Are you Euphonophile?

[Fitzcaraldo215]

I said nothing about the level of my own expertise in statistics. I do know a thing or two about it, about the hard physical sciences, math and also about Economics.

Economics is about the study of certain aspects and tendencies of individual human, group, institutional, government and world behavior involving the actions of individual people on up through the complexities of many simultaneous, interacting human economic "actors". Like all science, it is imperfect and incomplete in its understanding of the complex behavioral phenomena it seeks to address. Therefore its predictive models are only as good as the body of understanding currently available within the discipline.

That body of understanding, like all science, is ever expanding, but never does it encompass everything there is to know on the subject. And, especially dealing with humans and their groups, institutions, governments, etc., who do not always act prudently or rationally, specific cause/effect results or future outcomes may lack accurate economic predictability.

There are also unpredictable "externalities". A forecast of GDP, for example, may be far off due to unanticipated hot wars, trade wars, hurricanes, volcanoes, droughts, sun spots, strikes, etc. affecting economic activity such as farming and other commercial activity.

There are parallels to every other science in these regards, including audio science. Therefore, starting from known science with a top down model specified and derived even thoroughly and completely from that body of science can only be as comprehensive as the limits of current knowledge will allow, at best, even assuming perfect implementation of the model.

As Reagan said, "trust but verify". In audio, are instrument measurements derived from the same limited body of current scientific knowledge sufficient? Or, do we need to go further with controlled listening tests on samples of human subjects, messy as that is, in case we have missed something or introduced an inadvertent side effect?

 

[GGroch]

Audio is in my opinion supposed to be enjoyed first and foremost, that's why that old 1940s tube radio in the garage can rock me and I have fun with it.....Audio is a very personal hobby in the end.

I agree totally...and the reasons why we enjoy certain music, sounds, and gear are both personal and complex. I am old enough to have listened extensively to tube radios growing up. Interacting with and hearing tube radios brings back great memories and to me their sound is euphonic....but that is based not just on my ears, but also to my brain, history, and family, which try as I might I cannot disconnect.

No human can disconnect from their experiences, beliefs, and interests...they create our perspective. New studies have shown that in medicine....in many cases placebos actually work That is, sugar pills, if prescribed under the right circumstances, (through the right rituals...by a trusted expert in a white coat.) actually work, actually change our chemistry. They would not work if sneaked into our food unawares or with someone whom white coated physicians had no meaning.

I guess my point is that as a passionate music lover and audiophile for many years I should have MORE reason to doubt my objectivity...not less. I do have preconceived notions of how tubes sound, and what multi-bit means and I cannot totally discard those when critically listening to a new product. That makes me more subjective, not less.

 

[pwjazz]

The audiophile's valves or whatever may sound acceptable on girl-and-guitar, but will sound atrocious on a full blown symphony orchestra

I think that's a critical insight. I've found that some gear "synergizes" well with certain recordings and genres but sounds quite poor on others. Personally I like gear that works well with the majority of my recordings without having to worry about synergy.

Having now arrived at something like this through precise application of EQ, I'm able to enjoy the bulk of my music without even thinking about it. Occasionally, I'll encounter something unpleasant sounding and can be fairly confident that it's the fault of the recording and not my gear.

 

[watchnerd]

I think that's a critical insight. I've found that some gear "synergizes" well with certain recordings and genres but sounds quite poor on others.

This is one of the reasons why, when it comes to analog, I don't collect classical on vinyl, only reel to reel. Jazz is fine on LP.

And before people think I'm bonkers and engaging in fantasy:

1. Limited extreme dynamics and mono bass of LP more of an issue with classical than with jazz.
2. The benefit of playing individual songs, i.e. cue up different tracks at will (LP good for that), not an issue when listening to symphonies. Jazz albums are series of standalone songs grouped together nicely, as opposed to one continuous composed work broken into movements.
3. IGD PITA on LP side ends right during crescendos, the worst time. Yeah, Stevenson alignment helps a little, but not enough

 

[Sergei]

I'm a transparency person. If there's any euphonics to be applied for artistic reasons I'm happy for the record producers to do it.

