Live music, Tone and Presence: What most systems get wrong

[NorthStar]

This topic is making lite on the fact that the recording matters an awful lot.

For example Mike's Rostropovich cut that's a wonderful demo sounds soooo different from most other cello recordings. It was very well done and gives a presentation that exceeds who knows how many other recordings. Having something that good can even help a system produce something that may not be it's explicit talent.

If we wanted true to life productions, and I wouldn't say we don't, we'd have to put much higher demands on the recording, mixing, and mastering from the people involved to the equipment. Honestly though I enjoy an awful lot of albums that aren't life like, and never will be, just as they are since listening to choices in production can be a great experience depending on the type of music. With classical I don't feel that way, but for pop I can't deny that I can get more enjoyment out of something not as life like with many albums. Anyone who's spent time around musicians & live pop music can probably understand that.

You got it; the recording itself is all the spicy essence, the microphone and its ability to capture what the human ears hear (and doesn't hear), as best as a mechanical device approximation of the designed condenser itself.

And also true; with some type of music (many amplified inferior rock/pop/punk/rap/metal...) it just don't matter, what does is no microphones anymore but the sound effects from the mixing/console board and the human manipulation from the recording engineer. It is not a pure job but a robotic/mechanical/financial application for the auditive senses and fake emotions...just like fake news on TV and the internet. It's for the kids in search of an illusive world of domination through body impulses and vibrations...subwoofing rattling bones and excruciating subliminal pain of the mind through revolutionary brainwashing. :b

Simply put: Real versus Fake.
Peacefully pleasant and emancipation of the soul in a true natural pure form versus disturbingly atrocious and abomination of the mind in a fake degenerated escape.
...Some' like that.

Music is an acquired/evolved taste. I believe in a higher level of music listeners, irregardless of the gear used. One doesn't necessarily complement the other.
Most true ultra high end audiophiles have that higher music taste; of course because it's also part of what is a calibrated/mature audiophile...I think.

Lol, high res audio downloads (64-bit/352kHz) of Judas Priest, Metallica, Justin Bieber, Drake, Kanye West, tra-la-la ... yeah whatever.
_________

* If we ask two musicians, Ozzy Osbourne and Yo-Yo Ma, about what sound systems sound best (tone and presence) I bet we'd be in for some surprises...

P.S. Thanks to Carl for that previous link: http://www.newyorker.com/magazine/2012/02/06/flight-of-the-concord

 

[morricab]

@ exactly 8:39 and 10:19 what do you think it is?

* The 'Presence' of the instrument in the room is much more evident than in the recording with the room's reverberation added.
The recording is not accurate, it is colored. The capture by the microphone is 'imperfect'.
The 'Tone' of the cello? If the presence is better Live the Tone also is, because the recording not only colors the true 'Live' atmosphere but also the tonality of the instrument.
Brad was very correct in his first original post. And this is only a youtube video @ low resolution audio.

The musician, Vincent, has an expression that tells us a lot without saying a single word, when listening to the recording.

Interesting observations and I would mostly agree, although it is unclear for me what % of the difference between the live performance and the recorded/played back one was the recording or the gear.

Clearly the original recording had a different ambient field that immediately told you it was not live. The presence, as you noted, was reduced significantly and I found the tonality shifted downward in frequency. The cello sounded deeper on the playback. Also, the sensation of independent strings resonating and that interaction with the cello body was muddied up to some extent. Would be interesting to hear that setup again with different speakers that are setup differently. Might make a big difference. I think the sound though was fundamentally organic and non-artificial, which is a step up from most but it did not sound like the live really that closely...no world's away but for that kind of money I would have expected a bit closer.

The problem with making a recording and playing it back is the room influence. You have that acoustic on the recording no matter what you do. That is then overlayed with the room acoustic of the playback system. The live performance only has the playback room acoustic. Microphone distance is another factor. Too close or too far and the degree of presence will somewhat change...our brains are evolved to judge distance from sonic cues...not as accurately as sight of course but not too bad.

When I made recordings of my ex in my apartment they were very present even though I had the microphone setup literally at the stereo listening position (a bit over 3 meters) and she was stending in between the speakers. The projection of the Stradivarius she was playing was amazing...a real sonic cannon. It actually made it difficult to keep the gain in a good place so that I was not saturating the tape (I used a TEAC R2R from the mid 1970s that had been fully reconditioned). The projection/presence and basic tone of the recording are quite close to what I actually heard but the overlay of the acoustic space significantly dries out the ambience. Played back, you hear that it is not the same acoustic space you are currently in and this removes the "liveness" to some degree.

