The No.1 Website for Pro Audio
The Titanic, room curves and other GS style OT wanderings
Old 29th August 2018
  #121
Lives for gear
 

self-awareness

Quote:
Originally Posted by Northward View Post
...Read again...
...How is that mod power abuse?...
...if it's just another useless rant...
...I hope that's clear...
do you start seeing a pattern?!
1
Share
Old 29th August 2018
  #122
Gear Addict
 
ReDRuMx's Avatar
After carefully reading all the posts here, and all the linked papers, I think there are 3 main questions:



1. How important is the control room curve?

2. If it is important, which curve is the goal? -> I.e. can we assume we know what the majority of listeners are used to?

3. Are we comparing apples to apples? -> I.e. -> What is the difference between steady-state (long gate) vs. direct sound (short gate) measurements in domestic rooms vs. treated/control rooms.



I'll start with:

3. STEADY STATE / DIRECT SOUND / LARGE ROOMS / DOMESTIC ROOMS / TREATED VS. UNTREATED

After careful examination of the Toole paper (including some references) and my own measurements - my conclusion is this:

EDIT: It seem that this (a) is more wrong than right, but it doesn't really change much in regards to the rest of the post.

a) For any given room, the direct and steady-state curve shape is virtually identical at HF, regardless of room size and absorption properties. If present, the "tilt" downwards is explained by HF air absorption (if slightly larger distances) and off-axis measurements.

I know this goes against what some of you were saying, but I encourage you to double-check the paper, and your own measurements, if you have any of less-treated environments.

See below:

Quote:
Originally Posted by Toole
“… the thinking was that “if a room is tuned with pink noise as a test signal to have a 3 dB per octave slope from 2 kHz, the first-arrival signal will be closer to flat than the 3 dB per octave seen on an analyzer would suggest.“ This morphed into a common belief that seeing the X- curve high-frequency rolloff in steady-state measurements at the 2/3 listening distance ensured a flat direct sound. We now know that is not true with today’s loudspeakers in today’s dubbing stages and cinemas—direct and steady-state sounds above about 1 kHz are essentially identical

Quote:
Originally Posted by Toole
"The SMPTE/ISO method of calibration cannot deliver a flat direct sound. The high-frequency rolloff is identical in both the direct and steady-state sound fields. The low frequency direct sound will exhibit varying amounts of at- tenuation depending on the directivity of the woofers and the reflectivity of the venues—the shaded area. Only in dead rooms will the low-frequency direct sound be approx- imately flat."
Also:


NOTE: The superimposed topmost curves are directivity indices, NOT FR measurements.


So no, HF being equal in steady vs direct measurements is not true just for very treated rooms, it is true almost always, as can be deduced by carefully examining all provided ACTUAL (not only predictions) measurements in the Toole paper! Please do. All is explained by off-axis measurements and air attenuation (assuming flat speaker). Same steady vs. direct.

b) The progressive boost at LF is something that does happen in non (or less) treated environments. Most obvious in very large rooms, because the Schroeder frequency is low.

On the other hand, in smallish rooms this effect is still present, but it's hard to say how useful it is to even consider it while pondering the "curve shape" of home listening environments. And this is why:

Quote:
Originally Posted by Toole
"Below the transition/Schroeder frequency the room resonances and the associated standing waves are the dominant factors in what is measured and heard. These are unique to each room and are strongly location-dependent. Only on-site measurements can reveal what is happening and different loudspeaker and listener locations will result in different bass sound quality and quantity"
I'm sure you know how extremely erratic the LF response is in small, untreated environments. Let me illustrate with an example:



We see relatively typical low-end response for an untreated room. The slope/average curve is in black. Is that really what is always perceived for LF? I see attenuated bass below 110 Hz, with a huge bump at 50 Hz. Sure, for some material this will result in "more" low end, but for some material this curve will de-emphasize the bass.

So, for smaller rooms, it is debatable whether the area below the Schroeder can be considered as being simply "boosted". It most definitely is "messy and boomy", and it is tilted... But my point is that:

Saying domestic rooms exhibit a bass boost is true, but is an oversimplification of what is actually happening

To reiterate - For steady-state measurements -> there is no HF "cut" present only in the steady-state measurement. If it's there, it's there also in the gated measurement. There is a progressive boost towards lower frequencies, most obvious in the diffuse field (above Schroeder), and the amount off boost is a function of room size, reflective surface absorption, and air absorption. Even if I'm wrong about the HF, that doesn't change much, if anything, for further points.

So back to the original question. Can we compare treated room FRs and domestic room FRs, so that we bring the listening experience in one, closer to the other?

IMO, not really. Perhaps in the non-modal region. I'll further explain why in the follwing point:



2. WHAT IS THE HOUSE CURVE GOAL? CAN WE ASSUME WE KNOW WHAT THE MAJORITY OF LISTENERS ARE USED TO?

I would say most definitely not, I completely agree with the following:


Quote:
Originally Posted by scraggs View Post
whether the room has a 'flat' curve or your beloved B&K curve, the end user listening experience is an impossibly wide moving target.

cars, Beats, iphone speakers, laptop speakers, horrible Bose garbage with ridiculous bass, etc etc. and maybe there are some people out there who actually have a decent 'home system', but for sure it's in a room with no treatment and the FR is going to be all over the place.

