Hi Folks,
I was working on a project and really needed to see a copy of the original Thiele-Small papers on modeling dynamic speaker drivers from 1961 (Thiele) and 1971 (Small). These were not easy to find in a free format as they are typically cited as J. Acoust. Engr. Soc. papers. But I eventually found them. Very nicely written and amazing how they are still used to this day. A lot of similarity with the solution to the classic forced damped harmonic oscillator equation of motion.
Anyhow, it gave me an idea that once I find stuff, and if you guys find anything of use, let's post here for others to share.
PDF's:
Here is the one by Thiele from 1961:
https://diyaudioprojects.com/Technical/Papers/Loudspeakers-in-Vented-Boxes-Part-I.pdf
Here is where I found a scanned copy of the classic 1971 paper by Small:
https://diyaudioprojects.com/Technical/Papers/Direct-Radiator-Loudspeaker-System-Analysis.pdf
A nice treatment of the equations from above by engineers at ScanSpeak and put into MathCad format:
https://www.scan-speak.dk/datasheet/tech/Scan-Speak_Technote01_TS-measurement.pdf
How to use LTspice to sim T-S:
https://www.micka.de/en/download/spice-tsp_en.pdf
Threads / Websites:
There is a thread in Multiway on using LTSpice to model Thiele-Small parameters for speakers:
https://www.diyaudio.com/community/threads/thiele-small-reloaded-with-ltspiceiv.166467/
Videos:
Nice tutorial on analyzing stability of feedback circuits using LTSpice:
How to measure open loop gain in LTSpice:
Perhaps the BEST lecture I have ever seen on solving damped harmonic oscillators - by Prof Walter Lewin (MIT):
Prof Steve Brunton (U of Wash) also gives great explanations on damped harmonic oscillators (4 ways to solve):
I was working on a project and really needed to see a copy of the original Thiele-Small papers on modeling dynamic speaker drivers from 1961 (Thiele) and 1971 (Small). These were not easy to find in a free format as they are typically cited as J. Acoust. Engr. Soc. papers. But I eventually found them. Very nicely written and amazing how they are still used to this day. A lot of similarity with the solution to the classic forced damped harmonic oscillator equation of motion.
Anyhow, it gave me an idea that once I find stuff, and if you guys find anything of use, let's post here for others to share.
PDF's:
Here is the one by Thiele from 1961:
https://diyaudioprojects.com/Technical/Papers/Loudspeakers-in-Vented-Boxes-Part-I.pdf
Here is where I found a scanned copy of the classic 1971 paper by Small:
https://diyaudioprojects.com/Technical/Papers/Direct-Radiator-Loudspeaker-System-Analysis.pdf
A nice treatment of the equations from above by engineers at ScanSpeak and put into MathCad format:
https://www.scan-speak.dk/datasheet/tech/Scan-Speak_Technote01_TS-measurement.pdf
How to use LTspice to sim T-S:
https://www.micka.de/en/download/spice-tsp_en.pdf
Threads / Websites:
There is a thread in Multiway on using LTSpice to model Thiele-Small parameters for speakers:
https://www.diyaudio.com/community/threads/thiele-small-reloaded-with-ltspiceiv.166467/
Videos:
Nice tutorial on analyzing stability of feedback circuits using LTSpice:
How to measure open loop gain in LTSpice:
Perhaps the BEST lecture I have ever seen on solving damped harmonic oscillators - by Prof Walter Lewin (MIT):
Prof Steve Brunton (U of Wash) also gives great explanations on damped harmonic oscillators (4 ways to solve):
Last edited:
Here is a great article that describes how the human ear's pinna can discern phase shifts in the single microsecond range (giving us ability to locate direction with only one ear). Work funded by US military in 1964 and performed by Tufts Univ. This implies we hear amplitude to 20khz but phase to 300kHz. Hence, amps should be designed to be phase flat out to 300kHz for best stereo image and spatial soundstage.
Well, no wonder I've always been a phase-freak. Validates the ultrasonic or "super-tweeter" folks too. I wonder how that ability declines with age? Fortunately I can still hear a car comin up the road behind me.
Great read, thanks for posting it
I always thought our ears being slightly different in location on each side of our heads was what determined localization which might be a part of it, likely so it would seem but being able to determine location in mono is great to know about, love to keep learning, turned 72 yesterday, not stopping until my time in this body if finished.
Rick
I always thought our ears being slightly different in location on each side of our heads was what determined localization which might be a part of it, likely so it would seem but being able to determine location in mono is great to know about, love to keep learning, turned 72 yesterday, not stopping until my time in this body if finished.
