[sonitus mirus]

  I'm experimenting with loudness right now, hard to decide.
 
For what it's worth though, switching on the loudness and lowering the volume appropriately preserves that satisfying punch to the music yet makes the listening less fatiguing. The above post by the purple dragon about the fullness of sound may be onto something, it sure can sound very pleasant. Also the OPs observations about the output coming out of the mixing consoles being similar to loudness-enabled amps is interesting.

All loudness does is typically increase the dB of the lower and higher frequencies.  Usually about 6 dB.  When listening at lower volume levels, due to the Fletcher-Munson curves, an increase in these frequencies is typically beneficial.  Not a big mystery.  

 

[gregorio]

  Also the OPs observations about the output coming out of the mixing consoles being similar to loudness-enabled amps is interesting.

 
Mmm, depends on what you mean by "interesting". What comes out of a mixing console at a live gig depends entirely on the type of music and the sound system and acoustics of the venue. A standard rock gig in a smallish venue for example, would likely require a lot more amplification of the vocals than say most of the drum kit, parts of which may require very little or even none at all, because just the natural acoustic output of say the snare drum maybe enough to create an appropriate balance in the venue without having to feed much of the snare drum through the PA system. The mix coming out of the mixing console would therefore sound very substantially different to what the audience would be hearing, in this oversimplified example, plenty of vocals and almost no snare drum. At a very large venue, the mix out of the console all depends on the characteristics of the sound system, as the acoustic levels/balance produced by the musicians is effectively inconsequential. Although not always the case, the mix coming out of the console at a live gig would typically sound very or somewhat bizarre when listened to on it's own. Furthermore, the high output levels of the PA system in most live gig situations means that it's effectively impossible to know exactly what is coming out of the mixing console anyway, as headphones will only partially isolate the console mix from the PA system. For all these reasons, live sound engineers never (or exceedingly rarely) even bother to listen to the mix coming out of the console.
 
The OP's observations are therefore "interesting" because they don't make any sense!
 
G

 

[LayLay]

 
The OP's observations are therefore "interesting" because they don't make any sense!
 
G

The OP is talking about recording studios, not live shows.
 

  All loudness does is typically increase the dB of the lower and higher frequencies.  Usually about 6 dB.  When listening at lower volume levels, due to the Fletcher-Munson curves, an increase in these frequencies is typically beneficial.  Not a big mystery.  

 
I wish my amp had a loudness pot, not just a switch. I'll experiment now with EQing in some added bass, maybe I'll like that better than loudness.

 

[gregorio]

  The OP is talking about recording studios, not live shows.

 
"I've heard mixes directly out of large, mixing boards' headphone outs at live shows, into headphones such as ATH M50s and even through a pair Shure SRH-840s."
 
G

 

[LayLay]

   
"I've heard mixes directly out of large, mixing boards' headphone outs at live shows, into headphones such as ATH M50s and even through a pair Shure SRH-840s."
 
G

Well, he talks a whole lot. He mentions recording studios a few times and I guess I overlooked that bit.    I concede.
 
P.S. I think I'll be EQing a bit of bass in the end.

 

[Don Hills]

The trouble with the "standard" loudness switch, which added an amount of boost determined by the position of the volume control, was that most (pre)amps were designed for typical sources of the time. "Full volume" (usually onset of clipping) was designed to be about the "2 o'clock" to "3 o'clock" position of the control. The "loudness" tap of the control was around the "11 o'clock" position. Modern sources such as CD players, along with modern overcompressed CDs, meant that "full output" was often reached at 11 or 12 o'clock.  This meant that loudness compensation was engaged for all practical volume settings - too much compensation.
 
There's a fairly simple rule to apply for approximate loudness compensation:
Arrange the compensation so that, at a listening level where the audio sounds "right" or realistic to you, the response is flat. As you reduce the volume below that, progressively boost the bass (and optionally the treble). For every 10 dB you lower the volume setting, boost the lowest bass by about 5 dB and optionaly the treble by a dB or so.
 
You can experiment with this if you have an amp with tone controls. 

Stick a few paper labels on the faceplate so that you can mark temporary scales around the controls with a pencil.
You'll need:
A source of test signals (PC / smartphone generated or a signal generator.)
A voltmeter or other means of measuring the amp output level.
 
You've probably noticed that, for each genre of music, there's a volume level that sounds "right" or realistic to you. So play some music and adjust the volume control until it sounds "right". Attach the meter to the amp output and note the peak levels achieved.
Play a 1 KHz tone and adjust the generator (not the volume control) to about the same level on the meter. Now disconnect the speakers for the sake of your sanity.
 
