• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Let's discuss room correction

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,583
Likes
1,478
Location
Oslo, Norway
I struggle to understand this. I've seen some discussions of this on other hifi forums, but they quickly become very heated, and attract people who have vested interests in the issue. And they quickly become so technical that I end up being more confused than I was.

Here on the forum there are some different approaches. Cosmik thinks that room correction is "fundamentally not valid". Amir uses Dirac Live only in the bass region, but not higher up. While quite a lot of other forum members are using stuff like Acourate to correct for the room response all the way up. As for me, I'm using Dirac in the bass region when I'm listening in the near-field, as I find that it introduces something artifical higher up. In the far-field however, I find that it's a net improvement to correct everything.

Alrighty, then. What are the theoretical and objective arguments for and against different types of room corection? Hope I'll end up being more enlightened as a result of this discussion :)
 

edd9000

Member
Joined
Mar 20, 2017
Messages
66
Likes
49
I use it all the way up. Although really I'm only interested in what it does in the bass.
 
OP
oivavoi

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,583
Likes
1,478
Location
Oslo, Norway
Thanks, Edd9000. What do you think are the improvements? Any trade-offs, or is it all good?
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
34,958
Likes
129,824
Location
Seattle Area
Let me make this simple: below 200-300 Hz, room correction is mandatory. It just is. Without it you have boomy bloated bass as the room over emphasizes some frequencies.

Above that it is optional.
 

Fitzcaraldo215

Major Contributor
Joined
Mar 4, 2016
Messages
1,438
Likes
604
I struggle to understand this. I've seen some discussions of this on other hifi forums, but they quickly become very heated, and attract people who have vested interests in the issue. And they quickly become so technical that I end up being more confused than I was.

Here on the forum there are some different approaches. Cosmik thinks that room correction is "fundamentally not valid". Amir uses Dirac Live only in the bass region, but not higher up. While quite a lot of other forum members are using stuff like Acourate to correct for the room response all the way up. As for me, I'm using Dirac in the bass region when I'm listening in the near-field, as I find that it introduces something artifical higher up. In the far-field however, I find that it's a net improvement to correct everything.

Alrighty, then. What are the theoretical and objective arguments for and against different types of room corection? Hope I'll end up being more enlightened as a result of this discussion :)
I am still dangerously uneducated about it, and I am sure I have much more reading to do. I think much acoustic theory supports the notion that near field and far field listening are fundamentally different above the transition frequency. Below the transition frequency, near and far field listening are more alike.

The difference above the transition frequency is due to the comparative ratio of direct vs. reflected sound. This near/far field issue can be compensated for in the frequency domain via the target curve. Flat is good for near field, but downward sloping is preferred for far field. Also, the larger the room, the more HF attenuation - greater downward slope - there should be.

I personally use Dirac always in a far field situation. I am quite happy with their normal default, downward sloping default target curve applied full range in my room.

I know it is probably nutty on my part in this sane, objectivist forum, but my ultimate evaluation is one of subjective preference, involving my recollected perception of live concerts vs. recordings. On that score, I am quite happy with my system today, though seldom was I previous to DSP EQ.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
34,958
Likes
129,824
Location
Seattle Area
I know it is probably nutty on my part in this sane, objectivist forum, but my ultimate evaluation is one of subjective preference
No, that is actually the correct way to go about it. Room correction must be verified by ear. There is no architectural correction.

This is why I like systems with parametric EQ as opposed to full convolution because I can turn bands on and off.
 

dallasjustice

Major Contributor
Joined
Feb 28, 2016
Messages
1,270
Likes
844
Location
Dallas, Texas
There's no question that targeted EQ (only minimum phase issues) in the bass region is valid and results in improved listening experience. For those who attempt to correct for non-minimum phase bass issues (e.g. Allison effect), the results will be poor.

Correcting above 300hz or so is more case-by-case. Some like it and some don't. Whether you like it or not will depend on the target curve of your speakers and their off axis performance. All speakers already use EQ to a target. The question is whether you like that target compared with the one you've selected in the software. That's really a personal decision without a correct answer.

Ive always preferred full range DSP EQ targets.
 

edd9000

Member
Joined
Mar 20, 2017
Messages
66
Likes
49
Let me make this simple: below 200-300 Hz, room correction is mandatory. It just is. Without it you have boomy bloated bass as the room over emphasizes some frequencies.

Above that it is optional.

