• 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 Live's impact on soundstage & imaging

Joined
Feb 25, 2019
Messages
46
Likes
26
Location
Pennsylvania
#1
Hey everyone,

Looking for some input as to what I might be encountering with my recent implementation of Dirac Live 2.0 via miniDSP DDRC-24. Setup is as follows:

MacMini (macOS Catalina) running Roon->USB->miniDSP DDRC-24->RCA->KEF LS50W (w/Sub Out to Elac SUB3010)

My room is small at 10'W x 12'L x 9'H and slightly asymmetrical due to a doorway in the rear right and a shelving unit along the left side. The speakers are placed per KEF's guidelines to that they're 1/5 out into the room from the front and side walls and on 24" stands with about 20% toe-in. My listening position is 38% out from the rear wall. As you can imagine, this results in a near field setup where the speakers are 5' apart and also 5' from where my noggin lands.

The room has treatment in the form of soffit traps that run around the perimeter and floor to ceiling corner traps in the front corners.

After reading some positive reviews of the results that could be had with Dirac Live 2.0 I decided to try out the miniDSP DDRC-24, looking for the combo to be the "icing on the cake" so to speak.

Neither Dirac Live 2.0 nor the latest miniDSP software officially support macOS Catalina but I was able to get them to work and measured using the Chair position with success. I exported 4 different filter options from Dirac Live to the miniDSP:

  1. Dirac's default target curve.
  2. Slightly more tilted version of Dirac's target curve.
  3. A version of #2 that was the same below 250Hz but flat up to 10kHz.
  4. A version of #2 that was the same below 250Hz and only corrected below that frequency.
The Dirac target curve was as follows:

Freq / Correction
38 / +2
74 / +1.5
143 / +1.1
278 / +0.6
540 / +0.1
1050 / -0.4
2040 / -0.8
3970 / -1.3
7720 / -1.8
15000 / -2.3

The corrections to the bass are welcome, it's more even over those octaves now, but the impact to the soundstage and imaging is where I'm struggling. While it's shifted centered things to the center as opposed to slightly to the left, it's collapsed the soundstage to one that's very flat with no space around instruments and voices and there's a lack of detail like a blanket's been thrown over the speakers.

I'm not sure what the cause is but it's across all four filters.

Bill
 
Last edited:
Joined
Jan 17, 2020
Messages
95
Likes
112
#2
Are you using the latest Dirac version (2.5.3)? There have been some changes in the last couple of builds that have quite radically changed the sound of my setup (to the better IMO). You should probably remeasure after upgrading the Dirac version as well, although that should not be necessary, in practice it is. You should also experiment with different microphone arrangements - 'chair', 'sofa', wide etc. with 9, 12 or more positions. Although Dirac say the exact position of the microphones don't really matter they do, in my experience, make a subtle difference to the resulting calibration.

That said, in my experience, all the calibrations I have done over the last few months have resulted in better soundstaging and detail without exception.
 
OP
wgb113
Joined
Feb 25, 2019
Messages
46
Likes
26
Location
Pennsylvania
Thread Starter #3
@bluefuzz it is indeed v 2.5.3 that I'm using.

I read elsewhere that someone had better results using the Sofa settings for a single position but it seems counterintuitive to me to use measurements over an even larger area in such a small room and near field arrangement. The Chair measurement already took into account 9 positions.

I'm replacing my equipment rack and getting a secondary set of speakers out of the room this week so I'm going to have to remeasure again anyhow.
 
Joined
Jan 17, 2020
Messages
95
Likes
112
#4
it seems counterintuitive to me to use measurements over an even larger area in such a small room and near field arrangement. The Chair measurement already took into account 9 positions.
In my (limited) experience the bigger the area and the more points used the less extreme the resulting calibration. Conversely using too small an area results in a quite dead and flat calibration – almost headphones like. That would seem to be what you describe. Dirac recommend at least one metre between top back right and front bottom left even for the 'chair' position.
 
OP
wgb113
Joined
Feb 25, 2019
Messages
46
Likes
26
Location
Pennsylvania
Thread Starter #5
In my (limited) experience the bigger the area and the more points used the less extreme the resulting calibration. Conversely using too small an area results in a quite dead and flat calibration – almost headphones like. That would seem to be what you describe. Dirac recommend at least one metre between top back right and front bottom left even for the 'chair' position.
I guess that's where part of my confusion lies - in terms of amount of correction it doesn't seem to be doing much with it's default house curve, only +/- 3dB. It's also why I was surprised I was told to lower the Master Volume on the miniDSP to -10dB when it resulted in distortion at Max (I want to control volume only via the KEF).
 
Joined
Jan 17, 2020
Messages
95
Likes
112
#6
in terms of amount of correction it doesn't seem to be doing much with it's default house curve, only +/- 3dB
The default curve is just that: a default. Your own correction should broadly follow your speakers' natural slope for best results. Evening out any unwanted lumps and bumps if present.
I was surprised I was told to lower the Master Volume on the miniDSP to -10dB when it resulted in distortion at Max
If you have, say, a deep dip at 50 Hz Dirac's filter needs to cut all the other frequencies to match that dip - which in turn means a lot of make up gain after the filters. And that can lead to distortion if there isn't enough headroom.
 
OP
wgb113
Joined
Feb 25, 2019
Messages
46
Likes
26
Location
Pennsylvania
Thread Starter #7
Here's a shot of the measured response, default target curve and Dirac's estimate of the resulting curve post-adjustments. As you can see there's about a 20dB swing between 45-120Hz and my understanding is that 10dB of boost is the maximum amount that Dirac will attempt in it's correction.

