Subtractive vs additive EQ, pros and cons?

MF_Kitten

Registered
Thread Starter
Joined
Jun 24, 2018
Messages
5
I've been EQing studio monitors and subwoofers and headphones with REW measurements and filters for a while now, and I very early on devided I didn't like doing additive bell curve corrections to correct for dips. I haven't really checked it out in a long time, so I can't say I have a good understanding of it, but the idea is that boosting an EQ band often gives you a ringing sound that messes up everything. With speakers in a room, the acoustics will tamper your ability to correctly measure the frequency response, so while the graphs look nice, what you hear is way off. You might also be getting a phase cancellation at the microphone due to reflections.

Has anyone figured out a good way to correct for these issues when doing EQ to a system?
 

Wayne A. Pflughaupt

Moderator
Staff member
Joined
May 21, 2017
Messages
633
Location
Corpus Christi, TX
More  
Preamp, Processor or Receiver
Yamaha CX-A5000 A/V Preamp / Processor
Main Amp
Yamaha RX-Z9 AV Receiver (as multichannel amp)
Universal / Blu-ray / CD Player
Denon DCT-3313 UDCI Universal Disc Player
Front Speakers
Canton Karat 920
Center Channel Speaker
Canton Karat 920
Front Wide Speakers
Realistic Minimus 7 (front EFX speakers)
Surround Speakers
Canton Plus D
Surround Back Speakers
Yamaha YDP2006 Digital Parametric EQ (front mains)
Front Height Speakers
Yamaha YDP2006 Digital Parametric EQ (surrounds)
Rear Height Speakers
Yamaha YDP2006 Digital Parametric EQ (sub)
Subwoofers
Hsu ULS-15 MKII
Other Speakers or Equipment
Adcom ACE-515 (for power management)
Video Display Device
Yamaha DT-2 (digital clock display)
Screen
Pioneer PDP-6010FD 60" Plasma TV
Remote Control
Stock Yamaha Remote
Streaming Equipment
Roku Express
Other Equipment
Audio Control R130 Real Time Analyzer
I assume you’re talking about equalizing the main speakers. The trick to EQing mains is knowing what to address and what to leave alone.

If you had a bad experience equalizing, it may that you were making the mistake of peppering the speakers with a whole slew of filters, chasing the magical “flat graph.”

If you had an audible ringing from a boosted filter, chances are it was too narrow. You shouldn’t use filters tighter than about 1/3-octave. The idea is to address the broad issues, as those are the most audible, not tiny or narrow deviations of only few dB.

Also, you probably don’t want to boost a filter more than 4-5 dB tops, even if the graph “demands” more.

With speakers in a room, the acoustics will tamper your ability to correctly measure the frequency response, so while the graphs look nice, what you hear is way off. You might also be getting a phase cancellation at the microphone due to reflections.
If you can localize the speaker location with your eyes closed, then the acoustics aren’t all that bad, at least above the Schroeder frequency. Below that point I really don’t see the acoustics hindering the ability to measure – after all, the mic is only showing what you’re hearing anyway, acoustics and all. It doesn’t hurt to try to correct a direct + room signal that’s below the Schroeder frequency. The ultimate deciding factor, either above or below Schroeder, is whether or not the filter delivers an audible improvement. If the graph says a filter is needed, but it doesn’t get an improvement in sound - ditch it.

Regards,
Wayne
 

MF_Kitten

Registered
Thread Starter
Joined
Jun 24, 2018
Messages
5
Thanks!

I've seen some headphone measurements where everyone gets a big deep dip aomewhere, and then the CSD graphs show a peak that just gets phased out somehow. Very odd. That's the most obvious place I've seen this phenomenon.
 

Matthew J Poes

AV Addict
Joined
Oct 18, 2017
Messages
1,903
Wayne gives a lot of good advice.

If you want to eq above Schroeder I have a few suggestions.

A) take the speaker outside and measure its full polar response. Plug all of the data into a good eq modeling tool that can model a speakers power response, polar, DI, etc. I use VituixCad.

Ok a lot to take on so:

B) measure the speaker with the mic pointing at the speaker from several locations around the listening position. Look for anomalies that show up in every position and are significant. They should be high in amplitude and somewhat low in Q (meaning fairly wide). Then create eq to match this and correct. You can also apply shelf filters to change a speakers tonal balance from bright to warm.

What you shouldn’t do is attempt to correct a speaker flat using in room measurements. It’s very likely that your correction will make it better in one place and worse in another. It’s not just about the measurements you’ve taken. You could be causing more severe response issues off axis which causes the reflected sound to be very different from the direct sound.

Below Schroeder it is ok to eq a bit more liberally but you should still be smart about it. Like Wayne says, avoid adding a lot of boost. 3dB increase requires twice as much power. You could cause clipping. I’m fairly experienced with all this and even I have an occasional lapse in judgement. While testing a product under review I accidentally activated multiple eq boosts that were all sent to the sub. I clipped he output signal, input stage on the amp, and amp. It obviously sounded very bad. You also don’t want to use bass boost to extend the bass of your system below its natural roll-off unless you have very good data on the dynamic capability of your speakers and subs. You risk clipping your amp and damaging the woofers. I see people do this all the time in a desire to extend the response as low as possible.

Finally, you may want to consider using measurements to test success. Namely, take an area in the response that needs to be eqed. Say around 100hz. Apply a filter to the impulse in the filtered impulse tab at around 100hz. Start with 1 octave bad then look at 1/3 octave. Do it for a measurement pre and post eq. The impulse with eq should look cleaner with a smooth natural decay and less ringing. If it doesn’t the eq you applied has made things worse. You can see this in my article on eq and ringing.
 
Top Bottom