• 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!

What to do when Dirac won’t center the stereo image?

tjcinnamon

Addicted to Fun and Learning
Joined
Mar 20, 2021
Messages
519
Likes
193
I’ve done over 50 calibrations since getting new speakers with Dirac and only one has actually had the stereo centered. That doesn’t means the gains are equal, what I mean is that sound in the center is actually centered for MLP.

If I turn it off the sounds are right where they are “supposed” to be. I’ve obviously saved the project that is actually centered. The delays are correct but there’s usually a 2dB difference in gain which is okay if the EQ centers it again. This is a home theater so the speech should be in the middle of the TV.

I’m at a loss because my “good” project only has one measurement. I’m a little bewildered at how all over the map these calibrations are. It (should be) a pretty straightforward process and have some consistency. I’ve been using measurement mics, REW, and miniDSPs for years now. Not a pro but not a novice either.

Any ideas?
 

Bugal1998

Senior Member
Forum Donor
Joined
Mar 22, 2020
Messages
499
Likes
670
File a support claim with Dirac. They have an ongoing issue they need to resolve... And they've had it for years. Don't let them try to tell you otherwise.

Also, do make sure your MLP and mic location is at least a few feet from the rear wall.
 
OP
T

tjcinnamon

Addicted to Fun and Learning
Joined
Mar 20, 2021
Messages
519
Likes
193
File a support claim with Dirac. They have an ongoing issue they need to resolve... And they've had it for years. Don't let them try to tell you otherwise.

Also, do make sure your MLP and mic location is at least a few feet from the rear wall.
Wow!?! Out of all my times in the forums I haven’t seen this. If this works, I’ll owe you a huge beer.

I have 2 measurement mics, multiple stands, acoustic treatments, and multiple computers…. I’ve tried so many things and am to the point of looking for workarounds. Including a new AVR or a minidsp flex.
 

formula 977

Active Member
Joined
Apr 8, 2022
Messages
105
Likes
88
File a support claim with Dirac. They have an ongoing issue they need to resolve... And they've had it for years. Don't let them try to tell you otherwise.

Also, do make sure your MLP and mic location is at least a few feet from the rear wall.
Is this as big a deal as I think it is? Great job pointing out what can only be heard and not measured, tjcinnamon
 
OP
T

tjcinnamon

Addicted to Fun and Learning
Joined
Mar 20, 2021
Messages
519
Likes
193
File a support claim with Dirac. They have an ongoing issue they need to resolve... And they've had it for years. Don't let them try to tell you otherwise.

Also, do make sure your MLP and mic location is at least a few feet from the rear wall.
Any idea what the fix would be?
 

Sal1950

Grand Contributor
The Chicago Crusher
Forum Donor
Joined
Mar 1, 2016
Messages
14,073
Likes
16,609
Location
Central Fl
Great job pointing out what can only be heard and not measured,
What makes you say that ???
If Dirac is messing up the L&R balance, it could easily be measured.
 
OP
T

tjcinnamon

Addicted to Fun and Learning
Joined
Mar 20, 2021
Messages
519
Likes
193
Is this as big a deal as I think it is? Great job pointing out what can only be heard and not measured, tjcinnamon
Good point. Presumably it could be measure where if I measured right and left then one would be louder than the other.

I’ll measure a pre and post with a bad measurement. A bug in the software or the AVR is the only thing that makes sense. I’ve had plenty of other successful calibrations with my other speakers. That could incriminate the speakers BUT they sound normal with it off. Also, the measure close together without Dirac. And presumably Dirac would correct the issue not create it
 

Bugal1998

Senior Member
Forum Donor
Joined
Mar 22, 2020
Messages
499
Likes
670
Wow!?! Out of all my times in the forums I haven’t seen this. If this works, I’ll owe you a huge beer.

I have 2 measurement mics, multiple stands, acoustic treatments, and multiple computers…. I’ve tried so many things and am to the point of looking for workarounds. Including a new AVR or a minidsp flex.

I had a similar issue with my first Dirac enabled device in January 2019; the image would shift to the right as soon as the correction was enabled.

I always use the impulse display of Omnimic to center the mic to an accuracy of a mm or so, and then run Dirac (I admit to being obsessive).

Here's what would happen...

Pre calibration: I could see via the impulse display of both Omnimic and Dirac that the mic was centered and the channels were time aligned. The image was centered during listening tests.

Post calibration: Both omni mic and dirac impulse displays would show the right channel (I believe) was being delayed. In listening tests the image would shift to the right, but the levels were correct.

Dirac resolved the issue for me with a custom release file after reviewing my project files; thereafter, all was well... For a while.

I later had a similar issue with the same Dirac device in a different room, but a regular update from Dirac seemed to fix the issue.

Then over Thanksgiving 2021 I brought a pair of 708p speakers to an AirBNB to have music over the holiday week with family. There the issue returned with a new Dirac device, and this time the levels were off by a couple of dB AND the timing issue was back. It wasn't worth a support ticket for the week, so I did a manual calibration with Omnimic and the built-in speaker EQ.

If you search the various forums you will find people with similar off center imaging issues, and others complaining that Dirac sounds 'bad', 'off', 'hollow', 'out of phase', etc. etc. I suspect many or all of these users are suffering from some defect that Dirac has simply been unable to eliminate. They need to get there act together.

In the meantime, I've moved on to Trinnov (via JBL SDP-75).

But good luck with the issue... I bet Dirac can get you fixed up.

Edit: Dirac said the issue was a timing mis-match between the source clock and the playback clock.
 