How can we possibly figure out what's the "standard" from all this? One thing is certain, the following chart is not the answer:



If anything, this shows that there is a discrepancy between what pros prefer, and what casual listeners prefer as their playback environment.

But it doesn't tell us what the average listening environment is! And that is the main question here. Understanding that there is a difference between "taste in playback FR" and "actual average playback FR" is crucial.

Thinking that a "Hi-Fi listening room" study (B&K) from the 70's is somehow representative of the current "standard listening environment (which includes headphones, laptops...)" is, IMO, wrong (DanDan sorry to use your least favourite word ).

As a side-note, it is more than obvious, to anyone who carefully read the Toole paper, that even the famous cinema X-Curve is an obvious mistake! But ironically, it is kind of useful because dubbing stages mostly use it also... So both the cinema goer and the mix engineer hear the same "wrong" thing. I.e., - a standard. Funny!

Quote:
Originally Posted by Toole
If a flat direct sound was the objective in calibrating cinemas, the attempt failed.
Final question, perhaps the most important:



1. HOW IMPORTANT IS THE CONTROL ROOM CURVE AT ALL:

For all the reasons above, I agree that it is most likely not very important, and mostly comes down to taste. Of course, within reason. You don't want a certain frequency range boosted by 20 dB, so that it masks everything else. I still think "neutrality" (as in; no huge irregularities in FR), controlled decay times, and good stereo imaging are more important than the actual curve.

In conclusion, DanDan, to answer your question:

Quote:
Originally Posted by DanDan View Post

Can you offer a view on how a desired/achieved tonal balance established anechoically could possibly remain the same when listened to in a 6-10dB dulled environment?

Because (assuming a treated room), you can listen to your references before mixing, and calibrate your brain. The music of the past is your "control", your "taste", call it "style"; and it establishes the "frame" within which you draw (i.e. - mix). My opinion is that with proper references and an acclimatisation period, the FR becomes less important (within reason, of course).


The shape of the FR can be another personal "taste", but this layer is superimposed to what I desribed above. I.e., the FR doesn't change your style. But only if you have calibrated your brain properly.

On the other hand, if someone works a lot in various studios, and doesn't like using references... I can see how lack of standards would be a problem. Lucky movie sound guys with their flawed X-Curve (sarcasm alert)!

Also, I can totally understand that there are people who, for whichever reason, simply cannot stand a specific house curve. But claiming that somehow this should be true for everybody, and is a biological truth - is weird. Stubborn if you will.

Since you all are counting stripes, I went to check my uniform. I'm from a small country, so no Billboard top 100, sorry. But I earn a living from working in audio, and have mixed 800 songs in the past 6 years (some also recorded, some also produced). And I like my more-less flattish HF curve in my well treated room.

Finally, thanks to everyone for you contributions here, I really have learned a lot from this exchange. And it is really saddening to see the unnecessary bad vibes created. But that's the internet I guess, could be worse

P.S. - I would appreciate it, if you (anyone) find a flaw in my reasoning, to point it out. I'm trying hard to figure out what's "real" here, and would rather know the truth than (falsely) think I'm right.
Attached Thumbnails
The Titanic, room curves and other GS style OT wanderings-cccc.png   The Titanic, room curves and other GS style OT wanderings-screen-shot-2018-08-29-18.45.11.png   The Titanic, room curves and other GS style OT wanderings-screen-shot-2018-08-29-02.04.21.png  

Last edited by ReDRuMx; 29th August 2018 at 11:21 PM..
3
Share
Old 29th August 2018
  #123
Lives for gear
Quote:
Originally Posted by Deleted 4adc64a View Post
I'm not exaggerating when I say I found instances of you fighting in every single topic I researched while building my studio -- VPRs, ISD gap, diffusion, etc. And wading through that was pretty annoying to read
yep. i was on here for hours on end doing research before starting my new room. so many threads, so much needless arguing.

but i'm super thankful for the people here who do share helpful information. this place is a goldmine if you dig.

those people probably know who they are. if i'm ever fortunate enough to meet any of you, i hope you like bourbon.

anyway if i can get in a couple questions for Thomas before the ax falls on this:

1. how much attention do you pay to the FR above say 500hz? i ask because as i said earlier, simply moving my empty chair would change the measurement. and also, according to the graph, i have a general dip in the response between about 1 and 3k. but it doesn't sound like that at all, i've had the same speakers for 12 years and they sound the way they always have, just way better and more detailed. and actually any hi mid harshness jumps right out of the speakers and is really obvious.

so i don't think this is a problem at all, other than i'd like my graph to look prettier, just curious to hear your thoughts on this.

2. is off topic, but i'm not sure what the topic even was, other than a vehicle for dan to argue stuff he doesn't have any direct experience with, so here goes: in your rooms with the glass front walls, like J. LaPointe's, how do you handle the front wall trapping? is the lounge on the other side of the glass serving as a big bass trap? and/or there's stuff (secret or not) in the walls of that room?

thanks!
2
Share
Old 29th August 2018
  #124
Lives for gear
The room is not the majority area for the listening of the music.

First The headphone and the car.
4
Share
Old 29th August 2018
  #125
Lives for gear
 
Jens Eklund's Avatar
I think that anyone who reads post 91 and 112 quickly understands that some people are used to a FR (direct) that is not flat, and as long as these persons are aware of that fact, it´s perfectly fine.

It would however be very sad if the music industry decided to repeat the same mistake the movie industry did (the X-curve, that hopefully will slowly die away eventually thanks to Toole and others).

