Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[boggy]

Greetings everybody, 

Thanks to moderators, we can continue discussion from old forum topic "Acoustical Design of Control Room for Stereo and Multichannel Production — A Novel Approach" here.

AES Convention paper: Davidovic Z., Petrovic B.: Acoustical Design of Control Room for Stereo and Multichannel Production and Reproduction - A Novel Approach, 129th AES Convention 2010, San Francisco, CA, USA

TELFOR Conference paper (english): Petrovic Bogic, Davidovic Zorica: A new acoustical design of control room for multichannel production and reproduction, 18th Telecommunications Forum TELFOR 2010, Belgrade, Serbia. [NEW]

Our white paper about MyRoom Design can be found here or here.

We get couple of questions via email, and some responses can be good start for a topic, and possibly interesting for other people.


About type of wideband absorption behind air transparent diffusers...

We wouldn’t describe absorption realization in details, because we expect that reader (skilled acoustician) already have his own experience with “best” method. We always used Tom Hidley/Philip Newell principle (waveguides formed with hanged panels), because we believe this method is best fit to air transparent diffuser. But we don’t like to discourage others, with different experience, to use layers of mid density rockwoll or any other known method.


About depth of wideband absorption behind transparent diffusers...

Depth vary from place to place… and to be sure,... we fill with absorption panels any cavity we found in room,  with this, we formed one big “bass trap” around listener (half space above floor) . It’s substantial that we don’t have acoustically different “ends” of room… all surfaces except floor are highly absorptive. And yes, before you mount air transparent diffusers, room is completely “dead” and this was deliberately. We noted in our white paper, about volume of room we use for “bass trapping” or wideband absorption, and if you can’t figure out which depth is used… here is more precise information:
- depth vary from 30-80cm (12-32") in smaller room (Pressed Lizard studio), and from 10cm to 150cm (4-60") in bigger room  (RES Media studios).
Be aware that in “sub-compact” rooms, with room volumes below 100m³, you can’t exceed some "maximum needed" absorption for low frequencies, because room modes energy are much denser in audible low frequencies region AND, in same time, you have much smaller space (volume) for absorption… because this, low end absorption won’t be “ideal” anyway, whatever absorption method you use and whichever volume you spent for it!
Good thing is that most loudspeakers are omnidirectional below 200-300Hz and deepest bass traps works in some way in overall treatment even if you have it only behind and possibly around (soffit mounted) speakers.


About windows, doors, racks, ventilation ducts…

In our white paper we didn’t describe detailed plan of any particular studio, instead of this we described general design principles. This means that you can have windows to recording rooms, day light windows, doors for entrance, ventilation ducts, racks… etc. It’s only important to take care about symmetry, especially front side. Highly diffused rooms are very sensitive to non symmetrical obstacles. Or better, people are very sensitive to non-symmetrical arranged obstacles in highly diffused rooms. 
For example, if you have window to your front-left side, but not have one to your front right side, symmetrical to left, that may be a problem. Try to make “blind” window to your right side with same dimensions at your left side (not needed to be real window, may be some picture or illustration behind glass). This helps uniform distribution of reflections to your ears, and not causes frequency dependent position distortion in your stereo image.

to be continued...

NOTE:
I can't currently attach pictures where you can see rack/doors, and one with nice air-transparent fractal diffuser 2x7th order.
I hope this will be solved in meantime 
EDIT:
Here it is... I put images on my site  :
Rack, doors:

Fractal diffuser:


Best regards

[boggy]

About measurements...

All frequency response measurements are smoothed to 1/3 octave because International/European recommendations for control rooms ask this:

1. "Multichannel surround sound systems and operations", AES Technical Council, Document AESTD1001.1.01-10, New York.
2. “Methods for the subjective assessment of small impairments in audio systems including multichannel sound systems”, ITU-R Recommendation BS.1116 (rev. 1), ITU, Geneva, 1997.
3. W. Hoeg, L. Christensen, R. Walker, “Subjective assessment of audio quality– the means and methods within the EBU”, EBU Technical Review Winter 1997, pp. 40-43.
4. "Listening conditions for the assessment of sound programme material: monophonic and two–channel stereophonic",  EBU Tech. 3276 – 2nd edition May 1998, European Broadcasting Union, Geneva, Switzerland
...

