Tuning your subwoofer: How-To (setup and optimization procedure for a subwoofer and main speakers with REW)
This forum has served me well all these years, and I would like to give something back. There is plenty of advice hidden in a lot of different threads here regarding the use of REW to optimize your subwoofer settings. I only miss one of them gathering all the details together. Since I got my PB12Plus, I had to go through the whole REW/setup learing procedure. So hopefully I can list all the steps that you need to go through without ignoring some stuff because it is too "obvious".
I would advice against buying expensive test gear. If you have silly amounts of money available you can probably hire someone (your dealer?) to do all this for you professionally. Better test gear does not automatically imply better sound, because your room is as much involved in the sound quality as your whole system. I would also advice against buying an expensive AV-kit if you are not going to care about setup, for the reasons already stated.
Tuning your subwoofer to optimize its performance may seem like a difficult task at first and will probably require some hours to familiarize yourself with the software, the procedures and to take a lot of measurements. Then again, you have spent lots of hard-earned money on a high-quality subwoofer. So why should you drop it in the most convenient corner and leave it to boom away all of the time?
You may set your subwoofer “by ear” and think it sounds excellent. This is usually the first impression, since it opens to you the door to frequencies you had never heard before, that is, sounds in the movie/music album that you are now discovering. But you can get much more by tuning it properly. The golden rule is “if you like it, then keep it like that”. But you should first hear your system properly set-up before drawing conclusions.After that, it is up to you if you want to boost or lower the bass a little bit.
I am limiting my explanation to the integration of the subwoofer with your main speakers (front left and front right loudspeakers). Theoretically, you should consider all 5 or 7 loudspeakers in a surround sound system. Then study the integration of your subwoofer with all of them individually and in conjunction with each other. I can well imagine such a task goes well beyond the zest of the most ardent AV-fan.
Some of the settings discussed here may not apply in your particular case. Ohers may be available to you, but not mentioned in the following (you are welcome to comment them, and useful ideas can be added to this guide).
No particular level of knowledge is assumed. You always refer to your equipment manuals for detailed information regarding the connection of your amplifer/AV Receiver with your subwoofer, loudspeakers, CD-player, DVD-player, etc.
Only the most common and simplest case will be described here: an AV-receiver to which the loudspeakers are connected in the usual way, with the subwoofer connected to the low-level subwoofer output socket. If your setup differs from this, then you probably have either a package or a more complex system. If the former, then refer to your manual. If the latter, you know more about it than I do.
The best test microphone for the average AV enthusiast is actually a simple, inexpensive, analogue Radioshack SPL meter. If you have something better than this then you also know more about it than I do. But please note: Only the Radioshack SPL meters [and the Galaxy 140 SPL meter] have the essential REW calibration software files available. The calibration files correct the bass insensitivity of these relatively cheap SPL meters and the appropiate one to match your SPL meter must be used.
*NO* other SPL meter or microphone can be assumed to give a correct frequency response using the REW software. The frequency response becomes a completely unknown quantity and the results will probably be quite useless. The Radioshack meter cost about £30GBP from BK Subwoofers in the UK.
I. Test equipment required
1. Your active (amplified) subwoofer and main loudspeakers
2. An AV Amplifier or Receiver (AVR)
3. A personal computer (PC) or laptop having a soundcard with a line-in. The soundcard should meet or exceed the hardware requirements for the REW software.
4. A Radioshack analogue-type SPL meter. Also offered under different names. Tandy etc. Provided it looks *exactly* like the illustration it will be fine. The digital RS SPL meter is also suitable. Avoid alternatives!
5. REW software (Room Eq Wizard). Your browser will easily find it. REW is hosted on another forum. You will have to register on the forum to access the free REW Download area.
6. A tripod to hold the the SPL meter at ear height at the listening/film watching position. A tripod is not mandatory as the SPL meter can be placed carefully on the back of an armchair or sofa provided it won't fall off and break.
NOTE: To avoid endless repeats in the following text: RCA = Phono.These are simple mono plugs and sockets used for Line level (or low level) signals. Stereo connections require two plugs and two sockets. RCA plugs are male. RCA sockets are always female. Many are colour coded to aid correct identification and connection. Always pull on the plug itself rather than the delicate cable itself when removing an RCA plug. A slight twist might help if the plug is very tight. Always switch off the amplification equipment before making or breaking a connection!
Line level = Low level. These terms are interchanegable and mean exactly the same thing in the context of AV and Hifi. YOur equipment manuals and cables may be labelled as either. These cables carry only a fraction of a volt and are screened by a coaxial tubular mesh just inside the insulation to avoid hum pickup.
For simplicity it is assumed that you own an AV Receiver.(AVR) Surround amplifiers, processors etc. will each need their own settings but will still follow the same general instructions as an AVR.
