Why EQ Sucks - An Illustrated Story

[barefoot]

I experimented with a few PC based EQ's to illustrate their effects on sound. Attached is a file showing how 3 different GoldWave EQ's and an a Sonic Foundry EQ alter the waveform of a 100Hz square pulse. I used these applications because, well, they're the ones I have. I overlaid screen captures for easy comparison.

The pictures mostly speak for themselves, but here is a little elaboration.

Filters

First of all, all filters will alter the waveform. This is the nature of what they do. Beyond this, however, all filters add distortion as well (phase and transient distortion). That is to say, even a theoretically perfect filter adds things to the sound which were not there to begin with. Even if you simply pass your signal through an EQ with everything set flat you are still adding distortion (Software based EQ's can break this rule).

GoldWave Example

I wanted to show the effect of simply passing the signal through an EQ with completely flat settings. Software based EQ's can however have features in their algorithms which act like IF all settings equal zero THEN do nothing. This is the case with the GoldWave EQ's. When everything is set to zero no processing takes place. So I tried setting all the levels flat, but to -1dB. This worked. The EQ's processed the pulse signal and return the waveforms you see in the first example. The "Equalizer", which is described in the Help menu as being similar to those found on home stereos, produced a particularly horrible looking waveform. Remember, these are all FLAT settings!

Sound Forge Example

The SF "Graphic EQ appear to have an extra feature which acts like IF all settings are flat THEN just adjust the volume. I tried setting all the bands to -1dB like I did in GoldWave but only the volume was reduced and the shape of the waveform was not altered. So I tried a very soft midrange scoop centered around 1kHz. Looking at the overall wave, my guess is that the shape is probably very close to what theory would predict. Still, if you zoom all the way in on the corners of the wave you can see high frequency ripples which are certainly some sort of transient or aliasing distortion.

Contribute!

It would be great to see how other software and hardware based EQ's treat this wave. I think the effects of an analog EQ would be very interesting. I'll attach a copy of the pulse wave file so others can give it a try and hopefully post their results here!

barefoot

67KB attachment

 

[barefoot]

Here is a 16 bit 48 kHz version of the 100 Hz square pulse .wav file.

EQ the signal with everything set at zero. If this has not effect, try the settings I used above. Then tell us what you did and show us your results!

If you use an analog EQ it would be nice to compare the EQ'd wave with the non EQ'd wave which has also passed through the analog chain.

Thanks in advance for your posts!

barefoot

1KB attachment

 

[CyanJaguar]

wow barefoot, that was quite an eye-opener. Thanks for taking the time to do it. I was quite amazed at the malformed waveform resulting from eq. I will try to use as little as possible processing as I can from now on.

 

[tubedude]

Those arent very high quality EQ's though... if anyone wants to try this test with the demos of the SonicTimeworks V1 EQ or the Sony Oxford EQ plugs, be sure to do so and we'll see if they are any more true than the cheapies at that level. Also the Waves stuff would be good to try out, its pretty high resolution, too, but not as high as the Sonic and Sony stuff.
www.sonictimeworks.com for the V1 and mastering EQ demos...

 

[Gidge]

This is why Ed reccomended keeping boosts/cuts on a digital EQ to +/-6db MAX!.........

 

[MISTERQCUE]

This is why Ed reccomended keeping boosts/cuts on a digital EQ to +/-6db MAX!.........

Exactly Gidge! Even monitor manufacturers and engineers insist
upon no more than +/- 6db (read any monitor spec manual) and
if needed, parametric eq's with 5 or less bands provide more precise control than 15 band eq's with bands ranging from 60 and
above hz.
But as most of you already know, eq,whether plug-in/outboard,
graphic,parametric basically should be used to sparingly for control of frequencies that need adjusting.
However, being a former EQ freak, I've learned that our ears can be fooled into thinking that a particular frequency is incorrect and needs adj'ing. One of the reasons is caused by either incorrect placement of one's nearfield's or positioning yourself away from the "sweet-spot" of the monitors range position & axis (which is approx 24").
In a discussion (and beer-drinking fest) with Apple Studio's,Red
Trockley, the biggest mistake in freq'y monitoring is laying monitors on their sides. This will produce the largest number of sound variations caused by the physical/time offsets of the bass & hi driver,which induces changes in mid-range sound characteristics as you physically move either from side to side
(grabbing a smoke on your left side, picking up beer on the other)
Around the cross-over point ( the distance on the monitors baffle
between the hi&low drivers) the speaker will emit numerous sound lobes thus making you think that you need EQ correction on certain frequencies.
To verify this, try it ! Lay your nearfields on thier sides and move left to right; You will notice freq diffs that appear to be incorrect and needs adj'ing,when in actuality proper placement leads to
a more accurate monitoring of the audio spectrum.
Now don't get me wrong, their are plenty of times when EQ Band-Aids are needed to correct insufficient or overwhelming sound characteristics in a particular band,and monitor placement/critical listening position is not the only cause,but it is indeed 1 of the most overlooked!
Just my penny's worth
Peace
Mr.Q

 

[ausrock]

MisterQ,

Quote;".........the biggest mistake in freq'y monitoring is laying monitors on their sides".

You are right on there. Tannoy have an article covering monitor placement on their website and that is one thing they explain and illustrate.

Peace........ChrisO

 

[Gidge]

Yep, Tannoy does preach about that....the instruction booklet of my ProtoJ's goes into great detail about positioning, and a whole page is dedicated to why you should not lay them down.....

 

[drstawl]

My Events are laying on their sides. The manual sez this is preferred in certain installations due to the wider field produced in "the sweet spot".

MrQ: maybe you need some sort of head-positioning device like one that might be used for laser neurosurgery. Sounds way too S&M to me. And why do you have to reach so far to grab a brewski? Every PC now comes with a cup-holder built in.

 

[tubedude]

Gidge! Proto-J's eh? Whatcha think?

 

[Gidge]

I think for the price (they can be had for as little as $150/pr)they are a great monitor....very nice stereo imaging and surprisingly flat.....somewhat similar the the Yorkville YSM1's.....excellent for cheapskates like me......

 

[MISTERQCUE]

I think for the price (they can be had for as little as $150/pr)they are a great monitor....very nice stereo imaging and surprisingly flat.....somewhat similar the the Yorkville YSM1's.....excellent for cheapskates like me......

You got that RIGHT, home-boy!!!!