It's good to be aware that RT60 recommendation is even easier to reach, than frequency response in AESTD1001.1.01-10, because there may be, in RT60 recommended boundaries, “jumps” allowed below 125Hz. Instead of this, in frequency response boundaries we need “flat” response down to 40Hz. Peaks in measured RT60 are highly correlated with room modes, and because this, they are highly correlated with non-linearities in frequency response. This means that you need much better RT60 response if you like to have recommended frequency response. In older recommendations (ITU/EBU) there are wider boundaries at low frequencies, for frequency response, that was easier to reach, but nevermind...

To be honest, we didn’t really expect to build acoustics treatment in such small rooms and to get AES/ITU/EBU recommendations fully satisfied, especially for low frequencies below 125Hz. We are defined MyRoom method with consideration to listeners, mixing engineers, not to measurement microphones only, using all our knowledge and experience of acoustics and psychoacoustics, from books, papers and in the field. Nevertheless, we get a decent measurement results in regards to  recommendations.

Basic idea was to get best low end absorption as possible in “sub-compact” room, in all directions, spending largest possible volume of room for this, but still remain practical, and to bring back liveness in room later, with air transparent diffusers. That is our approach to get small rooms treated as best as can  be useful for music production.

Measurements came later, when we finished these rooms, and we publish it in our papers.

It’s true that room measurements results are not biggest quality of MyRoom treatment, because there are many, mostly psychoacoustic qualities that we can’t measure directly or objectively. Subjective impressions and mixing engineer experiences are one of our ways we used to define MyRoom method. And is also true that there are many different ways to reach recommended values for RT60 and frequency response, but all this methods don't sounds similar.


Origins...

Fundamental acoustical behavior we used for MyRoom Design is ordinary living room, where we live most of our time from birth. That ordinary room is mostly homogenous/chaotic reflective and absorptive in all directions. We believe that this acoustical behavior is a better way to get good and natural environment for mixing engineer, especially for mix translation and time needed for adaptation.

This means, for control room acoustics, an “ordinary” room with highly controlled low frequencies and decent and homogenous liveness for high frequencies, from all directions.

After all, added diffusion is a logical step to bring more details to listeners ears, and make control room for professional, dedicated purposes.

For all diffuse surfaces, except floor, inspiration came from Massenburg Blackbird Room.

For low end treatment inspiration came from Tom Hidley/Philip Newell Non-Environment design.

Construction that link all this and make it possible was "Air Transparent Diffuser" (image below)

Cross section of air transparent diffuser with applied wideband absorber (behind):


About phase grating and amplitude grating diffusers used in MyRoom principle...

Our air transparent diffusers are combined phase grating and amplitude grating construction.

Amplitude grating effect exists because space between (different depth) slats forms Helmholtz absorbers with different (random) working frequencies, and slats are already reflective. Because different working frequencies of slat absorber, there are amplitude grating effect (position dependent) for single frequency, in the working range.

Reflection from slat fronts (different depths/position) forming phase grating effect.

But, if someone don’t have a budget to build that type of diffusers, there is a way to build ordinary amplitude grating diffuser instead, binary MLS type, for example, used to bring back liveness in room after heavy and thick wideband absorption and maintain needed RT60. Even if amplitude+phase grating construction is better for diffusion, from our experience, bringing back liveness in a room is important, and when you do this with only binary diffusers (1D or 2D) you still may have nice and pleasing place for work, even without strong diffusion sound field.

To be continued...


Best regards

[Thomas@Northward]

Interesting

[boggy]

Thank you Thomas!

................
TELFOR Conference paper (english): Petrovic Bogic, Davidovic Zorica: A new acoustical design of control room for multichannel production and reproduction, 18th Telecommunications Forum TELFOR 2010, Belgrade, Serbia. [NEW]
.............

Last time i check, TELFOR web site give an "internal server error..." while someone try to download our paper, and because this I "mirrored" it on our site too, so you can get it here now:

TELFOR Conference paper (english): Petrovic Bogic, Davidovic Zorica: A new acoustical design of control room for multichannel production and reproduction, 18th Telecommunications Forum TELFOR 2010, Belgrade, Serbia. [mirrored]

Sorry for inconvenience.

Best regards,

[bblackwood]

Wild. How does it sound?

[boggy]

Wild. How does it sound?

Hmmm... last month we successfully got even better frequency response in a small Pressed Lizard room (described in white paper), using equalizing filters (IIR) in loudspeaker crossover
Here are the new measurement results, which obviously tell just one part of the story, and of course the most important element are the mix/master results you get from the room itself, and the way it leads your perception in a certain direction. But, having said that, I'm posting the measurement graphs, just so you can see the frequency response linearity (now it's easily in +/-3dB limits), and of course if you're more interested in design elements and/or even concrete work examples, I'm looking forward to sharing those details with you.