The standard (in the music industry at least) is a flat direct sound, as stated by Toole in the 2015 AES paper linked to in previous post (if not, every speaker maker is fighting an uphill battle trying their best to make their products measure flat on axis …) but if you’re accustom to a HF tilt; by all means use it, but be aware of what you do and also be prepared to readjust if you suddenly need to work with others. I have measured a lot of studios (not only the once I´ve designed) and they all measure flat (again; direct sound above about 1 kHz). The only time I´ve seen a tilt was in a home cinema situation … and that kind of makes sense; he was probably use to the X-curve.
4
Share
Old 29th August 2018
  #126
Lives for gear
 
DirkB's Avatar
Having worked for years in the typical bedroom/garage type of GS room and now more than 5 years and many dozens of mixes later in my proudly owned NA FTB CR I will add my biased $0.02 to the discussion.

State whatever you like, but the difference both in the final product and the road towards that final product in a CR where the response is void of all kind of lowend modal interference vs. your typically non treated room in the lowend part of the frequency spectrum is night and day.

Secondly, to get this flat respons in the lowend I know both from theory as well as having experienced the science (which I built myself completely) in praxis requires two things:
- a scientific design
- substantial space to implement the design.

Most professional studios (and like really "top of the spectrum" ones - that I visited myself) lack both and sound like sh$t in the CR.

Gr.
Dirk

Last edited by DirkB; 30th August 2018 at 09:54 AM..
4
Share
Old 29th August 2018
  #127
Moderator
 
Northward's Avatar
Quote:
Originally Posted by ReDRuMx View Post
After carefully reading all the posts here, and all the linked papers, I think there are 3 main questions:

1. How important is the control room curve?

2. If it is important, which curve is the goal? -> I.e. can we assume we know what the majority of listeners are used to?

3. Are we comparing apples to apples? -> I.e. -> What is the difference between steady-state (long gate) vs. direct sound (short gate) measurements in domestic rooms vs. treated/control rooms.
Thanks a lot for all the time you put into this!

I have to do a calibration/certification before end of the week, so I will have my measurement system at home with me when I come back in the evening. I'll keep it at home and document a couple things over the week-end. I live in a loft, very reflective and large space with high ceilings, no treatment whatosever, and lucky enough to have a pair of ATC SCM40 Active Hifi speakers in there - pretty solid reference.

I will do a proper, by the book old school Pink Noise/RTA measurement, loading the room properly, and a standard sweep measurement.

What this will *most probably* show you is that there is indeed a difference between the two... We'll see! I remember it being a lot more obvious that way. I used to use a pink noise + RTA system for some non-studio tasks in the early years.

Speaker directivity at HF has an influence on the results too. The ATC are wide dispersion.

That today there is no difference in much better treated modern Film dubbing stages is not a surprise. I have done only sporadic consulting for film dubbing stages and companies working in that field, so I don't have a lot of data about these.
1
Share
Old 29th August 2018
  #128
Gear Addict
 
ReDRuMx's Avatar
Quote:
Originally Posted by Northward View Post
Thanks a lot for all the time you put into this!
Hahah, writing is the best way of organising one's thoughts.

Anyway, I think that this "difference or no difference" at HF for domestic rooms for gated vs. non-gated measurements is (in context with other things I wrote) not that important.


Quote:
Originally Posted by Northward View Post
I have to do a calibration/certification before end of the week, so I will have my measurement system at home with me when I come back in the evening
Awesome, please share the results! If indeed the difference is obvious only by using the room-loading method, and not when using the sine sweep method, then we've shown (at least) that people need to be a lot more careful when equating the two methods (sine sweep long gate, and RTA + room loading). Still an interesting conclusion.

We (at least I) could also learn that one shouldn't expect a difference between gated and direct sound at HF when using modern FR measurement methods (i.e. - sine sweep). Cool.

Quote:
Originally Posted by Northward View Post
I live in a loft, very reflective and large space with high ceilings, no treatment whatosever,
Hopefully not too large and too reflective, that would make it non-average


Quote:
Originally Posted by Northward View Post
That today there is no difference in much better treated modern Film dubbing stages is not a surprise
Agreed, but the Toole paper shows "normal rooms" also:


Last edited by ReDRuMx; 29th August 2018 at 10:54 PM..
Old 29th August 2018
  #129
Lives for gear
 
Jens Eklund's Avatar
The Titanic, room curves and other GS style OT wanderings-gated-direct-vs-steady-state-un-gated-fr-untreated-room.-green-steady-state.png

Gated (direct) vs steady state (un-gated) FR in untreated room. Green is steady state. Speaker is 8250A (no DSP in use). Mic might not be exactly on axis but pretty close.


