Controlled Reflection Isolation Booth
—Arthur Noxon
Preprint No. 2531
Presented at the 83rd AES Convention
1987 October 16-19
New York
2.2 The Setup
These 1/2 and 1/4 round TUBE TRAPS are typically on 18 inch centers on the walls and ceiling of a small gyp board sound room. Their orientation on the walls is vertical. Starting at the ceiling they’ll often run all the way to the floor. This leaves a distribution of 6 to 7 inch wide strips of reflecting surfaces throughout the room. The ceiling is similarly set up.
The four wall/wall corners and the four wall/ceiling corners have the 1/4 round trap installed. This corner loaded trap has excellent absorption through 60Hz (10) and controls the lower frequency small room resonances (11, 12). Note the door and window are both covered with traps. As with the gobo system, visual access to the engineer or other players is maintained through the spaces between the traps.
The curved reflector in each trap serves to scatter mid and high frequency sound. The lower frequency range is not scattered by this specular reflection but by diffraction as the wavefront rebounds off the thin strips of hard wall surface left between the absorptive bodies of the traps. A more complete presentation of the performance of this unique class diffraction grating effect is presented in the addendum.
2.3 Vocal Test Setup
Here we look at the acoustic signature of a typical vocal setup in the QSF room. A full set of measurements is presented to develop an overall sense of the room’s performance.
2.3A (see below) illustrates the classic vocal setup used inside the QSF room. A small vocal hot spot speaker is positioned 14 inches away from the mic, some 5 feet off the floor. The mic/speaker center line is asymmetrically set in the room.
2.3B (see below) shows the 80ms ETC. At 6dB per division, the decay rate in the room is very steady, as evidenced by the flatness of the decay curve. Decay time, RT-60 is 0.11 sec. Except for the distinct 3ms gap between the direct and the first reflections, there are no spikes, no gaps, no time delayed kickers. The density of reflections is high and uniform.
2.3C (see below) retains the 6dB per division amplitude but has increased time resolution, only a 40ms ETC. The early decay rate is 0.09 seconds, just slightly faster than longer time averaged decay rate. The initial time gap is more evident. There is no loss in the smooth, dense fill of reflections that drop away in time.
2.3D (see below) shows the 20ms ETC at 6dB division. The regular features of the staccato of reflections continue to be observed. It is easy to count three significant reflections in almost any millisecond. The initial time gap is 42dB deep at 3ms after the direct signal. The subsequent diffusive fill begins at 4.5 ms, and each strike stays 24 to 30dB below the direct signal.
2.3E (see below) displays the 100 to 10K TEF waterfall over a 60ms period. The broadband smooth decay is obvious. The frequency axis is linear in this display. This type of decay in a small room is due to the balance struck between full range absorption and full range scattering. The vertical axis is 12dB/division.
2.3F (see below) shows the top view of the TDS waterfall. Again very regular, non-resonant decay is noted, evidenced by the high density of streaking straight down the time axis. Note the floor opens up at nearly the same moment, especially if the slight high frequency drop off by the speaker is taken into consideration.
2.4 Instrumental Setup
This series of ETC recordings was made with a sound source located 42 inches off the floor and 24 inches out from the face of a corner trap. The speaker faces across the room’s diagonal. Test measurements are taken along the diagonal as indicated in 2.3A. All ETC data is in 6dB per division and 40ms.
2.4B (see below) shows ETC for 40ms with a close mic setup. The speaker/mic separation is 6 inches. A high density of diffuse fill slopes down from the direct signal. The direct to reverb energy is about 19dB. Reverb time is 0.09 sec.
2.4C (see below) shows the ETC for 40ms with the mic moved back to 18 inches. Note the direct/reverb energy ratio drops to 15.5dB. The RT60 remains at 0.09 sec. The early diffuse signature is still strong.
2.4D (see below) shows a mic distance of 54 inches. The direct/reverb energy is down to 7.76dB with reverb time to 0.08 sec. Notice the development of an initial time gap, it is about 5ms wide.