Rick
Just posted these in a different thread. The early days of quarterwve resonators.
for reference (anecdote: scanned from the Blue Binder i got from KEF loaded up with aricles like these);
Bailey 1: http://p10hifi.net/planet10/TLS/downloads/Bailey_TLs_1.pdf
Bailet 2: http://p10hifi.net/planet10/TLS/downloads/Bailey_TLs_2.pdf
Letters: http://p10hifi.net/planet10/TLS/downloads/bailey_letters.pdf
Onley’s original 1937 article that predates Bailey: http://www.t-linespeakers.org/download/Onley-acoustic-labyrinth.pdf
dave
for reference (anecdote: scanned from the Blue Binder i got from KEF loaded up with aricles like these);
Bailey 1: http://p10hifi.net/planet10/TLS/downloads/Bailey_TLs_1.pdf
Bailet 2: http://p10hifi.net/planet10/TLS/downloads/Bailey_TLs_2.pdf
Letters: http://p10hifi.net/planet10/TLS/downloads/bailey_letters.pdf
Onley’s original 1937 article that predates Bailey: http://www.t-linespeakers.org/download/Onley-acoustic-labyrinth.pdf
dave
Yes, but phase flat to 300kHz. That is, no perturbations to the time accuracy down to 2usec or so.
Well a BW limit doesn't guarantee you will lose discrimination depending on what method you are using. For instance digital chips that use edge-triggering to increase BW. I would guess our hearing does all it can to be spectrum invariant since it has no control over which sounds will need to be localized.
It's important that this paper is dealing with localization from a single ear. This means it shouldn't be possible to reproduce this form of localization unless the speaker itself is moving around. Or by a feat of DSP virtually replacing your pinna with another one... So this falls in the realm of sound engineering of an effect that cannot be reproduced naturally by a speaker. And if your speakers don't "disappear" acoustically, maybe it is because they are aggressively triggering this kind of localization.
With that in mind, the reflection of the pinna can be analyzed like any other acoustic reflection. For example the paper shows a delay of something like 100us, which gives a 1/4 wavelength of 2.5KHz as the lowest peak or dip. That gives us a large portion of the audible spectrum as also affecting localization. Given most transducers can't play ultrasonics with any fidelity and most music doesn't contain it, I would say the ultrasonics should not be improved at the expense of the audible spectrum and my own experience indicates that there is a steep reward for getting 1KHz+ to be as clean as possible in my speakers. This would be respecting the possibility that the 2us sensitivity might also apply to say 3KHz sounds even if there is no ultrasonic content.
Given it needs to identify a sound and it's reflection, if it's only comparing the two then the spectrum itself might not matter (and an adaptible localization system would be spectrum invariant to some degree) until it falls below the threshold which can trigger localization, and that threshold seems to be very forgiving all considered. So maybe it's not so much how to trigger localization, but rather how to avoid confusing localization. We already know the largest factor in this regard is early reflections that fall within the same window as desired spatial cues. It would be very hard to get an amplifier to produce a similar effect unless it is grossly distorting. Or perhaps some form of quantization noise. So I suspect that a BW limitation is not usually the primary concern unless maybe there is ringing or overshoot.
It's important that this paper is dealing with localization from a single ear. This means it shouldn't be possible to reproduce this form of localization unless the speaker itself is moving around. Or by a feat of DSP virtually replacing your pinna with another one... So this falls in the realm of sound engineering of an effect that cannot be reproduced naturally by a speaker. And if your speakers don't "disappear" acoustically, maybe it is because they are aggressively triggering this kind of localization.
With that in mind, the reflection of the pinna can be analyzed like any other acoustic reflection. For example the paper shows a delay of something like 100us, which gives a 1/4 wavelength of 2.5KHz as the lowest peak or dip. That gives us a large portion of the audible spectrum as also affecting localization. Given most transducers can't play ultrasonics with any fidelity and most music doesn't contain it, I would say the ultrasonics should not be improved at the expense of the audible spectrum and my own experience indicates that there is a steep reward for getting 1KHz+ to be as clean as possible in my speakers. This would be respecting the possibility that the 2us sensitivity might also apply to say 3KHz sounds even if there is no ultrasonic content.
Given it needs to identify a sound and it's reflection, if it's only comparing the two then the spectrum itself might not matter (and an adaptible localization system would be spectrum invariant to some degree) until it falls below the threshold which can trigger localization, and that threshold seems to be very forgiving all considered. So maybe it's not so much how to trigger localization, but rather how to avoid confusing localization. We already know the largest factor in this regard is early reflections that fall within the same window as desired spatial cues. It would be very hard to get an amplifier to produce a similar effect unless it is grossly distorting. Or perhaps some form of quantization noise. So I suspect that a BW limitation is not usually the primary concern unless maybe there is ringing or overshoot.