Dirac on a 2.1 nearfield setup

[Peluvius]

I have been working on implementing Dirac on my nearfield desktop setup for the last week or so after my Minidsp flex arrived. I have not been able to find a Dirac profile that does not close down the vocals in particular. With slight tonal differences for different curves, the Dirac EQ reliably pushes the singer further back in the soundstage and thins it out. I am preferring by a significant margin an EQ export from REW actually.

I am using the focussed profile and tried tuning to a few different curves including the Harman.

Does anyone have any experience with this? Is Dirac always supposed to improve accuracy suggesting I am not calibrating it correctly or does it not necessarily work for all situations?

 

[ppataki]

I have been working on implementing Dirac on my nearfield desktop setup for the last week or so after my Minidsp flex arrived. I have not been able to find a Dirac profile that does not close down the vocals in particular. With slight tonal differences for different curves, the Dirac EQ reliably pushes the singer further back in the soundstage and thins it out. I am preferring by a significant margin an EQ export from REW actually.

I am using the focussed profile and tried tuning to a few different curves including the Harman.

Does anyone have any experience with this? Is Dirac always supposed to improve accuracy suggesting I am not calibrating it correctly or does it not necessarily work for all situations?

You can apply a target curve manually in Dirac
That way you can achieve a custom 'sound' and even fine-tune how Dirac behaves

 

[Peluvius]

You can apply a target curve manually in Dirac
That way you can achieve a custom 'sound' and even fine-tune how Dirac behaves

Thanks, I am aware of this capability. I am impressed with the user interface and the range of options available I must say. I am just not hearing an improvement.

 

[Aerith Gainsborough]

I have been working on implementing Dirac on my nearfield desktop setup for the last week or so after my Minidsp flex arrived. I have not been able to find a Dirac profile that does not close down the vocals in particular. With slight tonal differences for different curves, the Dirac EQ reliably pushes the singer further back in the soundstage and thins it out. I am preferring by a significant margin an EQ export from REW actually.

I am using the focussed profile and tried tuning to a few different curves including the Harman.

Does anyone have any experience with this? Is Dirac always supposed to improve accuracy suggesting I am not calibrating it correctly or does it not necessarily work for all situations?

I use Dirac for my 2.1 and 4.1 nearfield setups.

You most likely measure too small an area. Dirac gets extremely dry when you do that.

Also keep in mind: the "thinning out" is a desired effect to some degree. Dirac is not only smoothing the bass response but also trying to combat reverberation of your room. It will take time for your ears to adjust. Switching back and forth on/of is not recommended. Set the profile and live with it for a few days. Then make adjustments.

If you end up liking the reverb added by your room, I suggest sticking to corrections below the Schroeder frequency. Let Dirac handle the bass and let your system off the leash in the mids and treble.

Personally speaking: yes the system sounds warmer and fuller w/o Dirac but that also paints over details and results in a loss of clarity. I actually prefer the more precise, "thinner" version with Dirac but by no means does my system sound bad w/o Dirac, now that my Subwoofers DSP always handles the annoying 40 & 60Hz room modes. So it is a matter of taste.

 

[Peluvius]

I use Dirac for my 2.1 and 4.1 nearfield setups.

You most likely measure too small an area. Dirac gets extremely dry when you do that.

Also keep in mind: the "thinning out" is a desired effect to some degree. Dirac is not only smoothing the bass response but also trying to combat reverberation of your room. It will take time for your ears to adjust. Switching back and forth on/of is not recommended. Set the profile and live with it for a few days. Then make adjustments.

If you end up liking the reverb added by your room, I suggest sticking to corrections below the Schroeder frequency. Let Dirac handle the bass and let your system off the leash in the mids and treble.

Personally speaking: yes the system sounds warmer and fuller w/o Dirac but that also paints over details and results in a loss of clarity. I actually prefer the more precise, "thinner" version with Dirac but by no means does my system sound bad w/o Dirac, now that my Subwoofers DSP always handles the annoying 40 & 60Hz room modes. So it is a matter of taste.

Thanks for sharing your experience. I will try the wider zone. Your explanation sounds pretty close to what I am hearing. The closest thing I can relate the A/B to is switching between a solid state and valve amp (with Dirac EQ being the solid state).

 

[abdo123]

How about you turn the volume up a bit for these comparisons?

 

[Peluvius]

How about you turn the volume up a bit for these comparisons?

Listening at around 80db (slow/C).

 

[abdo123]

Try to volume match as much as you can when you're comparing the two.

 

[Lorenzo74]

I have been working on implementing Dirac on my nearfield desktop setup for the last week or so after my Minidsp flex arrived. I have not been able to find a Dirac profile that does not close down the vocals in particular. With slight tonal differences for different curves, the Dirac EQ reliably pushes the singer further back in the soundstage and thins it out. I am preferring by a significant margin an EQ export from REW actually.

I am using the focussed profile and tried tuning to a few different curves including the Harman.

Does anyone have any experience with this? Is Dirac always supposed to improve accuracy suggesting I am not calibrating it correctly or does it not necessarily work for all situations?

this set up present the challenge of desktop reflections. In my experience what make a difference is the distance between the speaker the desk and your head.
try to manage early reflections (front wall, desktop, lat walls,…)
can you share picture ( or drawings) , the step response and freq response at LP acquired with REW with and without Dirac?

