Testing audiophile claims and myths

[bigshot]

Yes, like live. But working the mic is imperfect and in the studio you can finesse it and balance levels perfectly. It’s OK to be imperfect live. The flaw goes by, you move on and the flaw never comes around again. But when you record you hear the same mix and performance every time you play the song. Imperfections can become irritating.

 

[castleofargh]

I have often wondered whether that processing (autotune/pitch correction & dynamic compression in particular) is really something that consumers wanted, or whether they have been made to want it and have become so used to hearing it that way that they now insist on it themselves. But that's a different topic.

For compression, it has been tested, and you find several papers that tell us mostly very expected things.
1/ At large, what you're used to is... what you like.
2/ direct result of 1/, the genres with more compressions are preferred with more compression.
3/ the compression settings and use(attack in particular) are just as significant as arguing for or against compression. Strong global compression is not appreciated, but per instrument compression is. it is hard to draw conclusions because 1/ of course, and because different experiments did things differently and got different results depending on how far they would push compression. There is also always the possibility of tracks being known for some tests and people just, again, being used to what they're used to. Most papers agree that people aren't good at telling when something is compressed. Correlation between impression of dynamic and compression are all over the place, except for one study I saw where there was some correlation in the most extreme cases of classical or rock music. Knowing that the maximum compression level was the thing you imagine when looking at the signal with everything smashed against 0dB. We never get classical music like that, and rarely do with rock(and we tend to whine when we do, hey there Metallica!). So even then, it's impossible to ignore the possibility that habits play the biggest part in our judgement.

Overall, some compression seems to be desired. And when noticed, people prefer mindful per track compression over global compression or heavy limiting. no surprise there.
Some of those papers tell the initial belief of the researchers about loudness war or compression in general as being detrimental to sound quality(subjective!), and then the results never are that clear-cut.

I don't have papers on autotune, but I imagine it's the same. Skilled subtle use wouldn't even register for a listener, and the over the top effect has just become a music genre of its own that please some and not others(I would expect a young VS old given what's popular nowadays VS what adults tend to say about it). Maybe it's also just a thing of our time when recording needs to go forever faster, and there is no need for 30 takes of a crappy singer if you're planning to make him sound like the 300 other fake rappers with the same final "voice"(oops my age is starting to show). Ok I just checked a top rappers' list, and now I'm mad.


About your main discussion, it is terribly difficult to say. We're all a mix of experience and predictions. It's hard to argue about a clear reference when anybody is able to recognize a piano for being a piano, no matter which one, who plays, where, or how it's recorded. Our brain has that ability, and it's one that's on autopilot. With training, you can learn to discriminate more, and I'm sure that going to listen to acoustic instruments has some impact on you, your experience, memories, who you are. But won't it also become a matter of liking what you're more used to(maybe some venue you go to more often), instead of recognizing what is "right" or accurately reproduced from some record in some other place with other instruments, musicians, mics, mixing...?
I get your view and sort of agree, but I also cannot fault gregorio's. We're only able to fool ourselves into associating very different sounds because of that ability to associate nearly anything with a vaguely similar envelope, some visual cues or preconceptions telling us what to expect/hear/memories. And from now on, that stuff will be associated to those groups of neurons instead of making a fresh memory, so of course we're going to find a link when that link is now probably real on a "circuit" level.
We can be ultra picky about playback, and we can still recognize and enjoy a catchy tune from a supermarket's PA system. We're the same person doing both.

 

[2leftears]

Mixers EQ instruments to remove overlapping frequency bands too. This increases separation between instruments.

Dynamic compression can be misused, but it is a very useful tool. Vocals are almost always compressed. If they weren’t, you wouldn’t understand the lyrics in quiet parts and balancing the voice to the band would be impossible. The singer would have to work the mic, which is a far more clumsy way to solve the problem.

Sure. But to what extent are you now also talking about the need to compensate for poor microphone technique and perhaps even vocal arrangements that aren't well thought-out?

Maybe I'm showing my age. What you refer to as a clumsy way to solve the problem I (as a music lover and sound engineering layman) would interpret as an experienced singer's technique to balance the timbre of loud singing vs soft singing with the need to have the vocals sufficiently audible against what the rest of band is doing. Keeping the mic at a constant distance regardless of whether the singer is singing softly or belting at full power I would consider a sign of a singer inexperienced in how to use a microphone effectively. If the singer uses the microphone effectively, they have more control over the vocal dynamics themselves, rather than the sound engineer having to fix it or putting their own interpretation over it.

But I realise that watching a live performance vs producing a commercial recording aimed at the mass market are two different things.

Yes, like live. But working the mic is imperfect and in the studio you can finesse it and balance levels perfectly. It’s OK to be imperfect live. The flaw goes by, you move on and the flaw never comes around again. But when you record you hear the same mix and performance every time you play the song. Imperfections can become irritating.

You hit on an interesting point here. I think even music fans fall into two categories here; those who like the live performance recordings made commercially available to have minimal post-production processing, keeping all those little flaws in it, vs. those that don't mind live recordings being cleaned up much like studio recordings are. I prefer the former, but I know many others prefer the latter.

For compression, it has been tested, and you find several papers that tell us mostly very expected things.
1/ At large, what you're used to is... what you like.
2/ direct result of 1/, the genres with more compressions are preferred with more compression.
3/ the compression settings and use(attack in particular) are just as significant as arguing for or against compression. Strong global compression is not appreciated, but per instrument compression is. it is hard to draw conclusions because 1/ of course, and because different experiments did things differently and got different results depending on how far they would push compression. There is also always the possibility of tracks being known for some tests and people just, again, being used to what they're used to. Most papers agree that people aren't good at telling when something is compressed. Correlation between impression of dynamic and compression are all over the place, except for one study I saw where there was some correlation in the most extreme cases of classical or rock music. Knowing that the maximum compression level was the thing you imagine when looking at the signal with everything smashed against 0dB. We never get classical music like that, and rarely do with rock(and we tend to whine when we do, hey there Metallica!). So even then, it's impossible to ignore the possibility that habits play the biggest part in our judgement.

