miniDSP Dirac Live causing imaging problems

[keks8430]

I believe the reason why my DIRAC imaging problems is due to my non symmetrical room. I have no technical proof.

No proof either, but made the same experience as

It's super important to have the mike positions symmetrical with respect to the listening position. I found that out the hard way.

My guess is that anything asymmetrical, the room or the listening positions messes the equalization up.
This might equally explain that the more focussed DL measurements sound "better".

Try to keep your measurement points halfway symmetrical around the MLP, so that you have an about same average response above Schrodinger (~300Hz).
Or limit correction to below Schroder, or not fully correct to average slope.

 

[Anthony LoFi]

I will just add that it is important to realize that the measurement positions in the diagram are "flipped" for the home arrangements. The person is facing you. The dots on the right are the listeners left.

View attachment 342230

The studio option is the opposite.

View attachment 342231

Confused me today. =)

Does anyone have a method to accurately place the microphone in the required position ?
I made up a large wire frame and then hung small weights in the area within the frame to identify the space in the 3D.
Bloody disaster, I dont know why I thought it would work! Had a Homer moment.

 

[Anthony LoFi]

No proof either, but made the same experience as

My guess is that anything asymmetrical, the room or the listening positions messes the equalization up.
This might equally explain that the more focussed DL measurements sound "better".

Try to keep your measurement points halfway symmetrical around the MLP, so that you have an about same average response above Schrodinger (~300Hz).
Or limit correction to below Schroder, or not fully correct to average slope.

I have given up on DIRAC and spent about 2 months using REW and other software to get as best I can from my room/speaker combination.
I have tried several methods (

) and have found an outcome that suits me fine.
I use a miniDSP SHD and HD 2X4 to achieve Room EQ and active crossovers.

I think DIRAC has a place in the market for those who have a "normal" room and dont want to go into the finite detail of trial and error.

 

[droid2000]

Does anyone have a method to accurately place the microphone in the required position ?
I made up a large wire frame and then hung small weights in the area within the frame to identify the space in the 3D.
Bloody disaster, I dont know why I thought it would work! Had a Homer moment.

I've had good results with a tape measure. Make sure the chair or sofa is in the correct spot, and then it's quick and easy to locate new points using symmetric features of the sofa or chair as reference points.

I found that it's critical that the distances between measurement points is large enough. Anything less than 3' did not result in good sound.

 

[Anthony LoFi]

I've had good results with a tape measure. Make sure the chair or sofa is in the correct spot, and then it's quick and easy to locate new points using symmetric features of the sofa or chair as reference points.

I found that it's critical that the distances between measurement points is large enough. Anything less than 3' did not result in good sound.

Yes that's a very good point which I have subsequently used. I think DIRAC mention this in their notes of use. I also add extra measuring points to the standard set. It takes longer but I believe a little more accurate.

 

[keks8430]

I think DIRAC has a place in the market for those who have a "normal" room and dont want to go into the finite detail of trial and error.

Hi Anthony,

what makes your room not-so-normal?
I cannot see that from the measurements in your video (hell of a work - wow!).

Were the first two REW curves those of L and R? If so, this would be a very symmetrical arrangement.
For the bass, a slight placement asymmetry can actually help, you would have nulls at different frequencies and the two speakers could fill each other.

Did you do post-correction measurements, at different locations? (1kHz = 34cm). Curious.

 

[SIY]

Does anyone have a method to accurately place the microphone in the required position ?
I made up a large wire frame and then hung small weights in the area within the frame to identify the space in the 3D.
Bloody disaster, I dont know why I thought it would work! Had a Homer moment.

I used tape marks on the floor (using a tape measure to place them) for setting the tripod that the mic was mounted on, then marks on the tripod vertical adjustment to indicate ear height, a position below ear height, and the position above ear height.

