Portable acoustic screens are designed to make it possible to record good–sounding vocals almost anywhere. We put this claim to the test!
When SE Electronics launched their Reflexion Filter Pro in 2006, they created an entirely new product category. The idea of a portable acoustic screen that can ‘dry up’ recorded sound by preventing unwanted room reflections from reaching the microphone has obvious appeal to the home recording market, where people often find themselves working in untreated recording spaces. The Reflexion Filter was a great commercial success, and though certain aspects of its multi–layered design are covered by an SE patent, the basic concept has been widely imitated by other manufacturers.
Eight years on, there are hundreds of thousands of these ‘portable vocal booths’ in homes and studios across the world. Yet there is no reliable information in the public domain on their effectiveness, and, as far as we know, none of them have ever been subjected to rigorous testing. Do they actually do what their manufacturers claim of them? And what, if any, are the negative side–effects of using them?
To find out the answers to these questions, Sound On Sound teamed up with the Acoustics Laboratory at the University of Salford. Thanks to their expertise and world–class test facilities, we were able to compare the performance of 10 rival products, including both the original Reflexion Filter Pro and SE’s latest RF Space. The test protocols were devised by Trevor Cox, Professor of Acoustic Engineering at Salford, and the tests carried out by doctoral student Nikhilesh Patil. Trevor then oversaw the analysis of the results.
(See the accompanying media page for the audio files) http://www.soundonsound.com/sos/oct14/articles/vocal-booths-media.htm
Screen One, Screen Them All
We tested 10 such screens, ranging in price from well under £100 to almost £300under $100 to almost $400. All are claimed to reduce the audible influence of room reflections on a recording, and many manufacturers make the additional claim that their screens can reduce the extent to which external sounds such as traffic, computer fan noise or air–conditioning hum are picked up.
The design is similar in almost all cases: a rigid outer shell is curved or angled to enclose a microphone, and lined with acoustically absorbent inner material. Design variations include the extent to which the rigid shell is perforated, and the choice of lining materials: while the SE screens have a complex multi–layered structure, many others use simple acoustic foam, usually shaped into wedges. There is also a surprising degree of consistency in the dimensions of these devices. With the exception of the Kaotica Eyeball — which is a completely different design — and the much larger Real Traps booth, all are within 5cm or so of each other.
The degree to which an absorber can attenuate sound is dependent on frequency. Most of these screens have an inner lining that is around 2.5cm thick. This, in theory, should achieve complete absorption of sounds where the wavelength is 10cm or shorter, which corresponds to frequencies of 3.4kHz and above. Below this, absorption will be incomplete, but will still be useful down to wavelengths of 25cm, which equates to a frequency of 1.36kHz. In principle, therefore, you would expect these screens to be most effective at mid and high frequencies, though in designs where the rigid shell is perforated, this should offer some bass absorption too.
The main way in which portable vocal booths reduce room reflections is by absorbing the sound before it can escape into the room, though they also prevent some returning reflections from reaching the microphone. However, as we have just seen, these screens cannot be equally effective across all frequencies, and will inevitably reflect some of the source sound back into the microphone, especially in designs where the shell is not perforated. Because the screens are designed to be placed in close proximity to the microphone, these reflections will arrive very shortly after the direct sound, and will thus be audible as coloration of the wanted sound rather than as reverberation or echo. The concave shape of most screens risks amplifying this effect through acoustic focusing.
Making A Noise
The two tests devised by Professor Cox are described in detail in the ‘Test Protocols’ box. The first, which took place in the small reverberation chamber, was designed to evaluate the claim that portable vocal booths can help to block unwanted sound such as traffic rumble and computer fan noise. The results, shown in Figure 1, are given as insertion losses, which is defined as the sound pressure level without the screen present minus the sound pressure level with the screen present. Positive numbers indicate that the screen is providing protection and attenuating the external noise; negative numbers indicate that it is actually increasing the noise incident on the microphone.
Apart from the Kaotica Eyeball, all the screens display similar characteristics. Above 2kHz or so, performance is largely independent of frequency, and between 2dB and 5dB of attenuation is achieved. This is brought about by acoustic shadowing. The greatest level of attenuation — nearly 8dB in some designs — occurs in the 800Hz–1kHz region, but further down the spectrum, all of these screens actually amplify unwanted noise within the 200–400 Hz region!