There is a major difference between record producers applying effects and the audiophile imagining their system can do it. A blanket effect over the whole composite, mixed-down recording is fundamentally different from effects applied separately to the individual elements - which the producers have access to. Maybe second harmonic distortion really can make a vocal sound more pleasing in the context of a studio creation. But second harmonic distortion applied to the entire composite track is just a source of unmusical intermodulation distortion, the level of which varies with the complexity of the recording. The audiophile's valves or whatever may sound acceptable on girl-and-guitar, but will sound atrocious on a full blown symphony orchestra.

I think you nailed what is objectively done by some (most?) of the audiophile gear. Subjectively though, some people only listen to a limited range of music genres, and prefer distortions advantageous to that genre, in combination with the peculiarities of their hearing systems.

 

[Sergei]

It apparently adds more richness, dynamics and space.

But starting with what sort of system? The lacklustre, deficient one I mentioned above, I should think...

In this study in which the distortion profiles of various opamps were modelled, the authors concluded (p. 176) that:



There is an in-depth discussion of this on pp. 176-179, including a review of past studies that have made similar findings, as well as hypotheses as to the reasons for the findings.

The author speculates that a preference for some level and type of distortion may result from the fact that distortion is a cue for loudness, and loudness is generally preferred:



But whatever the reason, there is decent evidence for the proposition that many listeners prefer a certain kind of "bent" transfer function to one that is not.

There is some truth to it, as distortions obviously add to the perceived loudness, because of the energy added to the critical bands. Yet the pleasing effects of some distortions persist even after careful adjustment for the altered loudness.

Another effect, arguably more significant, is that certain distortions in human voice correspond to expression of emotions. For instance, constriction of certain parts of vocal tract linked to, say, fear or joy, results not only in the change of fundamental frequency, but also in increased proportion of odd harmonics, as it makes the vocal tract behave more like an uneven-diameter tube closed on one side. Relaxation of the vocal tract makes it behave more like an ideal open tube, thus generating relatively more even harmonics.

Human auditory system processes music in ways similar to speech processing, or more specifically, to the processing of the aspects of speech relaying emotions. Thus, distorting sound in a way that generates more prominent odd harmonics makes the sound perceptually more exciting. Distorting it toward the abundance of even harmonics makes it perceptually more relaxing.

Obviously, there are more dimensions in the manifestations of emotions in speech than I just described. The point is that certain types of distortions perceptually color the sound with artificially introduced emotions, presumably desirable by the person who prefers said distortions.

 

[andreasmaaan]

Thus, distorting sound in a way that generates more prominent odd harmonics makes the sound perceptually more exciting. Distorting it toward the abundance of even harmonics makes it perceptually more relaxing.

What do you base these statements on please?

 

[watchnerd]

I think you nailed what is objectively done by some (most?) of the audiophile gear. Subjectively though, some people only listen to a limited range of music genres, and prefer distortions advantageous to that genre, in combination with the peculiarities of their hearing systems.

I split the difference, because:

1. For digital, it's capable of best in class transparency. So if I want to hear what's on the recording, that's my preference.

2. For analog, none of the options are transparent; they're all flawed. So I pick the set of flaws and compromises that sound okay because a) they're not transparent and b) if I want transparency, I listen to digital.

 

[Sergei]

The video world doesn't have these discussions even though the basic problem is pretty similar to that of audio.

But it does https://www.google.com/search?q="video+grading"

 

[Sergei]

What do you base these statements on please?

Direct experimentation. Tinkering with neutrally sounding phrases using non-linear DSP. Making them into fearful, joyful, angry, sad etc. Morphing a speaker's perceived gender. Making a neutral phrase sound like a question, a tentative guess, or a confident statement. One of the exercise themes in https://www.coursera.org/learn/audio-signal-processing.