5. Connection cables (see Foreword)
[a] One mono coax cable with an RCA male plug at each end (AVR LFE or Sub-out to subwoofer's low level input). Special subwoofer cables are available and may offer improved screening at higher cost.
Your usual loudspeaker cables connecting your Front L & R speakers to your amplifier or AVR.
[c] A Y-splitter to send the signal to both speakers simultaneously (RCA, 1 female to 2 male)
[d] One stereo low-level coax cable with 2xRCA male connectors on one end and a 3.5mm headphone jack on the other one (computer line-out or headphone out to cable [c], and this to one amplifier stereo line-level input).
[e] One mono coax cable with an RCA male connector on one end and a 3.5mm stereo headphone jack on the other end (to connect the SPL meter to your computer Line-in socket, a stereo coax cable will also work).
[f] A line-in stereo cable with a 3.5 stereo headphone jack on each end (to calibrate your soundcard).
Always refer to your receiver/amplifier and subwoofer and loudspeaker user manuals if in any doubt regarding the actual connections.
The test cables need to be long enough so that you can plug the various items into each other. A PC is usually a nuisance to move around so your test signal cable will need to reach from the computer to your AVR. Cables are easily availbale in 10 meter lengths if necessary.
The subwoofer cable should be long enough so that you can move the subwoofer to lots of different positions in your room while still connected to the AVR.
The cable from the SPL meter at the listening/film watching position must be able to easily reach the computer.
Cable [a] from the amplifier's subwoofer output (RCA type), to subwoofer line-in input (RCA type)
Cables from the amplifier's speaker outputs to the speaker's binding posts.
Cable [c] male RCAs to one of your amplifier's stereo line-level L and R inputs. The left (usually white) RCA male of cable [d] goes to the female RCA of cable [c]. The headphone jack of cable [d] goes to your computer line-out/headphone-out.
Cable [e] connects the SPL meter (RCA male end) with the line-in of your computer (headphone jack end).
Refer again to your manuals for connections between amp, loudspeakers and subwoofer!
If everything is fine, then switch on your sub at very moderate gain levels, same with your amplifier. You will turn them loud later.
It is assumed that you have REW downloaded along with the matching SPL meter calibration file for your SPL meter. REW should now be on your computer screen.
Use REWs test tones to check that the connections are done properly (sound coming from subwoofer and both loudspeakers, SPL meter signal is shown in REWs display. Remember that you turned down the volume in your AVR, adjust it now to a loud but comfortable level, and then turn up slowly the subwoofer gain to roughly match the speakers). Check the audio settings in your computer (receiving sound from line-in, delivering sound to line-out/headphone-out).
TURN THE VOLUME/GAIN RIGHT DOWN ON THE AVR BEFORE PULLING OUT ANY
HEADPHONE OR LINE-OUT PLUGS.
REW has a nice help describing how to use it, so I won't repeat that here. But I tell you what you need:
1. calibrate your sound card
2.calibrate your SPL meter (Follow the setup instructions of REW. For SPL meter settings, see next subsection. This can only be done after you set up everything else, so keep it in mind, and keep reading)
3. select correction curve for SPL-meter (in principle, only for Radio-Shack SPL meter)
4. Use two sweeps of 512k up to 200Hz (longer/more sweeps do not result in meaningful better resolution)
5. For the graphs, always use 1/3 octave smoothing
6. For graphs, scale from 10 to 200Hz and 45 to 105dB
As default, the left (usually white) channel is used for output. Using the Y-splitter, I send the signal to both channels simultaneously.
C. SPL meter
1. Placed at the listening position, at ear height and pointing towards the front speakers, at best on a tripod.
2. Connection to the computer are already explained.
5. 70dB scale
NOTE: Using the 70dB scale on the SPL meter means you will run the REW test sweeps quite loudly. I found that they were too much for the neighbours and now use the 60dB scale. Just make sure everything is quiet in the room while measuring at this lower level. .
1. Phase to 0 deg
2. PEQ disabled (if available)
3. Crossover/filter disabled or at maximum/bypass.
4. Placement in a corner near the main speakers is always a good place to start
5. As previously commented, little gain to start with
E. AV Amplifier/Receiver
1. Use a plain stereo setting in your amp, no ProLogicII or Equalization (we will make sure later that this is the case). The “direct mode” often bypasses also the crossover of the main speakers.
2. Subwoofer gain to 0dB, loudspeaker gain to 0dB.
3. Crossover to 80Hz, just to start with (unless your speakers' manual says otherwise!).
4. Set the loudness level so that you can make a reading in the SPL meter, but make sure it won't go out of range (watch out that needle in the SPL meter while taking measurements!). Note that this will affect the SPL calibration in REW.