Professor Cox thinks that both effects arise because sound waves diffract around the edges of the screen — principally over the top and under the bottom, where the path length is shortest. This gives rise to interference which is destructive in the 800Hz–1kHz region, but constructive in the lower frequency region. The uniformity of the results reflects the very similar size and shape of the screens (the only one with significantly different dimensions was the Real Traps Portable Vocal Booth, but our couriers failed to deliver it in time for it to be included in this test).
Figure 3 shows broadband figures for the attenuation of unwanted noise, calculated from the sum of intensity over the frequency spectrum from 100Hz to 10kHz. As can be seen, even the most effective of these products achieves a reduction of only 1.5dB overall, while the least effective provide barely 0.6dB attenuation of noise.
Overall, then, these screens can offer a little protection from external noise, but only in the mid and high frequencies. Where the external noise consists mainly of low and low–mid frequencies, by contrast, using a ‘portable vocal booth’ might actually increase the amount of noise that is captured! This is a real concern in relation to sources such as traffic noise, which usually contain prominent low–frequency components.
Reflect & Repeat
Ultimately, though, protecting the microphone from external noise is a secondary function of these portable vocal booths. Their main design goal is to minimise the capture of room reflections from the source being recorded. Professor Cox’s second test was set up in the listening room at the University of Salford, an environment which is deliberately constructed to be similar to a large domestic living room (albeit with rather more acoustic treatment!). This test measured the effectiveness of each screen at preventing room reflections reaching the microphone, and the extent to which the screens themselves introduce unwanted coloration.
Figure 4 illustrates the measured effect of one of these ‘portable vocal booths’. The black and red lines represent the impulse response with and without the screen in place. The big peak at the start of the impulse response is the sound travelling directly from the loudspeaker to the microphone. This is the same with and without the screen, so the (solid) black line is exactly overwritten by the (dashed) red line. Towards the middle of the graph, starting 6ms or so after the direct sound, the response without the screen (red) shows room reflections. These are clearly reduced by the presence of the screen.
To quantify how much protection was provided by each screen, the RMS pressure for the late part of the impulse responses (starting from 6ms after the direct sound) with and without the screen were compared. This is expressed as an attenuation value in dB, where the greater the attenuation value, the greater the protection.
Figure 5 shows how this insertion loss (averaged in this case from all the screens on test) varies with frequency. What screening there is against room reflections is effective only from 500Hz upwards. This is to be expected given the size of the screens, and the thickness of the foam: these devices are of the order of 30–40cm in size, and protection is seen from about half the wavelength upwards. This is what would be expected from the simple physics of waves interacting with objects. At lower frequencies, the screens are too small to significantly alter the sound.
Figure 6 shows the results for each screen represented as a single value for attenuation across the 100Hz–10kHz region. As you can see, the amount of broadband attenuation achieved ranges from approximately 2.5dB for the least effective, to almost 8dB for the most effective. The Real Traps Portable Vocal Booth achieves better screening against room reflections than the others mainly because of its size; conversely, although the Kaotica Eyeball almost completely encloses the microphone, its lack of mass means that it is only effective at high frequencies.
Colour Me Bad
No matter what materials are used in their construction, portable vocal booths will inevitably reflect sound as well as absorbing it. What’s more, the close proximity of the microphone to the screen, and the curved or angled shape that is common to most designs, make it likely that much of this reflected sound will be directed towards the microphone.
The early reflections introduced by one of these screens are clearly visible in Figure 4: the black line shows peaks at 2–3ms after the direct sound, which are not present in the impulse response captured without the screen present (red line). The very short time difference between the direct and reflected sound means that any effects will be audible as comb filtering, rather than echo or reverberation. In other words, these reflections will introduce an unwanted change in the tonality of the source sound. To visualise the coloration introduced by each screen, the impulse response taken with the screen present was converted into a spectrum using a Fourier Transform. This spectrum was equalised to remove the frequency response of the loudspeaker, as measured in Salford’s anechoic chamber. The plots are shown in Figure 7. Ideally, these spectra should be flat, but none of the screens on test achieves this. All of the traces show distinctive patterns of peaks due to comb filtering, though some introduce more coloration than others. The ones with a distinct regular pattern of peaks are likely to cause the most problematic change to the timbre. Some also show a broader change in frequency response: for instance, the Real Traps Portable Vocal Booth’s relatively good performance at screening out room reflections comes at the cost of a significant alteration to the tonality of the sound being captured, with a 5dB boost centred around 150Hz and a comparable cut at 1kHz.