Another room, same condition as above:

The Titanic, room curves and other GS style OT wanderings-gated-direct-vs-steady-state-un-gated-fr-untreated-room.-green-steady-state-2.png
Attached Thumbnails
The Titanic, room curves and other GS style OT wanderings-gated-direct-vs-steady-state-un-gated-fr-untreated-room.-green-steady-state.png   The Titanic, room curves and other GS style OT wanderings-gated-direct-vs-steady-state-un-gated-fr-untreated-room.-green-steady-state-2.png  
4
Share
Old 29th August 2018
  #130
Gear Addict
 
ReDRuMx's Avatar
Quote:
Originally Posted by Jens Eklund View Post

Gated (direct) vs steady state (un-gated) FR in untreated room. Green is steady state. Speaker is 8250A (no DSP in use). Mic might not be exactly on axis but pretty close.
Great, thanks for this! We need more Is this RTA or sine sweep? Room size?
Old 29th August 2018
  #131
Moderator
 
Northward's Avatar
Quote:
Originally Posted by scraggs View Post
anyway if i can get in a couple questions for Thomas before the ax falls on this:

1. how much attention do you pay to the FR above say 500hz? i ask because as i said earlier, simply moving my empty chair would change the measurement. and also, according to the graph, i have a general dip in the response between about 1 and 3k. but it doesn't sound like that at all, i've had the same speakers for 12 years and they sound the way they always have, just way better and more detailed. and actually any hi mid harshness jumps right out of the speakers and is really obvious.

so i don't think this is a problem at all, other than i'd like my graph to look prettier, just curious to hear your thoughts on this.
If you witness a room measurement (which always happens after a first critical listening session) you will see that the first thing I'm literally sweating about and can't wait to see is the ETC data and then look at the 20-200Hz frequency response. If that's where it's supposed to be - and I always double check the floor effect too - because of how these rooms work the rest is de facto OK too. So I'll concentrate on that first, and then later move up on the FR.

If the room is empty of furniture and equipment (which is always the case when we do a room certification) you will see some fairly gentle comb filtering happen in the MF, due to the floor. That's normal and it will change as you slightly move the mic around. The dip you have may be due to desk reflections. At least that would be the first thing I'd look into. But if it's not really an audible problem... Moving a chair near the mic definitely influences the measurement. In critical measurements, I will delay the sweep start and go hide in the studio's entrance door frame to prevent my body from altering the measurement by being too close to the mic.

So LF>MF>HF when measuring a room.

Quote:
Originally Posted by scraggs View Post
2. is off topic, but i'm not sure what the topic even was, other than a vehicle for dan to argue stuff he doesn't have any direct experience with, so here goes: in your rooms with the glass front walls, like J. LaPointe's, how do you handle the front wall trapping? is the lounge on the other side of the glass serving as a big bass trap? and/or there's stuff (secret or not) in the walls of that room?

thanks!
Ha!

The glass front wall basically being acoustically an extension of the ATC front baffle, in itself all it needs to be is very dense and non-resonant / dampened - and it also acts as a soundproofing interface between the two spaces. So the glass is really heavy and thick. I'm sure J still has fond memories of the install day and all the weight lifting. The detail that makes it work 100% is that speakers are fully decoupled in the nacelles you can see on photos, which are suspended on cables who are then are attached to a 2 axis push-pull spring system with an added dummy load to optimize the tension and natural frequency of the system in the Z axis, and a pre-constraint spring system in the Y axis. The outer box of the nacelles is simply holding in the glass and seals around the secondary box that is the one that's decoupled. So no vibrations are mechanically transmitted to the glass or the room. The very deep full range bass trap in the back (and side walls and ceiling) of the Mastering suite makes sure that basically no energy is bouncing back to glass wall at any point in time, it's a one way trip from speakers to any other surface (except floor).

You got it 100% right re: the lounge. It's a bass trap for whatever LF energy is still radiating from the back of the decoupling nacelles/speakers. But if you were to stand behind the nacelles in the lounge while J is playing music at high SPL, you'd realize that in fact not so much energy radiates anyway. While in the Mastering suite you'd have to scream to have a conversation, in the lounge, you could just speak with a normal voice. This is due to the nacelles being very heavy, using 2 thick layers of dense material. So it kills a lot of the residual LF from the get go. Which in turn makes my life a lot easier as it means I don't have to use heavy artillery in the lounge - unlike the bass traps in the Mastering suite itself which are very substantial. Backwall trap is often over 1.2m deep.

Down the line all I need to do in the lounge is control the modal range a bit, widen the Q of any resonance to tame it. From this point it is so attenuated on the way back to the glass wall, that it simply can't make it back through the heavy glass or 56 or 59dB R(W) door (can't remember which was used there). Even without the treatment it would not make it back. But we don't take chances.
3
Share
Old 29th August 2018
  #132
Moderator
 
Northward's Avatar
Quote:
Originally Posted by ReDRuMx View Post
Anyway, I think that this "difference or no difference" at HF for domestic rooms for gated vs. non-gated measurements is (in context with other things I wrote) not that important. But I really would be surprised if the differences were even remotely close to the B&K curve.

Awesome, please share the results! If indeed the difference is obvious only by using the room-loading method, and not when using the sine sweep method, then we've shown (at least) that people need to be a lot more careful when equating the two methods (sine sweep long gate, and RTA + room loading). Still an interesting conclusion.

We (at least I) could also learn that one shouldn't expect a difference between gated and direct sound at HF when using modern FR measurement methods (i.e. - sine sweep). Cool.
I do suspect that the old school method will yield a more obvious difference yes. Devil is in the detail.

We'll see! If it's the same: drinks on me!

PS: Jens beat me to it! I'll still do it, but expect the same results.

Last edited by Northward; 29th August 2018 at 09:06 PM.. Reason: added PS
1
Share
Old 29th August 2018
  #133
Lives for gear
 
Jens Eklund's Avatar
Quote:
Originally Posted by ReDRuMx View Post
Great, thanks for this! We need more Is this RTA or sine sweep? Room size?
Single channel, single 265k log sweep by ARTA.

