• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Dirac and similar

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
I use REW to measure and my current efforts involved creating filters with REW which I manually added to the Parametric EQ (PEQ) of my Behringer DEQ2496. It works pretty well.

Obviously if your hardware or software can implement the REW filters automatically that's great but it's an easy task to do it manually.

The Behringer has ten PEQ filters so I quickly ran out of them and it only got me to about 200Hz. I keep meaning to adjust higher up but the result is so good that I haven't got round to it (I should add I have lots of room treatment including many bass traps)!

REW creates PEQ filters while DRC-FIR creates FIR filters. Convolvers that support FIR filters allow for much more complex filters but with a little skill and a decent room acoustics 10 PEQ filters will make things much much better.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Certainly a lot of people seem to like what Dirac does full range and of course I am not criticising them at all and it's certainly worth trying.

Do the pro-audio industry correct for the room? Full range? The manipulation of a recording is a different thing of course.

From the room EQ perspective audible bandwidth is divided into 3 segments:

1. 20Hz to Schroeder frequency (usually somewhere between 200 and 300Hz) range

In this range mosts of room-related bad things happen. Peaks and dips can be high/low more than 15dB. Use high Q filters if your measurement is precise but be carefull not to compensate for dips more than 12dB as your speakers/amp may suffer. Be carefull to configure convolver to adjust for maximum fllter gain to avoid clipping (if note done automatically).

2. Transition range 200-300Hz (Schroeder frequency) to 500-600Hz

Room doesn't affect this range as much as lower range but it is still present. Use mild correction here with modest Q factors (3-5)

3. HF

Room doesnt' affect sound here in the terms of peaks/dips. Use small Q (1-2) corrections here to adjust for your favorite listening curve
 

audimus

Senior Member
Joined
Jul 4, 2019
Messages
458
Likes
462
Good categorization. The only other parameter is whether you are doing (1) close listening as at a desk, (2) a typical listening room at a distance of 6-8 ft without much space behind equipment or listener and possibly corners and hallway openings and finally (3) a serious listening room with recommended spaces and positioning, etc.

DRC for (1) isn’t really doing as much room corrections as correcting quirks and limitations of the audio chain and tonal preferences much like headphones eq.

DRC for (2) is where the maximum benefit is. Any DRC will be better than no DRC. At the lower frequencies you correct for room effects, at higher frequencies more for tonal balance. My experiments with the commercial Audyssey and Anthem DRCs suggest that is what they roughly do as well when doing full range correction rather than try to correct hills and valleys at higher frequencies. Mostly broad shelf filters or such to lift or lower high frequencies (except for that annoying dip Audyssey uses). These help if you have very position sensitive speakers but not enough room to get some optimal position. DRC is much easier to do here than room treatment.

DRC for (3) tends to be less of an impact as there is enough space to get the best out of positioning and room treatment and equipment tends to be much better in such setups as well. DRC can optimize some small idiosyncracies or substitute for problematic room treatment.

REW is great but way too many bells and whistles for any but the most ardent hands-on and technically adept that is willing to understand the concepts to pick what they need out of the zillion things you can do with it. Not a system to use following some blog somewhere. Too little understanding is a dangerous thing here since there are so many ways of doing things wrong with REW and never know it. Especially, if you are involving more than two speakers.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Good categorization. The only other parameter is whether you are doing (1) close listening as at a desk, (2) a typical listening room at a distance of 6-8 ft without much space behind equipment or listener and possibly corners and hallway openings and finally (3) a serious listening room with recommended spaces and positioning, etc.

DRC for (1) isn’t really doing as much room corrections as correcting quirks and limitations of the audio chain and tonal preferences much like headphones eq.

DRC for (2) is where the maximum benefit is. Any DRC will be better than no DRC. At the lower frequencies you correct for room effects, at higher frequencies more for tonal balance. My experiments with the commercial Audyssey and Anthem DRCs suggest that is what they roughly do as well when doing full range correction rather than try to correct hills and valleys at higher frequencies. Mostly broad shelf filters or such to lift or lower high frequencies (except for that annoying dip Audyssey uses). These help if you have very position sensitive speakers but not enough room to get some optimal position. DRC is much easier to do here than room treatment.

DRC for (3) tends to be less of an impact as there is enough space to get the best out of positioning and room treatment and equipment tends to be much better in such setups as well. DRC can optimize some small idiosyncracies or substitute for problematic room treatment.