2.4E (see below) displays a mic position of 66 inches. The direct/reverb ratio is now down to 5.7dB, RT60 holding at 0.8 sec. The initial time gap is being reduced to about 2ms with a strong fill in the first 8ms. Each early reflection is within 14dB of the direct signal but their density is packing sound power into the early reflection time period.
2.4F (see below) shows mic position of 78 inches. Very strong ambience is developing. Count nine reflections 15 to 23dB below the direct signal within the first 6ms.
2.4G (see below) is the diagonal opposite the speaker, 90 inches apart. Count four distinct reflections between 10 and 15dB down, and within 3.5ms, and another group 6dB down in the following 3.5ms. This is very similar to the Acoustic Island gobo signature.
Throughout this survey of the room’s acoustic performance, the decay rate in the room remained constant at .08 to .09 seconds. Nearfield and farfield mic positions had significant dir/rev ratio difference but otherwise had very similar, quickly dispersed sound fields.
2.5 Adjustable Setup
So far the ETCs presented are all variations of a strong direct signal immediately followed by a lower level set of diffuse reflections. As the mic moved back from the sound source, the dir/rev ratio was reduced. Additional adjustments to this ratio can be made. Here, the extreme case is reached where the direct signal is completely lost, leaving essentially only a pulse of diffuse signal.
2.5A and 2.5B (see below) show respectively the setup and the ETC which it produced. ETC scales are still 6dB per division with a 40ms view of the time. The speaker is placed in one corner facing the opposite wall. The mic is placed some 72 inches away just off the same wall near the other corner. We have the direct signal now weakened, due to its directionality features, lower by a couple dB than the early reflections. This signature has an initial delay gap of 3ms followed by a 6ms slug of diffuse signals. By count, 16 separate reflections are within 3dB of the direct signal. It is reasonable to expect the lower, omni-directional frequency range comprises the direct signal while its full range is found in the early reflections.
2.5C and 2.5D (see below) are similar setups, except the mic is in the diagonal corner. There is no initial time gap. The direct signal is immediately followed by 4ms of dense, equal level reflections (at least 8 by count) just 2dB down. Beyond that is a 10ms flood of signals just 6dB down from the direct. The mic, being more into the directivity pattern of the speaker, will show more midrange signals than the earlier setup.
2.5E and 2.5F (see below) show the extreme case of the direct signal being significantly below the diffuse early reflection group. The speaker faces into the corner and the mic is out in the open. The direct signal is easily 6dB below the peak of the early reflections. The early reflections remain stronger than the direct signal for over 12ms.
These setups are not necessarily being promoted. The desirability of their use rests within the ears of the engineer. The purpose served here is to illustrate in a first order manner how the QSF room handles off axis and directionality features of an instrument.
2.6 Real World Comparison
Information without a sense of reference is difficult to evaluate. The acoustical signature of the QSF type room needed a comparison. A series of tests were run that used approximately the same speaker/mic/room distances. By this, the distinctive feature of the QSF acoustical signature should become discernible. The speaker was generally 4 feet from the mic and both roughly centered in each room and 5 feet off the floor.
Three tests were taken in the QSF room, the first of which is the standard room. Then 2 sheets of plywood were installed to either side of the speaker/mic axis. The last test added two more sheets at either end of the room.
2.6A, 2.6B and 2.6C (see below) show the three stages of the QSF room. One measured difference is the RT60. It starts at 0.08 seconds, extends to 0.09 seconds with two sheets and to 0.12 seconds with the 4 sheets. The second feature measured is the decrease in dir/rev energy ratio. There is 6.7dB in the QSF room, dropping to 4.07dB when 2 sheets of plywood are added and again to 2.44dB with four sheets.
