the tensioning tend to deteriorate over time, either by stretching or the material, or loosening of its moorings?
Depends on your materials and process. We have put a ton of engineering into stability over time.