One of the nicest recordings I have is of Jim Hall "Impressions of Japan" on direct cut LP. The solo acoustic guitar on the first track has a real "He is there in the room" presence that blows me away every time I hear it. The judgement of the acoustic space in which that recording is made is superb and there is real genius in getting that to sound like live in a completely random room...with the right playback of course.

This recording will sound much less live if the gear does not convey that presence that is on the recording...it will sound "nice" but fall a bit flat.

I have only heard AN as a complete system and always found that it didn't live up to the potential from the ingredients...nice, even interesting but not "live". I tend to blame the speakers for this...herasy to an AN disciple I know, but that is where I think the weakest link in their chain lies.

 

[bonzo75]

I don’t like AN speakers either.

For those who do, they should try the hORNS fp10 and 15 which are similar but better.

That said, at the Windsor hifi show last month, the AN TT2 into their phono and SET with the AN IO was giving a stellar purity in the top and the midrange. Though I did not like the overall presentation.

 

[KeithR]

I have only heard AN as a complete system and always found that it didn't live up to the potential from the ingredients...nice, even interesting but not "live". I tend to blame the speakers for this...herasy to an AN disciple I know, but that is where I think the weakest link in their chain lies.

I've only heard AN at shows, but can't last 5 minutes in their demos. the downward shifted tonality you mention has reared its head at the same demos and i found them syrupy/colored. perhaps the electronics on different speakers would fare better.

I think you nailed a big component though - everyone's reference is in a room or hall of some sort and that changes the sonic cues. Mine is Disney Hall (like jtyo on this thread)

 

[YashN]

So, my luddite ass has been dragged into the chip playing era, so there's that. I may transcribe my digital music to solid state chips because they have decent storage capacity now.

I suggest you also try having the files on SDCards - some people swear by those. Most probably the circuitry that goes with reading from it and its inherent noise profile is much less than with most USB implementations.

Also, try wired rather than wireless as well.

 

[YashN]

I made the same mistake, and wrote an answer, but deleted it when I understood the situation. IMHO when re-posting something from another forum member we should post it as a quote.

Even better: don't post flawed experiments as reference.

 

[YashN]

You said, with SET on the ribbons, "played a cello sonata and heard all the SET magic but also in the big picture, missed sufficient definition within the mid bass structure and also it left the music less coherent than I prefer. " When I heard SETs on ribbons, I heard them missing aspects of the music. My suspicion is, not being a techie, the ribbon is not being sufficiently extended (because that is when the impedance drops and needs to be controlled, and when the ribbon comes back the impedance goes up leading to swings). Whatever that control is, the current from class A SS is producing the ribbon tone better.

SETs need a particular set of characteristics to operate their 'magic'. Some people say SET don't work well with most speakers. Well, that's because most speakers are now built not as highly-efficient as they used to be to match with SET Tube amps. All this because the SS amps manufacturers play the numbers game.

Indeed, to drive those inefficient speakers, you better have a lot of Watts.

So, generalising about ribbons sounds a bit incomplete, we'd need more details: the power rating of the SET, the power rating of the ribbon, the impedance profile, each manufacturer's recommendations, what SET amp, what Ribbon, in what configuration, size of the room, etc...

 

[cjfrbw]

My own use of SETs and ribbons relied on a few observations of ribbon performance.

Ribbons operate at flat impedance i.e. they do not by themselves have complex impedance, which makes them a rather benign load when driven directly by an amplifier. The amplifier views them as a static resistance. Because they do not have complex impedance, it is a bit easier to drive them directly even at a relatively low impedance i.e. 2.5 to 4 ohms because the amplifier just views them as a simple resistance.

Ribbons tend to work well with standard crossover slopes without requiring unusual compensations.

Ribbons have more complex impedances only when used with passive crossovers (caps and coils).

SETs have a large part of their distortion profile below 100 HZ. If you cross over an SET at 100 Hz, you have reduced a good bit of its distortion profile and its distortion above 100 Hz is much less.