Room size was 29,7 m2 for the first room and about 20 m2 for the second.


As an interesting side note; look at the second graph; guess the x-over for this speaker ... yes, 1,8 kHz. This demonstrates the effect of directivity on the steady state responce in an untreated (or poorly treated) room.

Last edited by Jens Eklund; 29th August 2018 at 10:03 PM.. Reason: Clarification
4
Share
Old 29th August 2018
  #134
Lives for gear
Thomas,

awesome, thank you! i figured that glass front wall must be super thick, and that is a really clever system with the nacelles.

Installing that glass seems like....not for the faint of heart.

Quote:
Originally Posted by Northward View Post
The dip you have may be due to desk reflections. At least that would be the first thing I'd look into. But if it's not really an audible problem...
i'm actually deskless. just a small rack to my left and a little mouse/coffeecup stand to my right. so who knows. as i said it doesn't sound like it looks and i'm within a reasonable +/- range, so...finish the fabric and take a picture already.

questions beget more questions, so if you don't mind:

1. can you talk about floor effect a little more?

2. way way off topic, but...clearly your rooms have a pretty specific design. how much does it change based on the location of the room? i.e. a mastering room in a standalone building out in the middle of nowhere in the country vs something like what they're currently doing at The Bunker, where they have 3 rooms right next to each other in a building in the city.

the latter obviously has way higher isolation requirements than the former, which i would imagine would then require much more treatment...yes/no/maybe/it depends?
Old 29th August 2018
  #135
Lives for gear
 
Jens Eklund's Avatar
Time to move this to a new thread perhaps; OT discussion from the thread OT discussion ...
2
Share
Old 29th August 2018
  #136
Gear Maniac
 

Despite the bickering, this thread should be a must read for those of us non-acoustician musicians who find acoustic design incredibly fascinating. Great to hear more insight into FTB.

DanDan has referenced Boggy quite a bit; I'd love to hear his opinion on this issue, as he represents another very different way of skinning a cat.

I think it is also important to note, and this is just my opinion, but different designs make sense for different purposes. If I was a top level mastering engineer, I unquestionably would use a FTB design. If I was a recording or mix engineer of real music played by humans on real instruments, I would no doubt go with a true RFZ design with Haas Kicker and ISD termination. As a producer who uses his room for everything from composing to recording to mixing to mastering to film/tv scoring, and who may need to expand to surround, I am going with the MyRoom concept.

However, all of these designs, despite differences, are (in a very basic way) fundamentally achieving the same result. Direct sound to listener unimpeded by the reflective surfaces, and modal control. So for those of us who are not acousticians and who are DIY oriented, I think any approach other than achieving those two fundamental goals is a waste of money and time. EQing to a certain curve while putting up insufficient porous absorbers, often via commercial "bass traps" does absolutely nothing to achieve those goals. Just my opinion.
8
Share
Old 29th August 2018
  #137
Lives for gear
 
Jens Eklund's Avatar
Quote:
Originally Posted by Boxtone View Post
However, all of these designs, despite differences, are (in a very basic way) fundamentally achieving the same result. Direct sound to listener unimpeded by the reflective surfaces, and modal control. So for those of us who are not acousticians and who are DIY oriented, I think any approach other than achieving those two fundamental goals is a waste of money and time. EQing to a certain curve while putting up insufficient porous absorbers, often via commercial "bass traps" does absolutely nothing to achieve those goals. Just my opinion.
+1
Old 29th August 2018
  #138
Gear Addict
 
ReDRuMx's Avatar
Quote:
Originally Posted by Jens Eklund View Post


Gated (direct) vs steady state (un-gated) FR in untreated room. Green is steady state. Speaker is 8250A (no DSP in use). Mic might not be exactly on axis but pretty close.


Another room, same condition as above:


The first measurement looks very similar to the prediction from the Toole paper, but the second is pretty much the B&K curve, definitely.

If I was to judge by only these two measurements, I do admit that it seems I was more wrong than right about HF being same in steady vs direct measurements. But I think that it is also very obvious that it depends a lot on the directional characteristics of the speaker.

Two things that confused me were: a) The curve I posted from the Toole paper (flat steady-state HF prediction). This one:



b) Some of my own measurements where I don't see a difference. But, to be honest, if I had to choose which measurements were more valid, mine or yours, it would be quite cocky of me to presume you did something wrong. It's much more probable I did But I will still explore this a bit more

In any case, this HF thing is not crucial for my original post, but it is still very interesting. And in case it wasn't clear, I completely understand that a flat direct curve is suggested. Nothing I wrote was intended to contradict that.

What's more important to me is the low end.. Check out the sub 100 Hz area on both of your measurements.... Would you call this "boosted", as the smoothed curves would probably suggest? Again, I think calling this a "boost" is not really correct. It'd better to call it a boomy mess

Anyway, thank you very much for exploring this, I really appreciate everyone taking the time.

