Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › Using an equalizer to achieve a flat response
New Posts  All Forums:Forum Nav:

Using an equalizer to achieve a flat response

post #1 of 9
Thread Starter 

Hi,

 

 
I know Head-Fi is generally against EQs, and due to the lack of a specific forum for equalizers or anything of the like this thread may be in the wrong forum, but hear me out. :)
 

I'm trying to get a flat response studio setup on a budget. Since the better monitors cost too much for my budget, I was wondering if I could use an equalizer to achieve it.

 

The current monitors are a pair of vintage passive Pioneers which I have salvaged and refinished to tip-top condition. It's a generic tweeter/woofer with a passive crossover, as expected of a passive monitor.

 

Upon (subjectively) analyzing the frequency response, I've noticed that each driver's peak frequencies have gigantic humps on them and the transition to the tweeter results in a volume drop around 10kHz. To find the exact equalizer setting required, I used a logarithmic sine sweep from 1 to 20kHz in Audacity and played around with the virtual equalizer there to see which settings get me the flat response I'm looking for. I've equalized subtractively as to avoid any clipping issues that may arise.

 

I've since moved onto a virtual 10-band systemwide graphic equalizer for Mac called Boom. I've noticed that such an equalizer is too generic for my needs; it can't make the fine adjustments that I need between, for example, 8kHz and 16kHz, since the tweeter hump is at around 14kHz. I'm looking into virtual parametric equalizer solutions at the moment.

 

I'm mainly concerned about the resultant sound quality, however. Will a virtual, digital or analog EQ affect the sound quality of the monitors? Is it even a good idea to use equalizers to achieve a flat response? If I'm to use a physical EQ with a DAC/amp, where in the chain should I place it, or should I go with a virtual graphic analyzer instead? Would it be more economical to get a DAC/amp with a virtual equalizer or a physical equalizer with the computer's built-in line out?

 

If I'm to go with a DAC and amp combo, I'd opt for an Apogee Duet 2 and an O2 amp; this combo seems feasible as a portable studio setup.

 

Thanks in advance!

post #2 of 9

There's nothing wrong in using EQ. A software EQ can work as good as a hardware EQ.

 

In fact I'd like to ask how did you determine that your EQ settings are 'flat' ?  Did you use a mic?

I want to do something similar with my setup.

post #3 of 9

A parametric or graphical EQ is going to be difficult to adjust precisely enough to compensate for the irregularities in a speaker's response. You can actually do a lot better... look into "digital room correction" and "BruteFIR" (you don't actually have to use BruteFIR... Foobar and other software can apply the necessary convolution filters, but it's one of the more popular tools). DRC can compensate for a speaker's frequency response very well, taking into account the room too. It's fantastic when set up properly. The only thing is you need a calibrated microphone, which adds some cost.

post #4 of 9
Thread Starter 

First to identify the general shape of the EQ I used a logarithmic sine sweep from 1 to 22050Hz.

Why 22050Hz, when I could have used a 96kHz sampling rate and gone up to 48kHz? Because the benefit of 24/96 is most seen not in the highs, but in the mids where oversampling can bring out previously unnoticed detail in the music. Also because the human hearing range is from 1 to 20kHz.

 

So anyway, the sine sweep had an equal amplitude throughout to keep the theoretical volume constant. When played back, there are peaks where each driver is at its best-performing frequency; these peaks are the places where the volume swells and the equalizer for that frequency needs to be toned down.

 

After first subjectively analyzing the frequency response I outlined the inverse equalizer slope necessary to result in a flat response in Audacity. It's really useful because you can plot points to form the EQ curve for a practically infinite-band equalizer! Where the volume was softest was the 0dB gain point and all higher volumes were reduced, thus resulting in a non-clipping subtractive equalizer. Applying this to the equal amplitude sine sweep resulted in a consistent volume throughout the sweep. If it didn't, I kept tweaking the EQ until it did.

 

Using this resultant EQ slope in Audacity, I intend to apply the same equalizer settings to the systemwide equalizer.