If you use active crossovers, then you can drive ribbons directly with the SET at their particular ideal frequency range. This not only improves the efficiency of the ribbon by about 3db (amplifier doubling) but does not engage the complex impedance of a passive crossover (caps and coils). The worse case scenario is that a very low impedance ribbon would require an impedance matching transformer.

With two 50 watt SET per stereo channel aka my Wavacs operated through an active crossover, say, to my Apogee Stage speakers, you would effectively have the same driving power of an 150 watt SET through a passive crossover, but without the complex impedance behavior of a passive crossover. Also, crossed over at 100 Hz to a subwoofer, the distortion of the SETs would be less driving the frequencies above 100 Hz. Also, crossing over to a subwoofer at 100 Hz further reduces the power demands on the SET amplifier driving the lower frequency ribbon.

This worked out quite well using active crossover with both Apogee Stage and Analysis Epsilon and subwoofer.

You can drive ribbons with SETs pretty successfully but it depends on the particular ribbon, its impedance, it’s efficiency, and how you work the crossover.

If you want to drive a ribbon speaker that has a passive crossover and multiple elements with a single amp, then you would need a whomping powerful SET as a rule, something like the big Wavacs, but even then, it might not be ideal. The exception might be some of the newer ribbons with very powerful magnets that can achieve efficiencies much higher than 90 db/ watt., in which case they would be even more SET friendly.

 

[bonzo75]

How much current does the Wavac deliver?

 

[morricab]

Even better: don't post flawed experiments as reference.

Exactly...the "study" seemed rather bogus to me.

 

[FrantzM]

Hi

In all this and I do understand the point about THD not being that much perceptible. I don't know for sure but I could understand the case for the distortion spectrum of a SET being mostly in the under 100 Hz region. That is fine my objection is in the construction of SET as being the end-of-it-all. How they present the musical information could well be prefered by a person but to make of this the absolute and only correct way to present musical information is to me wrong.
I , for one find many SET tend to beautify the music. Making it sound nicer when this is not called for. Can this be preferred? of course. Do they always do that? No Some do, not all .
The other side of the story is how do SET interface with other people preferred speakers. Some people, DDK , for example take the approach of getting the amp first then finding a speaker that would match with it. This is a hobby and we're talking about preferences here, so in this realm this is a valid approach. I don't subscribe to it. I chose my speaker first, i believe and this can be demonstrated BTW that a speaker impose more of if itself to the reproduction that electronics. I do thus go for speaker I like first ( these days horns ) and find the electronics to go with it. The "Tone", "presence". "body" notions are the listener's, personal characterizations of the reproduction and cannot be construed as proofs. They're subjective by nature being influenced by a plethora of inputs and stimuli ranging from knowing to what he/she is listening to (Great looking and known SET amplifers, speakers , front end) to his mental state at the time ( happy to be there, happy at the time or sad for that matter) or to his level of intoxication or lack thereof ...

While it is true that some system provide to some listeners a great level of believability. Something that they like and that they construe as being closer to the musical reality they perceive when experiencing Live Music. This is by no means universal. And in that SET or Tubes are no more valid than a person using Class A or D SS... it is what float that person's boat that counts at the end ... Let's not make of this debate that My preferences are superior because I have more exposure than another in listening to live music... Nothing could be further from the Truth.

 

[morricab]

My own use of SETs and ribbons relied on a few observations of ribbon performance.

Ribbons operate at flat impedance i.e. they do not by themselves have complex impedance, which makes them a rather benign load when driven directly by an amplifier. The amplifier views them as a static resistance. Because they do not have complex impedance, it is a bit easier to drive them directly even at a relatively low impedance i.e. 2.5 to 4 ohms because the amplifier just views them as a simple resistance.

Ribbons tend to work well with standard crossover slopes without requiring unusual compensations.

Ribbons have more complex impedances only when used with passive crossovers (caps and coils).

SETs have a large part of their distortion profile below 100 HZ. If you cross over an SET at 100 Hz, you have reduced a good bit of its distortion profile and its distortion above 100 Hz is much less.

If you use active crossovers, then you can drive ribbons directly with the SET at their particular ideal frequency range. This not only improves the efficiency of the ribbon by about 3db (amplifier doubling) but does not engage the complex impedance of a passive crossover (caps and coils). The worse case scenario is that a very low impedance ribbon would require an impedance matching transformer.