First picture is left channel only, blue is new response, red is old response
Second picture is right channel only, blue is new response, red is old response
Third picture is left and right channel, new responses, blue is left, red is right

NOTES:
1. All published frequency response graphs are smoothed to 1/3 octave (as is recommended in AES/EBU/ITU documments listed above)
2. All measurements are made at mixing position.
3. All measurements are made without desks, and any other studio furniture... floor/room is empty.
4. Loudspeakers are our SM-3, a three way active system. (now customized)
5. Loudspeakers are soffit mounted into bass trap with air transparent diffusers (please, see papers listed in first post)

Best regards,

[Ethan Winer]

All published frequency response graphs are smoothed to 1/3 octave (as is recommended in AES/EBU/ITU documments listed above)

Third-octave smoothing is useful for frequencies above 500 Hz or so, but for the bass range you really need to display the response at high resolution. Otherwise you miss seeing the full extent of peaks and nulls. No smoothing can look terrible! But it's the truth.

--Ethan

[boggy]

Third-octave smoothing is useful for frequencies above 500 Hz or so, but for the bass range you really need to display the response at high resolution.

I didn't try to define new standards and recommendations for displaying measurements.
There are three recommendations (EBU-European Broadcast Union, ITU - International Telecommunication Union, AES - Audio Engineering Society) listed above, and all of this papers agree in this one thing, smoothing should be 1/3 octave.
Two of these papers are international, one is European.
If you can tell me which document (international is preferable) define principle you recommend, I will be happy to read it.

Otherwise you miss seeing the full extent of peaks and nulls.

Someone skilled in reading smoothed graphs cannot miss anything.
I attached one graph where is displayed both unsmoothed and 1/3 oct. smoothed measurement, and there aren't any horror "surprise" below 500Hz, or I don't see well.
 

No smoothing can look terrible!

This is measurement only... no big deal. "Terrible" may be only what someone hear (or doesn't hear) in the room...
At least, any decent acoustician knows couple of different ways to treat same room and to get similar measurement results, but these treatments may not sound even similar in this same room.
 

But it's the truth.

From my point of view, and if we discuss (audio) studio acoustics, real truth we have only if control room and monitoring really helps or really annoy people in ambition to reach desired results without too much fatigue. All other things isn't much important.
 
Best regards,

[Ethan Winer]

Someone skilled in reading smoothed graphs cannot miss anything. I attached one graph where is displayed both unsmoothed and 1/3 oct. smoothed measurement, and there aren't any horror "surprise" below 500Hz, or I don't see well.

Looking at the unsmoothed response, you don't see the huge null just above 100 Hz, and just below 200 Hz, and the four peaks below 200 Hz? Those are completely hidden in the smoothed version. Looking at the smoothed version you'd think (incorrectly) that the response below 500 Hz is flat within a 3 or 4 dB window. Yet the unsmoothed response shows that the span is really more like 25 dB. I can't speak for you or others, but that seems very significant to me!

As for the various "official" recommendations, I suspect those are meant for rooms much larger than the typical bedrooms so many people mix in today. Newer thinking takes into account the size of the room.

--Ethan

[boggy]

Looking at the unsmoothed response, you don't see the huge null just above 100 Hz, and just below 200 Hz, and the four peaks below 200 Hz? Those are completely hidden in the smoothed version. Looking at the smoothed version you'd think (incorrectly) that the response below 500 Hz is flat within a 3 or 4 dB window. Yet the unsmoothed response shows that the span is really more like 25 dB. I can't speak for you or others, but that seems very significant to me!

I understand your opinion. But... I like to be slightly more practical...
These peaks and dips, you noted, are (very) high-Q artifacts that exist in any closed space and they aren't easily audible, in most cases they are 100% non-audible.
BTW, there was a (visible in 1/3 oct. smoothed graph), low-Q and annoying event in 40-50Hz band, and two low-Q dips from 100 to 200Hz, some of it are reported, some aren't, and you can see it at older graphs (non-equalized response). Room has near square horizontal cross section, btw.
We are born in reverberant space, we live most of our life in reverberant space, and we really can adapt to some degree using our brain.
1/3 octave smoothing is mathematical, and non-ideal, way to represent which anomalies in acoustic response, people can hear, in some practical and useful way. From my experience, only some of artifacts, visible in 1/3 octave smoothing, are audible, and most of it still aren't audible. I talk about skilled studio people, not ordinary people.