Overall, some compression seems to be desired. And when noticed, people prefer mindful per track compression over global compression or heavy limiting. no surprise there.
Some of those papers tell the initial belief of the researchers about loudness war or compression in general as being detrimental to sound quality(subjective!), and then the results never are that clear-cut.

For sure it is a matter of whether or not the compression has been applied sympathetically. But I now increasingly regularly come across recordings that have been very aggressively compressed to within an inch of their lives, and they sound awful. The impact of bass drums hits in particular simply have been wiped out all but entirely. No dynamic feel to it at all, flat as a pancake. Even worse is if the gain pumping becomes noticeable due to (I suspect) unsympathetic attack and release settings. It results in a very discomforting volume modulation effect (at least to my ears). It is hard to describe as I am not a sound engineer, but it sounds almost as if they compress the dynamics of the loud-peak instruments (like bass drum) to get rid of the volume peaks, but then to separate out the other instruments and get some artificial feel of dynamics back they briefly drop the volume of the other instruments right at the bass pedal hit, for them to very quickly recover/bounce back to their normal volume within a few (maybe milliseconds, can't tell) after the bass drum hit. You get this effect of e.g. the guitars volume bumping up and down briefly in line with the bass drum hits. It sounds absolutely awful. If this was the occasional poor skills of some sound engineers then I could live with it, but is seems to become the norm now... (at least in rock & pop)
If they can apply it sympathetically, then I am OK with it, but if they can't then I rather prefer not to have it at all.

I don't have papers on autotune, but I imagine it's the same. Skilled subtle use wouldn't even register for a listener, and the over the top effect has just become a music genre of its own that please some and not others(I would expect a young VS old given what's popular nowadays VS what adults tend to say about it). Maybe it's also just a thing of our time when recording needs to go forever faster, and there is no need for 30 takes of a crappy singer if you're planning to make him sound like the 300 other fake rappers with the same final "voice"(oops my age is starting to show). Ok I just checked a top rappers' list, and now I'm mad.

Much depends on the singer. Some need pitch correction/autotune, let's face it. But there are also examples of very good singers who don't need it at all, yet still get the filters applied because it is industry standard process by now. If done unsympathetically, that can kill the subtle tonal phrasing of those artists that makes their voice so special, and I hate to hear it when that happens.

About your main discussion, it is terribly difficult to say. We're all a mix of experience and predictions. It's hard to argue about a clear reference when anybody is able to recognize a piano for being a piano, no matter which one, who plays, where, or how it's recorded. Our brain has that ability, and it's one that's on autopilot. With training, you can learn to discriminate more, and I'm sure that going to listen to acoustic instruments has some impact on you, your experience, memories, who you are. But won't it also become a matter of liking what you're more used to(maybe some venue you go to more often), instead of recognizing what is "right" or accurately reproduced from some record in some other place with other instruments, musicians, mics, mixing...?
I get your view and sort of agree, but I also cannot fault gregorio's. We're only able to fool ourselves into associating very different sounds because of that ability to associate nearly anything with a vaguely similar envelope, some visual cues or preconceptions telling us what to expect/hear/memories. And from now on, that stuff will be associated to those groups of neurons instead of making a fresh memory, so of course we're going to find a link when that link is now probably real on a "circuit" level.
We can be ultra picky about playback, and we can still recognize and enjoy a catchy tune from a supermarket's PA system. We're the same person doing both.

I can absolutely see the argument from gregorio's side. That makes these discussions interesting; especially since there are a few sound engineers and producers on this forum/thread. It is very useful to get my own preconceptions challenged by hearing the arguments from the other side, esp. from those who work in the industry. Only way to learn really otherwise you get stuck in your own information bubble.

 

[gregorio]

I on the other hand would argue that as experience grows and the baseline widens, you are in a better position to anticipate how instruments will sound different depending on setting and how they are being played.

Surely, as your baseline widens, the range of instrument sounds that could be accurate also widens/increases and therefore, how do you know where in the baseline range is the actual sound intended on the recording, not to mention, what happens if your baseline is still not wide enough? For consumers it can rarely be wide enough and what happens is that they pick a sound within their baseline and compare to that, regardless of whether that’s the actual sound intended on the recording. So now they’re comparing different sound systems to their baseline (or choice within a baseline range) rather than to what is actually in the recording and therefore their judgement of “subjective fidelity” is nothing to do with fidelity but simply a preference that they’ve chosen. And also, it isn’t just “depending on setting and how they’re being played” but also on how and where they’re mic’ed and how they’re processed and that leads us to:

And doesn't the sound engineer also have to make a subjective decision when mixing, as to what sounds "right" and what sounds "wrong"?; don't they need extensive experience of real live instuments' sound for that?

Well exactly, the sound engineer (and producer) make subjective decisions but how do you know what those subjective decisions were? Again, you are assuming what the “real live instruments’ sound” is, and the goal is to achieve a recording which is “right” relative to that sound, but it is not! The aim is not a judgement of right or wrong relative to the original/real live instrument sound, it’s a judgement of “better” or “worse” and commonly or virtually always with rock and any of the popular genres, “better” actually means “more wrong” (less accurate to the actual sound that existed). In fact in many cases the difference is huge, the actual sound of say a drum kit in a studio is vastly different to the (manufactured) sound required in say Metal genres or virtually all electronic music and significantly different even to other types of rock and pop. Even in classical music it can be somewhat different.

While I can’t teach you all about recording and mixing in a single forum post, maybe this one example will help: All mixing desks (inc. software/virtual ones) feature a “Solo” button, which allows us to listen to an individual channel/instrument on it’s own. One thing we have to drill into students is not to spend much time applying processing (EQ, compression or whatever) in solo mode. The reason for this is that you can spend quite some time in solo mode, processing an individual instrument on it’s own to sound exactly “right” but then when you come out of solo mode and combine that instrument with all the other instruments in the mix, it no longer sounds “right”. So the bulk of the processing needs to be done out of solo mode (with the rest of the mix) and the result is that hopefully the instrument sounds “right” in the mix but probably “wrong” on it’s own (solo’ed)! And note again, that “right” and “wrong” in the above refers to closer to or further from what the engineer/producer wants, not more or less accurate to the sound that existed.