Mark the position of the volume control.
Now repeat the following steps:
- Play a 1 KHz tone and turn down the volume control by 6 dB on the meter.
- Play a 50 Hz tone and turn up the bass control until the level rises by 3 dB.
- Mark the position of the controls.
Repeat.
 
The aim is to have a set of marks to indicate the correct position of the bass boost for each position of the volume control.
Experiment with more or less bass for each level, and optionally some treble boost, to arrive at a curve that works for you.
A while back I designed a simple analogue circuit that did this automatically, but it needs a 6-gang potentiometer...
I believe Audyssey and some other home theatre processors offer the same function in DSP.
 
Edit:
Corrections made. References to film SPL removed, see my response to Gregorio.
 
The above procedure worked for me (for prog / rock / metal in the days before the loudness war.) YMMV.

 

[gregorio]

[1] The reasoning behind the 85 dB reference level is that it is close to the reference level for audio processing in film, and [2] some mastering engineers advocate it as a standard level for mastering music so as to achieve consistency across mastering houses.

 
There are unfortunately some very significant holes in your reasoning and the "facts" you are presenting:
 
1. 85dBSPL is not a reference, on it's own it's meaningless, to be a reference you actually have to reference 85dBSPL with something. In the case of film audio the actual calibration level is 85dBSPL(C) referenced with pink noise at -20dBFS. You are referencing 85dBSPL with "annoyingly loud", which could be almost anything depending on what an individual considers to be "annoyingly loud" plus, is variable depending on the genre of music as well! For example, with very compressed music (rock, pop, club, etc.) and assuming "annoyingly loud" does equal 85dBSPL, you're likely to be referencing 85dBSPL to somewhere around -5dBSPL. The reference level you're suggesting is not "close to the reference level for audio processing in film", at a rough guess it's probably out by somewhere between 10dB - 25dB!
 
2. Yes, I too have heard of some mastering engineers who advocate calibrating mastering suites to film audio levels. I don't know why though, I assume it's just ignorance of film audio. Firstly, because of the use of compression in music production and mastering, if they did actually calibrate their mastering suite to film audio levels, they would spend much of their working day monitoring at 95-105dBSPL, which is clearly impractical (and technically illegal in many countries)! Secondly, they are missing the point that film audio calibration levels are designed/intended for very large, cinema sized rooms and are specifically NOT recommended for smaller rooms! 85dBSPL at say 60ft from the speakers in a cinema is perceived as substantially quieter than 85dBSPL at say 6ft from the speakers in an average sized music studio (the reason for this isn't fully known/understood AFAIK). If those mastering engineers wanted to pick an existing calibration standard, the TV broadcast standard (78dBSPL = -20dBFS) would have been a far more logical/applicable choice, though still too high for music mastering IMO.
 
Despite all the holes in your facts/reasoning, it's possible that in some cases your suggested method could end up being not so far from moderately accurate, effectively through blind luck 

It's just as likely though, that the compensation method you are suggesting would end up being out by at least double, compared to what was intended. There is a potential solution on the horizon btw but it's likely to be a number of years before it filters down to anything which is accurately actionable by conscientious consumers and even then, only for new music releases.
 
G

 

[Don Hills]

That's what I get for trying to make it simple so as not to scare anyone off trying it.

In general, I agree with you, and I've edited my post to make it even simpler.
 
For the following, bear with me, I'm repeating a lot of stuff you already know, but not everyone here has your level of expertise.

 
On your (1):
The actual film reference is 83 dB SPL at the listening position per speaker, C weighted, using a pink noise source at -20 dBFS. You set your levels, then mix so that the result sounds right - where the loud bits are loud enough to have the intended effect (such as startle you), the vocals are natural and "subtle" cues are still clear. (If the cue is "a pin drop", you may need to mix it "louder than life" to make it audible.)
If you want to hear the result the way the engineer intended it to be, you need to listen at a similar level as they did. I agree that the levels for music may differ from that for film. For the music genres that I listened to back when I was experimenting with the technique (late 80s, with my shiny new Radio Shack analogue SPL meter

), about 85 dB SPL in the room sounded right (and quite loud) to me. (Large speakers, large room.) Over half a dozen friends' systems, the "right" level for the same music varied from 80 to 85 dB SPL. I grant your point that, for the rock/pop genres at least, that may be too high, especially for modern hypercompressed music.
 
On your (2):
 
Bob Katz has done a lot of research into mastering levels in an attempt to bring some science into the process.
He wrote an article published in the AES Journal: http://www.digido.com/how-to-make-better-recordings-part-2.html
He makes your point about relative room sizes, noting that most pre-loudness-war rock/pop was typically mastered about 6 dB below the "film" levels. That is, if you calibrated to the "film" spec then backed off 6 dB, the music usually sounded "right". He codified this in a metering scheme he called the "K-system". It's similar to, and pre-dates, ITU-R BS.1770. 
Some practical examples of this in action can be seen on his "Honor Roll" page: http://www.digido.com/media/honor-roll.html
It lists some albums and the monitor gains, relative to the aforementioned "film" reference level, at which they sound "right".
 