Subjectively, there are more “holes” in the ETC as plywood reflectors are added. There is also increased presence in sharp, strong spikes. Something more subtle is also visible, there are two parts to the bar graph ETC. The very dense, low level signals provide the solid, blacked out feature that seems to fill upwards from the bottom. Above this extremely dense signal set rises a series of distinguishable spikes. The more QSF the room is, the smaller and more frequent the spikes rise above the dense underfill. Weaker QSF rooms have their dense underfill cut into more often, the spikes become stronger and more separated.
2.6D, 2.6E and 2.6F (see below) show a new recording studio, drum room and vocal booth room. Decay rates are respectively 0.10, 0.10, 0.14 seconds and di/rev ratios are 7.71, 4.55, and 2.20dB. The good dir/rev ratio of the studio is due to its size (15 x 20 x 8’). In all three cases, the diffuse substrait is heavily eroded and crowned with strong, clean and spikes. These rooms are typical and not very “quick” sounding.
2.4A-F Compared Data
3.0 Subjective Reports
A recording room has two clients, the performing talent and the recordist. Both need to be satisfied and their requirements are not necessarily the same. The Quick Sound Field room with its unusual lively/dead quality seems to satisfy both. Here are comments of impressions made by actual users of the QSF room.
3.1 Vocal
The first, immediate impression is its quality of silence. You hear nothing but what you are doing, no residual noise. It is easy to get feedback and immediately adjust for improvements. This room seems to be the ultimate practice room because it doesn’t lie or make things more beautiful. It is an honest, clean, clear and yet dry room. The room is very comfortable to work in. I felt much sound all around me, and that is important, because in singing, I am so close to my instrument.
3.2 Cello
The cello in small rooms usually sounds muddy but not here. It has a natural, fat bottom-end. I heard the whole cello. Its sound came through the room clean, clear and comfortable.
3.3 French Horn
This room is impressive, dry and flat, not boomy but not dead. It seems very responsive to musical changes. A slight shift to emphasize upper partials really comes through and you can make immediate adjustments. I can hear a great deal, almost any detail, body movements, breathing, foot tap and fingering. It is an airtight room. I could hear all my imperfections which, while not fun, is good for me. For example, I thought I was ready for a demo tape with this piece until I had a session in this room. Now I know I need more work and no one had to tell me. Usually practice is hard because so much of the sound goes straight out the bell and I don’t get to hear it. Here, sound is very good, gives the whole sound of the horn. When I played, I didn’t feel like I had to hold back. It’s a very comfortable room.
3.4 Drums
Usually the drum booth is too dead. Here, you can’t tell you’re in a small room. It’s low and speaks very well. Sound is like a picture and this room takes an accurate one, good balance. Snare had a rounded, fat sound and I didn’t have to doctor the heads. Playback was exact with what I heard live, usually coloration ruins it. Clear accurate sound of the whole drum is played by the room. That’s what we want for sampling.
3.5 Recordist
The room is very interesting to work with, it is acoustically stable. I can move the mic anywhere and the room sound stays constant. It’s a big plus to have this kind of stability in a small room. It’s always hard to mic, placement and pattern selections are critical. Here I can make decisions based solely on the instrument and performer without concern for room color.
I used two mics on the French horn, one near and off axis while the other was set far across the room. There was no near wall reflection sound, no boom and no low end murky sound in the far field mic. The room saves the highs; I was able to get a good stereo pan between the two mic positions without seeing room color shifts.
Natural sound of the instrument is what I try to get. I often work in large halls and have to be 20 to 30 feet away to get the totality of the instrument. This room allows me to be just a few feet away and still collect the full sound without room color. This room does have strong ambience, but only of the instrument.
With drums, I usually close mic but here I could get 3 to 4 feet away from the mechanical noise. I usually roll off at 200Hz but here I can leave it fat and get the total drum sound, full, round and flat. I added reverb later and got a realistic concert hall snare. The bottom end is great. The room handles transients so well that they don’t mask the attack of the “whomp.” Hardstick on ride cymbals had fantastic ambience. I used only two mics on the snare, one 2 feet above and the other 2 feet below and to the side. I could mix very well, lots of isolation between the two signals.