Indeed room acoustics matter a lot, but one has to sort the obvious room acoustic problems first before judging speakers.

But that is typically not possible or possible only to a very limited degree. Consumers typically cannot afford an acoustic designer or to loose a substantial percentage of their room volume to acoustic treatments.

And YMMV, but I am quite sensitive to poorly matched spikes between the left and right capsules. Swapping capsules as well as swapping channels makes it quite clear that it is a capsule matching issue, and it can vary a lot between different samples of the same model headphones.

Sure but there you have an obvious reference, the left capsule to the right capsule. But when comparing audio systems with a music recording you are comparing (as you rightly said) the performance of each system based on the subjective decisions of the engineer/producer, which you do not know. So your reference is to what you think it should sound like or what you would prefer it to sound like, not necessarily what it actually does sound like.

In the world of rock and pop music the now ubiquitous heavy-handed use of dynamic compression and autotune/pitch correction are obvious examples. Consumers want it, I **** hate it.

No you don’t, I can absolutely guarantee that if all the compression were removed, it would sound like utter crap and you would hate it. The audiophile community have been saying this for years but it’s just another example of BS invented by those who have little/no idea how recordings are created. Compression is a fundamental tool that allows rock/pop music to exist, it would not exist without it! Unlike an orchestra or established acoustic ensemble, a rock/pop band does not inherently work, a human voice cannot compete (balance) with a rock/pop drum kit or the guitars and the bass guitar does not compete with the lead guitar, etc. Rock/pop evolved in response to the recording technology which enabled it and compression along with EQ are the two most fundamental tools which enable it. The issue is not the use of compression (which is essential) but the over-application of it and that is not as “black and white” as it may appear because the judicious application of compression for one genre (or recording) could be a serious over-application on another or not enough in another. For example, the judicious application of compression for say a 1970’s or 1980’s rock or pop recording would be insufficient for the vast majority of electronic and other genres since the 1990’s but a very serious (very ‘heavy handed”) over-application for almost all classical music recordings. And incidentally, the over-application of compression was the biggest topic of conversation amongst engineers when I first stated getting into the business in 1992, at least a decade or two before the audiophile community even noticed it.

Occasionally I come across a very simple live recording done by a small venue, just one or two mics, without all the consumer-preferred processing, and it sounds so much better to my ears.

That is possible and sometimes it results in excellent recordings. However, it typically results is worse recordings, a recording that could have been better, although of course consumers won’t ever know that, because they don’t know enough about recording/mixing/mastering to realise and they obviously do not have a better recording of that event to compare. The problem with “purist” recording, just using a stereo pair or other minimalist mic’ing, is that you have no option to change anything. You have to get it perfect to start with and if you later hear something that could be better or that you didn’t notice at the time, there’s probably nothing you can do about it. Taking your example of sensitivity to slightly too much or too little reverb, let’s say the engineer or producer notices too little or too much reverb when back in the high quality studio environment after recording in say a church with a poor monitoring environment; with a stereo/minimalist mic’ed recording you’re stuffed, there’s nothing you can do about it. Compare this with more typical mic’ing, one or two close mic’s which will therefore pick up far more direct sound and relatively little or no reverb, in addition to a couple of far more distant “room” mics, which will pick up little or no direct sound and almost entirely the reflections/reverb. Back in the studio you mix all the mics together and can increase or decrease the amount of reverb by increasing or decreasing the balance of the room mics. And this is of course just one example, multi-mic’ing would also allow us to alter the balance, EQ or whatever of individual instruments or groups of instruments, if it wasn’t recorded or performed perfectly at the time. There’s disadvantages too, for example the close mics will record freqs and sounds that would not exist or would not be audible to any audience members (or to a simple stereo pair) and also there’s likely to be some amount of timing/phase issues, so processing will sometimes/often need to be applied to compensate. However, the advantages greatly outweigh the disadvantages, which is why “purist” labels in the past have either failed to compete or competed for a few recordings and then given up and switched multi-mic’ing.

I have often wondered whether that processing (autotune/pitch correction & dynamic compression in particular) is really something that consumers wanted, or whether they have been made to want it and have become so used to hearing it that way that they now insist on it themselves. But that's a different topic.

It is a different topic but it is what consumers want. Particularly for rock and pop genres, consumers would rather have a beautiful male or female lead vocalist that can’t sing in tune very well and therefore needs to be auto-tuned, than an ugly lead singer with such a good voice they do not need auto-tuning. Opera singers are a bit different, they still need to be attractive but not as young and attractive and they’ve had many years of very intensive training which not only gives them excellent intonation and note production but also teaches them to effectively create the same effect as audio compression, which is essential for the acoustic sound projection necessary in quite large venues (such as Opera Houses). Pop/Rock vocalists get almost none of this training, certainly not the decade or more of intensive formal training before they even stand a change of being a professional, and they don’t need much of it anyway, they have mic’s, sound systems and compressors!

but I for one wouldn’t mind paying a few extra $$ for an “audiophile” (hate that term) edition …

That is what some/many SACDs give you. A more heavily compressed CD layer compared to the DSD layer but that’s only with some SACDs, commonly it’s the same master on both layers. The vast majority of us engineers would want that as well, not SACD but a higher quality/less compressed alternative 16/44 “audiophile” version, but budget/time restrictions, especially these days won’t allow it.

G

 

[danadam]

One thing we have to drill into students is not to spend much time applying processing (EQ, compression or whatever) in solo mode. The reason for this is that you can spend quite some time in solo mode, processing an individual instrument on it’s own to sound exactly “right” but then when you come out of solo mode and combine that instrument with all the other instruments in the mix, it no longer sounds “right”. So the bulk of the processing needs to be done out of solo mode (with the rest of the mix) and the result is that hopefully the instrument sounds “right” in the mix but probably “wrong” on it’s own (solo’ed)!