 

[ppataki]

I use Dirac for my 2.1 and 4.1 nearfield setups.

You most likely measure too small an area. Dirac gets extremely dry when you do that.

Also keep in mind: the "thinning out" is a desired effect to some degree. Dirac is not only smoothing the bass response but also trying to combat reverberation of your room. It will take time for your ears to adjust. Switching back and forth on/of is not recommended. Set the profile and live with it for a few days. Then make adjustments.

If you end up liking the reverb added by your room, I suggest sticking to corrections below the Schroeder frequency. Let Dirac handle the bass and let your system off the leash in the mids and treble.

Personally speaking: yes the system sounds warmer and fuller w/o Dirac but that also paints over details and results in a loss of clarity. I actually prefer the more precise, "thinner" version with Dirac but by no means does my system sound bad w/o Dirac, now that my Subwoofers DSP always handles the annoying 40 & 60Hz room modes. So it is a matter of taste.

I have the same experience. Actually when you measure for example only at 1 point (in the MLP) Dirac will compensate more aggressively vs when measuring in let's say 9 points
The wider/bigger space you measure the less compensation it will apply especially below the Schroeder frequency
I have some measurements to demonstrate this if anyone is interested (I am currently not at my PC)

Actually that is one of the reasons why I ditched Dirac (for the time being at least) and instead I design my own filters so I can compensate as much as I want where I want

 

[Absolute]

Dirac is tricky in the sense that you absolutely need to measure over a little wider area both in width and height than you would expect. It's also recommended to let the target follow the dips to avoid over-aggressive correction.

Aside from that, which speakers are you using? If they are good speakers chances are that it's a bad idea to let Dirac mess with the direct sound when listening to them in a near-field scenario.
Even in my horrendous room I restrict Dirac to about 300 hz because it doesn't sound right full-range.

Same story with Audiolense btw.

 

[Peluvius]

Try to volume match as much as you can when you're comparing the two.

Yes, I try to volume match carefully to minimise the chance of that having an impact.

this set up present the challenge of desktop reflections. In my experience what make a difference is the distance between the speaker the desk and your head.
try to manage early reflections (front wall, desktop, lat walls,…)
can you share picture ( or drawings) , the step response and freq response at LP acquired with REW with and without Dirac?

Plenty of reflections! That's what I was hoping Dirac could help me control. Placement is about as optimal as I will tolerate, the area still needs to function as a workspace.

I have the same experience. Actually when you measure for example only at 1 point (in the MLP) Dirac will compensate more aggressively vs when measuring in let's say 9 points
The wider/bigger space you measure the less compensation it will apply especially below the Schroeder frequency
I have some measurements to demonstrate this if anyone is interested (I am currently not at my PC)

Actually that is one of the reasons why I ditched Dirac (for the time being at least) and instead I design my own filters so I can compensate as much as I want where I want

Actually I tried sampling a wider area as Aerith suggested and it has had a marked impact. Still have some listening and fiddling to do but definitely appears to be the way to go.

Dirac is tricky in the sense that you absolutely need to measure over a little wider area both in width and height than you would expect. It's also recommended to let the target follow the dips to avoid over-aggressive correction.

Aside from that, which speakers are you using? If they are good speakers chances are that it's a bad idea to let Dirac mess with the direct sound when listening to them in a near-field scenario.
Even in my horrendous room I restrict Dirac to about 300 hz because it doesn't sound right full-range.

Same story with Audiolense btw.

Agreed the wider sampling area has had a big impact on what I am hearing it do. I am using Genelec 8030c which absolutely sound excellent with no EQ but you know how it is......my listening area is pretty hostile acoustically.

 

[abdo123]

By the way try to limit the correction to only below 300Hz.

 

[holbob]

Is that the 915 TKL? I picked one up this week, best one I ever owned. Love it!

 

[Peluvius]

Is that the 915 TKL? I picked one up this week, best one I ever owned. Love it!

Yes, it is excellent and the latency is acceptable for FPS gaming as well. I agree it is the best !

 

[Peluvius]

Well just an update on this implementation. I am very happy with the improvements I am hearing. I am getting better results when I only play with the low end (below). This is the EQ curve I have been using for the last few days. Voice intelligibility has improved at low volumes (youtube and such), nice tight bass at high volumes. Curve looks a lot messier than I would have thought it should, but it sounds great..... still plenty of playing to be had of course.

 

[abdo123]

Do you know how to take measurements with REW by the way?

 

[Peluvius]

Do you know how to take measurements with REW by the way?

I am no expert at REW but have used it on and off for a few years to tune this and that. Mainly EQing subs and looking at crossover options. I used it on this setup to do that. Is there anything particular you would suggest I do with it in this case?

 

[abdo123]

I am no expert at REW but have used it on and off for a few years to tune this and that. Mainly EQing subs and looking at crossover options. I used it on this setup to do that. Is there anything particular you would suggest I do with it in this case?

Just take two single point measurements with acoustic timing references (one for Left speaker and one for Right speaker) before Dirac correction and i will share some filters that i think will improve the situation above 200Hz that you can insert into the MiniDSP. upload the .mdat file incase that wasn't clear.

 

[Peluvius]

I am not familiar with using these data. Let me know if I have not taken the measurements correctly.