For a more academic treatment, you can read https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2683716/ , and the rich collection of other peer-reviewed papers it refers to.

Please note that the perception of even harmonics preponderance as relaxing may be related to the https://en.wikipedia.org/wiki/Missing_fundamental phenomenon, as such harmonics, in the relative absence of odd ones, perceptually add a fundamental subsonic in relation to the actual fundamental, which is indicative of more relaxed real vocal tract.

 

[Theo]

Another effect, arguably more significant, is that certain distortions in human voice correspond to expression of emotions.

Is the emotion conveyed by singers when they sing the blues, for example, an illustration of this? Usually, these guys (or ladies) have a colored voice, probably with lots of harmonics added (does anyone have examples of spectrum of singers voice?). Do you think that the blues sounds better when played on vinyl?

 

[Wombat]

Is the emotion conveyed by singers when they sing the blues, for example, an illustration of this? Usually, these guys (or ladies) have a colored voice, probably with lots of harmonics added (does anyone have examples of spectrum of singers voice?). Do you think that the blues sounds better when played on vinyl?

No. Nor acetate or wax cylinders.

 

[Sergei]

Do you think that the blues sounds better when played on vinyl?

I'm not qualified to answer blues+vinyl questions - no direct recent experience. Sorry.

I have a related experience with Pink Floyd's The Wall. While comparing a recording that was captured from vinyl with a version which was ostensibly digitized from a tape master for that vinyl, I liked some of vocals in some of the songs more on the vinyl, because they conveyed more emotion. The digitized master sounded a bit too calm in comparison.

Upon reflection, I realized that back in the day the vocals studio recording had to be calmer, because typical consumer gear would add enough distortions to cause the desired emotional coloration later on.

Or perhaps it wasn't actually the master for that particular vinyl, but a master mixed slightly differently, without putting the vocals through "enhancers" like http://www.thermionicculture.com/in...ure-vulture-1-17-192012-03-20-11-02-02-detail.

 

[andreasmaaan]

Direct experimentation. Tinkering with neutrally sounding phrases using non-linear DSP. Making them into fearful, joyful, angry, sad etc. Morphing a speaker's perceived gender. Making a neutral phrase sound like a question, a tentative guess, or a confident statement. One of the exercise themes in https://www.coursera.org/learn/audio-signal-processing.

For a more academic treatment, you can read https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2683716/ , and the rich collection of other peer-reviewed papers it refers to.

Please note that the perception of even harmonics preponderance as relaxing may be related to the https://en.wikipedia.org/wiki/Missing_fundamental phenomenon, as such harmonics, in the relative absence of odd ones, perceptually add a fundamental subsonic in relation to the actual fundamental, which is indicative of more relaxed real vocal tract.

I think these links all suggest that there might be a relationship between odd vs even harmonic distortion and emotional perception of a recording, but none of them appear to get anywhere near addressing this directly.

In the Halistone et al study, for example, the authors didn't investigate harmonics in isolation from envelope features, and didn't investigate odd vs even harmonics at all (they focused on high- vs low-order harmonics only, and did not add distortion to an existing music signal but used harmonics to synthesise musical "instruments"). I don't have access to all the other papers cited, but none of them seem to address the odd vs even question either.

The missing fundamental idea is interesting, but only speculative as far as I can tell in terms of whether it might contribute to a "relaxed" or "exciting" presentation.

Perhaps you could share some of your own experimentation? Audio files demonstrating the difference would be particularly interesting if you have any?

 

[Soniclife]

Upon reflection, I realized that back in the day the vocals studio recording had to be calmer, because typical consumer gear would add enough distortions to cause the desired emotional coloration later on.

I'm sort of convinced that this is behind some of the 'digital sounds sterile compared to vinyl' statements, when analogue was the only way of recording people strived for clean and great recordings, knowing they would end up grungy on playback at home, so when transferred to digital they end up more sterile than was expected. For recordings that were trying to accurately capture acoustic events that's not a problem, for lots of recordings going for a specific sound created in the studio it's still not a problem, but there do seem to be odd recordings that just don't quite work for some reason.