If you leave open windows or doors, it will probably change your frequency response. Big objects will absorb/reflect sound, changing the response too. Do not place anything between your loudspeakers and the SPL meter when testing.
III. Measuring Procedure
Run a sweep from 0 to 200 Hz (follow REWs manual). You got your first curve! Repeat the measurement and see that you get exactly the same curve, just to check that everything is fine.
If no sound comes from your sub/loudspeakers, check the connections/settings/levels.
If REW complains about too low levels, reduce the scale/increase the loudness.
If REW complains about too high levels, reduce the loudness.
I repeat, make sure that the needle stays in the range all the time.
If everything went fine, your curve should look more or less like this:
Figure 1. First curve with REW. The peaks and valleys are due to room-conditioned resonances and cancellations, which can be somewhat avoided by positioning, phase adjustment, etc. On the horizontal axis, the frequency is shown in Hz. For each frequency value, a corresponding sound pressure level (SPL) can be read out of the plot. For example, at 100Hz the SPL is 75dB. The blue line is the ideal response (without house curve).
You will probably see that your sub is too low/high with respect to your mains, and nasty valleys/peaks. You can change the subwoofer gain carefully to balance the subwoofer and loudspeakers relative levels. Then try another test sweep to see if you have achieved the correct level on the frequency response curve. Once this is achieved check the gain setting on the subwoofer knob for future reference.
a. You can give a name to each measurement (in the measurement info tab) and save them separately or as a set.
b. There is also a field (again in the measurement info tab) where you can add comments regarding the measurement: things like subwoofer placement, gain, speakers on or off etc. Use this! Otherwise you will end up with lots of curves which you cannot tell from each other.
c. You can change the color of each curve in the same tab.
You can also save test curves as JPEGS and they will appear automatically in your "My pictures" folder.
IV. Optimization method
If you aren't happy with your system's response let us examine which possibilities you have to improve things: The target for absolute perfection is a ruler flat frequency response (equal dB for all frequencies). Real life says +/-3 dB is excellent. While +/-5dB is reasonable (well, at least in my room...), note that +/-5dB means up to 10dB difference between different points of the frequency spectrum (!).
A rising frequency response into the deep bass is popular. Called a "house curve", this boosts the lowest frequencies a little givign more impact. So if you have such a response from your subwoofer in your room then you should probably be grateful. Usually this is only the case with large subwoofers. Smaller subwoofers tend to have a round shoulder on the response curve at lower frequencies as they run out of steam. Or rather lack the ability to move enough air to reproduce these frequencies at high levels.
The steps you should follow for the optimization are (ordered! In any case, change always only one parameter, and then measure to see its effect. And keep track of the steps, so that you can go back and/or do not repeat yourself):
- Location. With the speakers unplugged, make a curve for every subwoofer and sitting location that may be available. Small distances may play a crucial role here! Just check to which extent you are affected. Choose the location with the flattest possible response.
The relative location of subwoofer and speakers also plays a role (a complex one), so you may also look at the combined response for the different locations. But in principle, you should also adjust phase, gain, crossover..., for each location, in order to get the full picture. This is a lot of work, though.
Note: we know from long experience that moving the subwoofer to different positions in most rooms will change the frequency response. This is because walls (and corners in particular) reinforce bass frequencies. How near your subwoofer is to a wall or corner gives you a method of changing its natural response in a room. Most people prefer their subwoofer near a wall if not actually in a corner. Placed out in the room, the box will trip people up or impede natural foot traffic flow. If such a position happens to offer perfection in its response, then you may wish to drag the subwoofer to that position when watching films or listening to music. Most people won't want to do this regularly so we have to use the walls and corners as best as we can to optimise our subwoofer's response.
- Room. If your room is rather 4 naked walls, you will suffer much more the resonances (peaks in the response). Furniture can help to some extent, and then there are so called “bass traps”, to avoid reflections of the sound (which are the origin of the resonances). You can try adding stuff and see if the peaks/valleys remaining from the first step are removed. This may change your room's response, so you must revisit the first step again.
- Subwoofer response.
Run the sweep tone for several crossover settings (40, 60, 80, 100, 120 Hz).
Save this set of measurements as "subwoofer response alone". To see all curves at once, go to “all measured”.
They should look something like this (otherwise, you may have some setting wrong in the AVR):
Fig. 2. Subwoofer response alone for several crossovers. From the lightest to the darkest pink: 40, 60, and 80Hz crossover. The black line corresponds to a 200Hz crossover setting.
4. Loudspeakers response aloneRun the sweep tone for the same crossover settings and save the set of measurements as "loudspeaker response alone".
Switch off the amp and connect the loudspeakers. Switch off the sub, switch on the amp.
Switch off the amp and connect the loudspeakers. Switch off the sub, switch on the amp.