 

 

My next step in ensuring flat response would be to record the sweep with an external mic, such as my Shure PG58, which has a documented frequency response. http://recordinghacks.com/microphonesThis website seems to have a huge catalog of recording microphones with the frequency responses measured in their audio lab. If the studio monitor's frequency response is accurately calibrated, then the sweep recorded by the microphone should match the frequency response as documented in that website.

 

I'm trying to use Apple's 31-band graphical EQ (AUGraphicEQ) and/or the parametric EQ (AUParametricEQ), but I'm stumped as to how I can achieve this without having to resort to AULabs, which is $100 for the Apple SDK which I will most likely never use.

 

All of this is, of course, theoretical but it makes logical sense! :D

post #5 of 9
Quote:
Originally Posted by prodo123 View Post

I'm trying to use Apple's 31-band graphical EQ (AUGraphicEQ) and/or the parametric EQ (AUParametricEQ), but I'm stumped as to how I can achieve this without having to resort to AULabs, which is $100 for the Apple SDK which I will most likely never use.

 

All of this is, of course, theoretical but it makes logical sense! :D

 

XCode and the rest of the Apple SDK is free. (It's only $100 if you want a signing certificate.) You don't even actually have to download XCode or the whole SDK though... You can download AULab and HALLab directly. They're one of the optional bundles on the Apple Developer site... you can just download the bundle. It's called "XCode Audio Tools" or something similar and is 36 MB in size.

post #6 of 9
Thread Starter 

I didn't notice that before, thank you so much!

Will play around with the EQ and post back with results.

 

Also, just to make sure, software EQ is just as good as analog, right?

post #7 of 9
Thread Starter 

The AULab solution works a treat! 31 band stereo graphic and parametric EQs are both working well.

 

Only gripes I have are:

1. The lack of amplification; I have to boost things in the EQ to get some default amplification. I guess I might have to invest in a dedicated amp if I want to keep using it.

2. AULab must be open at all times.

post #8 of 9
Thread Starter 

After a lot of experimentation, I've come up with the solution.

Since I'm without an external amplifier, first a gain VST/AU is used to do a pre/post-amp. I used Blue Cat's freeware solutions; you can do channel-by-channel (L/R or if you're a surround sound person, all 6 channels individually!). Then the equalizer is shaped subtractively to fit the curve.

 

For Macs, http://osxdaily.com/2012/05/18/equalizer-for-all-audio-mac-os-x/ has the instructions to set up the Soundflower-to-output that you want. The best part is, if you want surround sound shaping you can use the 16ch Soundflower option for even more options. Pre- and post-gain AUs are necessary for operation without a dedicated amp; if not, you can set the output to your desired DAC, with the Soundflower sampling rate configurable in SysPrefs up to 192kHz which means unlike Boom, this solution will allow your DAC to run at its full potential.

 

Different VST/AU effects can be enabled or disabled in real time, which is an additional bonus.

 

I should write instruction on how to do this; I've set it up and wow, it's so surprising what the crappy built-in DAC of computers can do when tweaked right. I'm excited to hear how this would sound with a dedicated DAC/amp solution.

post #9 of 9
Thread Starter 

Here's the results I got for my headphones (the measly budgety venerable RH5Ma); they sound pretty darn flat except at the 3800-5300Hz bands where the volume dips severely at 4kHz and swells massively at 5kHz. As you can see I tried to fix this with multiple parametric equalizer adjustments (strangely enough Apple's para EQ won't let me do it in a single equalizer) and the effect has been minimized, but the 4kHz dip is still there and it's most likely the headphone's limitations.

Also, because of this, I was forced to undo the subtractive EQ to compensate, which is a pain. Even with these drops and swells, the bass response has been greatly improved and the headphones sound like they're worth five times more.

 

Does anyone know how I can fix this volume swell without introducing gain to other bands? Or if I can add more adjustments to a single parametric EQ?

 

https://dl.dropbox.com/s/xcnosgeiv7avms4/Screen%20Shot%202012-11-19%20at%2012.57.40%20PM.png


Edited by prodo123 - 11/19/12 at 10:11am
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Dedicated Source Components
Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › Using an equalizer to achieve a flat response