Unsurprisingly, then, all the screens on test introduce measurable coloration into the wanted sound. But are these undesirable side–effects actually audible? And do they outweigh the benefit of the wanted effects, namely a reduction in the amount of reverberation being captured at the microphone? This is a subjective judgement which needs to be taken by the producer or engineer, and which will depend on the circumstances in which a session is taking place. For the purposes of comparison, however, Professor Cox created ‘auralisations’ to compare the audible effects of using and not using the screens as they were set up in our tests.
To create the auralisations, Senior Lecturer at Salford University, Dr Sabine von Hünerbein, was recorded singing ‘The Green Dog’ by Herbert Kingsley in the anechoic chamber. An EQ was applied to the impulse responses with and without the booth to equalise the frequency response of the loudspeaker. The EQ settings were based on the measurement of the Tannoy loudspeaker in the anechoic chamber. Then the impulse responses were convolved with the anechoic singing.
In an ideal world, the use of a portable vocal booth would allow the anechoic recording to be recovered perfectly, but this is never achieved in this type of auralisation, even if the screen achieved its goal perfectly: some artifacts caused by passing the singer’s voice through a loudspeaker will remain even after equalisation. However, it is possible to compare the effects of each screen on the anechoic singing, both with each other and with the impulses taken without the screens present. To make your own comparisons, download the audio files associated with this article and listen to them yourself. It would be interesting to see whether the coloration effects can be compensated for with EQ.
Listening through, the effects of the booths are only heard on certain notes within the music. Dr von H nerbein is a soprano, and her voice is thus at the top of the human frequency range; the notes she sings in the audio example have fundamentals between about 300 and 750 Hz. Even so, it is clear from Figure 5 that most of the attenuation achieved by these screens takes place above these fundamental frequencies. Consequently, the screens mostly alter the strength of the harmonics (overtones) within the reflected sound. The coloration graphs (Figure 7) likewise show the most distinctive effects for most screens at 1kHz and above. Also, the audibility of the comb filtering will depend on whether the frequencies of the singing note combine with the peaks and dips of the comb filtering to significantly alter the harmonic series of the note. This won’t happen for every note. The lower notes, such as the opening few words of ‘If my dog was green’, seem to show differences most clearly.
There are measurable and audible differences in performance between the screens that were tested. Interestingly, though, these differences do not correlate in any obvious way with the cost of the screens! In terms of attenuation, for example, SE’s new RF Space filter offers only a limited improvement over the original Reflexion Filter Pro, despite its being nearly twice as expensive; and the Auralex Mudguard, one of the cheapest products on test, recorded better results than some of its much pricier rivals. Another outlier is the Real Traps Portable Vocal Booth: its significantly larger dimensions mean that it is the most effective at attenuating room reflections, but it also introduces the most coloration. The Kaotica Eyeball is a radically different design which is effective only above 1kHz or so.
The differences are subtle, and sometimes it is hard to say whether the sound with the screen in place is better or worse than without. In making this comparison it should be borne in mind that the Salford listening room is intended to provide a relatively pleasant and controlled acoustic environment: the benefits of the screens may be felt more strongly to outweigh their disadvantages in other rooms. What is clear is that none of the screens produce something that would be mistaken for anechoic singing. If you’re forced to record in a bad acoustic environment, portable vocal booths may provide a worthwhile reduction in the amount of unwanted room reverberation that is captured — but they won’t eliminate it.
What Else Can You Do?
A ‘portable vocal booth’ is, at best, a partial solution to the problem of unappealing room acoustics. If you are forced to record in less than ideal circumstances, there are other measures that should be taken too. The first is to ensure that you are using a microphone with a cardioid pickup pattern. Compared with an omnidirectional microphone used at the same distance, this will be more sensitive to the direct sound from the vocalist or other source, and less sensitive to reflected sound arriving at the sides and rear of the microphone. A large–diaphragm cardioid capacitor microphone is the most common choice for vocal recording, and this was what was used in our tests in Salford’s listening room.
The second measure that can be taken is to hang a broadband absorber behind the performer to trap reflections that might otherwise bounce off the rear wall and into the sensitive front side of the microphone. This should be as large as possible — in the absence of specialist acoustic treatment, a thick duvet often works well — and will be more effective if hung away from the wall. The subjective effect of a well-positioned duvet is often more obvious than that of a portable screen, and of course there is nothing to stop you using both.
