GuitarMasterClass Blog

As a first step to fixing up the acoustics of a studio, its wise to figure out what problems you have first. Measuring the characteristics of a room may sound daunting but it isn’t hard, and gives you a way to figure out if you have actually improved anything after the treatment is complete.

I have been using a program called Room EQ or REW to measure the existing acoustics of my studio and figure out what needs fixing. Its a free program and you can find it here. It was written by a home theater enthusiast - home theater setups share a lot of the same challenges as recording studios in terms of room treatment to get a good response so this is not at all surprising.

The idea behind this program is that you generate a signal out of your speakers and measure the resulting sound using a mic or sound meter. The response of the room is then calculated from this signal and can be visualized.

Before I discovered REW, I already suspected that I had a problem with modes, and had attempted to characterize it with a signal generator and microphone. This pointed out some peaks and troughs in the frequency response but didn’t help much with the time domain. REW is a lot more sophisticated and a lot easier to work with too!

Whilst you can use a microphone with REW to capture the generated signal, microphones themselves vary a lot in frequency response and in theory need to be calibrated to get accuracte results. Rather than worry about that, I bought a cheap SPL meter from Radio Shack:

It cost around $50, and most importantly it has a line level output. SPL meters such as this have a known frequency curve so are in effect pre-calibrated. Actually in acoustic terms there are a variety of frequency responses in use, and they are switchable on the meter - some are optiomized for human hearing ranges for instance. I will be using the so called C weighting profile as REW is setup for this.

So, after reading the instructions, as a first step I measured the frequency response of my soundcard by connecting its input to its output and running a test. This allows REW to factor this out of the equation and focus on the response of the the room and specific interactions between the room and the speakers. The curve was impressively flat which it should be for a high end sound card like the 1212M - some cheap motherboard soundcards won’t be nearly as flat. I saved this as a calibration file and used it moving forward in all future measurements.

Although the speakers are part of the equation here, they will not affect the underlying modes, however different speakers may highlight different modes to a greater or lesser degree depending on their frequency response, most notably, whether or not the low frequency roll off is steep enough to render some of the lowest frequency modes a non-issue due to the lack of acoustic energy at those frequencies.

Next I calibrated my meter using a test signal, then ran some measurements - I ran 3 in series to cancel out any anomalies, and they all turned out pretty much the same. The frequency response of my untreated room looks like this:

Its pretty ugly and shows a couple of serious dips at

• 34hz (not too serious as that is pretty deep bass and I would normally shelve that out of a mix anyway)
• 144hz (a bit more serious as that is in mid bass range)
• 353 & 855hz (not so problematic since these are higher frequencies and easier to treat)

In particular, that 144hz is going to be the one to focus on. As you can see, it takes a while to dip, and then rises again to a peak at around 161hz. what that does in effect is blanket out the first octave of the guitar, and then cuts back in with a mild peak at just below middle E. To compensate I play louder, and that subjectively sounds like a nasty resonance right in the middle register of the guitar. This is the frequency that has been bugging me and has made mixing of guitar extremely hard - this above all is the problem I want to fix, I need to flatten out that whole dip and as much of the rest of it as possible. The way to address this of course is to use bass traps - this is a fairly large deviation, so I am going to need a lot of them.

Another thing I could look at with REW is the time domain response of the room - that is to say how it behaves over time when a sound is made. In an ideal room, once a sound has been created, it should decay at a similar rate across the whole frequency band. When modes are involved however, certain frequencies will stick out and take longer to decay. That same set of measurements can be analyzed in a different way to produce what is called a waterfall graph - its a 3D plot with frequency on one axis, volume on a second, and time on the other. It ought to look like a pitched roof, with all frequencies falling away at the same rate. If it looks like a mountain range you have problems! Mine of course looks like a mountain range …

If you look at the front edge(the 300 ms mark) you can see this. That nasty 142hz frequency is in evidence - it is higher than the rest and looks like a ridge, meaning it falls away more slowly than the rest. We also see a ridge for that 161hz frequency showing that not only is it a peak in the frequency response, but it is also falling away more slowly meaning it is a mode as well - modes resonate and store energy so sound longer than other frequencies.

So there you have it - I knew my room sounded bad, now I can prove it! I’ll be tralking about the specific treatment I will be using in a future post.

Now a word on room EQ - REW has the ability to interface with a couple of different multiband graphic equalizers. The idea is that after you measure the room it can generate a set of filters to even things out. This is a deceptively simple approach and is gaining in popularity - who wouldn’t want to spend $150 on an Equalizer and have it solve all those studio problems without putting up any room treatment? Seductive as this approach is, it is ultimately not the right way to do this in my opinion, here is why.

By their very nature, modes which is really what this technique sets out to cure, vary all across the room. A low in one area will be compensated for by a peak in another area. Treatment by EQ can only really help in a specific location, or can compromise by working less well in a wider area. If you EQ for your mixing position that is fine, but remember, this room will be used for tracking as well as mixing. There is a good chance that fixing the mixing position using EQ will make things worse elsewhere in the room.

Also, working this way only affect the frequency response, it does not help at all with the time domain response. Reflections will not be addressed, which leads to problems in imaging, and various unwanted sonic artifacts.

Finally, whilst EQ can boost a dip, it can’t do anything about nulls - if you have a null whilst you are recording, there is zero acoustic energy for that particular note. No amount of boosting will fix that.

All in all, when you look at this in an informed way, EQ is not the solution, so its back to bass traps and tiles! At best, EQ can be used to fine tune an already treated room but on its own its not the solution in my opinion.

And a final word on room measurement - I had a lot of fun doing this, but ultimately it told me what I already knew; I will need bass traps, and acoustic tiles. If you don’t want to go through the hassle of measuring your room like this, chances are that it too needs bass traps and acoustic tiles. You certainly can’t have too many bass traps - you can have too few. For mixing you almost certainly want to hit the mirror points with tiles (more later), and treat some of your vocal area as well. If you do this, regardless of whether or not you have measured your room, the chances are pretty high that you will improve matters, and you certainly won’t make them worse!

This entry was posted on Monday, May 26th, 2008 at 9:08 pm and is filed under A Studio is Born. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.