 

[Sergei]

The missing fundamental idea is interesting, but only speculative as far as I can tell in terms of whether it might contribute to a "relaxed" or "exciting" presentation.

Perhaps you could share some of your own experimentation? Audio files demonstrating the difference would be particularly interesting if you have any?

I guess I could dig up that old Linux VM image, and run again the Python code I wrote for the exercises. Yet there are simpler means to demonstrate the difference using audio plugins. Like this one:

.

Yes, the missing fundamental theory is speculative as it relates to even harmonics, because in a perfectly linear system the missing fundamental would be at not perceivable 0 Hz. But cochlear isn't linear, perception of pitch is shifted in relation to fundamental frequency, and the shift depends on sound intensity, so my contention is that we might be getting some fantom low frequency instead, perceived as "fullness":

. I believe I can hear it as something significantly lower than 100 Hz.

It is much less controversial in relation to odd harmonics. The perceived missing fundamental is supposed to be twice the frequency of the real fundamental (5f-3f=2f), but due to cochlear non-linearity not at exactly the frequency of second harmonic naturally generated by cochlear. This can be perceived as dissonant harshness. Additional dissonance comes from higher odd harmonics, as explained in the same video:

. Per theory of music composition, dissonance is linked to perception of emotional tension.

 

[andreasmaaan]

I guess I could dig up that old Linux VM image, and run again the Python code I wrote for the exercises. Yet there are simpler means to demonstrate the difference using audio plugins. Like this one:

.

Yes, the missing fundamental theory is speculative as it relates to even harmonics, because in a perfectly linear system the missing fundamental would be at not perceivable 0 Hz. But cochlear isn't linear, perception of pitch is shifted in relation to fundamental frequency, and the shift depends on sound intensity, so my contention is that we might be getting some fantom low frequency instead, perceived as "fullness":

. I believe I can hear it as something significantly lower than 100 Hz.

It is much less controversial in relation to odd harmonics. The perceived missing fundamental is supposed to be twice the frequency of the real fundamental (5f-3f=2f), but due to cochlear non-linearity not at exactly the frequency of second harmonic naturally generated by cochlear. This can be perceived as dissonant harshness. Additional dissonance comes from higher odd harmonics, as explained in the same video:

. Per theory of music composition, dissonance is linked to perception of emotional tension.

Thanks for the thorough explanation, and I understand the theory and agree in principle insofar as pure tones are concerned.

Where I believe the theory becomes speculative is with regard to the effect on complex signals rather than pure tones. Complex signals already contain huge numbers of frequencies that are dissonant with each other. Regardless whether even or odd harmonic distortion is added to to these signals, huge numbers of additional frequencies will be created, which will in turn create additional dissonance.

I understand that the conventional wisdom is that even order (especially lower order) harmonics are "musical" (or "calmer" as you put it) for the reasons mentioned in the video, but I'm not convinced that this is the case when any type of distortion is applied to a complex signal.

What I'm thus hoping for is some experimental evidence that people subjectively prefer the kinds of harmonic distortion generally described as "musical" under controlled conditions, as is often claimed by SET advocates. etc.

As I've mentioned in previous posts, I'm willing to imagine that this may be the case, but I don't know of any evidence of it other than speculation and sighted anecdotal experience.

The demonstration with that distortion effect unit in the video doesn't really convince me, partly because there are no measurements to verify what the unit is doing, partly because I wouldn't describe the effects that I hear from either type of distortion as "calm" or "exciting" necessarily (although I agree the effect is "fuller" sound), partly cos it's all sighted (visual cues and suggestions are very persuasive), and partly cos the guy's voice is way louder than the music, making it hard to listen critically.

 

[Sergei]

Thanks for the thorough explanation, and I understand the theory and agree in principle insofar as pure tones are concerned.