This is pretty much my feeling. Where the room becomes modal, I use EQ, and have done in my setup for over 10 years, starting with a behringer feedback destroyer and a subwoofer. If I had the space I would use distributed bass and three subwoofers, but this would still need correction, it would just correct for a wider area than the listening spot.

I'm not keen on correcting above 300hz, other than say correcting the speaker response quasi-anechoic. My solution to this was to build speakers with a constant directivity horn and as even off axis as possible, the result when toed in 45 degrees is less room interaction so when I do correct above 300hz, its more just to match the target curve than anything else, and I don't feel it degrades the sound.

Here is my measured and corrected response:
drc-before-after.jpg
 

RayDunzl

Master Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,506
Likes
12,129
Location
Riverview FL
What are the theoretical and objective arguments for and against different types of room corection?

History:

Tone controls (bass/treble) - that's all any regular person had when I was a kid.

31 band Tascam eq used by ear with a rocking teen combo in a few clubs... No measurements like we can do today, as that was 1982 or so. They saw some duty at home, but weren't really part of any system...

For a long time went all purist - CD player, a couple of potentiometers, amp, speakers. I didn't have any measurement tools so it was 'good'.

Bought this house, and had time and money to play around. Bought a DAC and new (used) amps and preamp, and then, having an external DAC, could try out digital EQ - Behringer DEQ2496. Digit manipulation is noiseless - unlike the somewhat noisy EQ above. Distortion doesn't seem to be a problem, either.

Liked and disliked the Behringer, bought miniDSP OpenDRC-DI, again, in the the digital line, did manual FIR filters with rePhase, liked and didn't like that, bought AcourateDRC, and use that for filter generation now, and am still in the "like" phase. AcourateDRC is a lite version of Acourate tuned to the OpenDRC series DSP, minimal interface, very good automated single measurement position results.

Improved my measurement capabilities and skills during the last two paragraphs.

---

WTF is it, anyway?

I could argue it's not room correction, because it does nothing to the room. It's not speaker correction, because it does nothing to the speakers. It isn't signal correction, because there was nothing wrong with my signal. So, what should it really be called?

---

In my case single-point correction of (some) of the time and (some) of the frequency amplitude at a single spot in the room, by changing the signal presented to the speakers, rather severely, as compared to amplifier or other electronic device specs. I see +/- 12dB in the adjusted range, vs +/- 0.1dB or so for the raw electronic signal.

Example of the "corrected" signal (FR and and Phase) coming out of the preamp:
green/blue left/right - red uncorrected

Frequency response upload_2017-3-20_16-54-53.png Phase upload_2017-3-20_16-55-44.png

Any trade-offs, or is it all good?

I find the trade-offs to be minimal to nonexistent, audibly, to my feeble ears that have always been immune to HF content.

Let me make this simple: below 200-300 Hz, room correction is mandatory. It just is. Without it you have boomy bloated bass as the room over emphasizes some frequencies. Above that it is optional.

As for correcting the higher frequencies, I don't find it objectionable, and the only other person who listens here attentively doesn't either. I can turn off the FIR filters, which mainly affect the higher range, if someone does object, leaving the IIR filters on for room boom correction.

Example preamp out frequency response:
Green - left with automated IIR and FIR filters, Red IIR only with FIR bypassed upload_2017-3-20_17-30-44.png

You lose the phase correction when you do that with this setup upload_2017-3-20_17-32-21.png
 

Blumlein 88

Master Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
12,459
Likes
18,776
I use Tact room correction all the way up. A friend does Dirac all the way up. I haven't heard terrible downsides to using at higher frequencies. Now I do refine and customize the target slope. If there is some small negative to correction above 300 hz it seems smoothing out the overall response at the listening seat at least compensates for it. People I have done it for or who have heard mine say when it is turned on you take a big step closer to the music. Or it seems like some cloudy impediment has been pulled out of the system.

Before buying into it, my test abilities consisted of measuring one third octave warble tones off the Stereophile test disc with a RS sound level meter. This was about when Audacity became available. I started EQ'ing mainly bass problems, but also general slope of response with Audacity. It was obviously a benefit, so I was convinced doing more would be worthwhile.

Something like REW (free) and Umik or other $100 or less measuring microphones can be very educational about what you have going on in your room. Certainly a few PEQs for the low end can be the biggest help for the least effort.

As for the theoretical part, below the Schroeder frequency, resonances and reflections are sparse and you have some nulls of nearly no response in the room. No one argues against fixing that I don't think.
http://www.acousticfrontiers.com/2012314the-schroeder-transition-frequency-explained-html/ The linked two part article explains in more though still simple detail about the Schroeder frequency. Which is usually a lower frequency as room size increases.