Image 5-13-20 at 8.59 AM.jpeg


I wish there was an easy/reliable way to see the actual results post-adjustment.

Still uncertain as to why I'm experiencing the differences in soundstage and imaging with the filters activated.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
2,926
Likes
2,290
#8
Still uncertain as to why I'm experiencing the differences in soundstage and imaging with the filters activated.
Have you tried limiting Dirac to correct only 20-500Hz region? You should also try to correct only 20-200Hz as in that case it shouldn't affect soundstage/imaging at all.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
2,926
Likes
2,290
#10
Btw, you shouldn't worry about that dip at 120Hz. it is relatively narrow and Dirac would do what it can with it so I doubt it will be audible, except in rare cases of bass guitar or drum solo when it hits that note.
 

thewas_

Addicted to Fun and Learning
Joined
Jan 15, 2020
Messages
935
Likes
1,815
#11
Don't try to correct midrange and the treble.
The LS50W he uses have quite a good tuning and directivity which can be seen at his "before" curve which is almost identical to his target curve and thus in this case the difference of limiting the correction to the bass or not wouldn't be significant.
 

Dimifoot

Active Member
Forum Donor
Joined
Mar 27, 2019
Messages
262
Likes
329
Location
Greece
#12
While it's shifted centered things to the center as opposed to slightly to the left, it's collapsed the soundstage to one that's very flat with no space around instruments and voices
What you are describing as “soundstage & space around instruments” might be slight phase mismatch because of your assymetric room.
Dirac will correct that, leaving you with this impression of collapsed soundstage.

Also, have you considered getting a digital- only minidsp version, in order to avoid double analog/digital conversions?
From what I know your speakers have distal input, and they convert analog incoming signals to digital.
(Correct me if I am wrong)
 
OP
wgb113
Joined
Feb 25, 2019
Messages
46
Likes
26
Location
Pennsylvania
Thread Starter #13
What you are describing as “soundstage & space around instruments” might be slight phase mismatch because of your assymetric room.
Dirac will correct that, leaving you with this impression of collapsed soundstage.

Also, have you considered getting a digital- only minidsp version, in order to avoid double analog/digital conversions?
From what I know your speakers have distal input, and they convert analog incoming signals to digital.
(Correct me if I am wrong)
I was beginning to wonder the same thing with regards to phase and the soundstage. Some others have said what @bluefuzz mentioned, that they’ve gotten better results with the Sofa measurements and to spread them out as much as possible so that Dirac will not attempt as much correction so I may five that a go this weekend.

As for miniDSP units I wanted one with analog input so that my turntable would benefit from the correction as well as my CD player via Toslink. The only two Dirac models that fit the bill from an input standpoint were the DDRC-24 and the SHD, the latter being ideal with a digital out but it’s 3x the price and non-returnable. Too much to gamble on IMO.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
2,926
Likes
2,290
#14
The LS50W he uses have quite a good tuning and directivity which can be seen at his "before" curve which is almost identical to his target curve and thus in this case the difference of limiting the correction to the bass or not wouldn't be significant.
What do you propose he do? Correction in LF range certainly didn't affect imaging/soundstage as it can only be related to MF/HF so what else can he try?
 

thewas_

Addicted to Fun and Learning
Joined
Jan 15, 2020
Messages
935
Likes
1,815
#15
What do you propose he do? Correction in LF range certainly didn't affect imaging/soundstage as it can only be related to MF/HF so what else can he try?
What Dimifoot wrote above, his "wide soundstage" is probably due to his L/R asymmetries and when reducing those, it doesn't sound that "impressive", the question is what does he prefer, something spectacular that isn't in the recordings or the opposite.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
2,926
Likes
2,290
#16
What Dimifoot wrote above, his "wide soundstage" is probably due to his L/R asymmetries and when reducing those, it doesn't sound that "impressive", the question is what does he prefer, something spectacular that isn't in the recordings or the opposite.
Asimmetry in frequency response between L and R channel? He didn't show measurement for that but used only the description "slightly" so we can only guess that is the reason. Besides, balancing channels shouldn't really result in losing imaging/soundstage.
 

thewas_

Addicted to Fun and Learning
Joined
Jan 15, 2020
Messages
935
Likes
1,815
#17
Asimmetry in frequency response between L and R channel?
It doesn't have to be necessarily amplitude but also phase differences can cause this effect which is often reported and I also have witnessed it quite some times.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
2,926
Likes
2,290
#18
It doesn't have to be necessarily amplitude but also phase differences can cause this effect which is often reported and I also have witnessed it quite some times.
That is indeed true, but again it would be the phase error above 200Hz and that is why I suggested he tries limited correction as I was hoping Dirac wouldn't correct the phase above 200Hz when you limit frequency correction range.
 

thewas_

Addicted to Fun and Learning
Joined
Jan 15, 2020
Messages
935
Likes
1,815
#19
Yes, the question remains though does he really want to keep the "wrong phase" when he knows now its just an "erroneous artefact"?
Also I think Dirac always corrects the full phase but only used to have a test version of the old 1.x so I cannot test it.
 

QMuse

Major Contributor
Joined
Feb 20, 2020
Messages
2,926
Likes
2,290
#20
Yes, the question remains though does he really want to keep the "wrong phase" when he knows now its just an "erroneous artefact"?
Also I think Dirac always corrects the full phase but only used to have a test version of the old 1.x so I cannot test it.
As he didn't provide any measurements we didn't establish he has a wrong phase that produces better imaging/soundstage. In fact he shown only measurement and Dirac's estimated corrected response of his left channel, so we are deeply in speculating territory, and that is why I proposed different correction strategies.
 

Similar threads

Top Bottom