REW is great but way too many bells and whistles for any but the most ardent hands-on and technically adept that is willing to understand the concepts to pick what they need out of the zillion things you can do with it. Not a system to use following some blog somewhere. Too little understanding is a dangerous thing here since there are so many ways of doing things wrong with REW and never know it. Especially, if you are involving more than two speakers.

I don't agree. Correcting the 1st range (below transition frequency) is by far the most important role of room EQ.

At the lower frequencies you correct for room effects, at higher frequencies more for tonal balance.

This is true.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Nothing I have said contradicts that.

This what you said is wrong.

"DRC for (1) isn’t really doing as much room corrections as correcting quirks and limitations of the audio chain and tonal preferences much like headphones eq. "
 

audimus

Senior Member
Joined
Jul 4, 2019
Messages
458
Likes
462
This what you said is wrong.

"DRC for (1) isn’t really doing as much room corrections as correcting quirks and limitations of the audio chain and tonal preferences much like headphones eq. "

It just says the things it is correcting (in case 1 of close listening) for are more likely to come from the audio chain than the room (ymmv might vary since every desk is in a different location). That says nothing about whether correcting LF range is more important or not. Think of a typical non-audiophile grade desktop or PC speakers with awful bass reproduction with peaks and/dips and very often exaggerated LF that a “DRC” algorithm can correct for as easily as room induced ones. Correcting for tonal balances to satisfy personal preferences would contribute greatly in this case as well.

Think of the headphones as a logical extreme with very little if any room induced dips and peaks (yes, there is some “room” effects within the confines of the headphones itself but that is much better controlled by manufacturer in a good pair of headphones). But yet, the same algorithms which don’t know if the dips, peaks and roll-offs if any are caused by room issues or the audio chain will still work to create improvements.

There should be a term better than “room correction” for these eq algorithms. Room correction is one application of them and possibly the most commercially adopted. Causes a lot of confusion. May be Digital Response Correction. :)
 
Last edited:

digitalfrost

Major Contributor
Joined
Jul 22, 2018
Messages
1,521
Likes
3,086
Location
Palatinate, Germany
"DRC for (1) isn’t really doing as much room corrections as correcting quirks and limitations of the audio chain and tonal preferences much like headphones eq. "
Yeah okay, I can get with that. The thing is, most speakers are too bright. And I think especially at this distance, this plays out even more.

e: Btw, I even using convolving to adjust my headphone EQ to my preffered room curve using https://github.com/jaakkopasanen/AutoEq . This makes my headphones and my speakers sound pretty identical. The only thing I'm missing is a solution for my car :)
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
It just says the things it is correcting (in case 1 of close listening) for are more likely to come from the audio chain than the room (ymmv might vary since every desk is in a different location).

In the 20Hz- app 300Hz range loudspeaker response is contaminated dominantly with room modes compared to their anechoic response.

Think of a typical non-audiophile grade desktop or PC speakers with awful bass reproduction with peaks and/dips and very often exaggerated LF that a “DRC” algorithm can correct for as easily as room induced ones.

Applying room EQ with such speakers is kind of pointless as DRC cannot correct such flaws in speakers design. Investing in a decent speakers is the only solution to that issue.
 

audimus

Senior Member
Joined
Jul 4, 2019
Messages
458
Likes
462
In the 20Hz- app 300Hz range loudspeaker response is contaminated dominantly with room modes compared to their anechoic response.
True but that does not logically imply contamination also does not happen because of limitations of typical speakers (which have most likely never seen the insides of an anechoic chamber) in real life.

Applying room EQ with such speakers is kind of pointless as DRC cannot correct such flaws in speakers design. Investing in a decent speakers is the only solution to that issue.

In this I wholly disagree. Not to imply the logical extremes of any speaker flaw can be corrected. But anecdotal evidence does exist of people liking DRC in such situations where room is not necessarily the factor. There is one just above in this thread.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
True but that does not logically imply contamination also does not happen because of limitations of typical speakers (which have most likely never seen the insides of an anechoic chamber) in real life.



In this I wholly disagree. Not to imply the logical extremes of any speaker flaw can be corrected. But anecdotal evidence does exist of people liking DRC in such situations where room is not necessarily the factor. There is one just above in this thread.

