PC Musician
Technique : PC Musician
The subject of studio acoustics can seem
like black magic - but don't let it scare you off. You can make easy
changes, that could vastly improve your recording and mixing
environment, with the help of free or cheap PC software utilities, as we
discover this month.
|
While it would be wonderful if you could just
connect your audio interface to your PC, amp and speakers and get on
with recording great music that sounds exactly as you want it to, often
it's not quite that easy, as the room you work in also significantly
affects what you hear. I and other Sound On Sound writers
regularly talk about the benefits of acoustic treatment, but we remain
surprised at how many musicians do nothing to treat their rooms, and
don't even consider acoustic issues before piling the gear in and
connecting it up.
If you want a sobering experience after installing
your expensive new monitor speakers, try moving them to several
different positions in your room and playing back the same CD tracks.
Even if you listen from the same spot in the room each time, the bass
end will probably sound hugely different with each speaker location,
while if you walk around the room listening you'll probably notice even
bigger variations, with huge amounts of bass at some spots and almost
none in others.
This sort of variation makes a mockery of critical
mixing decisions, and explains why so many mixes don't 'travel' (don't
sound good on lots of different playback systems), yet many musicians
still avoid acoustic treatment because they think it's too complicated.
Fortunately, there are various PC utilities that can help you decide
what to install and where to put it, and you needn't wait until your
gear is in place before you consider acoustic issues, either, since
other utilities can help you decide whether or not a particular room
will be a good shape acoustically or will be a pig to work with. Even if
you're considering building a studio from scratch, you can benefit from
such utilities, since they can suggest slight changes in dimensions
that will benefit you in the end.
A good goal for the 'in situ' frequency response of
your monitoring loudspeakers is +/-3dB over the range 50Hz to 16kHz.
However, there are probably some professional studios that don't meet
this spec, while, without suitable acoustic treatment, speaker systems
in home studios could exhibit peaks and troughs as large as 30dB.
The reason for this is that each of the room
dimensions (width, depth and height) will encourage certain frequencies
whose wavelengths are related to them. For instance, if your room is 8
feet in length you'll get a set of 'modes' at 70Hz and the harmonics of
70Hz (140Hz, 210Hz, 280Hz and so on), a second set for the room width,
and yet another for its height. To stand a reasonable chance of
achieving good acoustics, you should try to choose a room whose combined
modes are reasonably well spaced across the frequency range,
particularly below about 300Hz (above this it's comparatively easy to
deal with them using acoustic tiles).
You can calculate your room modes by hand (each
lowest Mode Frequency = 6720/Dimension in inches), but why bother, when
Ethan Winer has written the freeware PC Modecalc utility (www.ethanwiner.com/Modecalc.exe),
which displays the first 10 modes for each dimension of your room and
then plots their frequencies across the bottom of the window?
|
The single-dimension 'axial' modes displayed by Modecalc
are the strongest ones, but you also get tangential modes that involve
two dimensions at half the energy, and oblique modes that involve three
dimensions at one quarter of the energy. If you're interested in
exploring these too, you'll find another handy Room Mode Calculator that
uses Javascript at McSquared (www.mcsquared.com/modecalc.htm).
When you use either of these utilities, just look
for any areas where several modes pile up at similar frequencies. That's
a recipe for a particularly bad 'honk' that will be more difficult to
treat. The very worst room you can choose for your studio is a cube,
since all three sets of modes are identical, resulting in an extremely
lumpy frequency response. So if you're lucky enough to have a choice of
rooms for your studio, or are in the fortunate position of being about
to build a studio from the ground up, it's well worth doing some
preliminary checks to find the one with the most suitable dimensions —
after all, while you can install loads of acoustic treatment to tame
most problems after the fact, choosing a room that's as acoustically
neutral as possible in the first place should make your life a lot
easier.
There are various 'Golden Ratios' of width, depth,
and height for rooms, that space the modes more uniformly and therefore
sound better, and Modecalc displays some of these. However, RPG's Room Sizer utility (www.rpg-europe.co.uk/Products/software/software.htm),
which runs on Windows 95/98/2000/NT/XP, offers a rather more
sophisticated approach with more accurate predictions. After you enter
the smallest and largest dimensions available for your room, it will
search for the best room within these limits and then display the
frequency response results between 20Hz and 300Hz for the best and worst
dimensions in your range. If you're choosing between several existing
rooms, you simply input identical smallest/largest dimensions for each
room in turn and Room Sizer will predict how they will perform,
so that you can decide which is going to be easier to work with.
Obviously, modal pile-ups are bad, but this utility cleverly tells you
how much they will affect frequency response.
