Old yet still interesting James Boyk piece: the upper octave extension of instruments
- Thread starter MylesBAstor
- Start date
I hesitate to chime in on this thread, but I will anyway. 
Also, in the big picture, does it really matter if some people can barely perceive certain very high frequencies? Is that really needed for a satisfying musical experience? It seems to me that a great sounding piece of recorded music will still sound great through a low-pass filter set to 15 KHz or even 10 KHz. I'm not arguing for limited bandwidth! Just questioning out loud how important some of this minutiae really is.
--Ethan
My understanding is the Oohashi test used a single tweeter whose IM distortion created alias tones down in the audible range. I thought I read somewhere that subsequent tests used separate tweeters, and then nobody could perceive the "ultrasonic" content. Can anyone confirm the details of this debunking?
Also, in the big picture, does it really matter if some people can barely perceive certain very high frequencies? Is that really needed for a satisfying musical experience? It seems to me that a great sounding piece of recorded music will still sound great through a low-pass filter set to 15 KHz or even 10 KHz. I'm not arguing for limited bandwidth! Just questioning out loud how important some of this minutiae really is.
--Ethan
Steve williams
Site Founder, Site Co-Owner, Administrator
Ethan
IIRC, what Myles was trying to show, was the extension of many of these instruments not that we couldn't hear them
IIRC, what Myles was trying to show, was the extension of many of these instruments not that we couldn't hear them
That also is my understanding, that the loudspeaker intermodulation distortion resulted in sub-harmonics in the audible range, and it goes to show that any test must have proper controls in place or the results of the test are, well...I hesitate to chime in on this thread, but I will anyway.
My understanding is the Oohashi test used a single tweeter whose IM distortion created alias tones down in the audible range. I thought I read somewhere that subsequent tests used separate tweeters, and then nobody could perceive the "ultrasonic" content. Can anyone confirm the details of this debunking?
As to the Boyk article itself, while it was known that the various instruments produced ultrasonic content, Boyk's illustrations/graphs are indeed wonderful.
Greg, I think my wife is a low pass filter. Since she has better hearing than me she hears some of those high freqs that I do not and as a consequence I get: "Honey, can you turn it down a little?"
I understand, and I was just asking about the Oohashi "debunking" mentioned.
As a side story, many years ago (late 1970s?) I had a friend who was an engineer at Hewlett-Packard. So he had access to all sorts of cool test gear. One time he brought a spectrum analyzer to my recording studio, which we connected to a condenser microphone. I jiggled a set of keys directly in front of the microphone while we watched the analyzer display. I was surprised to see a fairly flat line all the way out to the 50 KHz limit of the analyzer! So yeah, there's lot of ultrasonic stuff prevalant in nature. Doesn't mean we can hear it though!
--Ethan
That's actually an interesting comment and some of the reasons for the differences between female and male hearing has been discussed! Perhaps that even worthy of further exploration.
But I tell you what: I wouldn't be w/o the input of my ex-wife or my present girlfriend when listening to equipment. In fact, I've been trying to educate Heidi, who's a trained soprano, about the joys of high-end audio, and get her feedback about the sound of different components. OK, she is slightly biased as most of my gear is tubes
On TV, the commercial goes "Chicks dig the long ball." Do chicks dig tubes as well?
Yeah they love it when the room hits 100^oC
hearing and sensing arent the same
we may not distinctly hear a sound or supersonic vibration but we have the ability too sense these phenomena. if i am not mistaken, some of the higher freqs are detected by hair/follicles (perhaps why juliian hirsch was the reviewer he was).
also, the completeness of a waveform may help describe the accuracy of its reproduction.
well, enough of my babbling.
we may not distinctly hear a sound or supersonic vibration but we have the ability too sense these phenomena. if i am not mistaken, some of the higher freqs are detected by hair/follicles (perhaps why juliian hirsch was the reviewer he was).
also, the completeness of a waveform may help describe the accuracy of its reproduction.
well, enough of my babbling.
Stereophile measurements of Class D amplifiers in their reviews
I believe that Atkinson is now putting in a high frequency filter - is it at 50kHz? - to eliminate ultrasonic noise from the review of Class-D amplifiers. Thus noise and distortion data cannot be fully depicted or elucidated. His early reviews did not. I recall he documented the characteristics of the amplifier, and perhaps his test equipment - with and without the filter. With the filter many AUTs look great. Without, many did not test well even interfering with 10kHZ square waves and so on. (Even I can hear "hash" at 10kHz.) He got some stick over this approach from the commentators. Indeed, run full band the output of the Class-D AUT resembles the products of a clipping conventional amplifier.
At first blush it looks like the high level of ultrasonic noise characteristic of the Class-D amps might inter-modulate, mask, and otherwise interfere with the instrumental overtones at very high frequencies such as documented in the Boyk paper. Indeed many speakers (Sound Labs, some Wilsons, etc.) can recreate the ultrasonic range. Going further, perhaps this explains that many listeners do not like Class D amps with these speakers compared to a well-designed traditional A/B solid state or tube amp - even from the same manufacturer. Knowingly or unknowingly they are hearing "hash" on ordinary recordings that is exacerbated with the kinds of recordings Boyk describes.
Is the ultrasonic noise from these class D amps interfering with the audible spectrum in the amplifier hooked up to a test load? Repeat question using "speaker" rather than test load. Since the speaker, cable, and amp is a system, what effects do we have going on here to produce instability and interaction with the impedance angle at these high frequencies as the speaker returns part of its electrics to the amp? Are new speaker cable designs called for with the amp/speaker combinations to "damp" or roll-off the noise and distortion products of these amps in situ? Are there any class-D amplifiers that, when run full-band, do not produce these artifacts?