I remember I watched a video (sadly I can't find it, I want to say it was Rick Beato's, but I'm not sure) where they played solo'ed guitar parts from some hits from the past (e.g. one was something from Michael Jackson, I think) and it sounded awful, but it worked in the mix.

 

[2leftears]

Sure but there you have an obvious reference, the left capsule to the right capsule. But when comparing audio systems with a music recording you are comparing (as you rightly said) the performance of each system based on the subjective decisions of the engineer/producer, which you do not know. So your reference is to what you think it should sound like or what you would prefer it to sound like, not necessarily what it actually does sound like.

That is a very good point I hadn't actually considered; the left vs right capsule matching is indeed going to reveal sensitivities of the human hearing much more readily than a comparison against a subjective sound reference would ever do.

No you don’t, I can absolutely guarantee that if all the compression were removed, it would sound like utter crap and you would hate it. The audiophile community have been saying this for years but it’s just another example of BS invented by those who have little/no idea how recordings are created. Compression is a fundamental tool that allows rock/pop music to exist, it would not exist without it! Unlike an orchestra or established acoustic ensemble, a rock/pop band does not inherently work, a human voice cannot compete (balance) with a rock/pop drum kit or the guitars and the bass guitar does not compete with the lead guitar, etc. Rock/pop evolved in response to the recording technology which enabled it and compression along with EQ are the two most fundamental tools which enable it. The issue is not the use of compression (which is essential) but the over-application of it and that is not as “black and white” as it may appear because the judicious application of compression for one genre (or recording) could be a serious over-application on another or not enough in another. For example, the judicious application of compression for say a 1970’s or 1980’s rock or pop recording would be insufficient for the vast majority of electronic and other genres since the 1990’s but a very serious (very ‘heavy handed”) over-application for almost all classical music recordings. And incidentally, the over-application of compression was the biggest topic of conversation amongst engineers when I first stated getting into the business in 1992, at least a decade or two before the audiophile community even noticed it.

I will have to take your word for it. Whenever I hear recordings that sound better to my ears, they have less compression but I cannot tell if they don't have any compression at all. So it does indeed likely come down to over-application.

But your assertion does leave me with a question to ask you. As a sound engineer, do you think live blues-rock and jazz-rock music performances generally sound good or bad? I mean as they are experienced by a member in the live audience, not as experienced as an engineer sitting somewhere behind a mixing desk. And I am referring in particular to very small venue settings, say 50 to 100 audience capacity max. Are compressors then still applied to the instruments during the live performance, so that what the audience hears coming out of the stage speakers is already compressed, or would that generally be uncompressed? In my layman capacity I would assume that an engineer present at the live venue would adjust the mix levels of the various microphones at the start of a song (or maybe at the start of the whole set?) based on settings noted down during a sound check session, so that it sounds good when amped over the stage speakers. But I would also assume (and I may well be wrong about that!) that there would not be any compression involved. Those sessions usually sound pretty good to my ears compared to less dynamic sounding studio versions.

Now, I understand that capturing the live experienced sound on a commercial recording is going to be difficult. Taking multiple mic feeds you can't just add them all together and be done with it. It may be that that is where the disconnect is between audio engineers on the one hand, and audiophiles demanding zero compression on the other. Audio engineers have to work with different feeds from the different mics, some of which are taken very close to the instruments, some of which capture the instruments' sounds, some of which capture more ambient sounds. An audiophile in the audience on the other hand would probably think if they placed a single microphone right here where I am standing with zero post-processing and zero compression it would sound great. But the two simply are incompatible, as I understand it.

That is possible and sometimes it results in excellent recordings. However, it typically results is worse recordings, a recording that could have been better, although of course consumers won’t ever know that, because they don’t know enough about recording/mixing/mastering to realise and they obviously do not have a better recording of that event to compare. The problem with “purist” recording, just using a stereo pair or other minimalist mic’ing, is that you have no option to change anything. You have to get it perfect to start with and if you later hear something that could be better or that you didn’t notice at the time, there’s probably nothing you can do about it. Taking your example of sensitivity to slightly too much or too little reverb, let’s say the engineer or producer notices too little or too much reverb when back in the high quality studio environment after recording in say a church with a poor monitoring environment; with a stereo/minimalist mic’ed recording you’re stuffed, there’s nothing you can do about it. Compare this with more typical mic’ing, one or two close mic’s which will therefore pick up far more direct sound and relatively little or no reverb, in addition to a couple of far more distant “room” mics, which will pick up little or no direct sound and almost entirely the reflections/reverb. Back in the studio you mix all the mics together and can increase or decrease the amount of reverb by increasing or decreasing the balance of the room mics. And this is of course just one example, multi-mic’ing would also allow us to alter the balance, EQ or whatever of individual instruments or groups of instruments, if it wasn’t recorded or performed perfectly at the time. There’s disadvantages too, for example the close mics will record freqs and sounds that would not exist or would not be audible to any audience members (or to a simple stereo pair) and also there’s likely to be some amount of timing/phase issues, so processing will sometimes/often need to be applied to compensate. However, the advantages greatly outweigh the disadvantages, which is why “purist” labels in the past have either failed to compete or competed for a few recordings and then given up and switched multi-mic’ing.

As I alluded to in my previous paragraph, I think that is where there is a disconnect between sound engineers and audiophile music lovers. The latter (myself incl.) will have little understanding about the limitations of just placing a stereo microphone at a "good" position in the audience. Whilst being aware of it, I have e.g. a very incomplete understanding about the sensitivity and directional characteristics of microphones. Placing a stereo microphone at a "good" position in the audience may seem like an obvious thing to do, but it may not be so easy at all. Still, I have heard great recordings done that way, so it is at least possible given the right considerations, even though it might not be practical.

