Dirac Experts - Assistance Needed

[tmtomh]

This actually doesn’t represent a change, this is how I thought it worked from the get go. That each of the four slots represents a complete configuration. What threw me for a loop was people saying that it was wrong that I had Dirac 1 assigned to both Left and Right.

I did two sets of measurements, Slot 1 was with the 0degree mic cal and Slot 2 was the 90degree cal. I wanted to see if using a different mic cal would clear up my issue and it didn’t.

People were saying that because it is indeed wrong. Here it is direct from MiniDSP:

Under the heading of each Output - which is one channel - the green box says which input is being routed to that output. So if you have two channels, Input 1 should be routed to Output 1 and Input 2 should be routed to Output 2, as I've outlined in red here.

I understand your setup says "Dirac 1" in the green boxes instead of "Input 1" - but see the bottom note from MiniDSP, which I've also outlined in red: Dirac automatically renames the input channels to "Dirac" instead of "Input."

The same concept is referenced in MiniDSP's instructions for doing Dirac calibration - during the calibration, the left speaker is listed as Dirac 1 and the right speaker is listed as Dirac 2. Here's a sample screenshot from the volume calibration step - note how they refer to Dirac 1 and Dirac 2. both are part of the same stereo calibration process and the L channel/speaker is clearly labeled as Dirac 1 and the R channel/speaker clearly labeled as Dirac 2 - the whole point there is that you need to ensure that the output of each speaker/channel - Dirac 1 and Dirac 2 - is being picked up by the UMIK mic at roughly the same volume level.

So "Dirac 1" should only be listed under Output 1. Dirac 2 should be what's listed in the green box under Output 2.

Finally, regarding PEQ - yes, that's independent of Dirac. However, if you create a Dirac filter/setting, and you also apply some of your own manual PEQ through the MiniDSP's PEQ functionality, that Dirac filter and those PEQ adjustments will both be saved to whichever of the four MiniDSP presets you choose - and only to that preset.

If you disable Dirac but stay on that preset, then the PEQ settings will still be active. If you disable Dirac and also switch to another preset, then both Dirac and the PEQ will be disabled.

 

[Dougey_Jones]

People were saying that because it is indeed wrong. Here it is direct from MiniDSP:

View attachment 306533

Under the heading of each Output - which is one channel - the green box says which input is being routed to that output. So if you have two channels, Input 1 should be routed to Output 1 and Input 2 should be routed to Output 2, as I've outlined in red here.

I understand your setup says "Dirac 1" in the green boxes instead of "Input 1" - but see the bottom note from MiniDSP, which I've also outlined in red: Dirac automatically renames the input channels to "Dirac" instead of "Input."

The same concept is referenced in MiniDSP's instructions for doing Dirac calibration - during the calibration, the left speaker is listed as Dirac 1 and the right speaker is listed as Dirac 2. Here's a sample screenshot from the volume calibration step - note how they refer to Dirac 1 and Dirac 2. both are part of the same stereo calibration process and the L channel/speaker is clearly labeled as Dirac 1 and the R channel/speaker clearly labeled as Dirac 2 - the whole point there is that you need to ensure that the output of each speaker/channel - Dirac 1 and Dirac 2 - is being picked up by the UMIK mic at roughly the same volume level.

View attachment 306536

So "Dirac 1" should only be listed under Output 1. Dirac 2 should be what's listed in the green box under Output 2.

Finally, regarding PEQ - yes, that's independent of Dirac. However, if you create a Dirac filter/setting, and you also apply some of your own manual PEQ through the MiniDSP's PEQ functionality, that Dirac filter and those PEQ adjustments will both be saved to whichever of the four MiniDSP presets you choose - and only to that preset.

If you disable Dirac but stay on that preset, then the PEQ settings will still be active. If you disable Dirac and also switch to another preset, then both Dirac and the PEQ will be disabled.

When I have a block of time, I will clear all of my presets and start over from scratch.

Currently:

Would appreciate everyone's feedback on how the room may be interacting with my setup. The room is no longer a bare, reflective box. It does have a deep pile rug, a fairly large couch on the back wall and some curtains on the 96" wide sliding glass doors. But I admit, it is still a fairly live room between the brick wall, 8ft sliding glass doors, walls and ceiling.