With two 50 watt SET per stereo channel aka my Wavacs operated through an active crossover, say, to my Apogee Stage speakers, you would effectively have the same driving power of an 150 watt SET through a passive crossover, but without the complex impedance behavior of a passive crossover. Also, crossed over at 100 Hz to a subwoofer, the distortion of the SETs would be less driving the frequencies above 100 Hz. Also, crossing over to a subwoofer at 100 Hz further reduces the power demands on the SET amplifier driving the lower frequency ribbon.

This worked out quite well using active crossover with both Apogee Stage and Analysis Epsilon and subwoofer.

You can drive ribbons with SETs pretty successfully but it depends on the particular ribbon, its impedance, it’s efficiency, and how you work the crossover.

If you want to drive a ribbon speaker that has a passive crossover and multiple elements with a single amp, then you would need a whomping powerful SET as a rule, something like the big Wavacs, but even then, it might not be ideal. The exception might be some of the newer ribbons with very powerful magnets that can achieve efficiencies much higher than 90 db/ watt., in which case they would be even more SET friendly.

Nice post but I have a few comments/questions:

You mention a 3db increase in sensitivity bypassing the passive crossover...that would be double the power and not triple (you mention from 50 to 150 watts), no? Have you measured this 3db increase or is this somehow theoretical (please show me, I am curious about it).

What active crossover were you using for the Stage and Epsilon?

You are right about the more efficient ribbons but we found even at around 87db or so it works as well up to probably about 100db.

I had really good results driving a Bohlender Graebener planar driver with KR VA350 using an Accuphase F-25 crossover at 300Hz. That was 87db and 4 ohms. I had only a passive notch filter to control a resonance in the driver. The bass was a 10 inch woofer in a slightly overdamped (Qts = 0.6) sealed box that was also directly driven by a second KR VA350. Very good sound from that and it would play very loud. KR is one of the few SETs that doesn't have significant core saturation in the bass (something that WAVAC has an issue with based on the Stereophile measurements of their biggest monos) so it is very tight and slightly dry in contrast to most other SETs. The only ones I have heard as good or better for bass (on conventional speakers...horns are another story) are from Ayon and especially Aries Cerat.

KR on the Apogee Centaur Major is also a treat.

 

[morricab]

Hi

In all this and I do understand the point about THD not being that much perceptible. I don't know for sure but I could understand the case for the distortion spectrum of a SET being mostly in the under 100 Hz region. That is fine my objection is in the construction of SET as being the end-of-it-all. How they present the musical information could well be prefered by a person but to make of this the absolute and only correct way to present musical information is to me wrong.
I , for one find many SET tend to beautify the music. Making it sound nicer when this is not called for. Can this be preferred? of course. Do they always do that? No Some do, not all .
The other side of the story is how do SET interface with other people preferred speakers. Some people, DDK , for example take the approach of getting the amp first then finding a speaker that would match with it. This is a hobby and we're talking about preferences here, so in this realm this is a valid approach. I don't subscribe to it. I chose my speaker first, i believe and this can be demonstrated BTW that a speaker impose more of if itself to the reproduction that electronics. I do thus go for speaker I like first ( these days horns ) and find the electronics to go with it. The "Tone", "presence". "body" notions are the listener's, personal characterizations of the reproduction and cannot be construed as proofs. They're subjective by nature being influenced by a plethora of inputs and stimuli ranging from knowing to what he/she is listening to (Great looking and known SET amplifers, speakers , front end) to his mental state at the time ( happy to be there, happy at the time or sad for that matter) or to his level of intoxication or lack thereof ...

While it is true that some system provide to some listeners a great level of believability. Something that they like and that they construe as being closer to the musical reality they perceive when experiencing Live Music. This is by no means universal. And in that SET or Tubes are no more valid than a person using Class A or D SS... it is what float that person's boat that counts at the end ... Let's not make of this debate that My preferences are superior because I have more exposure than another in listening to live music... Nothing could be further from the Truth.

Well, if it was only sound that influenced people's choices in systems I think there would be a LOT less choice and subjective blah blah out there. The fact is that people are just as heavily influenced by appearance, what they "know" and have been taught as they were inducted into this hobby and their friends than what they actually hear. I am sad to say that very few know how to really listen critically and make the connection between what they hear in live, unamplified concerts or recitals and how that translates into what is heard from playback.