It is a different topic but it is what consumers want. Particularly for rock and pop genres, consumers would rather have a beautiful male or female lead vocalist that can’t sing in tune very well and therefore needs to be auto-tuned, than an ugly lead singer with such a good voice they do not need auto-tuning. Opera singers are a bit different, they still need to be attractive but not as young and attractive and they’ve had many years of very intensive training which not only gives them excellent intonation and note production but also teaches them to effectively create the same effect as audio compression, which is essential for the acoustic sound projection necessary in quite large venues (such as Opera Houses). Pop/Rock vocalists get almost none of this training, certainly not the decade or more of intensive formal training before they even stand a change of being a professional, and they don’t need much of it anyway, they have mic’s, sound systems and compressors!

I can't remember where, but I seem to recall some recent (informal, likely unscientific) research that concluded younger audiences have become less tolerant of singers singing slightly sharp or flat. They find the vocals on older recordings to sound "pitchy" (if I recall correctly). Maybe it is a generational difference. I have recordings where the singers don't always hit the notes spot-on, but I still prefer that to unsympathetically applied or over-applied autotune.

 

[The Jester]

Surely, as your baseline widens, the range of instrument sounds that could be accurate also widens/increases and therefore, how do you know where in the baseline range is the actual sound intended on the recording, not to mention, what happens if your baseline is still not wide enough? For consumers it can rarely be wide enough and what happens is that they pick a sound within their baseline and compare to that, regardless of whether that’s the actual sound intended on the recording. So now they’re comparing different sound systems to their baseline (or choice within a baseline range) rather than to what is actually in the recording and therefore their judgement of “subjective fidelity” is nothing to do with fidelity but simply a preference that they’ve chosen. And also, it isn’t just “depending on setting and how they’re being played” but also on how and where they’re mic’ed and how they’re processed and that leads us to:

Well exactly, the sound engineer (and producer) make subjective decisions but how do you know what those subjective decisions were? Again, you are assuming what the “real live instruments’ sound” is, and the goal is to achieve a recording which is “right” relative to that sound, but it is not! The aim is not a judgement of right or wrong relative to the original/real live instrument sound, it’s a judgement of “better” or “worse” and commonly or virtually always with rock and any of the popular genres, “better” actually means “more wrong” (less accurate to the actual sound that existed). In fact in many cases the difference is huge, the actual sound of say a drum kit in a studio is vastly different to the (manufactured) sound required in say Metal genres or virtually all electronic music and significantly different even to other types of rock and pop. Even in classical music it can be somewhat different.

While I can’t teach you all about recording and mixing in a single forum post, maybe this one example will help: All mixing desks (inc. software/virtual ones) feature a “Solo” button, which allows us to listen to an individual channel/instrument on it’s own. One thing we have to drill into students is not to spend much time applying processing (EQ, compression or whatever) in solo mode. The reason for this is that you can spend quite some time in solo mode, processing an individual instrument on it’s own to sound exactly “right” but then when you come out of solo mode and combine that instrument with all the other instruments in the mix, it no longer sounds “right”. So the bulk of the processing needs to be done out of solo mode (with the rest of the mix) and the result is that hopefully the instrument sounds “right” in the mix but probably “wrong” on it’s own (solo’ed)! And note again, that “right” and “wrong” in the above refers to closer to or further from what the engineer/producer wants, not more or less accurate to the sound that existed.

But that is typically not possible or possible only to a very limited degree. Consumers typically cannot afford an acoustic designer or to loose a substantial percentage of their room volume to acoustic treatments.

Sure but there you have an obvious reference, the left capsule to the right capsule. But when comparing audio systems with a music recording you are comparing (as you rightly said) the performance of each system based on the subjective decisions of the engineer/producer, which you do not know. So your reference is to what you think it should sound like or what you would prefer it to sound like, not necessarily what it actually does sound like.

No you don’t, I can absolutely guarantee that if all the compression were removed, it would sound like utter crap and you would hate it. The audiophile community have been saying this for years but it’s just another example of BS invented by those who have little/no idea how recordings are created. Compression is a fundamental tool that allows rock/pop music to exist, it would not exist without it! Unlike an orchestra or established acoustic ensemble, a rock/pop band does not inherently work, a human voice cannot compete (balance) with a rock/pop drum kit or the guitars and the bass guitar does not compete with the lead guitar, etc. Rock/pop evolved in response to the recording technology which enabled it and compression along with EQ are the two most fundamental tools which enable it. The issue is not the use of compression (which is essential) but the over-application of it and that is not as “black and white” as it may appear because the judicious application of compression for one genre (or recording) could be a serious over-application on another or not enough in another. For example, the judicious application of compression for say a 1970’s or 1980’s rock or pop recording would be insufficient for the vast majority of electronic and other genres since the 1990’s but a very serious (very ‘heavy handed”) over-application for almost all classical music recordings. And incidentally, the over-application of compression was the biggest topic of conversation amongst engineers when I first stated getting into the business in 1992, at least a decade or two before the audiophile community even noticed it.

That is possible and sometimes it results in excellent recordings. However, it typically results is worse recordings, a recording that could have been better, although of course consumers won’t ever know that, because they don’t know enough about recording/mixing/mastering to realise and they obviously do not have a better recording of that event to compare. The problem with “purist” recording, just using a stereo pair or other minimalist mic’ing, is that you have no option to change anything. You have to get it perfect to start with and if you later hear something that could be better or that you didn’t notice at the time, there’s probably nothing you can do about it. Taking your example of sensitivity to slightly too much or too little reverb, let’s say the engineer or producer notices too little or too much reverb when back in the high quality studio environment after recording in say a church with a poor monitoring environment; with a stereo/minimalist mic’ed recording you’re stuffed, there’s nothing you can do about it. Compare this with more typical mic’ing, one or two close mic’s which will therefore pick up far more direct sound and relatively little or no reverb, in addition to a couple of far more distant “room” mics, which will pick up little or no direct sound and almost entirely the reflections/reverb. Back in the studio you mix all the mics together and can increase or decrease the amount of reverb by increasing or decreasing the balance of the room mics. And this is of course just one example, multi-mic’ing would also allow us to alter the balance, EQ or whatever of individual instruments or groups of instruments, if it wasn’t recorded or performed perfectly at the time. There’s disadvantages too, for example the close mics will record freqs and sounds that would not exist or would not be audible to any audience members (or to a simple stereo pair) and also there’s likely to be some amount of timing/phase issues, so processing will sometimes/often need to be applied to compensate. However, the advantages greatly outweigh the disadvantages, which is why “purist” labels in the past have either failed to compete or competed for a few recordings and then given up and switched multi-mic’ing.