Where I believe the theory becomes speculative is with regard to the effect on complex signals rather than pure tones. Complex signals already contain huge numbers of frequencies that are dissonant with each other. Regardless whether even or odd harmonic distortion is added to to these signals, huge numbers of additional frequencies will be created, which will in turn create additional dissonance.

I understand that the conventional wisdom is that even order (especially lower order) harmonics are "musical" (or "calmer" as you put it) for the reasons mentioned in the video, but I'm not convinced that this is the case when any type of distortion is applied to a complex signal.

What I'm thus hoping for is some experimental evidence that people subjectively prefer the kinds of harmonic distortion generally described as "musical" under controlled conditions, as is often claimed by SET advocates. etc.

As I've mentioned in previous posts, I'm willing to imagine that this may be the case, but I don't know of any evidence of it other than speculation and sighted anecdotal experience.

The demonstration with that distortion effect unit in the video doesn't really convince me, partly because there are no measurements to verify what the unit is doing, partly because I wouldn't describe the effects that I hear from either type of distortion as "calm" or "exciting" necessarily (although I agree the effect is "fuller" sound), partly cos it's all sighted (visual cues and suggestions are very persuasive), and partly cos the guy's voice is way louder than the music, making it hard to listen critically.

Indeed, perceived distortion or enhancement of complex signals is much more difficult to explain. I don't have all answers. I have some though:

The "huge number of frequencies", and its counterpart in the time domain "huge number of reflections", effectively raises the noise floor. If a piece of music contains a consonant set of strong harmonics above that noise floor, the noise will be perceptually filtered out. This is the mechanism explaining why dipole speakers such as Linkwitz's Orion are subjectively decent, despite objectively generating a very messy signal at the ear.

When Orions were new and hot, I auditioned one in a house of a fellow Audio DIY member. After listening to it for over an hour, we noticed that on every piece of music there was a puzzling presence of a frequency around 1KHz, seemingly only correlated with the overall music fragment loudness.

After some experimentation, we realized that the annoying sound's wavelength corresponded precisely to the width of the dominant horizontal beam supporting the room's roof. We were unwillingly echolocating the room's geometry ! The lesson was that uniformly distributed random components of sound are perceptually filtered out, whereas the correlated ones stick out.

SET is a complex subject. Here's an interesting take on it: http://img2.tapuz.co.il/forums/1_133166698.pdf. I'm not convinced by that article. Long time ago, back when I was on the audiophile treadmill, I owned and evaluated a number of SET amps. My final conclusion was that in one way or another, SET perceptually generates a pleasant "smiley" frequency response, characteristic of Hi-Fi gear: "full" and "sparkly".

I don't think you'd find, experimentally or anecdotally, that all people always prefer a specific distortion applied to the whole music signal, for every piece of music. Perceptually enhancing music tracks, especially vocals, via a moderate amount of distortion specific to a track's time slice, is a well-established practice. As is the removal of unwanted frequency components appearing in the original signal, for instance when a vocalist was too anxious.

In the video, the mixing engineer demonstrates how he can perceptually make the record to sound more Hi-Fi through controlled application of distortions. Unlike audiophile gear vendors, he gives us the exact recipe. As to what the plugin actually does - you can buy it and measure it if you are interested, just like you'd measure physical gear, using your own methodology.

 

[andreasmaaan]

Once again I basically agree with all you say I’m maybe just a bit less convinced of the idea that there’s a significant difference in emotional impact of even vs odd aharnonuc distortion.

I don't think you'd find, experimentally or anecdotally, that all people always prefer a specific distortion applied to the whole music signal, for every piece of music.

No you certainly wouldn’t find that There is some evidence that many people prefer audible distortion for many signals, but what types of audible distortion is an area that needs more investigation IMO.

Very interesting to hear about your experiences with SET amps and dipoles etc BTW...

 

[VMAT4]

Well, I don"t know hifi but, I know what I like !