• 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.

Dirac and similar

tuga

Major Contributor
Forum Donor
Joined
Feb 5, 2020
Messages
1,853
Likes
1,368
Location
Oxford, England
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
I'm using REW. How much smoothing should one use on the curve when applying room correction filters below Schroeder?

What do you mean by "Use high Q filters if your measurement is precise"?

Is a single measurement from the listening spot enough or should I average 2 or 3 curves at slightly different mic positions?
Or would the moving mic technique produce better results?
 

Pio2001

Active Member
Joined
May 15, 2018
Messages
199
Likes
261
Location
Neuville-sur-Saône, France
Hi tuga,

Is a single measurement from the listening spot enough or should I average 2 or 3 curves at slightly different mic positions?
Or would the moving mic technique produce better results?
Use both.
The moving mic provides a better measurement of the average level. It also naturally provides the best smoothing of the curve in high frequencies. The multipoint technique is less consistent, unless you make more than 10 takes.
The multipoint measurement, on the other hand, gives you the variability of the peaks and dips of your frequency response as you are moving along your listening area.
I take my moving mic measurements with 1/48 octave RTA, 65536 FFT, and display them without any smoothing.
Here is how I display my multipoint measurements :
24-HarmonicOverlay.png



I'm using REW. How much smoothing should one use on the curve when applying room correction filters below Schroeder?
In this example, the corrections are based on the above measurement.
But after some listening sessions and extra measurements, the positive corrections at 140 and 180 Hz (filters 7 and 8, the curve is displayed upside down) are excessive. Too sharp and too strong.
All the corrections below 110 Hz sound excellent.
The corrections above 200 Hz are barely audible, except with specific songs, where they are quite good.
And last, the correction at 200 Hz is better than nothing, but imperfect. We can see on the above multipoint measurement that the curve is unstable in this area.


32_201706_filtres.png
 

tuga

Major Contributor
Forum Donor
Joined
Feb 5, 2020
Messages
1,853
Likes
1,368
Location
Oxford, England
Hi tuga,



Use both.
The moving mic provides a better measurement of the average level. It also naturally provides the best smoothing of the curve in high frequencies. The multipoint technique is less consistent, unless you make more than 10 takes.
The multipoint measurement, on the other hand, gives you the variability of the peaks and dips of your frequency response as you are moving along your listening area.
I take my moving mic measurements with 1/48 octave RTA, 65536 FFT, and display them without any smoothing.
Here is how I display my multipoint measurements : View attachment 56034




In this example, the corrections are based on the above measurement.
But after some listening sessions and extra measurements, the positive corrections at 140 and 180 Hz (filters 7 and 8, the curve is displayed upside down) are excessive. Too sharp and too strong.
All the corrections below 110 Hz sound excellent.
The corrections above 200 Hz are barely audible, except with specific songs, where they are quite good.
And last, the correction at 200 Hz is better than nothing, but imperfect. We can see on the above multipoint measurement that the curve is unstable in this area.


View attachment 56036
Thanks for the comprehensive reply.
 

Bombadil

Active Member
Forum Donor
Joined
Jan 28, 2017
Messages
144
Likes
120
Location
Davis CA
We certainly need some objective comparison testing of the systems (DIRAC, Audessey etc). I'm very accustomed to using ARC and I think we could all agree it modifies the sound we hear. In my case generally positive, noticed primarily in the bass. ARC Genesis measures by default up to 5000 hz.
apologize if it has already been noted by others but Audioholics has produced a nice set of discussion videos on the topic, released just this past week. They really get into the weeds and go over pluses and minuses, worth a listen IMO.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
2,995
Likes
2,373
[QUOTE = "Igor Kirkwood, post: 358178, member: 12826"] Obtaining good results for speaker measurements in an anechoic room or with a Klippel system is satisfactory for the manufacturer and for the customer.
But to optimize his system, the auditor needs real measures in his
room.

Some software such as Dirac or hardware have been discussed here.
I would like to present another solution used in my system.

As a sound engineer for classical music, I generally work with a minimum recording configuration:
2 Neumann TLM50 microphones, no compression, no DSP,…

So my listening configuration is used to check my recordings:

A 130 cubic meter room, the walls are treated with almost 5 cubic meters of glass wool
The listening distance is 3.1 m, the distance between the front speakers is 2.9 m The stereo front channels are Yamaha NS1000X in an active 3-way configuration, the original tweeter has been replaced by a 27 mm Focal Be, foam on the front wall to minimize diffraction
-2 surround with passive Yamaha NS1000X
-4 SVS PC-2000 subwoofers in the middle of the walls
-QSC Q-Sys 110F performs FIR filtering for subwoofers (FIR filter at 90Hz, 70dB / oct), crossovers for NS1000X (375 and 1800Hz, 70dB / oct), EQ FIR and IIR correction, delays, .. .
This processor also takes care of the choice of source, stereo / multichannel, phantom center, LFE and bass management,…
The management of the subwoofer is also special: you can choose between mono, stereo, M / S, Griesinger (improved LR),… and possibility of modifying L + R
(in phase) and LR (out of phase) bass levels
For listening tests, all configurations can be switched and compared instantly without "clicks" on a computer with a user-friendly interface

The configuration of the system was carried out by Jean-Luc Ohl (who introduced MMM on www.ohl.to )

About the correction (that's the subject): it is based on MMM for the amplitude (therefore min phase only) and several averaged scans for the excess phase;
DRC-FIR, thanks to Denis Sbragion software, and loaded into Q-Sys with 8192 FIR taps plus IIR
(all programming and measurement analyzes are carried out remotely).