It is a different topic but it is what consumers want. Particularly for rock and pop genres, consumers would rather have a beautiful male or female lead vocalist that can’t sing in tune very well and therefore needs to be auto-tuned, than an ugly lead singer with such a good voice they do not need auto-tuning. Opera singers are a bit different, they still need to be attractive but not as young and attractive and they’ve had many years of very intensive training which not only gives them excellent intonation and note production but also teaches them to effectively create the same effect as audio compression, which is essential for the acoustic sound projection necessary in quite large venues (such as Opera Houses). Pop/Rock vocalists get almost none of this training, certainly not the decade or more of intensive formal training before they even stand a change of being a professional, and they don’t need much of it anyway, they have mic’s, sound systems and compressors!

That is what some/many SACDs give you. A more heavily compressed CD layer compared to the DSD layer but that’s only with some SACDs, commonly it’s the same master on both layers. The vast majority of us engineers would want that as well, not SACD but a higher quality/less compressed alternative 16/44 “audiophile” version, but budget/time restrictions, especially these days won’t allow it.

G

For sure,
It’s only when an album is hugely popular is there an opportunity to “cash in” by making a “remastered” edition that can be sometimes better quality .. 16/44.1 CD is fine …don’t really need the HiRes or SACD versions unless they do a decent surround or Atmos remix as well …

 

[bigshot]

Sure. But to what extent are you now also talking about the need to compensate for poor microphone technique and perhaps even vocal arrangements that aren't well thought-out?

None. It doesn’t matter how mics are placed. To achieve optimal balance, you have to mix. The balances are constantly changing all through the song. One mic position wouldn’t get you there.

Maybe I'm showing my age. What you refer to as a clumsy way to solve the problem I (as a music lover and sound engineering layman) would interpret as an experienced singer's technique to balance the timbre of loud singing vs soft singing with the need to have the vocals sufficiently audible against what the rest of band is doing.

I think your confusion comes from never having recorded and mixed music yourself. You have an idea of how it works, but since you’ve never done it, your idea isn’t based on anything.

 

[2leftears]

None. It doesn’t matter how mics are placed. To achieve optimal balance, you have to mix. The balances are constantly changing all through the song. One mic position wouldn’t get you there.



I think your confusion comes from never having recorded and mixed music yourself. You have an idea of how it works, but since you’ve never done it, your idea isn’t based on anything.

I'm confused now. Then what did you mean with "working the mic" in your original post? (I wasn't talking about microphone placement, but about how the singer manipulates/handles the microphone distance). Maybe we are talking cross-purposes?

 

[bigshot]

Working the mic works, but it’s not perfect. On repeated listenings the variations that weren’t corrected would stand out. The only way to fix that is compression and careful level riding. Mic placement wouldn’t correct for frequency overlap at all. A mixer also ducks levels up and down to weave fills into open spaces and to make sure all the sound elements blend rather than compete with each other. Everything is built in a hierarchy. Drums require a lot of adjustments. There are thousands of small corrections performed in a mix, even for classical and jazz, that make it sound better than just sticking a mic in front of the band and hitting the red button.

A mix is organized and properly balanced. You can’t to that without the tools… equalization, compression, reverb, levels, etc.

 

[gregorio]

As a sound engineer, do you think live blues-rock and jazz-rock music performances generally sound good or bad? I mean as they are experienced by a member in the live audience, not as experienced as an engineer sitting somewhere behind a mixing desk. And I am referring in particular to very small venue settings, say 50 to 100 audience capacity max.

I can’t really answer that question, some are good, some aren’t. It also depends on exactly what you mean by those sub-genres you listed. Blues and early jazz were acoustic ensembles, while as I mentioned previously, rock isn’t, it requires mixing/balancing in order to work. Therefore, more mixing/processing is required with rock and relatively little with say blues, particularly in a small venue. So exactly how much with say blues-rock depends on the balance of blues to rock and is also influenced of course by the small venue.

Are compressors then still applied to the instruments during the live performance, so that what the audience hears coming out of the stage speakers is already compressed, or would that generally be uncompressed?

For live performances “yes”, compression would virtually always be applied. Even with opera singers in particularly large venues there would be some compression applied, let alone rock and jazz singers who don’t have that operatic training. There maybe the odd exception, particularly in very small venues but even then it would be the exception rather than the rule.

Audio engineers have to work with different feeds from the different mics, some of which are taken very close to the instruments, some of which capture the instruments' sounds, some of which capture more ambient sounds. An audiophile in the audience on the other hand would probably think if they placed a single microphone right here where I am standing with zero post-processing and zero compression it would sound great. But the two simply are incompatible, as I understand it.

I can answer this and the next quote in one go:

As I alluded to in my previous paragraph, I think that is where there is a disconnect between sound engineers and audiophile music lovers. The latter (myself incl.) will have little understanding about the limitations of just placing a stereo microphone at a "good" position in the audience. Whilst being aware of it, I have e.g. a very incomplete understanding about the sensitivity and directional characteristics of microphones. Placing a stereo microphone at a "good" position in the audience may seem like an obvious thing to do, but it may not be so easy at all.