Please note that I am raising questions for discussion. I have no dog in the fight at this time and hope to hear from my new forum friends and experts. I do applaud Stereophile for doing measurements of reviews equipment. It comprises a database of good information, on the whole, that I would not want to be without. But I, for one, would like to see the amp tests - including the results into the "simulated speaker load" run with and without the filter.
I believe that Atkinson is now putting in a high frequency filter - is it at 50kHz? - to eliminate ultrasonic noise from the review of Class-D amplifiers. Thus noise and distortion data cannot be fully depicted or elucidated. His early reviews did not. I recall he documented the characteristics of the amplifier, and perhaps his test equipment - with and without the filter. With the filter many AUTs look great. Without, many did not test well even interfering with 10kHZ square waves and so on. (Even I can hear "hash" at 10kHz.) He got some stick over this approach from the commentators. Indeed, run full band the output of the Class-D AUT resembles the products of a clipping conventional amplifier.
At first blush it looks like the high level of ultrasonic noise characteristic of the Class-D amps might inter-modulate, mask, and otherwise interfere with the instrumental overtones at very high frequencies such as documented in the Boyk paper. Indeed many speakers (Sound Labs, some Wilsons, etc.) can recreate the ultrasonic range. Going further, perhaps this explains that many listeners do not like Class D amps with these speakers compared to a well-designed traditional A/B solid state or tube amp - even from the same manufacturer. Knowingly or unknowingly they are hearing "hash" on ordinary recordings that is exacerbated with the kinds of recordings Boyk describes.
Is the ultrasonic noise from these class D amps interfering with the audible spectrum in the amplifier hooked up to a test load? Repeat question using "speaker" rather than test load. Since the speaker, cable, and amp is a system, what effects do we have going on here to produce instability and interaction with the impedance angle at these high frequencies as the speaker returns part of its electrics to the amp? Are new speaker cable designs called for with the amp/speaker combinations to "damp" or roll-off the noise and distortion products of these amps in situ? Are there any class-D amplifiers that, when run full-band, do not produce these artifacts?
Please note that I am raising questions for discussion. I have no dog in the fight at this time and hope to hear from my new forum friends and experts. I do applaud Stereophile for doing measurements of reviews equipment. It comprises a database of good information, on the whole, that I would not want to be without. But I, for one, would like to see the amp tests - including the results into the "simulated speaker load" run with and without the filter.
Last edited:
It seems that some hi-fi equipment (not just Class D amp) don't handle ultrasonic frequencies well. They generate intermodulation distortion (IMD) who's product is in the audio band.
The opposite
Great point. I vaguely remember the Audio Critic review of the Bedini 25/25 Class A amp in the mid-80s. When they measured it with a resistor, it became unstable at high frequencies. When they hooked it up to a speaker (Quad 57?) it measured and sounded just fine. No more instability.
What testing equipment is required to measure the amplifier characteristics hooked up to an actual speaker?
Great point. I vaguely remember the Audio Critic review of the Bedini 25/25 Class A amp in the mid-80s. When they measured it with a resistor, it became unstable at high frequencies. When they hooked it up to a speaker (Quad 57?) it measured and sounded just fine. No more instability.
What testing equipment is required to measure the amplifier characteristics hooked up to an actual speaker?
Now that's strange. Multi-driver loudspeakers are a lot more challenging than resistor dummy loads. But then a Quad 57 is not your typical loudspeaker.
I'll have to look thru my old Audio Critic mags.
Same test equipment. Stereophile magazine has a dummy load that as the same curve as a speaker.
About 15 years ago, Cyril Bateman showed the some reasonable amplifiers, thru reasonable speaker cables, into reasonable loudspeakers may ring or even oscillate in the 5MHz range.
I'll have to look thru my old Audio Critic mags.
Same test equipment. Stereophile magazine has a dummy load that as the same curve as a speaker.
About 15 years ago, Cyril Bateman showed the some reasonable amplifiers, thru reasonable speaker cables, into reasonable loudspeakers may ring or even oscillate in the 5MHz range.
Last edited:
Thanks for the reply.
I believe the SPhile load is a "typical dynamic speaker." (Probably an LS3 5/A ?) I'm more interested in Wilsons, hybrid and full range (CLX and SL) electrostatics. I know many of the last are said to run 100 Ohms at 30 Hz and <1 Ohm at 20kHz or 50kHz. What kind of amp can drive that? Or do you need two amps, one for high voltage (bass) and one for high current, and thus bi-amp the speaker? How many (watts or volts) do you need to drive a 100 Ohm - 30 load to the same level as a high current load invariant-amplifier can drive 1-3 Ohms in the high frequencies?
I'd like to see the Bateman article. Was this in HFN&RR?
What do you mean by "same test equipment?" I'd be looking for either something software based and inexpensive, say for an iPAD or PC, or old HP or TTronics test equipment that I could buy for a song. I don't need to measure jitter or anything digital, I think.
As for measurements, there might be 100 different meaningful tests so you have to decide which one you want to do. Testing could be a long thread in itself.
You could start at the top with an Audio Precision $25000 instrument or for less money a good sound card and maybe a battery powered RMS AC volt meter. A couple of O-scopes would be nice also.
You could start at the top with an Audio Precision $25000 instrument or for less money a good sound card and maybe a battery powered RMS AC volt meter. A couple of O-scopes would be nice also.
Bateman links (down the page).
http://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=6&t=153&hilit=bateman
*************************
https://linearaudio.nl/cyril-batemans-capacitor-sound-articles
http://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=6&t=153&hilit=bateman
*************************
https://linearaudio.nl/cyril-batemans-capacitor-sound-articles
Similar threads
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |