Today I learned how to use Room EQ Wizard to tune my audio room. I had already done room tuning on my own and was happy with the results. But REW enabled me to get it even better.

Here’s the final FR measured from the listener position, psychoacoustically smoothed. The Grey line is without EQ, the red is with EQ. You can see that the EQ is only a few bands from 500 Hz and lower.

It’s linear and smooth, with a typical tapering response. This is a treated room having big tube traps, bass traps, bass resonators and acoustic foam. The parametric EQ is mild with gentle amplitudes and slopes. I’d rather have a few little bumps in the response, than perfectly flat response with bloated phasey sound from extreme EQ settings. Don’t let the cure be worse than the disease!

Overall, this smoothed response throughout the range. During test listening I can switch curves instantly while the music is playing. My ears like the difference, especially noticeable on good acoustic music recordings.

Equipment & Details

• Test audio files created by REW version 5.2 beta 4, burned to DVD-A
• Oppo BDP-83 toslink PCM output
• Behringer DEQ2496 digital EQ, toslink input and output
• Corda Soul DAC-preamp, toslink input, XLR output
• Adcom 5800 amp (28 years old), XLR input
• Magnepan 3.6/R speakers (20 years old)
• Room treatments (floor-ceiling tube traps, RPG acoustic foam, etc.)
• MiniDSP UMK-1 measurement mic, and Rode NT1A mics, both calibrated
• Recorded from the listener position

Here are the rest of the REW plots:

Total distortion averaged about -50 dB (0.3%); higher in the bass, lower in the treble. That seems surprisingly low, considering it’s measured at the listener position and includes all distortion from the power amp, microphone & recorder. Many headphones, even some tube amps, have more distortion than this. These speakers reveal that the NT1A mics have lower distortion than the UMIK-1.

The bad news is that distortion at 40 Hz is about 10%. Yikes! But it’s down to 1% by 50-60 Hz, which would be great for headphones, quite rare for speakers.

I’ve always been happy with the bass response in this room after I treated it. 25 Hz is audible, if attenuated. But seeing these measurements, it seems that getting a subwoofer to handle everything below 60 Hz could “unload” the Magnepans and reduce overall distortion. I don’t want more bass, but tighter cleaner bass is always A GOOD THING.

Group Delay is pretty flat. Rises in the bass as usual. But it’s 10 ms or less from 60 Hz on up, a near perfectly flat zero for most of the range. This seems typical of planar speakers.

Initial impulse response is near zero in about 3 milliseconds, and you can see the reflections at 5 and 10 ms.

Total impulse energy is about -40 dB in the first 100 ms, from the listener position which includes room reverb. Room treatment damps the room, but it’s not completely dead. The grey is minimum phase IR, which is very close to the actual response.

The CSD looks linear (no obvious ringing frequencies above the bass region) and decently fast. The room treatment certainly helps here:

Here’s the Spectrogram. There’s some mild ring around 64 Hz and rising decay time below 50 Hz. Overall pretty flat and even. That’s room treatments doing their job.

Since I treated this room I’ve been happy with the sound. With these measurements I was able to apply EQ to fine tune some things I couldn’t fix with room treatments.

Generally speaking, balanced and differential signaling are two different things. They’re often (but not always) used together, and in audio, the term “balanced” refers to this.

Speakers and Headphones

A speaker or headphone responds to the voltage difference between its 2 input wires. It doesn’t assume either is ground, though one might be, it doesn’t matter. So connecting a speaker or headphone to a balanced output is easy. Just wire (-) to (-) and (+) to (+) whether or not the (-) is a ground (unbalanced output) or carries a signal (balanced output). If the unbalanced output has a common ground for both channels (like a headphone), you can split it to both L and R (-) in parallel.

Converting a balanced speaker or headphone output to an unbalanced connector is not as simple. An unbalanced headphone cable (a standard 1/4″ or 1/8″) has 3 wires: L (+), R (+), and a single wire that is a common ground for the L and R. You can’t connect a balanced output’s (-) wires to this ground. That would mix the channels, and allow the amp’s output stages to drive each other, which is bad because they usually have very low output impedance, so it can overdrive the output stages. Also, you can’t just ignore the output’s (-) wires and connect the headphone (-) wires together; this will give a common floating ground. In short, you need a transformer to do this conversion.

Components

If the balanced/unbalanced conversion is between components like a preamp (not a speaker or headphone), it gets more complex because unbalanced components assume the (-) is a ground, but the balanced (-) carries a signal and its ground is a separate (3rd) wire. You can’t connect a balanced output (-) signal to ground; it will overdrive the balanced output as it tries to swing a voltage over a 0 ohm load. Also, you need to ensure the (-) wire has the same impedance to ground as the (+) wire.

So the best way to convert unbalanced to balanced between components is to use a transformer.

However, you can wire unbalanced output directly to balanced input. Connect the unbalanced (-) output to both pins 1 and 3 on the balanced side (negative & ground), and the unbalanced (+) output to pin 2 on the balanced side (positive). That is, carry the unbalanced source ground through to the balanced input. Since unbalanced (consumer) output is at a lower voltage than balanced (pro), the downstream balanced component will be receiving a lower level signal than it expects. This may or may not be a problem, depending on how clean is the input signal and the balanced device’s input voltage sensitivity and gain.