Above this frequency reflections become dense and may represent as much or likely more energy than the direct sound is. There are ways of windowing some of that out. But then the reflections can effect your perception of frequency balance, soundstage and timbre even if the speaker output is perfect. So what is the theoretically correct in room response with reflections? There is also the real fact about ringing in FIR filters. The ringing people worry about with digital playback ring in a sense, but at too high a frequency to be heard. You start doing deep or steep FIR filtering well down in the audible band and such ringing may be genuinely audible. So I believe Diracs solution is to use mixed phase filters. In the midband I believe they use minimum phase filtering rather than FIR filtering. I may have that wrong if so someone can correct me. There is the issue that correcting for uneven response in linear phase systems results in proper phase, while correcting for non-linear systems may not do anything for phase and could cause problems.

At higher frequencies measuring becomes problematic. You move the measuring microphone an inch and get altogether different peaks and valleys though the overall trend is likely the same. Plus your ears are hearing at two points in space not one like the measuring microphone. Some work around this by measuring several locations and averaging in various ways. Also your ears aren't staying perfectly still. So over-correction can cause strange results.

Some measure with sweeps, some with impulses and some with a special pseudo random signal called a Maximum Length Sequence or MLS. Theoretically frequency response sweeps can be mathematically equivalent to impulse response as you can derive one from the other. If your measurements have phase and magnitude correct.

Some anecdotal practical experience. Tact uses an impulse response they repeat. The problem with impulses is poor signal to noise ratio. Many tried Tact and didn't like the results. I think they didn't realize how important noise was. You could pick how many impulses the Tact gear used to average and get results. If you only used 5 impulses during measurement, the results varied pretty widely. Some sounded very odd. They could be effected by a loud truck driving buy or airplanes crossing over or trains you barely noticed going down nearby tracks (nearby being a mile away). I didn't think results were good until you got up around 15 impulses to average. Much nicer results happened at 25 to 30. Further your results would be better if you turned off the fridge, turned of the HVAC, tried quiet times at night and things like that. I used to use one of 50 impulses done at 3 am on December 26 th (the best I can tell the quietest time of the year).

Now using a relatively slow sweep like REW does, it isn't immune to noise, but it also isn't greatly effected by it. I do think there is a trade off in how tightly you can window out room reflection however (and I may be wrong about how that is done in REW).

So you have theoretical concerns about what is the perfect response and how to measure what response you are getting which will obviously impact what correction works best.

Now before getting too deep into it, I have heard several different correction systems and all help some if you learn how to apply them. Most help more than just some. Most help in an obviously beneficial way to my ears. I have not heard Acourate though it is apparently very good. Tact worked well to me, Dirac is very good. Early Audyssey never worked worth doing in my opinion as the results seems weird and bright. Later versions do seem beneficial. Even some digital AV pre-amps with several PEQ bands you can select and use can be a worthwhile thing to do.
 

RayDunzl

Master Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,506
Likes
12,129
Location
Riverview FL
Theoretically frequency response sweeps can be mathematically equivalent to impulse response as you can derive one from the other.

One day I tried some square waves sweeps just to see what they would look like before and after correction, recorded the waveforms with Audacity.

Then I realized a low frequency square was like a repeated "step" - 10Hz square gives a 50ms long step.

Then I was looking at the step response display in REW which is calculated from a frequency swept sine wave.

And to my continuing surprise, the calculated step response of a 10-20kHz sweep was essentially identical to the recorded pseudo-step waveform recoding in Audacity from playing a 10Hz square wave alone.

At that point my appreciation for REW's analytical abilities jumped up notch or two.

Top: Room EQ Wizard calculated step response from sine sweep test tone
Bottom: Audio recording waveform of 10Hz square wave (pseudo-step signal) - first 12ms of what would be a 50ms "step"

index.php


"Evelyn, a dog, having undergone
Further modification
Pondered the significance of short-person behavior
In pedal-depressed panchromatic resonance
And other highly ambient domains . . .
Arf, she said." Frank Zappa - Evelyn, a Modified Dog
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
34,958
Likes
129,824
Location
Seattle Area
Sweep sines (as used in REW) have the advantage that they don't stress the equipment like real impulses do. And they provide huge gain in signal to noise ratio.

