Studio monitor reference frequency CD

S

Serratus

New member
Hi guys,
I have a small home studio with KRK nearfield monitors and a KRK sub.
I borrowed a reference frequency CD from a friend (which goes through the 1/3rd octave bands from 20Hz to 20kHz) and I set it up with a condensor mic in the normal listening position to see the dB at the different frequencies.
So now I have the levels for each frequency and there aren't any major peaks or troughs, but I want to know what amount (in terms of dB) peaks I should be worried about. There are a couple of peaks at around 400Hz and 1K that a about 3dB higher than the average level for most of the other frequencies. However, the frequencies from 63Hz to 8K are all within a 6dB band.
Obviously this could also be down to the mic, but I'm not sure whether I should worry about it and try to EQ it out, or is this a 'normal' amount of difference ('normal' being a relative word!) that I will just get used to as I learn the sound of my monitors?
Any advice/opinions appreciated.
 
Serratus said:
...goes through the 1/3rd octave bands from 20Hz to 20kHz ... there aren't any major peaks or troughs ... I'm not sure whether I should worry about it and try to EQ it out
Click to expand...

Your real room response is nothing like what you think you measured. :D

Seriously, you need real room software like ETF or the freeware Room EQ Wizard to see your room's true response. Peaks and nulls are often much closer than 1/3 octave apart, and real room software can resolve to less than 1 Hz. Most small untreated rooms have peak/null spans of 30 dB. Much more here:

Using ETF

As you can see in the graph below, the 1/12th octave response (blue line) reveals the room's peaks and nulls far more accurately than the exact same data plotted at 1/3rd octave resolution (red line). When you measure correctly using room testing software, the disparity is even larger.

As for suitable microphones, see this article:

Comparison of Ten Measuring Microphones

Finally, EQ is not suitable for fixing most room problems. Bass traps and reflection absorbers are the correct solution for that.

--Ethan

art_mon3.gif
 
Listen to the test CD with your ears,

The test CD should have a track that sweeps through the full frequency range
Sit in the mixing position with your eyes shut.
Play the test CD sweep track on repeat.
If it's got the same track backwards play that too on repeat.
Listen closely, are there any frequencies that jump out at you or are lost (apart for the high hi's and the low low low's)?

If there are find out what the frequency is by using the frequency tracks of the CD, these are the frequencies that need treatment.

After all it is your ears that will be mixing the music, not the test microphone.

Cheers

Alan.
 
I agree with Alan in part, but Ethan is dead on. The problem it total reliance on your ears is that they get used to ambient conditions so fast, just like your eyes. When you walk into a room your eyes adjust to the lighting conditions very quickly even when they vary by a great deal. It's exactly the same with your ears. You'll get used to the problems very quickly. That's why important to at least get some input in terms of analysis.

The other thing is that 1/3 octave is *really* course. The lowest I'd trust would be more like 1/12, and only for a general look at the room acoustics.

Frank
 
(First, let me say that I respect Ethan Winer's vast knowledge and professional opinions on acoustics. I take most of what he has taught us as gospel and I am grateful he is here helping us. I only challenge this one point because the little gears-of-logic in my head seem to always question everything.)

I am not sure I understand what testing with 1/12 octave accuracy will accomplish. When it comes down to room treatments our choices are very generalized as far as frequencies are concerned. Short of building a tuned Helmholtz Resonator or a tuned slat system, we can only treat with bass traps, mid-bass traps, broadband absorbers, high frequency absorbers, and various reflectors.

Using Ethan's chart above, even the 1/3rd octave testing reveals a few general nulls in the lows, mid-range, and again in the highs. So we throw in some bass traps and some broadband absorbers and test, adjust, test again, adjust again, etc.. Most of us cannot treat the room with much more accuracy than that.

Am I missing something?

Go ahead...call me "young grasshopper" if you must.
 
I suppose my suggestion that you listen with your ears and not your eyes is that most home studio people don't have the understanding of what the test charts actually tell them.

If you set up your room, listen to a collection of CDs that you know and like the sound of and they sound good in the room, you are on the right track. If the CDs sound boxy, bassey not bassey, bright, not bright, etc you need to tailor the room with some treatment.

Take CDs of your mixes to other systems and rooms, if they don't sound right it also means you need to fix the room to be the opposite to what is wrong, i.e too much bass on CD, not enough bass in your room.

When you think you have it almost right, there is no harm in running a test for any hidden problems, which will then stick out as their wont be as many to deal with as the room is on the path to being right, but you must have the knowledge to understand what the readings are telling you.

Cheers

Alan.

p.s: EQing of the monitors should never be used to fix room acoustics in a studio. This is done for live sound PA's as you are stuck with the room sound.
 
Alan, believe it or not I completely agree with you in the sense that I don't think that testing should in any way *replace* your ears. I think it can and should be used in conjunction with them to get the best possible solution for every room. I also agree that EQ is not a great solution, especially 31-band filters and the like. Solutions like ARC or ERGO can actually be very helpful once the room is at least moderately treated though.

Why use 1/12 or greater? Because 1/3 won't reveal everything that's sonically important. This is actually a point Ethan and I diverge on: I don't think it's helpful to work a room using anything much finer than 1/12 because many of the problems that are revealed at 1/48 aren't acoustically significant. Even a 30dB null doesn't matter if it's only 2Hz wide. At 1/12 you can definitely hear everything you see on the plot. If you've got a -30dB null at 1/12 you most definitely need to address it. I'm not saying that 1/3 isn't useful, because it is. That's your zoomed-out view...very useful for getting a general room curve.

Make sense?

Frank
 
^^^ I know, but I've never seen nulls with a Q that high. Again the point is that a deep null is a problem if it aligns exactly with a note in the music, even if the null is only 0.001 Hz wide. :D

--Ethan
 
You must log in or register to reply here.
Back
Top