None of that accounts for the fact that with Dirac disabled, the center image is stable, centered and correctly positioned from an image height standpoint. All that goes out the window when Dirac is enabled, but I do like what Dirac does for the FR. It seemingly fixes a lot of "congestion" in the mid bass through midrange and adds a little sparkle to the treble.

Edit: The PS Audio ICE Power based amps are being replaced by the Apollon NCx500 based units. One is already installed for the front L/R and the other unit is with Amir for testing.

 

[Anthony LoFi]

Excellent result. You have now discounted one variable. I think this is important when your trying to solve a problem where there are no obvious solutions.
To answer your NOTE from above.
I do think a clear path of space is created around the sphere of the mic is important, hence why Dirac has a time delay before each position of the mic is measured. I believe this is to give you time to exit the general area. Another aspect could be that the boxes could duplicate the human body so we could negate that obstruction from interfering with the measurement results. As you say it would be difficult to use a floor mounted mic stand because you would have to move the furniture so why not use your existing method of boxes. We must leave the couch and other furnishings place as the reflective and sound absorbing properties of you room is what Dirac is all about.

I suspect the unsatisfactory results you are experiencing could be very complex boundary reflections of unequal balance from each speaker.
You were going to attach a plan of your room?
I would try and simplify the measuring process by using the single person narrow listening position. You haven't mentioned which of the 3 measuring positions you are using (narrow, single person, medium, two persons or wide, 4 persons)

Could I suggest doing two complete measurements in 1 Narrow and the another in Wide and save them into slot 1 and 2.
Then you could perform your reference listening test to see where there is an improvement.

If you find the Narrow measurements gives the best results to you, then try the following.
Leave the first Narrow 9 point measurement in slot 1.
I would then do a second measurement exactly the same but with additional measurements that you can add more test points to the narrow listening position. After completing the standard 9 test spots, the tile on the screen display has an extra button that you can with your mouse select another measurement place in the 3D area around the listening positions head. These new extra measuring positions would be within the existing boundaries shown on the screen but between the first 9 points. Then save this in slot 2.
Then you can assess the the two measurements while using your test recordings.
You may find extra test points may increase the accuracy of Dirac to provide better room response.

You haven't mentioned if you have done a thorough speaker placement review? as mentioned in my earlier post.
I find this is critical to getting the right results. Dirac can only do so much. You need to give the Dirac system a chance to work its algorithms.
Also, when first setting up the levels in the second Dirac calibration screen, both left (Dirac1) and right (Dirac2) measurement vertical meters should read the same.
If left speaker is louder than the right, try moving the mic horizontally to the right so both speakers are equal in distance from the listening position.
Dirac says the first centered measurement is the most critical so I believe you should get out the tape measure out to ensure the distance from each speaker to the mic test place (your heads listening position) are equal.

Keep us posted.

 

[Anthony LoFi]

Sorry just saw you photos.
Yes you do have a challenging room. Glass on the left, semi absorbent brick on the right.
Now just a question, how are you feeding your sub?. You said you only use your flex 1 & 2 outputs, not using 3 & 4 for sub (crossover etc)
You have time delays for sub to take into acount and straight firing down the left hand side.

 

[OCA]

When I have a block of time, I will clear all of my presets and start over from scratch.

View attachment 306548
View attachment 306549

Currently:

View attachment 306550

Would appreciate everyone's feedback on how the room may be interacting with my setup. The room is no longer a bare, reflective box. It does have a deep pile rug, a fairly large couch on the back wall and some curtains on the 96" wide sliding glass doors. But I admit, it is still a fairly live room between the brick wall, 8ft sliding glass doors, walls and ceiling.

None of that accounts for the fact that with Dirac disabled, the center image is stable, centered and correctly positioned from an image height standpoint. All that goes out the window when Dirac is enabled, but I do like what Dirac does for the FR. It seemingly fixes a lot of "congestion" in the mid bass through midrange and adds a little sparkle to the treble.