Yes, most audiophiles do think that you could effectively just put a mic or stereo pair right where they are standing in the audience. This is because they are not only ignorant of how humans perceive sound but also of how mics and acoustics work and of history as well! Stereo had been around for a long time before it was released to the public, the first experiments were done in the 1920’s and stereo theory was formally published in 1932 (by Alan Blumlein) but it wasn’t released to consumers until 1957, 25 years later. There were various reasons for this, including the war of course, but there were stereo recordings being made long before it’s release to consumers and during that time a great deal of experimentation occurred, including the most obvious, just putting up a pair of mic’s at an ideal distance from the ensemble. In fact Blumlein had been doing exactly that with EMI at Abbey Road studios and concert venues for several years starting around 1933. Unfortunately, the audiophile community is great at coming up with “an obvious thing to do” and then believing that numerous actual professional recording engineers, across numerous different countries, were collectively somehow a lot less bright than they are and had not thought of the “obvious”. That’s stereotypical Dunning-Kruger! By the early 1950’s, still about 5 years before stereo was released to the public, some of the more advanced engineers were already working on techniques that solved the issues of “just placing a couple of mics in a good position”!

So, what are those issues? It’s an issue of acoustics and how humans perceive sound, and of the compound effect of both. Where “I am standing” (and let’s assume it’s an ideal position for the sake of argument) is obviously in the audience and the audience in terms of acoustics is a whole bunch of sound absorbing bodies, this significantly reduces the room acoustics (reflections/reverb). In addition, of all animal species, we are particularly good at isolating what we’re concentrating on (called “the cocktail party effect”). In other words, listening to the music at a gig and being somewhat oblivious of other sound/noise which interferes with that. These are all the pieces of the puzzle! Namely, due to human perception we will be somewhat oblivious to the room reverb and the audience noise at that location at a gig, they will appear to be much quieter than they actually are, especially the audience noise. Mics obviously don’t have human perception and will pick up that audience noise very strongly, because they would actually be in the audience and therefore much closer to the audience than to the musicians. We can obviously solve that problem by recording in a studio or venue without any audience. However, not only will the mic still obviously pick up more reverb than we would perceive (due to our perception reducing it) but we’ve now got more reverb because we don’t have an audience absorbing some of it. The obvious solution is not to place the mics in that ideal audience position but a lot closer to the musicians, thereby picking up more of the musicians (the direct sound) and less of the reverb. Problem solved but it introduces a new problem; we’ve now have the mics much closer to some musicians than others, by a much bigger differential than at our ideal audience position, so the balance will be significantly different. Take for example an orchestra: If we put the stereo pair at the right distance to reduce reverb appropriately, that’s most likely to be quite near the conductor (commonly just above) but of course that means the left mic is only a few feet from the front desk violins but many feet (maybe 30) from the back desk violins and the right mic will be only a few feet from the front desk of cellos but double that distance from the back desk of cellos and even further from the basses, resulting in a significant different and inappropriate balance. In addition, we’re likely to have some imaging or balancing issues in the centre position (violas and/or woodwind) depending on the exact positioning/angles between the two mics. This was widely researched leading up to and during in the 1950’s and various optimal setups (mic patterns and angles) discovered, the ORTF pair in 1960 becoming probably the most widely used, but they still had the balancing problem. That’s why more innovative techniques were tried in the early 1950’s, the “Decca Tree” solved the centre image issue but not the balance issue so shortly thereafter “outrigger” mics were introduced; a very wide stereo pair, one placed near the back desk violins and one near the back desk cellos and basses. So already we’re up to 5 mics (3 for the decca tree and 2 outriggers) and this is still just before stereo was released to the public. While it solved the most obvious string balance problem, there were still often an issues with horns, other brass and sometimes percussion. Over time the number of mics increased, as mixing desks could handle more mic channels and we could record more channels, which allowed more fine tuning of the balance with these other instruments as well as more control over the reverb. Today, we typically record an orchestra with 30-50 mics but still almost always use a Decca Tree or stereo pair at the heart of the setup (the main array).

G

 

[2leftears]

I can’t really answer that question, some are good, some aren’t. It also depends on exactly what you mean by those sub-genres you listed. Blues and early jazz were acoustic ensembles, while as I mentioned previously, rock isn’t, it requires mixing/balancing in order to work. Therefore, more mixing/processing is required with rock and relatively little with say blues, particularly in a small venue. So exactly how much with say blues-rock depends on the balance of blues to rock and is also influenced of course by the small venue.

For live performances “yes”, compression would virtually always be applied. Even with opera singers in particularly large venues there would be some compression applied, let alone rock and jazz singers who don’t have that operatic training. There maybe the odd exception, particularly in very small venues but even then it would be the exception rather than the rule.

I can answer this and the next quote in one go:

Yes, most audiophiles do think that you could effectively just put a mic or stereo pair right where they are standing in the audience. This is because they are not only ignorant of how humans perceive sound but also of how mics and acoustics work and of history as well! Stereo had been around for a long time before it was released to the public, the first experiments were done in the 1920’s and stereo theory was formally published in 1932 (by Alan Blumlein) but it wasn’t released to consumers until 1957, 25 years later. There were various reasons for this, including the war of course, but there were stereo recordings being made long before it’s release to consumers and during that time a great deal of experimentation occurred, including the most obvious, just putting up a pair of mic’s at an ideal distance from the ensemble. In fact Blumlein had been doing exactly that with EMI at Abbey Road studios and concert venues for several years starting around 1933. Unfortunately, the audiophile community is great at coming up with “an obvious thing to do” and then believing that numerous actual professional recording engineers, across numerous different countries, were collectively somehow a lot less bright than they are and had not thought of the “obvious”. That’s stereotypical Dunning-Kruger! By the early 1950’s, still about 5 years before stereo was released to the public, some of the more advanced engineers were already working on techniques that solved the issues of “just placing a couple of mics in a good position”!

