Good ESL speakers can be brutally honest about your electronics. Don't shoot the messenger.
Oops: iPad self-typo. Should read 'a 2-2/3' adds nasality'.
This is what organists do all day long: mix harmonics, on the only instrument where you can do it systematically. Yet nobody ever asks them. Instead engineers run redundant perceptual experiments on apparently tone-deaf subjects, and devise incorrect rules like 'even harmonics consonant, odd harmonics dissonant', the n-squared/4 weighting, etc.
False. Many organs have 'synthetic' 32' stops which are really a 10-2/3 stop, i.e. the 3H of the missing 32' fundamental. It beats with the 16' tone to produce a 32' tone by, err, intermodulation.
There is “tube sound”, and then there is “realistic sound”.
I do remember being amazed at how lousy my friend’s Radio Shack system sounded compared to my 1960s sound lab surplus Scott 299 tube amp in 1983.
Would only hope that things would be better by now, and any comparison to tubes would be long past via better analysis.
Seems like the audio push towards better sound has slowed down, eh?
I do remember being amazed at how lousy my friend’s Radio Shack system sounded compared to my 1960s sound lab surplus Scott 299 tube amp in 1983.
Would only hope that things would be better by now, and any comparison to tubes would be long past via better analysis.
Seems like the audio push towards better sound has slowed down, eh?
Who knows exactly how it happens, but the detector is the ear itself. If not by intermodulation/demodulation, how else would those frequencies be generated?
How about a little science experiment. If you’re like most of us you can’t hear a thing above 14k or so anymore. Take two completely independent audio systems, and play an 18 kHz tone through one and a 19kHz tone through the other. Give ‘em a good 90dB - just shy of letting magic smoke out of the tweeters. See if you can hear a 1KHz tone.
How about a little science experiment. If you’re like most of us you can’t hear a thing above 14k or so anymore. Take two completely independent audio systems, and play an 18 kHz tone through one and a 19kHz tone through the other. Give ‘em a good 90dB - just shy of letting magic smoke out of the tweeters. See if you can hear a 1KHz tone.
Okay. Let's try something else. Take four frequencies, and apply a fixed phase shift to two of them. No new frequencies are created. What might that sound like? One fixed phase shift setting may sound like the volume is being turned up and down, while the other may sound like there is frequency modulation. The envelopes are certainly different, and so are the time domain waveforms.
Files are available for listening at: https://purifi-audio.com/2019/12/07/amfm/
To further complicate things, there are always four frequencies added together. How many distinct IMD products should one expect to hear in that case?
Files are available for listening at: https://purifi-audio.com/2019/12/07/amfm/
To further complicate things, there are always four frequencies added together. How many distinct IMD products should one expect to hear in that case?
Last edited:
Doesn't necessarily follow. That its to say, the observation that distortion varies with amplitude neither proves nor disproves the hypothesis that there is more than just distortion. It could be either way. In part it depends on how distortion is being defined. In Bruno Putzey's F-work essay, he seemed to be describing nonlinear distortion as some curvature in the transfer function of, say, an amplifier. If the curvature is fixed, then we already have long established EE models for small-signal and large-signal operation which result in different levels of observed signal distortion. Yet the nonlinear transfer function can be the same in both cases. So in that sense the fixed nonlinearity is all there is.
When thinking about why tube amps sound the way they do, I thought it was interesting to look at how tube circuits behave when pushed towards an extreme, yet not all the way to so extreme as hard clipping. One area where more pronounced tube circuit effects are studied is for instrument amplifiers. Of course for hi-fi use certain tendencies will be much smaller, but maybe not all the way to zero. One possible resource: https://www.amazon.com/Designing-Tube-Preamps-Guitar-Bass/dp/0956154522
Last edited:
We should also consider what hearing and psyche are doing. Hearing is first and foremost there to perceive differences. These can be interpreted, as a further step, e.g. as a "frequency mix" caused by an "instrument". Or, if differences are not in a context that can be interpreted as frequencies of an instrument, as e.g. noise or something else. Thus the evaluation of the sound of an amplifier is more complex than that of an instrument: (heard) differences can be caused by the instrument, by recording circumstances as well as recording technique, by playback circumstances as well as playback technique. And conceivably further;-)
So, at least for the Tripath amp, the trick is to find an amp today with the TA2020 chip.
https://www.parts-express.com/Lepai...-Fi-Audio-Mini-Amp-w-o-pow-310-301?quantity=1 This might not be so easy since Cirrus Logic bought Tripathi's patent and has since phased out the chip. Here's one such amp.
https://www.parts-express.com/Lepai...-Fi-Audio-Mini-Amp-w-o-pow-310-301?quantity=1 , though I'd prefer one with at least twice the power.
Meanwhile, could you recommend a few model Class A solid-state amps which you may have heard?
OK, good to know because I don't know how to simulate it, and it would be a handy thing to do. Can you demonstrate the filter shift with changing amplier impedance? What software are you using?
I'd like to know how much intermodulation the ear causes. Hearing isn't absolutely clean and distortion-free, but I think it's at way too low a level to explain this phenomenon. Here's a few quick quotes:
I dare say the Wikipedia article is more accurate here, or else Britannica's words "is used by the ear to create the fundamental" is more figurative than literal.