Edit: The PS Audio ICE Power based amps are being replaced by the Apollon NCx500 based units. One is already installed for the front L/R and the other unit is with Amir for testing.

I'd rather be in that pool but if you're inside, this information from a respected German magazine might help you with the Revels:

If your room reverbration time is below 600ms and the speakers are 90cm or more away from the front walls, it seems to be fine. AK 62 is quite high btw, these speakers need proper juice, an AVR wouldn't cut it. I cannot see your amplification but my Kantas are also 62 and they required a Michi to bring the best in them.

 

[witwald]

I'm at a loss at this point, I really like the FR adjustments that Dirac makes, so from that perspective it's doing a great job, but it totally dismantles the otherwise coherent center image that my system creates.

Maybe that is just the very nature of the processing that Dirac introduces, and there is no workaround. When you mention "FR adjustments", as far as I can tell Dirac does not adjust the on-axis response alone of the loudspeakers. If that were something that it concentrated on, then Dirac would be a worthwhile broadband system. As it stands at present, those "FR adjustments" are an amalgam of changes based on room acoustics, loudspeaker radiation pattern, and loudspeaker frequency response shape. Dirac is proposed to be a room correction system that is a tool to get the sound that you like, because in the end it is subjective and there isn't a one-stop solution here. Hence, in your case, removing the "effects" introduced by Dirac appears to be the logical approach to take. It appears doubtful that there is any better solution available to you.

 

[OCA]

When I have a block of time, I will clear all of my presets and start over from scratch.

Post-calibration off-center behavior is a known Dirac issue. You can raise a ticket to their tech support and they will probably solve it. They will try to tell you your MLP is too close to the rear wall, etc. but it's due to timing mis-match between the source clock and the playback clock. Some MiniDSP units introduce too much delay for Dirac to cope with.

 

[tmtomh]

Post-calibration off-center behavior is a known Dirac issue. You can raise a ticket to their tech support and they will probably solve it. They will try to tell you your MLP is too close to the rear wall, etc. but it's due to timing mis-match between the source clock and the playback clock. Some MiniDSP units introduce too much delay for Dirac to cope with.

Can you explain this bit further? Are you talking about DSP latency, or something else? Thanks!

Also, I am intrigued by @Dougey_Jones 's statement at the beginning of this thread that with Dirac the perceived image also moves upward vertically. I've never heard of that before - and I'm thinking if there is a way to exert some intentional control over the height of the soundstage image, that could be very handy. Does anyone have any insight into that?

 

[OCA]

Can you explain this bit further? Are you talking about DSP latency, or something else? Thanks!

Also, I am intrigued by @Dougey_Jones 's statement at the beginning of this thread that with Dirac the perceived image also moves upward vertically. I've never heard of that before - and I'm thinking if there is a way to exert some intentional control over the height of the soundstage image, that could be very handy. Does anyone have any insight into that?

Troubleshooting: Off-Center or Biased Imaging - Dirac Live Support - Confluence

confluence.dirac.services

the one before last starting with "buffer size..." is the most common cause AFAIK.

 

[Dougey_Jones]

Sorry just saw you photos.
Yes you do have a challenging room. Glass on the left, semi absorbent brick on the right.
Now just a question, how are you feeding your sub?. You said you only use your flex 1 & 2 outputs, not using 3 & 4 for sub (crossover etc)
You have time delays for sub to take into acount and straight firing down the left hand side.

Sub is not connected to my stereo setup, it’s only connected to the RZ50 for Home Theater. The Left/Right preamp outputs from the RZ50 get fed to the Adcom preamps processor loop, so when I use my HT, I just hit processor on the remote and that lets the RZ50 pass through my preamp to the Hypex amps.

When I want to listen to my stereo only setup, I just disengage processor and it’s like the HT setup doesn’t exist.

 

[Dougey_Jones]

Can you explain this bit further? Are you talking about DSP latency, or something else? Thanks!

Also, I am intrigued by @Dougey_Jones 's statement at the beginning of this thread that with Dirac the perceived image also moves upward vertically. I've never heard of that before - and I'm thinking if there is a way to exert some intentional control over the height of the soundstage image, that could be very handy. Does anyone have any insight into that?