It has been demonstrated that THD has NO correlation with sound quality...none. Geddes et al. demonstrated this in a couple of papers. AND if it was the case then it would be clear that SET has no chance in delivering even acceptable sound quality. When they were first being reintroduced to the world that is exactly the kind of knee jerk reaction that the audio press had about them. There is NO WAY they could possible sound good...8 watts with 5% THD??? Outrageous!!! And yet, it gained a foothold and now a stronghold because it rather sounds more realistic in many implementations (not all but many).

What does correlate with sound quality is the harmonic structure of the distortion that is generated...the order of the harmonics and their relative and absolute levels. This has been known at some level since the 1930s but was largely ignored once negative feedback pushed all naysayers away into a THD arms race.

Both Geddes and Cheever have put forth their pyschoacoustic models that should do a better job of predicting sound quality from electronics (it applies not only to amps but also preamps and output stages of DACs and phono as well). Cheever looks at the ear/brain and self-generated harmonics and how that affects masking. It turns out to be SPL dependent, which has implications for amp/speaker matching.

Based on this work, it is clear that in the absence of a truly distortion free linear amplifier, there are psychoacoustically preferrable ways of designing electronics. This inexorably leads to a limited number of topologies that will get you the desired distortion pattern that can "hide" in the blind spots of perception. The point is that there will always be people who prefer something else and not necessarily because of sound quality, despite what they claim because they don't even know their own psychological blind spots, as this area is not and never will be exact science. What can be done though is to show a strong correlation with a certain types of distortion and then design products to mimic or avoid certain patterns. Of course some people will continue to drive towards zero distortion...who knows maybe they will get there and if they do maybe it will sound the best...but the current way of doing it with negative feedback doesn't seem to get there because of the other problems that this path creates.

Some designers openly say they are using a psychoacoustic model for optimizing their designs...Lamm comes to mind. It is no wonder then that their top tube amps are SET and that they have disotortion patterns that largely agree with the papers I have mentioned. Lamm claims not to do listening tests, just optimization towards a model of hearing. Not sure if it is true but I wouldn't be surprised. The ML2 (and its successors) are really good sounding amps...I haven't heard the ML3 but I can't imagine it being worse.

DDK, is IMO, largely right to start with the right electronics and then find the best speaker to convey the signal it is given and not the other way around. That is not to say that optimizing the speaker is not important, it is to get max realism but you can never get realism from a system without the best electronics you can afford. I have heard so many times that it is hard to count a great set of electronics elevating a speaker that most would consider ok or mediocre to something really great and, conversely, a great speaker completely ruined by the sound of the electronics behind it. That is not to say that the great electronics/mediocre speaker could not be further elevated by replacing with a great speaker...it can of course. But you can go to the very pinnacle of loudspeakers and not get realistic sound of any kind with mediocre electronics. That is because distortions from electronics are wholly unnatural and outside our evolutionary experience. Speakers are largely mechanical things that makes distortions that our evolutionary experience can still relate to and accept more easily. Therefore, getting that right, is both harder and more important, IMO, than the speaker choice. To get to a very high level requires both of course.

A SET, like all tube amps except OTLs, can have issues in the bass due to saturation of the output transformer core. This makes a lot of distortion and it is clearly audible and goes right up through the midrange to color the sound. It is also output power dependent. SETs in particular are difficult this way because of the static bias on them (a Push/pull has zero net current running through it) that reduces the linear headroom. This means the core needs to be BIG but a big core makes having a wide bandwidth difficult without clever winding ideas. For example, Aries Cerat Diana has 18Kg output transformers! That is to ensure no core saturation within the rated power. The amps sounds very clean and uncolored as a result of NOT having significant core saturation. If the SET or PP amp has taken care of this issue then the < 100Hz performance can be stellar.

So, you can claim pure relativism an that there is no preferred way to do these things, but it seems that the science that has been done in this area would not really support that at least for a majority of listeners and if you remove the other non-sound quality factors from the equation. Would Lamm sell more if their amps looked more sexy or had lower THD?? Probably but at least for the latter point it might not sound better.

 

[microstrip]

(...) What does correlate with sound quality is the harmonic structure of the distortion that is generated...the order of the harmonics and their relative and absolute levels. This has been known at some level since the 1930s (...)