Last edited by ReDRuMx; 29th August 2018 at 11:45 PM..
Old 30th August 2018
  #139
Gear Guru
Out The Gate

Olive or Toole, and indeed anybody, not use short gated measurements?
The Lingua Franca is the normal gating.
Because that represents the HEARD sound.
A great loudspeaker will broadcast a Flat Direct Field over an angle of about 30 degrees only. This 'wedge' of pure sound will not travel far before the Room Tone overwhelms it. Let's say less than 1 Metre.
So only a solitary soul placed on axis and less than 1 M from both speakers could possible hear the On Axis Direct Field. Everyone else hears is a composite of DF and RT, and Eq if applied.
The measurements and graphs presented by those Scientists represent the sound HEARD by the majority of domestic listeners.
Rigorous science aside, this phenomenon is easily audible by anyone in a domestic room with a pair of speakers. i.e. In extremely close they sound remarkably brighter.
People that want to measure the OADF response of Loudspeakers use Anechoic Chambers, or Outdoors, up on poles.

Now if someone treated their room as much as is reasonable domestically, or prosumer, and happily sat in just one seat which experiences the OADF, they would find it necessary to tame the top and boost the bottom a little. To hear music as it was intended to be heard, but with much clarity due to the treatment.
jim1961's Listening Room is such. A reasonably well treated room, one listener position, seems remarkably like an average DIY CR to me. The Listening Curve also looks remarkably like those in Olive, BK, Harman, etc. etc.
My Listening Room

Red, you are both welcome and thank you. If you wish to explore further with me please use direct email, as on my websites.
Ditto any other truth seekers.


DD
Attached Thumbnails
The Titanic, room curves and other GS style OT wanderings-jim.jpg  

Last edited by DanDan; 30th August 2018 at 12:42 PM..
Old 30th August 2018
  #140
Lives for gear
 
Jens Eklund's Avatar
It is clear that you still don't understand how tne auditory system works ... your post above proves that. Please read the posts by Thomas and Mike again ... or any paper on the topic.

The direct sound is key (and again to stay clear from misunderstandings; talking about mid and high frequency range).

Also, do note that the FR you think looks good have almost no tilt above 1 kHz ...

Last edited by Jens Eklund; 30th August 2018 at 06:59 AM.. Reason: Clarification
3
Share
Old 30th August 2018
  #141
Lives for gear
 
Jens Eklund's Avatar
Quote:
Originally Posted by ReDRuMx View Post
What's more important to me is the low end.. Check out the sub 100 Hz area on both of your measurements.... Would you call this "boosted", as the smoothed curves would probably suggest? Again, I think calling this a "boost" is not really correct. It'd better to call it a boomy mess

Anyway, thank you very much for exploring this, I really appreciate everyone taking the time.
Remember that these are measurements from completely untreated rooms. The low end will be a mess.

In a treated room, the low end will be controlled but there will still be a slight “boost” in the low modal range, how much depends on the level of treatment.

What ´s important to understand is that for the upper range, it makes no sense to look at the steady state when doing corrections since it´s the direct sound that governs how we perceive timbre in the upper frequency range.

The historical use of steady state response to shape the target curve is the source of the problem:

Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms - Floyd E. Toole - Google Bocker

In a well-treated control room, the steady state and gated measurement won´t differ (again; in the upper range), but in an un-treated or poorly treated room; they will. Regardless, if using the gated direct sound when shaping the upper frequency range; you´ll get it right (assuming you make in flat …), so it makes no sense to look at the steady state response when dealing with the upper range.

The lower end is another story since we cannot distinguish the direct sound from room contribution and thus a slight boost is not incorrect to allow since this is what will happen if you were to measure a lot of rooms with a lot of different speakers (measuring flat in anechoic condition as most good speakers do) and made an average of all these measurements.

So what I do when making corrections, I make sure that the direct sound (gated) is flat in the range above about 1 kHz, and for the lower range I use the steady state (or I actually often use a dual gate in order to se both at once (gated for upper range and steady state for the lower part) and apply the tilt that the client feels comfortable with, usually a +3 to +7 dB boost at 30 Hz compared to the upper range. This bass tilt is gradual and not a sudden jump at a specific frequency. Not very different from the tilt seen in the FR in jim1961's room shown above.
5
Share
Old 30th August 2018
  #142
Gear Addict
 
ReDRuMx's Avatar
Quote:
Originally Posted by Jens Eklund View Post
Remember that these are measurements from completely untreated rooms. The low end will be a mess.

In a treated room, the low end will be controlled but there will still be a slight “boost” in the low modal range, how much depends on the level of treatment.
My point was only that it's an oversimplification to say that untreated rooms exhibit a "boost" in the LF, because it's a bit more complicated than that. It's at least useful to understand how that "boost" behaves - especially when one is comparing treated to untreated rooms.

Everything else you wrote I understand and agree with. Also, thanks for sharing your fine-tuning procedure / experiences!
Old 30th August 2018
  #143
Lives for gear
 
Jens Eklund's Avatar
Quote:
Originally Posted by ReDRuMx View Post
My point was only that it's an oversimplification to say that untreated rooms exhibit a "boost" in the LF, because it's a bit more complicated than that, as I wrote in my original post.
Absolutely: in a specific room, there will be boosts at some modal frequencies and dips at others. But as stated in previous post, if looking at the average of many different measurements; a general boost in the modal range is expected.
2
Share
Old 30th August 2018
  #144
Moderator
 
Northward's Avatar
Quote:
Originally Posted by scraggs View Post
1. can you talk about floor effect a little more?
It's the last ER that remains in the room, especially in rooms where there is little or no furniture to mitigate it a bit, or furniture designed to have a minimal acoustic print.
What you will measure at the sweet spot is the delta of (/the difference of) distance between direct speaker sound and the geometrically reflected sound off the floor, happening roughly about half way between sweet spot and engineer.