Finally, it should be pointed out that the position of the mic in relation to the screen can make a difference. Our tests positioned the mic at the centre of an imaginary line drawn across the edges of the screen, as recommended in several of the product manuals including that of the SE Reflexion Filter Pro. However, James Ishmaev–Young of SE Electronics says that moving the mic further into the screen should increase the attenuation of room reflections, and will make differences in coloration between the better and less good screens more audible. We did not test this claim.
The Screens On Test
SE Electronics Reflexion Filter Pro (£175)($249)
The original, iconic ‘portable vocal booth’ remains a best–selling product more than eight years after its launch, and our tests show that it is still one of the most effective. It differs from most other designs in that multiple layers of different materials are used in its construction.
SE Electronics RF Space (£299)($399)
The latest iteration of SE’s Reflexion Filter design is the most highly engineered of the screens on test. Its mounting hardware is improved over that of the Reflexion Filter, but our tests suggest that it achieves similar levels of attenuation of room sound. The coloration it introduces is also similar in level, but shows a different spectral pattern.
This design features a rigid frame with an internal lining of foam. Unlike many of the others, its shape can be adjusted using hinged ‘wings’. It adds the least coloration to the sound, but is also the least effective at protecting the microphone from external noise and room reflections. This was one of the more difficult screens to position, thanks to a mounting system which tended to droop.
Auralex Mudguard (£79)($109.99)
Another design that uses a hard plastic outer shell lined with acoustic foam, the Mudguard provided probably the best price/performance ratio in our room reflection test, offering useful attenuation of reverberation without excessive coloration of the wanted signal. Late arrival meant we were not able to test its ability to attenuate noise.
Real Traps Portable Vocal Booth ($299.99 plus import duty and shipping)
Much larger than the other screens, the Real Traps design resembles a conventional acoustic treatment panel which is hinged in the middle. Its size means that it provides the most effective protection against room reflections, but it also imposes the most obvious coloration on the wanted signal of all the screens. In particular, the auralisation and graphs both show a noticeable bass boost, though this could perhaps be corrected with EQ.
Editors Keys Studio Series Portable Vocal Booth Pro (£169.99 including stand)($299.99 including stand)
This is a fairly generic design consisting of a curved metal shell lined with foam. Unlike most, it is supplied with its own purpose–built stand, which makes mounting easier. However, our tests suggested that it was not particularly effective at attenuating room reflections, and adds significant coloration to the wanted signal.
Primacoustic Voxguard (£105.60)($99.99)
This screen employs high–density acoustic foam within a shell made out of hard plastic, with ports which are said to “prevent excess bass buildup and resonance commonly associated with stand–mounted absorbers”. Its attenuation of room reflections was roughly comparable to that of the SE Reflexion Filter, but coloration of the wanted signal is more obvious.
This design employs a “complex sandwich construction of five differing materials” within a perforated metal shell. Our tests suggested that it provides reasonably good attenuation of room reflections, but also introduces some coloration of the wanted signal.
Vicoustic Flexi Screen Ultra (£147.60)
This screen looks a little different, thanks to its wooden shell. This, again, is slotted, presumably with some acoustic goal in mind. It was one of the better performers in our tests, offering worthwhile attenuation of reverberation with relatively little coloration.
Kaotica Eyeball ($199.99 plus import duty and shipping)
A completely different design from the other screens in the test, the Kaotica Eyeball resembles an overgrown microphone windshield. Unlike all of the other portable vocal booths we examined, the Eyeball is made entirely from foam, with no rigid parts or shell.
Because the design almost completely encloses the microphone, the effectiveness of the Eyeball at high frequencies is very good, both for protection from external noise and room reflections. However, as it is intrinsically lightweight, it provides very little protection in the mid-range, and no protection at all at low frequencies.
Prices include VAT where applicable.
The Test Protocols
To evaluate the effectiveness of the SE Reflexion Filter Pro and other portable vocal booths, Professor Trevor Cox devised two test protocols. The first is similar to those used for testing the isolation provided by acoustic enclosures placed around noisy machines, and was intended to evaluate these screens’ effectiveness at attenuating external noise. This test was carried out in the small reverberation chamber at the University of Salford.
A loudspeaker was placed in the corner of the reverberation chamber emitting noise. This creates a diffuse field where sound is incident on a microphone in the room from all possible angles. It is important to ensure the sound is arriving from all directions, because this means weak spots are tested. While someone using a screen in a real–world context will tend to place the booth between the noise source and the microphone to shield the mic, this isn’t going to be possible in all cases, for example when traffic noise is coming through the structure of the building.