You may also be interested in seeing the response of the loudspeakers without crossover settings (set mains to large and no subwoofer in your av amp, and use a longer sweep for better resolution).
They should look something like this
Fig. 3.Loudspeaker response for several crossovers. From lightest to darkest green, 40, 60 and 80Hz. The black line corresponds to the response without crossover.
If you see too deep valleys/peaks, you should consider re-positioning your loudspeakers (away from walls). Changing your listening position will also help. Make sweeps at different listening positions and you will see if you can improve something. In any case, change always only one parameter, and then measure to see its effect. The position of the loudspeakers will also be determined by other effects, like imaging. Play also with some toe-in, it may help.
5. Gain and Choice of crossover
Once you found the best spot for your speakers and subwoofer, and have all the graphs described for this positions, measure the combined response of sub and speakers. Start with a 80Hz crossover. If the sub is too low/high, turn the gain up/down in small steps, measuring until you see it is even. You can use your receiver later if you want to run it a bit "hotter". I like to run a sweep up to 20kHz to see how loud the bass is with respect to the rest of the spectrum.
Then, run sweeps for various crossovers. The capabilities of your loudspeakers will also play a role, but usually 80Hz is adequate.
Fig. 4. Black, purple and green curve have -5dB, 0dB and +5dB in the subwoofer gain. The crossover at 80Hz is noticeable
There is usually one optimal phase setting for each crossover, so you may want to go to next step for each crossover.
6. Tuning the phase
By changing the phase, you will see that the sub and speaker signals at the crossover region add up or cancel each other. Make a set of measurements with 10 or 15 degree steps. Choose the one that makes the smoothest transition.
Fig. 5. Effect of the phase.For a crossover of 40Hz, the phase can add up the combined outputs of subwoofer and loudspeakers around the crossover region, or cancel them out.
7. Parametric Equalizer (PEQ)
If your subwoofer has a PEQ, you can bring down the most dominant peak and make the response more even. Refer to the instructions from your sub's manual. This must always be done at the end!
Fig. 6. Effect of the PEQ. In blue, curve without PEQ, in black, PEQ applied to the peak around 55Hz. Light pink and light green show the separate response of the subwoofer and loudspeakers at the crossover setting used (80Hz).
Note that I included the response of subwoofer and loudspeakers alone as light lines (always for the settings of the combined response measured!). I would recommend always including them in the graph.
V. Conclusions, or "the bass-nightmare room"
So, you have followed all the above, but cannot avoid having strong hills and/or valleys in your frequency response. What to do next?
The definite solution is a multi-band parametric equalizer. A clear affordable favourite is the Behringer Feedback Destroyer (BFD) 2491. The easiest way is the Velodyne SMS-1, which has auto-setup, but does not come cheap.
IMPORTANT NOTE: dips caused by room-conditioned cancellations cannot be cured by equalization. Any attempt to do so can damage your subwoofer. You may boost some frequencies, but never much. Keep in mind that boosting will reduce the headroom of your subwoofer, limiting its maximum SPL.
In principle, subwoofers are like loudspeakers and are subject to break-in to some extent. This means that your subwoofer may change its characteristics during the first hours of use. (or dozens of hours or even hundreds). Some re-adjustment might be needed after some weeks/months if your sub is new on the first REW tests. Any change is likely to be small but it would be interesting to go back and test your subwoofer again after a period of use to see if the frequency response has changed at all.
Acknowledgements: I would like to thank Chris (Nimby) for his valuable input and corrections, and for proof-reading this guide.
Concluding notice: a new release of REW (v.4.11) is now available. I have written this guide based on the version 3. Since it only deals with the most basic REW features, it should be ok also for v4.11
VI. Appendix: AVR and player settings
This guide assumes that you will feed either plain analog stereo signals or digital ones into your AVR. In this case, you should disable any crossover/bass management in your player(s), since it will interfere with the one in your AVR. AFAIK, some players even apply those settings to the digital output. Always apply bass management at only one place.
If you use 5.1 analog inputs, you need to be careful, since full bass management in 5.1 analog inputs is only available in the most expensive models, AFAIK.
In certain cases, it may be advantageous to use the 5.1 analog outputs of your player. I can think of three:
1. The AVR bass management is inexisting (I wonder if this can happen...)
2. The digital to analog converters (DAC) of your player are better than those in your AVR.
3. It just sounds better by doing so (golden rule...)
Despite this, keep in mind that using the bass management in your AVR is the only way to apply it to all your sources. If you use the 5.1 analog inputs, the AVR may not apply the crossover to your loudspeakers (you can use REW to find out, PM me for details). In this case you'd need to use the bass management of your player and switch off the one in your player, as I previously said.
Using the player's bass management or the AVR's one should not make a difference, I assume. Thus, if you optimize your sub's response using the AVR's bass management and then use the player's one, it should be ok.