It’s definitely possibly to change image height via DSP, because there are AVR manufacturers who have this feature. I think Sony and Yamaha? Not sure, but I’ve 100% seen this before as a feature.

 

[Dougey_Jones]

Troubleshooting: Off-Center or Biased Imaging - Dirac Live Support - Confluence

confluence.dirac.services

the one before last starting with "buffer size..." is the most common cause AFAIK.

How/Where do I change buffer size? Is bigger or smaller better?

 

[OCA]

How/Where do I change buffer size? Is bigger or smaller better?

I don't have Dirac at the moment but if I remember correctly you can find it by selecting properties of Dirac live processor under Windows Sound settings or in ASIO settings. I'm not really sure. Someone with Dirac should help.

 

[Dougey_Jones]

Post-calibration off-center behavior is a known Dirac issue. You can raise a ticket to their tech support and they will probably solve it. They will try to tell you your MLP is too close to the rear wall, etc. but it's due to timing mis-match between the source clock and the playback clock. Some MiniDSP units introduce too much delay for Dirac to cope with.

Just to clarify, if all of my other options fail and I end up lodging a ticket with MiniDSP, is it the Flex Balanced hardware itself that’s at fault?

Basically a hardware failure or QC problem which would involve them replacing the unit?

 

[vicenzo_del_paris]

Just to clarify, if all of my other options fail and I end up lodging a ticket with MiniDSP, is it the Flex Balanced hardware itself that’s at fault?

Basically a hardware failure or QC problem which would involve them replacing the unit?

I would be very surprised that the source of your issue is linked to minidsp hardware.
So far, your flex configuration was faulty (wrong routing) and I am not surprised that it could lead to the described off center issue.

 

[Elgrosso]

It seems clear the process was faulty, especially the fact that you didn’t use the right routing.
Each preset are independent, routing/crossover/peq etc + dirac adjustments.
If you want to compare only different Dirac calibrations then the same pre-Dirac settings need to be uploaded in all presets.
And don’t change anything on the presets after each dedicated calibration (unless you know what you’re doing).

Mic calibration file would not have such an impact, it’s mostly on highs and quite subtile.
Room and speaker placements are fine, they would not impact imaging that much.
Current preset 1 is mono, so of course center is strong, with or without dirac applied.
Current preset 2 use a stereo feed, but probably with the wrong dirac calibration, or the one done with preset 1, so messy FR adjustment, timing off, etc.

 

[Deleted member 48726]

I actually quite like the effect that Dirac has full range, but I will try limiting the project's scope to see if it makes a difference in the imaging.

No need to limit what it can do. I do full range as well. How am I else supposed to create the sound profile I like?

It takes time thoug. Be patient and exxperiment with the target curves. I've had succesful EQ by pulling the target so it is below the response curve. This means no boosting of frequencies are applied.

 

[Dougey_Jones]

I would be very surprised that the source of your issue is linked to minidsp hardware.
So far, your flex configuration was faulty (wrong routing) and I am not surprised that it could lead to the described off center issue.

Believe it or not, the issue existed prior to me creating a second (90 degree) profile to try and fix the issue. The control panel mixup was just me futzing around. Changing it back didn’t fix the off center or image height issues.

 

[vicenzo_del_paris]

Believe it or not, the issue existed prior to me creating a second (90 degree) profile to try and fix the issue. The control panel mixup was just me futzing around. Changing it back didn’t fix the off center or image height issues.

Did you redo calibration with correct routing ?

 

[Dougey_Jones]

Did you redo calibration with correct routing ?

Just so we’re on the same page. In order to do the calibration correct with the correct routing, I should:

1) Delete the data from all of my presets.
2) Click on profile 1 on the left
3) Have routing set to input 1 Dirac 1 and input 2 Dirac 2.
4) Launch Dirac from the control panel.
5) Run calibration for stereo
6) Set target curve and export results to slot 1

I plan to make two calibrations, another attempt at a full range cal, and a separate stereo cal only correcting 0-350 or 0-500hz.