I didn't ever think about acoustical treatment, that can achieve absolutely flat, non-smoothed, response from 20Hz to 20kHz in closed space used for control room, because this isn't really possible (at least because floor which must be reflective).

As for the various "official" recommendations, I suspect those are meant for rooms much larger than the typical bedrooms so many people mix in today. Newer thinking takes into account the size of the room.
........

Room volume is taken into account in formula for calculating desired RT₆₀ in room. Here it is:

T_m=0.25*(V/V_r)^1/3

Where V is useful volume of the room and V_r is reference volume (100m³).

Smaller rooms are much more complicated for acoustical treatment, because much smaller space for bass trapping and because much higher level of reflected energy... and I don't see any logical reason not to use available standards for small(est) rooms, because these standards are even much more stringent for it...
It's not easy job to reach +/-3dB response in space with only 30m³ (1060 cubic feet) available room volume... and this is actual (starting) volume of "Pressed Lizard room"

I don't see where is a benefit in displaying non-smoothed and high-Q artifacts which we cannot hear and which we even cannot fix with any known method, especially in small rooms...

I hope that we discuss about engineering... not magic


Best regards,

[franman]

I'm all for looking at hi resolution measurements, and we typically make higher resolution tests below 500Hz when evaluating room/ speaker systems.... but I have to agree with boggy that both AES and EBU standards call for 1/3 octave smoothing when displaying full bandwidth room response for listening rooms. Also, we find that most very high Q dips and peaks are rarely audible. Viewing them can be informative when trying to locate 'cause and effect' for the acoustician, but they are of little use when presenting final room response to clients or when trying to adjust monitor systems. They are certainly never anything we would address with any electronic adjustment and even when deciding on tuned trapping we typically want to see wider Q 'trends' instead of very high Q anomolies...

I've worked with many of the most respected room tuning 'gurus' and I can tell you for fact that pretty much everyone uses 1/3 octave or 17% smoothing (for TEF). This IS a standard... (my ten cents)

FM

[Thomas@Northward]

We only use 1/24th here - and the design result wrt to FR is based on such a resolution.

We've had many clients hear pretty narrow notches with an amazing accuracy, especially in mastering. So it is what we use when showing them the room response.

1/3rd we only look at for the 'general' trends in the room - to make sure the overall response is balanced between LF, MF and HF.

[Ethan Winer]

These peaks and dips, you noted, are (very) high-Q artifacts that exist in any closed space and they aren't easily audible, in most cases they are 100% non-audible.

I have to agree with boggy that both AES and EBU standards call for 1/3 octave smoothing when displaying full bandwidth room response for listening rooms. Also, we find that most very high Q dips and peaks are rarely audible.

I knew this would be the response, and I'm prepared.

The notion that very narrow notches are not damaging or even audible goes back to a 1981 AES paper by Roland Bucklein. This article explains an important failing of those tests, and includes audio examples proving that very narrow cuts can be audible:

Audibility of Narrow-Band EQ

As explained in the article, one problem with Bucklein's research is he did not correlate the frequencies being cut with frequencies present in the music. If your audition music is in the key of B, you can cut 110 Hz by 40 dB or more and not affect the music. But do that with music in the key of A and I promise you will definitely notice! Now, it's true that the response can change drastically over very small distances, even at very low frequencies. So a deep null at one ear may be partially filled in at the other ear. But not all nulls are highly localized.

Also, the response in a typical bedroom is very different from a larger pro-size control room you professionals work with. The most common problem I hear about is that mixes sound great in the bedroom, but sound tubby and boomy elsewhere. This is due to one or more deep nulls that are almost always present in small rooms. The nulls are often very narrow, but they still have a huge effect.

--Ethan

[boggy]

We only use 1/24th here - and the design result wrt to FR is based on such a resolution.

I attached same graph as above, but smoothed 1/24 octave.

We've had many clients hear pretty narrow notches with an amazing accuracy, especially in mastering. So it is what we use when showing them the room response.

I have experience about their amazing accuracy too, but only for some of notches, not all... that is interesting for me.
 Also, different people detect (a bit) different anomalies...

1/3rd we only look at for the 'general' trends in the room - to make sure the overall response is balanced between LF, MF and HF.

Of course I used non-smoothed graphs in design process for Pressed Lizard (as always).