Fresh From the Bench: NAD M10 Masters Series Streaming Amplifier - page 2 - p 14207

In this article, Stuart Yaniger reviews the M10 Masters Series Streaming Amplifier from NAD Electronics, a compact but powerful Class-D home audio amplifier and network streamer that surprises positively for multiple reasons. One of those is the integration of the latest Dirac Live room...

audioxpress.com

 

[LIΟN]

The first starting point of the impulse is important, regardless of whether the room is asymmetric or not.
The distance from each speaker to the microphone must be the same, as well as the distance from the front to back (to be precise, diagonally).
And this is actually something that should work without a big error if you're distant from listening, but users who are distant (approximately 2 to 3 meters) usually listen in their living rooms.
Then we saw that the ultra-initial reflection of the back wall also affected this.
Not only that, but some people have seen the resolution after changing from umik microphone to earthworks.

To prevent this, measure the first measurement in mono when measuring 9 points (both left and right are active).
Then measure the rest separately.
It prevents stereo image degradation or sound images from being tilted to one side in Dirac Live.

 

[Davide]

The first starting point of the impulse is important, regardless of whether the room is asymmetric or not.
The distance from each speaker to the microphone must be the same, as well as the distance from the front to back (to be precise, diagonally).
And this is actually something that should work without a big error if you're distant from listening, but users who are distant (approximately 2 to 3 meters) usually listen in their living rooms.
Then we saw that the ultra-initial reflection of the back wall also affected this.
Not only that, but some people have seen the resolution after changing from umik microphone to earthworks.

To prevent this, measure the first measurement in mono when measuring 9 points (both left and right are active).
Then measure the rest separately.
It prevents stereo image degradation or sound images from being tilted to one side in Dirac Live.

Sorry but where would these instructions come from?
Dirac's official guides explain that microphone position is anything but essential. The processing algorithm should be able to adequately weight the results.

The basic principle of microphone placement is that any additional measurement improves the
correction. More points = more accurate filters. However, depending on your room's acoustics and equipment, the benefit from more measurements may diminish faster. We recommend that
you complete every measurement point in your chosen arrangement.
• The measurement points should have a distance of at least 30 cm (12 in) between one
other.
• Avoid making measurements in too small a space. Even for the "Tightly focused"
listening environment, it is important to spread out the microphone positions in a
sphere of at least 1 meter in diameter. Too small space will result in over-compensation,
which sounds sounds very dry and dull.
• Measure some points outside the listening area. E.g. for a sofa, it is recommended to do
a few of the measurements outside the couch by 20-30 cm in any direction.
• Remember that you are measuring a three-dimensional volume rather than a two-
dimensional plane, so be certain to take measurements in different vertical positions
instead of in a single horizontal line. Consider depth as well.
• Point the microphone toward the ceiling (90°) when measuring to ensure that the
microphone's coloration is similar for both the wall reflections and the direct wave from
the speaker. A 90° microphone calibration file is needed in this case.
• The positions specified in the "Select Arrangement" view the act as a guide. You may
deviate from them as needed in order to emphasize or deemphasize particular spaces.

Then in practice I also find a reduced soundstage with the correction... but apparently it is precisely the effect of the greater phase coherence between the two speakers.

 

[LIΟN]

Sorry but where would these instructions come from?
Dirac's official guides explain that microphone position is anything but essential. The processing algorithm should be able to adequately weight the results.

It's not a guide.
I've seen the experiences and solutions of different users, not the instructions. I've experienced them, too.

 

[theREALdotnet]

How does one measure with more than one speaker active at a time in Dirac?

I have not found a degradation in soundstage, on the contrary. The centre position needs to be accurate (measure it!), the other 8 spots around the head are less critical in my experience.

Also, since I’m not trying to optimise for multiple listeners spread out on a couch, I keep the measuring spots closer together than what Dirac recommends. My “cube” around the head is no bigger than 50cm.

 

[Anthony LoFi]

Hi Anthony,

what makes your room not-so-normal?
I cannot see that from the measurements in your video (hell of a work - wow!).