Once you're happy that your room isn't going to be
terrible to treat, you need to find the best position for both you and
your loudspeakers. As mentioned earlier, the room modes will result in
various frequency peaks and troughs at different positions around the
room, so what you hear from your speakers will depend to a large extent
on where they are placed (unless you've already installed lots of bass
trapping to damp down the modes and flatten the response of the room —
more details later).
|
To complicate matters, the speaker/boundary and
listener/boundary distances also produce comb filtering. For instance,
if your speakers are spaced two feet from the front wall, there will be a
'null' at 141Hz (the frequency at which this distance is a quarter
wavelength, and the direct sound and reflected sound bouncing back from
the wall therefore end up exactly out of phase). There are further
nulls, for example at three-quarters of a wavelength.
This 'boundary interference' is responsible for many
frequency-response dips between 50Hz and 200Hz in smaller rooms, but
once again Ethan Winer has a handy PC utility you can download (www.realtraps.com/sbirlbir.htm). The DIY-er can use SBIRLBIR.EXE
to find out whether a particular dip in room response is due to
boundary problems. You can enter a problem frequency and the utility
will tell you what distances it relates to, or enter a particular
distance (between your head and the rear wall, for instance) and it will
give you the series of related frequencies. Once you've found a
correlation with your room's response (most easily measured by playing a
vanilla sine-wave synth patch from a keyboard and listening for those
notes that are significantly louder or quieter than the others) you can
try moving your speakers or changing your listening position to reduce
the problem.
The height of your speakers from the floor needs to
be such that the tweeters are pointing at your ears for the best
high-frequency response, so this depends to some extent on whether
you're sitting on a sofa or at a mixing desk. Meanwhile, the horizontal
spacing of the speakers should be symmetrical about an imaginary line
drawn down the middle of your room from front to back. Their spacing
will also depend to some extent on how far away from them you're
sitting. Most experts recommend an equilateral triangle arrangement,
with the speaker distance from your ears the same as the distance
between the speakers.
As you can see, just plonking your ears and your
speakers at the most convenient spot in your studio can seriously affect
what you hear, and therefore your mixing decisions. Even once you
understand all the rules it's still difficult to decide on the best
positions for everything. Good rules of thumb are to place your ears 38
percent of the distance from the front or back walls, since this avoids
the major modal null points, and to place your speakers either several
metres into the room or very close to the front wall, to avoid the worst
boundary effects.
Having said all this, there's a much easier and more
thorough way to determine the best position for you and your
loudspeakers, and that's RPG's Room Optimizer utility (URL as above). Like its Room Sizer partner, Room Optimizer
runs on all Windows versions, is designed for rectangular (cuboid)
rooms, and can cope with stereo, surround and more complex setups with
up to 20 loudspeakers, including subwoofers. When you've entered your
room dimensions and user preferences, it carries out the
number-crunching for loads of different positions for you and your
loudspeakers, shows you the best and worst results, and suggests what
additional acoustic treatment would benefit the room
and where to place it. Room Optimizer
calculates both speaker-boundary interference and the modal response
(both of which are available in graph form for further study), aims for
the smoothest and flattest low-frequency response between 20Hz and
300Hz, and can help with decisions about the correct height for speaker
stands and seating, as well as speaker positioning. It's an extremely
clever utility that should prove very useful to anyone who owns a
project studio, or for 'home theatre' users who simply want to find out
the best place to site everything without getting involved in lots of
practical measurements. This is particularly true of surround setups,
where the possibilities multiply alarmingly. However, the software could
also save more serious studio owners a lot of time!
You can download demos of both Room Sizer and Room Optimizer from RPG's web site. RPG Europe (www.rpg-europe.co.uk)
told me that they are happy to sell both of the utilities on a single
CD-ROM for just £60 plus VAT and delivery, which seems very reasonable
to me.
While prediction software is extremely useful, when
it comes to the crunch there's nothing better than actually measuring
the acoustic response of your room — and, contrary to what many people
think, it's not terribly difficult. Essentially all you need are your
loudspeakers, an audio interface, a suitable microphone/preamp and
suitable software, connected as shown in the diagram opposite.
|
Doug Plumb's $150 ETF (Energy Time Frequency) utility, which you can find at www.etfacoustic.com,
has been extremely popular with PC users over the last few years,
because it lets you see your acoustic problems and how they benefit from
bass trapping, diffusion, and other treatment, using almost any PC
soundcard with a stereo line input and output. When you're running ETF,
the left output is always connected to the left input, both to
synchronise the test signals and to correct for any tone-control or
other frequency-response anomalies in the audio interface, while the
right output goes to your loudspeakers and the right input is connected
to the test microphone/preamp.
Once you've run a level check and then the actual
test signal (which lasts just a couple of seconds), an Impulse Response
measurement appears in the main window. It's tempting (particularly for
the novice) to leap straight into the various frequency-response views,
but the Impulse Response display can tell you a lot about what's
happening in your room: the main spike is the initial direct signal from
the loudspeakers, while any other (hopefully much lower-level) spikes
that arrive later on are reflections arriving after the sound has
bounced off your walls, ceiling, mixing desk and so on.