After we finished second phase, I asked which frequencies they find to be problematic (without seeing graph). Strong correlation with measurement results was only for some of frequencies visible in 1/3 octave smoothing, but not all. Second opinion was that Lizard is a bit darker than someone expect, and I think that this is because:

1. there still isn't mounted all (decorative) reflective surfaces,
2. side, ceiling and back diffusers are constructed with raw OSB,
3. cutted edges of slats didn't veneered, but only lacquered.
4. front diffusers are builded from raw MDF, edges didn't veneered too, only lacquered.

I knew this would be the response, and I'm prepared.

The notion that very narrow notches are not damaging or even audible goes back to a 1981 AES paper by Roland Bucklein. This article explains an important failing of those tests, and includes audio examples proving that very narrow cuts can be audible:

Audibility of Narrow-Band EQ
..........

Writing a new AES Convention paper about that topic is a better way to "Dispelling common audio myth" if such myth even exist, IMHO.

........
Also, the response in a typical bedroom is very different from a larger pro-size control room you professionals work with......

True.

.......The most common problem I hear about is that mixes sound great in the bedroom, but sound tubby and boomy elsewhere. This is due to one or more deep nulls that are almost always present in small rooms. The nulls are often very narrow, but they still have a huge effect.
.....

I don't agree that narrow notches in acoustic frequency response can be a "huge" reason why mixes, mixed in a small rooms, sounds "tubby and boomy". AFAIK this is a common result when someone mixed in a small room with (usually) mobile and "instant" acoustic solutions... I mean couple of 100mm thick broadband panels (foam or rockwoll), and some weak corner bass traps. There still exist a strong notches (wideband, low Q) in bass and low mid response, that are easily visible even with 1/3 octave smoothing. Mixing engineer trying to compensate it with EQ, not aware that this is control-room "sound", not his music. That mix may sound good in room where is mixed, but only in there. BTW, this phenomenon is named "mix translation".

From my experience, reasons for boomy and tubby sound are:

 
1. non existence of serious bass treatment, instead of this we have more wide-band (low-Q) low frequency acoustic anomalies.
2. non existence of strong diffuse sound field from 1kHz and above.
3. non existence of bigger speakers with stronger and lower bass in a small rooms (only "near-fields" is popular)


But I believe that some people can notice narrow notches in acoustic response.

Boomy and tubby sound also can become from beginning in recording process, when using a non adequate (recording) room.

I'll try to attach some examples mixed in Pressed Lizard, if you interested to check if they sound tubby and boomy (elsewhere).

Using "bedroom studios" is widely spread today, mostly because economical reasons, then some compromises is expectable, logical and reasonable.

My (practical) conclusion about that may be:
1. We cannot do much with room modes in a small(est) rooms,
2. we can fill with (wide-band) absorption all practicaly available volume (30-50%) of room, in all directions,
3. we can build and mount air-transparent diffusers in front of absorbers,
4. with all above we can "kill" most of room non-minimal phase behavior.
5. then we can do IIR type of room equalization before monitoring system, to correct low-Q anomalies that still exist.

And I'm pretty sure that such small rooms may be more useful than they cost, even if still has some of high-Q, narrow, notches, in non-smoothed response.
 
 

Best regards,

[Ethan Winer]

I attached same graph as above, but smoothed 1/24 octave.

Yes, that's much closer to the true response. Again, whether narrow nulls are audible depends entirely on the frequencies present in the music. It also depends on how localized the nulls are, since a null in one ear can be filled in at the other ear.

Writing a new AES Convention paper about that topic is a better way to "Dispelling common audio myth" if such myth even exist, IMHO.

That's a lot of work, and I already have my hands full. A proposal I submitted for a comprehensive book about audio was just accepted by a major publisher. So I'll be dispelling audio myths of all types for the next six months. 

I mean couple of 100mm thick broadband panels (foam or rockwoll), and some weak corner bass traps. There still exist a strong notches (wideband, low Q) in bass and low mid response, that are easily visible even with 1/3 octave smoothing.

Sure, but again we don't really know how bad things are when the response is glossed over using coarse smoothing at 1/3 octave. I totally agree that the problem with boomy mixes made in small rooms is due to inadequate bass trapping rather than how a graph is displayed.

BTW, this graph shows Before and After in a small (11x16) room, and both traces are unsmoothed:



So it's possible to make a small room flat, and have it look flat at high resolution. But it takes a lot of bass traps!

--Ethan