Now let's share some measures:

1 / no equalization on the NS1000X before, just crossings and delays


2 / NS1000X front with equalizer


3 / NS1000X front with 4 subwoofers:

This configuration is optimized for
improve LR (stereo signals on low frequencies, see Griesinger) [/ QUOTE]




This seems to be done by the book, which is to be expected from the professional insider. I also very much like the idea of checking your recordings on the home system like this.
 
Last edited:

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
2,995
Likes
2,373
No man, really, dude deserves respect, if only for getting out of his wife's or decorator's way!
You're making it sound like he really had a choice but to move out of her way! :D

No respect for that as it falls into "self preservation" cathegory - every clever animal can do that. :D
 
Joined
Jan 28, 2020
Messages
67
Likes
95
Oh c'mon, don't give him credit for the room asethetics as well - it was probably his wife that decorated it! :D
My wife help me...:)
"Paradise" is high celling 5,3 meter maximum, 5 cuber meter of wool... and the work of Jean-Luc Ohl, managing the QSC and 4 subs :)

View with 3 subwoofer and one surround passive Yamaha NS-1000X .

20200201_111735.jpg
 
Last edited:
Joined
Apr 4, 2020
Messages
4
Likes
1
Location
Algarve - Portugal
That´s really beautifull and interesting setup. bet it sounds amazing!!! Cheers to you for not following the norm. You´re living room seems very similar to the house where I went to get my main stereo speakers .

The main Yamaha´s are tri powered or using internal crossover ? And also I wouldn´t be brave enought to change the tweeter in those Yamaha´s , but I guess you changed it to one of the better available brands.
Were you afraid of the radioactive berylium ? ;)))

regards and would love to know more about your system! Do you have another topic ?
 
Joined
Jan 28, 2020
Messages
67
Likes
95
That´s really beautifull and interesting setup. bet it sounds amazing!!! Cheers to you for not following the norm. You´re living room seems very similar to the house where I went to get my main stereo speakers .

The main Yamaha´s are tri powered or using internal crossover ? And also I wouldn´t be brave enought to change the tweeter in those Yamaha´s , but I guess you changed it to one of the better available brands.
Were you afraid of the radioactive berylium ? ;)))

regards and would love to know more about your system! Do you have another topic ?
Thank you algarvehifi:)

Yes a topic on the French HCFR
https://www.homecinema-fr.com/forum...-q-sys-core110f-4-subs-ns1000x-t30094413.html

Perhaps Jean-Luc Ohl can explain the FIR active filter with QSC Qsys Core 110f

For the very good Beryllium tweeter FOCAL look the plan to put in on a Yamaha NS- 1000X loudspeaker (or with change to any loudspeaker)

You can give this plan to a specialist.....easy !


20200405_100445 2.jpg
 
Last edited:
Joined
Sep 2, 2019
Messages
81
Likes
78
I don't have a time-stamp because I listened to this while I was on the subway, but what I found interesting was Matthew Poes worked on implementing Dirac Live with NAD and claims that people most of the time couldn't actually hear a difference between full-range room correction and correction up to 200hz only with Dirac, in other words EQing the bass made so much difference in the sound that small differences in correction above 200hz were way too subtle for most people to spot (when using speakers that had a near perfect spinorama) . He also said you should gate the top of the correction to around 10KHz because most mics are completely unreliable above that frequency.

 
Last edited:

rxp

Member
Joined
Jun 19, 2020
Messages
42
Likes
30
I saw that video - Poes is really an exceptional communicator. I love how he summaries the science and the way he tests/demonstrates information.

There is another recent video that had the Audyssey engineer (
) that confirms that the AVR's have a built in correction curve to the standard mics that get them to 2db (I think) to 24khz.

For what it's worth I know one of my Audyssey mics measures much hotter than the other - but I've seen measurements on AVS that show the amplitude shape was near identical - only level differences.

Still - I've tried full range and partial - I cannot hear the difference. Like if you asked me to tell which one it was 20 mins after I'd have no clue. Which also makes me skeptical about how terrible high frequency correction is. How can it be if I can't tell the difference? However, I could damn well tell you if the bass was to my liking. That doesn't mean flat, but just the level/slope.

The only area I can see full range correction being exceptionally valuable is in stereo imaging with one sweet spot. I believe that surround users prefer "center spread" because they're too used to blurred center images due to weakness of 2 speakers creating a virtual image (and the dip your HRTF creates). With full range correction you can really improve L and R producing similar responses.
 
Joined
Sep 20, 2019
Messages
75
Likes
18
What do you guys think of this guy's measuring method?


He puts the mic next to his ears while measuring. Also, for the left speaker measurement he measures 3x from the left ear and 1x from the right ear and then averages it. Same process for the other channel.
 

Lbstyling

Active Member
Forum Donor
Joined
Jan 2, 2019
Messages
231
Likes
154
What do you guys think of this guy's measuring method?


He puts the mic next to his ears while measuring. Also, for the left speaker measurement he measures 3x from the left ear and 1x from the right ear and then averages it. Same process for the other channel.
:D
Laughable.
:facepalm:
This guy kills me.
 

Similar threads

Top Bottom