The speed of sound is approximately 1000 feet per
second, so if, for instance, there's a secondary spike on the impulse
response display at 5ms, the reflected sound that caused this has
travelled an extra five feet, so you should be able to guess at the
reflection point. (It will probably be one of the so-called 'mirror
points' on the side walls and ceiling, where, if a friend holds up a
mirror, you can see a reflection of one of your loudspeakers). When
you've found the causes of any IR spikes you can fit some acoustic tiles
at those points to remove the reflections, which will tighten up your
stereo imaging no end.
When you do start examining the frequency response
of your room, you'll probably be shocked by how lumpy it is,
particularly the low-frequency response between about 20Hz and 300Hz.
However, the low-frequency 3D waterfall response displaying the same
information over time will be even more of a shock: you'll see how the
various room modes 'ring' like organ pipes, the worst ones possibly
ringing for several hundred milliseconds. Any bass notes that use these
frequencies will not only sound louder than the rest, but also boom like
crazy!
Now you can start adding bass traps in the corners
of the room to see how much they 'dry up' the ringing and smooth the
low-end response by damping peaks and raising troughs. For me, one of ETF's
best features is its 'Sequential Data Acquisition'. Any of its displays
can be updated in real time every couple of seconds, so you can move
your bass traps around the room while watching the frequency response of
your room change, to find their most effective position, or shift your
speakers or listening position a little to see if the response becomes
flatter.
From the same developer as ETF, there's a new utility called RPlusD (Response Plus Distortion) that is more comprehensive and easier to learn and use. It's a direct replacement for ETF5, although not all of the latter's features are yet in place. Existing ETF5
users can request free registration details for this new product and it
costs $150 for newcomers, so if you're interested in either ETF5 or RPlusD, I would suggest you buy ETF5 and then get RPlusD free as well.
RPlusD makes taking measurements a lot easier than ETF5
does, using a completely different test signal that provides a readout
of input signal level, its signal-to-noise ratio and the distance
between mic and loudspeaker. The beauty of the S/N readout is that you
can make the test signals coming from your loudspeakers a lot quieter
and still be sure that you're not compromising your results. Only if
your studio has a high background noise level will you need louder test
signals. The mic/speaker distance readout is handy if you need to
measure the combined output from left/right or surround speakers,
because if your mic isn't exactly the same distance from each speaker
there will be anomalies in the measured high-frequency response.
Once you've captured one or more room responses, you
can start to experience the benefits of the sophisticated
click-and-drag zoomable display of this software. It's possible to
continue taking extra measurements and get them overlaid on the same
display, which makes comparisons a lot easier than with ETF5,
and you can also average up to 32 measurements, so you can move the
microphone around to get more of a feel for room trends. Like ETF5, RPlusD
has various smoothing options that modify its graphs to be closer to
what the human ear hears, as well as an intriguing new psychoacoustic
option, which takes that idea even further.
The 'R' of RPlusD stands for 'Response',
and this part of the utility is designed to emulate various DSP hardware
EQ units, such as Behringer's DSP1124P, with the aim of helping you set
them up to flatten your room response. It does this by identifying
response peaks and then performing some clever calculations to generate a
set of parametric filter curves, for frequency, Q and level, that
should flatten the peaks. These can then be fed into your equaliser.
Unfortunately, equalising the direct signal instead of installing
suitable acoustic treatment is rarely a good idea in the studio, as it
causes both direct and reflected sounds to end up with skewed frequency
responses. However, I take author Doug's point that hi-fi and
home-theatre users, who may find that extensive bass trapping isn't
practical in their listening environment for aesthetic reasons, can
achieve cost-effective improvements by carefully positioning their
loudspeakers and using a little corrective EQ.
Also new in RPlusD are three-band Energy
Time Curves that enable you to separate out the reflections from
different drivers (woofer and tweeter, for instance) in your
loudspeaker. RPlusD is useful for loudspeaker designers, since
its filter response features can be used in the design and optimising of
DSP crossover filters. Loudspeaker distortion can also be measured,
using a special test signal and a microphone.
Overall, RPlusD is an extremely capable
utility with plenty of new features, and there's no denying its power in
experienced hands. However, it's still complex for the novice, so you
can download an ongoing series of short SWF (Shockwave Flash Format)
movies that cover specific aspects of the utility and demonstrate how to
get the best from it.
Microphones For Acoustic Tests
You'll need a suitable microphone to capture the
'sound' of your room for acoustic tests. While you can obtain
rough-and-ready acoustic measurements using a typical cardioid model as
your test mic, the well-known bass proximity effect will skew the
low-frequency results. More accurate results require a mic with an
omnidirectional response that captures the sound coming from all around
the listening position. However, you don't need an expensive one — as
long as its frequency response is relatively flat compared with the huge
peaks and troughs of most rooms, it will be accurate enough for
acoustic measurements. One popular model is Behringer's £35 ECM8000
measurement mic, which apparently also works well for recording acoustic
guitar and vocals!