A half–inch Bruel & Kjaer omnidirectional measurement microphone was placed in the body of the reverberation chamber at least 1m from any surface and at least 2m from the loudspeaker. A B&K HATS (Head And Torso Simulator) was placed in the singer’s location to simulate shielding and reflections from the singer’s head and torso. The sound pressure level (in dB) on the microphone was measured in third–octave bands using an acquisition time of 30 seconds. The measurement of sound pressure level was then repeated with the portable vocal booth in position around the microphone.
To confirm that these results were not unduly influenced by the position of the screen and microphone within the room, the original SE Reflexion Filter was also tested in six different positions. While the fact that the soundfield in the room is not completely diffuse means there is some variation in these results (see Figure 2), the overall characteristics are very similar, suggesting that the test protocol is fair.
Professor Cox’s second test was designed to measure how well the screens achieve their main design goal: to prevent room reflections from the source being picked up by the microphone. This test also allowed us to measure the amount of unwanted coloration each screen introduced.
These tests took place in the listening room at the University of Salford. This is a noise–free environment with a controlled reverberation time of about 0.3 seconds in the mid-range. To make the room configuration more like a typical home studio, an area of acoustic treatment was removed, giving a clear specular reflection from the wall behind the microphone. A Rode NT2A cardioid capacitor microphone was set up on a stand 2m in front of this wall, and a Tannoy System 6 dual–concentric monitor speaker was positioned 22.5cm in front of the microphone — a typical distance for close vocal recording. (A dual–concentric design was chosen to ensure that the high- and low-frequency components of the sound arrived at the mic at the same time and from the same direction.)
The general principle is to measure the impulse response between the signal driving the loudspeaker and the electrical signal picked up by the microphone. This was done with and without the screen in place, allowing its effect to be seen. For each test, the room configuration, the microphone and loudspeaker positions, and settings on the acquisition system were identical. For each product the impulse response with and without the screen present was measured. Where possible, the microphone and loudspeaker were not moved between the two measurements (this wasn’t possible for the Kaotica Eyeball).
Most of the screens on test permit a range of possible positions in relation to the microphone. Where a particular position was recommended by the manufacturer, that position was used in the tests; otherwise, the screen was positioned such that the microphone was in the middle of an imaginary straight line drawn across the edges of the screen’s arc. Positioning was complicated by the mounting systems supplied with the screens, many of which did not work very well!
The impulse response was measured between the loudspeaker and the microphone using a swept sine wave with sufficient duration to capture both the early reflections and room reverberance. A one-second signal was used. The sampling frequency was 48kHz. The loudspeaker was driven at a typical listening volume level to minimise any distortion artifacts created by the loudspeaker.
In addition to the above measurement, a calibration was undertaken in the Acoustic Laboratory’s anechoic chamber, where the impulse response between the loudspeaker and microphone was measured using the same spacing as used in the listening room tests. These measurements were used to equalise the frequency response of the loudspeaker, where necessary, during analysis. In addition, a piece of vocal was recorded in the anechoic chamber for auralisation.
Studying Acoustics At Salford
The Acoustics Laboratories at the University of Salford include anechoic chambers, a listening room and reverberation chambers, all of which were used in testing the portable vocal booths for this article. Acoustic research has been carried out at the university for nearly 50 years. The first Acoustics Laboratories were established in 1965, and in 1975 the first degree (Electroacoustics) was taught. Current courses include the BEng (Hons) Audio Acoustics and MSc Acoustics, which are aimed at budding engineers who might go on to develop new audio technologies or scientists who might carry out research in acoustics. Other courses, such as the BSc (Hons) Professional Sound & Video Technology and MSc Audio Production, attract students looking to work as sound engineers producing high–quality audio content. The Acoustics Research Centre is the primary partner for acoustics research in the BBC Audio Research Partnership.
Current major projects in audio include examining the future of spatial audio in the home, assessing the audio quality of user–generated content, and new mathematical models for simulating room acoustics. In recent years, the establishment of MediaCityUK at Salford Quays has led to an expansion of the University’s commercial work with the audio industries, including testing many of the acoustic products used in the fit–out of the new studios at MediaCityUK.
To download and listen to Professor Cox’s ‘auralisations’ of each screen, surf to http://www.soundonsound.com/sos/oct14/articles/vocal-booths-media.htm. In each case, you can hear the same piece of anechoic singing convolved with an impulse response captured with, and then without, the screen present behind the microphone. The files are in mono, 48kHz WAV format.