Last edited:

Bugal1998

Senior Member
Forum Donor
Joined
Mar 22, 2020
Messages
499
Likes
670
Is this as big a deal as I think it is? Great job pointing out what can only be heard and not measured, tjcinnamon

It can absolutely be measured, and it is a big deal. It sounds like #&*@ when Dirac does this nonsense. See my post above.

It was my measurements unambiguously showing the issue that enabled the resolution of the issue by Dirac.
 
Last edited:

formula 977

Active Member
Joined
Apr 8, 2022
Messages
105
Likes
88
What makes you say that ???
If Dirac is messing up the L&R balance, it could easily be measured.
Good point. Presumably it could be measure where if I measured right and left then one would be louder than the other.

I’ll measure a pre and post with a bad measurement. A bug in the software or the AVR is the only thing that makes sense. I’ve had plenty of other successful calibrations with my other speakers. That could incriminate the speakers BUT they sound normal with it off. Also, the measure close together without Dirac. And presumably Dirac would correct the issue not create it
I guess what made me say that was an attempt to stir up controversy. :)
Maybe you're both right?
 

Bugal1998

Senior Member
Forum Donor
Joined
Mar 22, 2020
Messages
499
Likes
670
Incidentally, the fact that both the timing (or possibly phase correction?) is off in addition to a channel level imbalance (at least in my case) means that making a manual adjustment to the levels still results in a wonky sounding calibration.
 

Bugal1998

Senior Member
Forum Donor
Joined
Mar 22, 2020
Messages
499
Likes
670
I found the video I sent to Dirac tech support in Feb. 2019 (prior house)... It's long, but documents the initial issue I experienced, including measurements.

5/30/2023: Edited to remove link for privacy purposes now that OP's issue is resolved.
 
Last edited:
OP
T

tjcinnamon

Addicted to Fun and Learning
Joined
Mar 20, 2021
Messages
519
Likes
193
Incidentally, the fact that both the timing (or possibly phase correction?) is off in addition to a channel level imbalance (at least in my case) means that making a manual adjustment to the levels still results in a wonky sounding calibration.
That is good to know. One thing I was thinking was to add a MiniDSP Flex and manually add a dB or 2 to balance the image.

The FIR/IIR element is what intrigues me. My plan was to run Dirac just on my LR and have my AVR run vanilla so I could adjust the dB that way. Then when I get a new AVR I could get maybe an Anthem with ArcG or denon with Audyssey but use the Flex pre-cal for the impulse adjustments… a lot of expensive moving parts and points of failure.

The issue there is I don’t have Dirac on my
Center channel (for non 2 channel sources)

In fairness to myself, the Denon measures extremely well for an AVR, has a more recent SDK of Dolby DSU, HDMI 2.1, and I loved Dynamic EQ back when I had a Marantz. My NAD 758v3 measures notoriously bad SINAD-wise. So perhaps my rationale isn’t terrible.

Anyways, these are the ideas of a desperate man.
 

formula 977

Active Member
Joined
Apr 8, 2022
Messages
105
Likes
88
Incidentally, the fact that both the timing (or possibly phase correction?) is off in addition to a channel level imbalance (at least in my case) means that making a manual adjustment to the levels still results in a wonky sounding calibration.
Exactly the point. Simple measurements confirming accurate levels and thinking the problem is solved. The incorrect delay/phase, more difficult to measure, is probably what causes the problem. Even in a strictly analog system where I have tried to correct for an off center soundstage with a balance control it never worked properly and I always ended up with a 'wonky sounding calibration'.
 

Rottmannash

Major Contributor
Forum Donor
Joined
Nov 11, 2020
Messages
2,969
Likes
2,606
Location
Nashville
It can absolutely be measured, and it is a big deal. It sounds like #&*@ when Dirac does this nonsense. See my post above.

It was my measurements unambiguously showing the issue that enabled the resolution of the issue by Dirac.
If the issue is resolved why is it occurring for @tjcinnamon ?
 

Bugal1998

Senior Member
Forum Donor
Joined
Mar 22, 2020
Messages
499
Likes
670
If the issue is resolved why is it occurring for @tjcinnamon ?

No idea... As I said above, the issue later occurred on multiple occasions for me as well. The custom release Dirac sent to me--and subsequent public releases--resolved the issue for me, but only for a while.

Here's the public release for the issue I reported:

Dirac Live v2.2.1 Changelog
2019-05-09
...

Fixed
Improved UPnP device discovery
Issue that could cause off center stereo image
...
 

Rottmannash

Major Contributor
Forum Donor
Joined
Nov 11, 2020
Messages
2,969
Likes
2,606
Location
Nashville
I've never noticed this (RZ50) but now will be listening for it. Thanks.
 

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,386
Likes
24,749
Location
Alfred, NY
I ran into this myself. I redid the measurement/calibration with measured and marked mike locations and a reasonable distance between mike and wall. Cleared things right up.
 

Bugal1998

Senior Member
Forum Donor
Joined
Mar 22, 2020
Messages
499
Likes
670
I ran into this myself. I redid the measurement/calibration with measured and marked mike locations and a reasonable distance between mike and wall. Cleared things right up.

I've read that a mic position too close to the rear wall or extraneous noise can cause issues that are easily resolved with minor tweaks and/or remeasurement.

There's another subset of issues that are reproducible with great precision and are caused by a defect in the software or algorithms, such as the issues I and others have experienced.

If a dirac user gets funky results (offset center image, hollow or diffuse sounding soundstage, mismatched levels, etc.) and minor tweaks and remeasurement don't bring satisfaction, then they need to open a support ticket.
 
Top Bottom