In a good empty room, it will show up as a narrow Q, deep hole in the FR. Usually between 120Hz and 200Hz, with another one, much less pronounced with a slightly wider Q at harmonic. As soon as there is some furniture in there, it tends to be less deep and have a wider Q.

You can easily spot it on a full range ETC. From there you can check software measured vs. physically measured distance in the room and correlate to confirm.

See Graphs attached.

As you move forward towards the speakers it will lower in frequency. As you move away, it will go higher in frequency. The type of speaker used and its height plays a large role in where the hole will be frequency wise. Single or dual woofer will also make a difference (vertical or horizontal dual woofers etc).

What's difficult to assess with this phenomenon is if whether or not it's going to be audible, especially since moving your head / mic a bit will drastically change the effect. In the case of the studio in the attached graphs, it's not audible. In other studios it is. I can't yet explain exactly why that is.

The other issue is when we decide to take care of that reflection (in the few cases where it is audible). The graph flattens, the ETC sees a much reduced spike. But as it turns out, every time we got rid of that reflection a few days or weeks later, the engineer would call us and say he's removed the treatment because it sounded somehow a bit off, even if the graph looked better and at first it seemed like a small improvement.

It's also hard to explain clearly what happens there, but current consensus is that the floor plays a major role in how we perceive spaces and very particular cues, which for the engineers may down the line be more important than the FR impact of the effect:

On the design of canonical sound localization environments
Eric J. Angel1,2, V. Ralph Algazi1, and Richard O. Duda


"Several key research papers served as motivation for the design of the canonical localization environments. These papers focus primarily on the perceptual effects of reverberation and head tracking.
Hartmann [8] and Rakerd and Hartmann [2] studied the effect of early reflections on azimuth localization accuracy using acoustically adjustable rooms. They found that reflection from the floor and ceiling gave more accurate azimuth judgments for sounds in the horizontal plane, while lateral early reflections blurred the location of the source."

"With substantial evidence that lateral reflections may be detrimental to azimuth localization, and following the work of Hartmann and Rakerd who found that a "room" comprised solely of a physical floor degraded localization accuracy by only a modest amount, we concluded that for a virtual environment the externalization provided by a floor reflection would be of value to localization. Thus a room with only a virtual floor was chosen as the basis for a new canonical localization environment. Such a floor would reinforce the direct sound azimuth cue and may not be detrimental to elevation localization."

" 7. DISCUSSION AND CONCLUSIONS
The results reported lead to the following conclusions with respect to canonical localization environments for a broadband noise stimulus:
1. When head tracking is not possible, an environment with a single floor reflection is a canonical environment in that it reduces reversal rates by more that 40%, decreases the bias of azimuth errors on the average by 30% and does not increase elevation errors. Adding reverberation to that environment does not significantly change the localization accuracy.
2. When head tracking is used, then an environment that includes a floor reflection as well as head tracking is canonical. For most subjects, the reversal error rate was the lowest in such an environment, and on the average was 65% lower than for dry sound and 40% lower than for the canonical environment without head tracking. The bias in azimuth errors was also reduced further by 15% from the canonical environment without head tracking and the elevation errors were slightly reduced.
We expect these results to hold for any broadband sound source and not only for noise. The results are likely to be different for speech or any other sound source with no high frequency content, in that the key contributions of the pinna to the determination of elevation will be absent or reduced. In that case we expect that head tracking may lead to a greater improvement in elevation localization."

Since studios are real spaces and not virtual like in the paper so 'head tracking' is naturally in full effect via HRTF, the current consensus is that floor plays indeed an important role, likely even more for a subject with trained ears. Getting rid of it seems more damageable than not.

See attached PDF.

Quote:
Originally Posted by scraggs View Post
2. way way off topic, but...clearly your rooms have a pretty specific design. how much does it change based on the location of the room? i.e. a mastering room in a standalone building out in the middle of nowhere in the country vs something like what they're currently doing at The Bunker, where they have 3 rooms right next to each other in a building in the city.

the latter obviously has way higher isolation requirements than the former, which i would imagine would then require much more treatment...yes/no/maybe/it depends?
The shells are always heavy shells as we need to reach NR15 for FTB, a very quiet environment. In very rare instances we won't float (like at J's since he's really in a super quiet place, studio is in a separate building far away from any road or noise and he does Mastering, so no instruments next door). But otherwise floating is the de facto norm to be able to achieve such quietness.

For bunker, the shells are a bit heavier since they are close to one another. But the real main difference in soundproofing between the rooms will be the fact that they are floated to low frequency. Bunker is under 8Hz. Sterling Sound NYC... ca 3.5Hz! A lot of people think that adding wall layers after layers will increase isolation, but quite rapidly it won't as most of the noise heard at this stage is structure born and re-emitted by the non-floated studio shell and floors. You can keep adding all you want, it won't reduce the LF content much. You need to stop the mechanical transmission.

In terms of inner acoustic treatment, the limited shell mass variations in our designs have little impact, we disregard it and design treatment based on a standard heavy shell no matter what.
Attached Thumbnails
The Titanic, room curves and other GS style OT wanderings-floor-effect-fr-example.jpg   The Titanic, room curves and other GS style OT wanderings-floor-effect-etc-example.jpg  
Attached Files
File Type: pdf AES113_5714.pdf (532.6 KB, 45 views)
4
Share
Old 30th August 2018
  #145
Gear Addict
 
ReDRuMx's Avatar
Quote:
Originally Posted by Jens Eklund View Post
Absolutely: in a specific room, there will be boosts at some modal frequencies and dips at others. But as stated in previous post, if looking at the average of many different measurements; a general boost in the modal range is expected.
Exactly. But no listener actually experiences this average (in domestic rooms).