As I originally said, it's only an approximate scheme. It relies on the observation in the equal loudness curves that the perceived loudness at low frequencies varies at about half the rate compared to midrange. Droppping the midrange level by 10 dB results in the same perceived level drop as 5 dB does for the bass. The perceptually correct compensation varies with the dynamic qualities of the music. Apparently the loudness compensation algorithms in Audyssey and other DSP implementations perform dynamic compensation.

 

[gregorio]

 
On your (1):
The actual film reference is 83 dB SPL at the listening position per speaker, C weighted, using a pink noise source at -20 dBFS.

 
No, it's 85dBSPL(C) at the listening position per front speaker, 82dBSPL(C) per rear speaker and the LFE channel at +10dBSPL of in-band gain relative to the front speakers. This is as specified by Dolby, the company who actually make, install and certify cinemas and theatrical dubbing stages. The confusion here is due to how we define and measure "a pink noise source" and therefore most pink noise sources are actually different from each other (are not actually -20dBFS rms). The solution is to use Dolby's own -20dBFS pink noise test files (which are freely available online) and reference each of your speakers as per the theatrical standards above. I re-iterate though, this is the calibration for cinemas/theatrical dub stages and is NOT applicable to a consumer/home listening environment!
 

 
[1] If you want to hear the result the way the engineer intended it to be, you need to listen at a similar level as they did. [2] I agree that the levels for music may differ from that for film. [3] For the music genres that I listened to back when I was experimenting with the technique (late 80s, with my shiny new Radio Shack analogue SPL meter

), about 85 dB SPL in the room sounded right (and quite loud) to me. (Large speakers, large room.) Over half a dozen friends' systems, the "right" level for the same music varied from 80 to 85 dB SPL. [4] I grant your point that, for the rock/pop genres at least, that may be too high, especially for modern hypercompressed music.

 
1. Exactly! My point being, that your suggestion does not result in a similar perceived level as the engineer experienced.
 
2. Yes, massively! Theatrical film mixes typically employ no compression at all.
 
3. Here we run into 3 problems. 1. I obviously don't know what music you were listening to in the late '80's but even rock/pop of the late '80's was employing a significant amount compression. 2. I don't know what test signal or electrical output you were using to reference your 85dBSPL. If it was the theatrical standard (85dBSPL = -20dBFS pink, where -20dBFS 1k tone = 1.228V) it would not have been comfortable or quite loud, it would have been ridiculously loud, depending on what you mean by 3. "Large room". For theatrical standards to be applicable, we are talking of a large room being roughly 100ft x 60ft x 40ft or more, with the listening position being 60ft or so from the front speakers.
 
4. Again, yes! And don't forget that article of Bob's was published 16 years ago.
 

 
Bob Katz has done a lot of research into mastering levels in an attempt to bring some science into the process.

 
Yes he has and he's gained a lot of respect in the industry for doing so. However, Bob's K-20 system (the one you effectively quoted) was not only about bringing science into the process but also to further his personal (IMO, laudable) agenda of addressing the loudness war. Hence the K-20 system which forces a higher than film spec monitoring level (due to small room sizes) and therefore a lower mixing level. In practise, the K-20 system is unusable for music mixing. Practical room sizes for music studios/homes (rather than dub stages) dictate the level should be somewhere around 76dBSPL - 79dBSPL = -20dB pink. Bob's K-system also allows for K-14 and K-12, 6dB and 8dB below film standards respectively, to allow for this but still doesn't address the level of compression employed in practice. Most rock/pop music of the last 20 odd years actually comes out at what would roughly be about K-6, that's about 14dB higher! Bob's "honour roll" is testament to this. Only 3-4 of his honour roll listings actually equate to his K-20 system, all the others are between 1 and 9dB louder and of course there are countless releases which don't make Bob's honour roll because they are more heavily compressed (more than 9dB louder than his K-20 system).
 

 
As I originally said, it's only an approximate scheme. It relies on the observation in the equal loudness curves that the perceived loudness at low frequencies varies at about half the rate compared to midrange. Droppping the midrange level by 10 dB results in the same perceived level drop as 5 dB does for the bass. The perceptually correct compensation varies with the dynamic qualities of the music. Apparently the loudness compensation algorithms in Audyssey and other DSP implementations perform dynamic compensation.