Great paper on this is here:
Transfer Function Measurement with Sweeps
SWEN MÜLLER, AES member,
Institut für Technische Akustik, RWTH, 52056 Aachen, Germany
AND
PAULO MASSARANI
Laboratório de Ensaios Acústicos (LAENA), Instituto Nacional de Metrologia, Calibração
e Qualidade Industrial (INMETRO), Av. Nossa Senhora das Graças, 50 – Xerém/Duque de
Caxias – RJ, Brazil

"7 CONCLUSIONS
FFT techniques using sweeps as excitation signal are the most advantageous choice for
almost every transfer function measurement situation.
They allow feeding the DUT with
high power at little more than 3 dB crest factor and are pretty tolerant against time variance
and distortion. Choosing an adequate sweep length allows complete rejection of the
harmonic distortion products. Moreover, these can be classified into single frequencydependent
harmonics, allowing a complete and ultra-fast distortion analysis over the whole
frequency range together with the evaluation of the transfer function. When it comes to
capturing room impulse responses for auralization purposes, there is no alternative to sweep
measurements: The high dynamic range in excess of 90 dB required for this purpose is
unattainable with MLS or noise measurements.
But even in standard RT measurements that
don’t require such a high dynamic range, sweeps are helpful as they allow easily increasing
the dynamic range up to 15 dB
compared to MLS-based measurements, using the same
amplifier, loudspeaker, and measurement length."
 

RayDunzl

Master Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,506
Likes
12,129
Location
Riverview FL
Sweep sines (as used in REW) have the advantage that they don't stress the equipment like real impulses do. And they provide huge gain in signal to noise ratio.

Last words of the conclusion in the paper:

"Measuring with sweeps is also more natural. After all, bats do not emit MLS to do their acoustic profiling."
 

Nightlord

Active Member
Joined
Mar 15, 2017
Messages
207
Likes
99
Location
southern Sweden
I used to run Canton Digital 1.1 speakers, so I've been down that route.

I never run Audyssey on the cinema receivers, but I do run MCACC in the tv-room which I feel works better.

I was trying to run a DSPeaker DualCore 2.0 in the cinema, but as it turned out not to be transparent even when in passthrough mode (=off), it had to go. With the treatments done to the room, there's fortunately no big fundamental modal issue to deal with... blocking off the floor reflex would be more needed.

If I sum where I am... I can very well approve of room correction in the low end, but if you like me care about more than one listening position - I'm also a bit hesitant on that as a general statement.
I do think room treatment has to have higher priority.

It's also a bit hard to verify the transparency of equipment like this... do they damage more than the change you're requesting? So if you are like me and have put a lot of effort in having a straight-wire-with-gain setup... then plopping in a machine like this is a bit uncomforting to say the least. I would definitely not want it to do it's final D/A itself, but leave that for an external, verified ok, DAC.
 
OP
oivavoi

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,583
Likes
1,478
Location
Oslo, Norway
Thanks guys, excellent input.

Btw, has anybody here tried the freeware "multi-sub optimizer"? I'm thinking of trying that out and see if can do as good a job as dirac in the bass region. Right now I have a small "how cheap can you go" project, to try and see if I can get better sound than most audiophiles while paying less than they do for their power conditioners. Therefore interested in freeware.
 

Nightlord

Active Member
Joined
Mar 15, 2017
Messages
207
Likes
99
Location
southern Sweden
Thanks guys, excellent input.

Btw, has anybody here tried the freeware "multi-sub optimizer"? I'm thinking of trying that out and see if can do as good a job as dirac in the bass region. Right now I have a small "how cheap can you go" project, to try and see if I can get better sound than most audiophiles while paying less than they do for their power conditioners. Therefore interested in freeware.

If you generalize it - the more points where you 'activate' the room from, the more calmed the ripples on the room ocean will become on average. One of the reasons I'm a strong advocate of using large number of subwoofers. I would not want any fewer than 4. (Have 18 in total myself over my three setups)
 
OP
oivavoi

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,583
Likes
1,478
Location
Oslo, Norway
If you generalize it - the more points where you 'activate' the room from, the more calmed the ripples on the room ocean will become on average. One of the reasons I'm a strong advocate of using large number of subwoofers. I would not want any fewer than 4. (Have 18 in total myself over my three setups)

Yap, thats my impression as well. Never really tried it out though. I'm considering to try out four of the 10-inch dipole subs that Linkwitz has designed to go along with his Lxmini in a multisub configuration. But I'm not sure whether the multi-sub approach works for dipoles? (but this is OT in this thread)
 
Top Bottom