I am tempted to be a believer in this theory. I appreciate the amplifiers designed by people who tell us they use it in their work. However most of my tube equipment sounds decent after a five minute warmup, reasonable after 30 minutes, good after one hour and great after 2 or more hours - the values change somewhat according to the equipment. I have measured the distortion spectra of some of these amplifiers at all these phases and it is exactly the same. IMHO we need a dynamic parameter to correlate with sound quality, not just the harmonic structure. Just looking at Stereophile measurements is not enough. We pick what pleases our findings, but little else. Just MHO, YMMV.

 

[microstrip]

(...) How they present the musical information could well be prefered by a person but to make of this the absolute and only correct way to present musical information is to me wrong. (...)

I think no one is interpreting this thread in such sense. We all know since long what is preference, individual opinion and statistical meaning. But we (not all, it seems...) enjoy and use a free, perhaps sometimes too professorial style.

While it is true that some system provide to some listeners a great level of believability. Something that they like and that they construe as being closer to the musical reality they perceive when experiencing Live Music. This is by no means universal. And in that SET or Tubes are no more valid than a person using Class A or D SS... it is what float that person's boat that counts at the end ... Let's not make of this debate that My preferences are superior because I have more exposure than another in listening to live music... Nothing could be further from the Truth.

Again, if someone his able to justify his preferences with examples of the real, that we can understand, question and openly rebate, I consider his opinion more valid (in the sense of being of insterest, not the Truth) than those that can not present anything other than science and measurements.

BTW, I also do no own SETs, but really appreciate the debates!

 

[morricab]

I am tempted to be a believer in this theory. I appreciate the amplifiers designed by people who tell us they use it in their work. However most of my tube equipment sounds decent after a five minute warmup, reasonable after 30 minutes, good after one hour and great after 2 or more hours - the values change somewhat according to the equipment. I have measured the distortion spectra of some of these amplifiers at all these phases and it is exactly the same. IMHO we need a dynamic parameter to correlate with sound quality, not just the harmonic structure. Just looking at Stereophile measurements is not enough. We pick what pleases our findings, but little else. Just MHO, YMMV.

When you say measured, what do you mean by that? How did you measure? Would you be willing to share that data? Most amps bias does not settle right away and parts change values slightly during warm up etc. etc. I would not be suprised to find that most active devices are not at their designed operating points until they have warmed and stabilized. This will change the distortion generated.

 

[microstrip]

When you say measured, what do you mean by that? How did you measure? Would you be willing to share that data? Most amps bias does not settle right away and parts change values slightly during warm up etc. etc. I would not be suprised to find that most active devices are not at their designed operating points until they have warmed and stabilized. This will change the distortion generated.

Unfortunately it is not organized, as I did not collect it systematically. I used the SectraPlus software coupled with an USB EMU tracker soundcard, that shows the total THD and full spectrum of the measurements. As it has the capability of keeping the six successive spectra in overlays, differential comparison it is an easy and fast job. I will try repeating it with the conrad johnson LP2575m and saving the results. The example I attach is just a photo screen when analyzing tubes for lower noise and distortion in my tube jig that reproduces the operational condition of the ARC preamplifiers.

Bias points usually settle in a few minutes - less than five for most tube electronics, being minimal after two minutes.

 

[miniguy]

This a failed demo because the recording was improperly done by including the room in the recording. The proper way to conduct a live vs recorded demo is to record the live instrument under anechoic conditions as was done by Edgar Villchur of Acoustic Research over 50 years ago with the Fine Arts quartet. He recorded them outdoors.

 

[miniguy]

This a failed demo because the recording was improperly done by including the room in the recording. The proper way to conduct a live vs recorded demo is to record the live instrument under anechoic conditions as was done by Edgar Villchur of Acoustic Research over 50 years ago with the Fine Arts quartet. He recorded them outdoors.

This post was in reference to the Audio Note Cello demo mentioned in an earlier post.

 

[Al M.]

But you can go to the very pinnacle of loudspeakers and not get realistic sound of any kind with mediocre electronics. That is because distortions from electronics are wholly unnatural and outside our evolutionary experience. Speakers are largely mechanical things that makes distortions that our evolutionary experience can still relate to and accept more easily. Therefore, getting that right, is both harder and more important, IMO, than the speaker choice. To get to a very high level requires both of course.

Speakers have crossovers that are not mechanical things and introduce distortions, e.g., phase shifts, that our evolutionary experience also cannot relate to.