So, what are those issues? It’s an issue of acoustics and how humans perceive sound, and of the compound effect of both. Where “I am standing” (and let’s assume it’s an ideal position for the sake of argument) is obviously in the audience and the audience in terms of acoustics is a whole bunch of sound absorbing bodies, this significantly reduces the room acoustics (reflections/reverb). In addition, of all animal species, we are particularly good at isolating what we’re concentrating on (called “the cocktail party effect”). In other words, listening to the music at a gig and being somewhat oblivious of other sound/noise which interferes with that. These are all the pieces of the puzzle! Namely, due to human perception we will be somewhat oblivious to the room reverb and the audience noise at that location at a gig, they will appear to be much quieter than they actually are, especially the audience noise. Mics obviously don’t have human perception and will pick up that audience noise very strongly, because they would actually be in the audience and therefore much closer to the audience than to the musicians. We can obviously solve that problem by recording in a studio or venue without any audience. However, not only will the mic still obviously pick up more reverb than we would perceive (due to our perception reducing it) but we’ve now got more reverb because we don’t have an audience absorbing some of it. The obvious solution is not to place the mics in that ideal audience position but a lot closer to the musicians, thereby picking up more of the musicians (the direct sound) and less of the reverb. Problem solved but it introduces a new problem; we’ve now have the mics much closer to some musicians than others, by a much bigger differential than at our ideal audience position, so the balance will be significantly different. Take for example an orchestra: If we put the stereo pair at the right distance to reduce reverb appropriately, that’s most likely to be quite near the conductor (commonly just above) but of course that means the left mic is only a few feet from the front desk violins but many feet (maybe 30) from the back desk violins and the right mic will be only a few feet from the front desk of cellos but double that distance from the back desk of cellos and even further from the basses, resulting in a significant different and inappropriate balance. In addition, we’re likely to have some imaging or balancing issues in the centre position (violas and/or woodwind) depending on the exact positioning/angles between the two mics. This was widely researched leading up to and during in the 1950’s and various optimal setups (mic patterns and angles) discovered, the ORTF pair in 1960 becoming probably the most widely used, but they still had the balancing problem. That’s why more innovative techniques were tried in the early 1950’s, the “Decca Tree” solved the centre image issue but not the balance issue so shortly thereafter “outrigger” mics were introduced; a very wide stereo pair, one placed near the back desk violins and one near the back desk cellos and basses. So already we’re up to 5 mics (3 for the decca tree and 2 outriggers) and this is still just before stereo was released to the public. While it solved the most obvious string balance problem, there were still often an issues with horns, other brass and sometimes percussion. Over time the number of mics increased, as mixing desks could handle more mic channels and we could record more channels, which allowed more fine tuning of the balance with these other instruments as well as more control over the reverb. Today, we typically record an orchestra with 30-50 mics but still almost always use a Decca Tree or stereo pair at the heart of the setup (the main array).

G

Wow, thanks @gregorio! I wan't expecting such a detailed reply; I feel a bit guilty now taking up that much of your time, so I really hope this is useful to others too! Very informative and well-explained

I have noticed myself how perception can be influenced even by something as simple as watching a music video accompanying the music. Watch the MV on-line, and the music sounds great. Extract the exact same audio (or even listen to the MV with the eyes closed), and all of a sudden you notice many more imperfections in the sound. Human cognitive processing is complex...

 

[gregorio]

Wow, thanks @gregorio! I wan't expecting such a detailed reply; I feel a bit guilty now taking up that much of your time, so I really hope this is useful to others too! Very informative and well-explained

No problem, glad it was helpful. TBH, it didn’t take that long, I started and then it all just came out. It really is the most basic/fundamental recording knowledge that every audiophile should know, especially if they are using recordings as reference and/or actually commenting on recording/production techniques.

I have noticed myself how perception can be influenced even by something as simple as watching a music video accompanying the music. Watch the MV on-line, and the music sounds great. Extract the exact same audio (or even listen to the MV with the eyes closed), and all of a sudden you notice many more imperfections in the sound. Human cognitive processing is complex...

Yes, it is complex, in many ways it’s very easy to fool and in others quite difficult. For example, why does our perception filter out the audience noise and some of the reverb at a live gig but not when listening to a recording at home? Presumably the brain recognises you’re in your sitting room, rather than a concert venue and are not surrounded by an audience. So this dissonance/cognitive contradiction keeps us more aware of it than if we were actually at the gig.

G

 

[bigshot]

I’ve heard the opinion that there is a disconnect between engineers and music lovers a lot at Head-Fi and I don’t know what that is based on. All the engineers I’ve worked with love the music they’re working on, and a lot of them are musicians to one degree or another themselves. I’m wondering if the people that say that actually know any engineers.

Music has purposes, and it isn’t all engineered the same. Some music is designed to be played at low volumes in the background on tv shows, and some is designed to be played loud on high end stereos. If engineering is wrong, it’s more likely you’re playing it in a way that wasn’t intended. People have a tendency to think that their own way is what everyone wants, but audiophiles are just part of the marked for recorded music, and they aren’t always typical of the overall market for the music they listen to.

 

[2leftears]

As I alluded to in my previous paragraph, I think that is where there is a disconnect between sound engineers and audiophile music lovers. The latter (myself incl.) will have little understanding about the limitations of just placing a stereo microphone at a "good" position in the audience.

I’ve heard the opinion that there is a disconnect between engineers and music lovers a lot at Head-Fi and I don’t know what that is based on. All the engineers I’ve worked with love the music they’re working on, and a lot of them are musicians to one degree or another themselves. I’m wondering if the people that say that actually know any engineers.

I'm not sure about what others may have meant, but I didn't mean a disconnect in the sense of loving music; I meant a disconnect in the level of understanding about sound engineering. But I guess you think the contrast I made here is a flawed one, which it is indeed. Sound engineers are also music lovers. Let me rephrase it as a disconnect between sound engineers on the one hand, and audiophile music lovers who aren't also sound engineers on the other hand.

My point was that audiophile music lovers who aren't sound engineers will often reason from a limited and (flawed) intuitive understanding (or maybe misunderstanding).