You'll also need to amplify the output of your mic
to line level, in order to use it with your audio interface and PC
software. If your audio interface already has a mic input, plug it into
that; if not, you'll need to pass it through an external mic preamp
first.
If you're intending to buy an omnidirectional mic
primarily for acoustic measurements, consider instead buying a cheap
portable battery-powered SPL (Sound Pressure Level) meter, which is also
useful for other tasks in the studio. This includes an integral omni
mic that's quite flat enough for acoustic purposes, as well as a
built-in preamp whose output you can connect directly to a line-level
input in your audio interface. The most popular model by far is the
analogue version of the Tandy/Radio Shack SPL meter. Although Tandy no
longer have shops in the UK, you can still buy their products from
companies such as BK Electronics (www.bkelec.com), who sell the analogue version for just £26.
The advantage of buying one of these is that you
can also use it to balance the levels from the various speakers in a
surround system, check sound levels at gigs, and calibrate the level of
your monitors. If you mix at low levels, you'll be tempted to add more
bass, while if you listen very loud you're likely to end up with a
bass-light mix at more normal listening levels, so adopting a
standardised volume will help you.
Film mixers adopt an calibrated level of 85dB SPL
with a -20dBFS test signal, but in the smaller confines of home or most
professional recording studios this is likely to be too loud, so a more
practical reference is 79dB SPL. To calibrate your speaker levels to
this standard you simply need to play back a pink-noise test signal
whose RMS level is -20dBFS. You can get one of these from the download
page of mastering engineer Bob Katz's web site (www.digido.com).
Place your SPL meter where your head normally is at the mixing
position, and if you have weighting and ballistic controls on the meter
you need to set them to 'C' weighting and Slow response. Then adjust
your speaker levels one at a time so that each one measures 79dB SPL.
|
Room EQ Wizard
RPlusD's main competitor was also primarily designed to set up room equalisers. Room EQ Wizard
(see screen at start of article) started life as a Java application
written by the Technical Director of TAG McLaren Audio, to assist users
of their room-equalisation filters, but then evolved into a more
comprehensive acoustic-testing utility. It runs on Windows, Mac OS X and
Linux and is a free download from the Home Theater Shack forum (www.hometheatershack.com/forums), which you'll need to register with first.
The help file is easier to follow than those of ETF and RPlusD,
and it guides the newcomer through setting up audio inputs and outputs,
the test signal level through your speakers, your interface's input
sensitivity, matching the software's SPL meter reading with that of your
hardware SPL meter, and then calibrating the audio interface itself so
that you can compensate for any frequency-response failings in your
acoustic measurements. It also includes compensation curves for the
various Radio Shack SPL meters (see 'Microphones For Acoustic Tests'
box).
|
Once you've captured a measurement, you can study
the Impulse Response, Low-Frequency Spectral Decay and Waterfall
responses, just as you can with ETF5 and RPlusD. I found such measurements taken in my studio very consistent across all three utilities, which is encouraging. While REW doesn't let you load multiple measurements in different windows, like ETF5, or overlay dozens of measurements and average them in the same window, like RPlusD, I soon found a way around this: REW
does cater for up to eight channels of surround speakers, so you can,
if you wish, store sequential results from the same speaker in these
various channels and rename them, to make before-treatment and
after-treatment comparisons easier.
Room EQ Wizard may lack a few of the more sophisticated features of ETF and RPlusD,
but its user interface is far easier and more rewarding to use (in
particular, the flexible zooming capabilities of the main display
windows and the 3D rotation of the waterfall plots). As with RPlusD,
you can (and probably should) simply ignore its filter-matching
features for studio purposes and concentrate on measuring the benefits
of fitting bass traps and acoustic tiles, although if you do have a
31-band parametric EQ available I'm sure you'll be tempted to try the EQ
features.
Although I've already used ETF5 for several years, the more experienced I got with Room EQ Wizard, the more I found myself liking its approach. I did miss ETF5's
looped real-time capture mode, which is so useful when you're marching
round the room holding a bass trap in different positions, but if you
want a free utility that can help you improve your studio's acoustics, Room EQ Wizard is pretty much unmissable.
Overall, this collection of PC utilities can help a great deal with a variety of studio acoustic problems. Although you'll need to be patient to get the best from them, and you'll probably have to spend some money on acoustic treatment, the reward is that your speakers and sound sources will all sound more expensive, while you'll be able to hear details in your mixes that you've never noticed before, and be more confident that they will sound just as good whatever the playback system. Sounds good to me!
No comments:
Post a Comment