Well, the truth is DRC was designed primarilly to correct for room infuence, not to correct for serious speaker flaws. Whether you agree or not will not change that fact.
 

audimus

Senior Member
Joined
Jul 4, 2019
Messages
458
Likes
462
Since this is a SCIENCE forum

I think there should be a name other than “room correction” for the collective systems that is based on the following paradigm.

1. There is a measurement of sound in a transducer system.
2. There is a target output response expected/desired.
3. A deviation from the target response is corrected by modifying the inputs sufficiently for output to match the target as much as possible.

It encompasses a whole range of products including active equalizers within speakers regardless of where the deviation from the target come from. Or to limit what it is used for.

Commercial room correction fit the above and it was a continuation of those algorithms but combined with additional factors induced by a “room” including multiple listening areas. But the commercial systems also include tone balance correction (which don’t always have to do with the room, could be brightness or lack thereof in speakers) to meet a target curve because people prefer different balances in hearing. That translates to commercial success.

Any application and situation such a system is capable of helping meet the target response is fine by me. Not to get into debates of whether trains were first designed to carry goods or to carry people and so should only be used for one or the other because of it.
 

jonfitch

Senior Member
Joined
Sep 2, 2019
Messages
479
Likes
532
In the 20Hz- app 300Hz range loudspeaker response is contaminated dominantly with room modes compared to their anechoic response.



Applying room EQ with such speakers is kind of pointless as DRC cannot correct such flaws in speakers design. Investing in a decent speakers is the only solution to that issue.

What about electrostatic or planar speakers? Those tend to have awful frequency response. Are you saying people should just give up on those and not even bother to EQ them?
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
What about electrostatic or planar speakers? Those tend to have awful frequency response. Are you saying people should just give up on those and not even bother to EQ them?

Main problem with such speakers is not non-linear on-axis response but bad directivity index which results in a narrow sweet spot and that cannot be corrected with DSP.

I think you guys should be aware that although modern DSP technology seems like it can do miracle with sound in terms of EQ the truth is not quite as bright: speaker has to have decent anechoic response, not only in terms of on-axis response but in terms of parameters measured and shown in spinorama graphs. If the speaker is showing less than stellar spinorama response there is not much you can do with room EQ to make things better.
 
Last edited:
OP
W

Willem

Major Contributor
Joined
Jan 8, 2019
Messages
3,658
Likes
5,276
I am not sure about other electrostats, but last weekend I measured the in room response of my Quad 2805s and the waterfall graph was excellent with only minor peaks and dips above 100 Hz and a pretty quick and smooth decay. Yes my old ELS 57s had a narrow listening position but these modern ones not nearly so much. My current issues are restricted to the sub 80 Hz region where I also have a subwoofer and resonating windows and doors. I am currently trying to isolate those various culprits.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
I am not sure about other electrostats, but last weekend I measured the in room response of my Quad 2805s and the waterfall graph was excellent with only minor peaks and dips above 100 Hz and a pretty quick and smooth decay. Yes my old ELS 57s had a narrow listening position but these modern ones not nearly so much.

Waterfall graph is showing acoustic characteristic of your room, not your speakers.
 
OP
W

Willem

Major Contributor
Joined
Jan 8, 2019
Messages
3,658
Likes
5,276
I know but the response was also beautifully flat, not to mention their low distortion. I really do not understand the denigration of these wonderful speakers. It is not for nothing that they are used by some of the world's leading recording engineers of classical music. As dipoles they drive the room differently, but that includes a lower suspectability to room modes.
If you prefer more traditional speakers that is fine with me. I have Harbeth P3ESR in my study and they are very nice as well, but cannot quite replicate the disappearing act of these modern Quads.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
I know but the response was also beautifully flat, not to mention their low distortion. I really do not understand the denigration of these wonderful speakers. It is not for nothing that they are used by some of the world's leading recording engineers of classical music. As dipoles they drive the room differently, but that includes a lower suspectability to room modes.
If you prefer more traditional speakers that is fine with me. I have Harbeth P3ESR in my study and they are very nice as well, but cannot quite replicate the disappearing act of these modern Quads.

Oh, I actually like the sound of modern MLs - as long as you are sitting at the right spot they are presenting magical sound that only the best traditional speakers can match. :)
 
Top Bottom