 
I appreciate it was only an approximate scheme but it was too approximate IMHO and just as likely to result in a more significant compensation error than not applying any compensation at all. At least 10dB lower than your suggested calibration levels would also be a very approximate scheme but probably a more accurate one in most cases (IMHO). There is an additional factor here which I haven't mentioned previously and which also has a strong baring on the matter. While I can't speak for all mastering engineers (or those who market themselves as a mastering engineer) I am well aware of the equal loudness contours and so too are those mastering engineers I personally know. While I stick to a specific calibration for the majority of the mastering process, I also monitor/check my master at much lower levels to see how much bass I'm loosing and (briefly) at much higher levels, to see how much bass I'm gaining. I'll then adjust the master accordingly and typically that means having a little more bass than I'd ideally intend at normal calibrated level, on the basis that generally a little too much bass is far preferable to not enough (although it does depend on genre as to the degree). It's not quite this simple in reality though, because I'm often looking at/employing ways to increase the perception of bass without actually increasing the level of the bass. So in addition to the calibration issues mentioned above, I'm also often already applying some of the compensation you are suggesting adding! I've no idea how many other mastering engineers do likewise but I presume a fairly considerable number, if not a significant majority.
 
I understand that tweaking (equipment and settings) is integral to most audiophiles enjoyment of audiophilia and therefore the exchange of some "fairly simple rules" of tweaking is very attractive to many in the audiophile world. Unfortunately though, there are many variables and factors at play which audiophiles do not account for, simply do not realise exist or massively over generalise and which frequently either completely invalidates the "fairly simple rules" or makes them applicable in a far more limited number of scenarios than they assume. I don't doubt that tweaking commonly ends up with a sound which the individual audiophile may personally prefer but rather than getting them closer, it commonly takes them further away from their stated aim of higher fidelity, transparency and/or experiencing the intention of the artists/engineers!
 
G

 

[Don Hills]

   
No, it's 85dBSPL(C) at the listening position per front speaker, 82dBSPL(C) per rear speaker and the LFE channel at +10dBSPL of in-band gain relative to the front speakers. ...

 
What's this new-fangled "5.1" stuff? We only have 2 ears. This is Head-Fi. 

Thanks - I was working from what was explained to me during setup for an ADR session some years back. (I was chaperoning the young actress. 10 years ago now... time flies.)
 

... 3. Here we run into 3 problems. 1. I obviously don't know what music you were listening to in the late '80's but even rock/pop of the late '80's was employing a significant amount compression. 2. I don't know what test signal or electrical output you were using to reference your 85dBSPL. If it was the theatrical standard (85dBSPL = -20dBFS pink, where -20dBFS 1k tone = 1.228V) it would not have been comfortable or quite loud, it would have been ridiculously loud, depending on what you mean by 3. "Large room". For theatrical standards to be applicable, we are talking of a large room being roughly 100ft x 60ft x 40ft or more, with the listening position being 60ft or so from the front speakers.
 

 
1: Individual track compression, yes. 2-bus compression, not so much unless you were Phil Spector or worked for Motown.
2: No test signal, just measuring the in-room SPL of albums like Fleetwood Mac's "Rumours" at a volume which sounded "right". So referenced to 20 micropascals...
3. A 2-storey enclosed courtyard, Speakers with Altec 15 inch drivers. But I acknowledge it was still a lot smaller than the above mentioned dubbing stage.

 
Thanks for the detailed explanations.

 

[gregorio]

  1: Individual track compression, yes. 2-bus compression, not so much unless you were Phil Spector or worked for Motown.
2: No test signal, just measuring the in-room SPL of albums like Fleetwood Mac's "Rumours" at a volume which sounded "right". So referenced to 20 micropascals...
3. A 2-storey enclosed courtyard, Speakers with Altec 15 inch drivers. But I acknowledge it was still a lot smaller than the above mentioned dubbing stage.

 
1. Master-buss compression/limiting was pretty standard practise even in the '70's, obviously no where near as much as today though. Typically in film there is no compression or limiting at all, either on individual tracks or on the master or sub busses!
 
2. It's impossible to know, probably 85dBSPL was somewhere around -10dBFS rather than -20dBFS at a guess.
 
3. Size does matter as far as theatrical specs are concerned but it's not only about size, acoustic characteristics are also important. Controlled/damped reflections are part of the specs, much more damping than is usual in music studios. An enclosed courtyard probably has several times the amount of reflected audio energy than a cinema.
 
G

 

[Don Hills]

   
...  
3. Size does matter as far as theatrical specs are concerned but it's not only about size, acoustic characteristics are also important. Controlled/damped reflections are part of the specs, much more damping than is usual in music studios. An enclosed courtyard probably has several times the amount of reflected audio energy than a cinema.
 
G

 
We had decorative throw rugs hanging from the ceiling, They helped a lot. The sound also changed significantly if we opened all the room doors around the space...