I have another. This is an in-box measurement, absolute calibrated (accurate to +/- 0.4dB) however -10dB has been subtracted to show the 1 metre equivalent SPL reading value. The fundamental has 1/24 octave smoothing (fixed) and the harmonics are based on a log chirp aka exponential sine sweep:
The legend marks the cursor position.
The last study that I looked at into ultrasonic detection is some time ago. The method is one where the experimenters allowed participants ie. listeners to perform ABX tests, to see if people could reliably tell the difference between high-rez audio, and audio that had a low pass filter at 22KHz, and then at 11KHz inserted.
However, they also attached electrodes to the subjects skull, to measure electrical changes in the electroencephalogram (EEG). The music was played through speakers with high frequency response, not earphones, to allow for detection of other possible modalities of ultrasonic perception eg. head related transfer functions, tactile response. This was in case it may be detected by any other means, in other parts of the brain.
The study is almost 5 years old now (Dec 2020) which in my field of training is starting to get old and considered “almost out of date”
Reference:
https://pmc.ncbi.nlm.nih.gov/articles/PMC7730382
The legend marks the cursor position.
The last study that I looked at into ultrasonic detection is some time ago. The method is one where the experimenters allowed participants ie. listeners to perform ABX tests, to see if people could reliably tell the difference between high-rez audio, and audio that had a low pass filter at 22KHz, and then at 11KHz inserted.
However, they also attached electrodes to the subjects skull, to measure electrical changes in the electroencephalogram (EEG). The music was played through speakers with high frequency response, not earphones, to allow for detection of other possible modalities of ultrasonic perception eg. head related transfer functions, tactile response. This was in case it may be detected by any other means, in other parts of the brain.
The study is almost 5 years old now (Dec 2020) which in my field of training is starting to get old and considered “almost out of date”
Reference:
https://pmc.ncbi.nlm.nih.gov/articles/PMC7730382
Last edited:
@tktran303 I read the article diagonally and it appears one aspect of its summary is what i learned some 45years ago from earlier mentioned meeting with Michelle Reverchon and his engineers. That just the tone is not audible but when part of a audible transient its contribution is audible. Yet how our (total) sensory system detects and notices it is not yet fully understood. ( i understand my references a waay out of date according to your study standards ;-))
Anyhow even while the breakup is well above 20kHz i will place an rlc to dampen its output as it is energy i am not interested in ;-)
Anyhow even while the breakup is well above 20kHz i will place an rlc to dampen its output as it is energy i am not interested in ;-)
Why not performing some objective AB and ABX tests yourself?
Just a reminder that there's no actual data on 99% of recordings above 20kHz.
Sometimes even above (roughly) 15kHz.
Depending on the hardware, many DACs have quite a steep LP around 22-30kHz.
So the only thing that could be left, is some additional very minor harmonics.
Just a reminder that there's no actual data on 99% of recordings above 20kHz.
Sometimes even above (roughly) 15kHz.
Depending on the hardware, many DACs have quite a steep LP around 22-30kHz.
So the only thing that could be left, is some additional very minor harmonics.
Hi all,
it's only in German language available, but DeepL may help, some builds with the new Purify tweeter here
https://spendormania.blog/2025/06/29/weltklasse-diy-im-sauerland/
Have more fun, Stefano
it's only in German language available, but DeepL may help, some builds with the new Purify tweeter here
https://spendormania.blog/2025/06/29/weltklasse-diy-im-sauerland/
Have more fun, Stefano
My GAYA2P is electronically and filterwise the same. I use 2 x 8" sealed, 4" sealed and preparing upgrade to ptt1.3.
I wished i could visit Markus'ses place on June 21, and also meet Uli in person.
I wished i could visit Markus'ses place on June 21, and also meet Uli in person.
Perhaps lateron, bit complex but i have the hang loose convolver to instant switch between filters, but the on off switch of the RLC is the difficulty.
For the Midrange i did do a test, be it by prolonged listening to A and then to B and repeat. Again here with midtange filters designed for each case, as the passive filter introduces some change in time domain.
My conclusion: there was a difference , especially on voices in choirs (onset of notes/words) of and sound space. But not so prominent, and if done blind i doubt it it will be a perfect score.
Never the less it is now standard in place.
May be if done mono on one 1 box, and instant blind switching, may be a better score, not high on my checklist. I am more focused on clarity and soundstage.
So here is a Two Tone FFT plot at 4 volts rms input at both frequencies. So the only thing that could be left, is some additional very minor harmonics?
Note the IM Distortion sidebands on either side of the 17k test tone.
Thanks DT
Here here ! In my less-than-humble, probably-biased opinion, military applications are the most challenging of any engineering project. I spent my whole career on military aircraft.I also doubt that humans can hear it, or at least I doubt that I can hear it ... I did an experiment a few years ago with an SB26CDC, which as a breakup at about 24k. I tried notching it out with DSP, and suppressing it with an inline passive notch filter.
I could not hear a difference with 24 bit/96k recordings. So I am skeptical that a resonance at 34k has any significance...
I also doubt that humans can hear it, or at least I doubt that I can hear it ... I did an experiment a few years ago with an SB26CDC, which as a breakup at about 24k. I tried notching it out with DSP, and suppressing it with an inline passive notch filter.
I could not hear a difference with 24 bit/96k recordings. So I am skeptical that a resonance at 34k has any significance...
pity this big pond is between us.I too doubt wrt to the tweeter it is audible in A-B.
While I have no clue about other materials, I seriously doubt it will be beryllium. The main manufacturer of the beryllium domes shut down its production. It's the main reason as far as I know why suddenly a lot of tweeters have been discontinued and these carbon composite tweeters have become suddenly really popular (i.e. textreme series from SB).
Such materials are dirt cheap and easy to apply in series production, plus very consistent. A driver manufacturers dream. Audiomania plus -philia do all the rest.
I agree, but for some reason we only see the TPCD tweeters recently.
Creating a new design is not free either. Hence I figured it was due to the stop of beryllium manufacturing.
I don't have insight in OEM pricing or whatsoever. However as far as I know there is only Oxeon that manufactures them and there is no competition. I feel like they can charge whatever they want (provided it is competitive with other solutions). It may be very cheap to manufacture, but was time consuming/expensive to develop. Same could be argued for new purifi tweeter. I doubt the cost is in manufacturing given the amount of cheap WG tweeters. I feel like the real cost was in the immense amount of R&D effort that went into finding the proper shape.
On another note lars, are there still plans for rearmounted variants of the tweeter (i.e. no visible screws)?
On another note lars, are there still plans for rearmounted variants of the tweeter (i.e. no visible screws)?
@lrisbo beat me to the punch... I cannot imagine that a thin ply carbon diaphragm is in the same price range as a polypropylene, aluminum, or paper diaphragm.