Were the first two REW curves those of L and R? If so, this would be a very symmetrical arrangement.
For the bass, a slight placement asymmetry can actually help, you would have nulls at different frequencies and the two speakers could fill each other.

Did you do post-correction measurements, at different locations? (1kHz = 34cm). Curious.

Hi Keks,
The video is from another ASR member.
I just followed the methods and applied them to my room which is rectangular but has a large open room adjourning it on one side.

Im using a stereo pair with separate subs which also double as stands for the main speakers.
The cones of both main and subs are not perfectly inline so this program allows you to measure the impulse of both speakers and allow for the difference.
There are a number of other ways and methods that OCA has trialed and put up on his Youtube page.

Im not sure what you meant about post correction?

What I can say factually is:
My present system can produce a 3D sound stage using some of the methods used by OCA and verified by test audio tracks.
The 3D sound stage is collapsed when I use DIRAC as an alternate to the above.
I am not saying DIRAC is flawed. I believe my rooms irregularities may be the reason why DIRAC doesn't work for me.
Thousands of DIRAC users love it so it works for the majority.

 

[keks8430]

Im not sure what you meant about post correction?

Can you show REW measurements with all corrections applied, for several points?

I have been through some rounds and are happy now, maybe your room/speakers are not a lost cause for Dirac either.
I do not want to sell Dirac, but think we need a list of gotchas or lessons learnt.

Could you show some details about the Dirac measurements (frequency and impulse) with target curve? Best with L and R on top.

 

[Anthony LoFi]

Can you show REW measurements with all corrections applied, for several points?

I have been through some rounds and are happy now, maybe your room/speakers are not a lost cause for Dirac either.
I do not want to sell Dirac, but think we need a list of gotchas or lessons learnt.

Could you show some details about the Dirac measurements (frequency and impulse) with target curve? Best with L and R on top.

The frequency curves for both channels do not indicate spatial relationships so dont understand why this would be of interest.

My simple test was to play a number of audio tracks that have known 3D effects.
With DIRAC they vanished (still had some stereo separation).
When then switching back to "normal", the effects were there.

Too understand why would be very good knowledge but would need to employ the services of a qualified engineer to dissect my specific audio chain of components and the room to determine whats is happening acoustically.
I dont have the resources or energy for such as I believe that audio acoustic behavior is so complicated that I will accept the best sound I can achieve and spend the future enjoying it.

 

[keks8430]

I dont have the resources or energy for such as I believe that audio acoustic behavior is so complicated that I will accept the best sound I can achieve and spend the future enjoying it.

Fair enough - enjoy!

 

[Eldus]

Before I run Dirac, I make sure to get my speaker placement "good" I measure the distance to the main listening position and try to place the whole thing equal in distance to all near boundaries to at least create the SBIR for each of the speakers, moving in inches and cm. The trickiest was getting the toe-in angle equal and "right". This took months to get right for me. One day, it all fell into place. While doing this, I used REW to find and pull down the worst bass peaks only. Then Dirac just elevated the whole presentation. Transients are much better with Dirac.

 

[keks8430]

Just made a pink noise test.
I cannot tell if the phantom center is wider with Dirac Live, but it sounds definitively different.
With Dirac I get a much more audible sound variation when moving my head laterally.
For me this means that L/R comb filtering is stronger with DL.
Could make sense given DL tries to reduce reflections, which blur two-speaker combing.

 

[Davide]

It's not a guide.
I've seen the experiences and solutions of different users, not the instructions. I've experienced them, too.

I tried redoing my room measurement following a more precise and symmetrical microphone placement, and indeed the Dirac correction improved a bit this way.
The measured frequency response is practically the same, in fact I used the same target, but the stereo imaging has improved.
Furthermore, before the center was slightly off-centre, but now it is decidedly better.
As a microphone I already use an excellent Beyerdynamic MM1, which I had already noticed improvements compared to the UMIK1 (I don't know why, perhaps coincidence).

However, this contrasts somewhat with Dirac's guidelines...

 

[pogo]

DL issue