In any case, I'm neither correcting you, nor contradicting anything you are saying. Just expanding it a little bit. It goes in favor of the fact that the "average consumer listening experience" is a moving target.
1
Share
Old 30th August 2018
  #146
Gear Guru
Dichotomy

Red, you have put an awful lot of work in this exhibiting a curious honest spirit.
It is for the likes of you that I posted the question. Thank you for respecting both intent and content.

Quote:
you can listen to your references before mixing, and calibrate your brain. The music of the past is your "control", your "taste", call it "style"; and it establishes the "frame" within which you draw (i.e. - mix). My opinion is that with proper references and an acclimatisation period, the FR becomes less important (within reason, of course).
Absolutely, this is indeed what Ninja Engineers do when necessary. But the acclimatisation wears off.
It is horrible and tiring forcing yourself to make a bright tonality completely alien to what you are used to hearing in the home and car.
But a question, why try to force Calibrate your ear/brain, why not Calibrate the Speakers?

Quote:
the "average consumer listening experience" is a moving target.
No doubt, what with Earbuds relating all the centre info, Bass, Kick, Vocal, -6dB down compared to speakers, Cars......

But what is curious is that despite decades of changing room and equipment styles, recent SonarWorks surveying of the average DOMESTIC summed response of room+ speakers is remarkably similar to B&K, Harman, Olive etc.
The 1dB per octave slope is ubiquitous.


A Flat CR or Anechoic Chamber or Outdoor playback measure and sound similar to each other using exactly the same Lingua Franca Measurements, no short gates.

Quote:
they are very far from your average living room.
Word.

The vast majority of listeners in domestic rooms experience an added room tone leading to a summed response as seen in the work of Olive and Toole. Very far from a Flat CR.


DD

Last edited by DanDan; 30th August 2018 at 01:15 PM..
Old 30th August 2018
  #147
Gear Addict
 
ReDRuMx's Avatar
Quote:
Originally Posted by DanDan View Post

The vast majority of listeners in domestic rooms experience an added room tone leading to a summed response as seen in the work of Olive and Toole. Very far from a Flat CR.


DD
Yes but... Everyone agrees that a professional CR shouldn't have FR irregularities (peak / dips), which are a result of reflections interfering with the direct sound.

I.e, there's no way to have a regular FR without removing the reflections

A domestic environment has a specific "room tone", which is a result of reflections + direct sound.

What I'm saying is that no frequency curve correction will bring the CR experience close to the "home environment" experience, where the room tone (reflections) are a major factor. I'm also saying that I think it's not necessary to try and come close to the home environment experience in pro CRs. I've explained why I think so in point nr. 1 in my long post.

A major reason is that no individual "average listener" listening experience is very close to the average. In domestic rooms, this is especially true in the sub area. I.e. - the variance is huge - And there's no way to find a useful representative FR.
Old 30th August 2018
  #148
Gear Addict
 
ReDRuMx's Avatar
This doesn't mean I think that LF boost in pro CRs is a bad thing. It obviously makes sense.

Having a great room is major facilitating factor, but I still believe that understanding what I wrote in my long post (point nr. 1) is equally, if not more important.
Old 30th August 2018
  #149
Gear Guru
Representative

Flat vs 6-9dB HF sloped is a very wide dichotomy.
We can simulate the Room Tone and add it in the CR.
This is enveloping but not particularly useful.
We can simulate the average FR HEARD (and preferred) by domestic and Car Listeners, very easily and very well.

This aids translation immensely for many of us.

I would recommend anyone curios for truth and with the resources at hand to try it. All you need is Ears!

Hi Fi Living Room. Listen to music very close to speakers. Way bright right. Listen in the regular seats, normal tonality.
Same Music, Flat or Pretty Flat CR and Eq. Adjust the Eq so that the tonality is the same as at the Couch.


Quote:
This doesn't mean I think that LF boost in pro CRs is a bad thing. It obviously makes sense.
Obviously.
You know, the translation curve is often expressed as a 1dB/Oct SLOPE DOWNWARDS TOWARDS HF.
Rephrasing, a 1dB/Octave SLOPE UPWARDS TOWARDS LF. Same thing no?

We are bouncing between using Eq curves and HEARD SPECTRA.
If I may draw focus specifically for a moment on the tonality/spectrum of actual music in actual places.
We all love bass and are extremely tolerant to excesses of it. A huge variance indeed. But not so with HF. We have resonant and biomechanics hearing amplification in the 3-5K region, where our hearing mechanisms can be easily physically damaged. The body will protect itself. The spectra of music we like to ear are remarkably sloped irrespective of the listening environment. Why should our favoured tonality change with location?

DD

Last edited by DanDan; 30th August 2018 at 01:58 PM..
Old 30th August 2018
  #150
Gear Addict
 
ReDRuMx's Avatar
Lol Dan, you changed your post while I was replying, now it makes less sense

EDIT: What I meant was: my reply now makes less sense

Last edited by ReDRuMx; 30th August 2018 at 01:47 PM..
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump