TOPIC:

New MiniDSP (SHD Studio) / Dirac user here...

I did the 9 measurements for Dirac's Tight Focused Arrangement and was careful that the central spot was equidistant to both speakers.
Next to the right speaker is a subwoofer connected to the amp with speaker level inputs, so not on a separate MiniDSP channel. I.e.
- Output 1 = Mains-L + Sub (L channel signal) combined
- Output 2 = Mains-R + Sub (R channel signal) combined

After applying the Dirac filters, the stereo image is clearly off, with the left channel being approx. 4 dB louder than the right channel.
The strange thing is that this remains the case when I deactivate Dirac (top right) in MiniDSP.
Only when I do a full configuration restore, the stereo balance is again restored.
NO PEQ or xover are applied.

Any ideas what could be the reason / solution?

P.S. I will get a second subwoofer of the same type in a few weeks and will then redo the measurements.

Hi,
I came across this video on YouTube a week ago where they say that sometimes after Dirac you have to tweak the balance because mics and ears measure/hear differently. As a result although the balance may measure equally it doesn't sound equal. 



Appreciate if you have measured and your mic/rew is telling you there is a 4db difference this isn't the same scenario you have but might help others reading this thread with a similar imbalance.

Thanks, interesting webinar!
In my case I am pretty sure it was an SPL thing rather than a timing issue as I remember that the initial REW SPL measurements of the L and R channels were different. So perhaps Dirac got confused because of this.
Alas I did not save these measurements, but today I started with a clean slate (fully resetting the MiniDSP SW) and measured again, this time with a 90 degree UMIK-1 and calibration file. Now the L and R SPL's where very similar, see graph.
First I used REW EQ up to 300 Hz to do some prior LF smoothing and then I ran Dirac with a focused image arrangement (other than the previous tightly focused one). I like the result better than the initial one, even the stereo issue aside.
I made one warmly toned version with elevated LF and subdued HF over the full spectrum and three more transparant versions only up to 1.500, 300 an 200 Hz respectively. I tend to prefer the latter three for transparency as my room and speakers (TAD ME1) seem to be well behaved and the KEF KC62 sub sounds really controlled after Dirac filtering. The full range version is very pleasant though for instance for classical recordings with strident highs.

So far I always lost some of the holographic experience when applying a full range Dirac target curve. Is this a known thing or am I doing something wrong?

 

That the same level difference is apparent even with Dirac set to off makes sense, since doing this doesn't disable the level and time delay adjustments. Much as I wish it would. 

If you look in the plugin, is it showing a 4 dB difference on the inputs?
 

Thanks, good to know that there are also behind-the-scenes changes that remain when Dirac is turned off. Sounds like a bug though?

I am afraid I was not enable to reproduce the old issue with stereo as I did not save the configurations. But luckily it did not return after rerunning Dirac as described above.
Now taking some time to get used to the 5 sound signatures (4 with Dirac and one without) so that I will be more aware how to define the target curve after making a third set of measurements once my second sub is in place.

bodiebill wrote:

Thanks, good to know that there are also behind-the-scenes changes that remain when Dirac is turned off. Sounds like a bug though?

While I'd like it to behave differently I suspect it is intentional rather than a bug.

Have you now found where the Dirac delay and level adjustment are shown?

> Have you now found where the Dirac delay and level adjustment are shown?
I must admit no. I remotely access the NUC with the MiniDSP/Dirac SW from my Chromebook and the view is very blurred (see pic). Do you have a pointer how I can find these?
 

bodiebill wrote:

Do you have a pointer how I can find these?
 

You need to look in the SHD plugin, not Dirac Live..

Make sure the sub works for both left and right measurements..... if it’s only playing on one channel during calibration it may do goofy stuff

You might also try taking the sub off during the 1st measurement.... it won’t correct the sub phase , but it will do eq work as long as it’s in the rest of the measurements....

but it should not come to that 

just make sure sub works for both left and right measurements...

if it still does it try making the sub only in the right measurement, then do your target for both , then split your target after making it what you want and curtain off the left up to where it rolls off.... 


just some ideas to try 


if your mains play all the way down 
maybe take sub off during calibration, just do the mains if they can play all the way down with no crossover...

then add the sub back in and use dsp in the sub to manually tune to match your desired target 

 

denver8me4dinner wrote:

Make sure the sub works for both left and right measurements..... if it’s only playing on one channel during calibration it may do goofy stuff

You might also try taking the sub off during the 1st measurement.... it won’t correct the sub phase , but it will do eq work as long as it’s in the rest of the measurements....

but it should not come to that 

just make sure sub works for both left and right measurements...

if it still does it try making the sub only in the right measurement, then do your target for both , then split your target after making it what you want and curtain off the left up to where it rolls off.... 


just some ideas to try 


if your mains play all the way down 
maybe take sub off during calibration, just do the mains if they can play all the way down with no crossover...

then add the sub back in and use dsp in the sub to manually tune to match your desired target 
 

Fortounately I believe the original issue has now gone away. Have I got that right @bodiebill?

Correct, playing a mono recording, left and right now have the same volume again. I appreciate the recommendations though!
Now I am waiting for the second sub and then I will measure again, in which case I could try my planned method, i.e. measuring Main-L + Sub-L as one full range speaker and same for R, or alternatively measuring Main-L and Main-R separately and then adding and DSP-ing the subs by ear afterwards. Although I can imagine that in the latter case I may not get rid of room modes in the lower regions. On the other hand, using 2 subs could already help in that respect. 

Is it possible for you to connect things up so that each sub gets the summed left and right signals? This would generally be better if it is but I appreciate it may not be.

Having two subs myself I’d agree with @ultrasonic about having them both mono. The subs are the same make and model but due to room positioning one is completely flat 20-40Hz the other is completely flat from about 40Hz upwards. Running them stereo wouldn’t give even bass for each channel. Theyre on purpose in diagonally opposite corners which doesn’t lend itself to stereo either but does provide me with flat bass response from 20 to 100hz

Thanks @ultrasonic and @asx77, I will try both setups.
As I am planning to place both subs not too far from the amp, there should be a way to sum the outputs of the amp to both subs.
I can then compare the result to that of a stereo sub setup. (By the way, the latter enables me to run the cables for the subs directly from the corresponding main speaker.)

If you run the subs from the mains (presumably high level) you can’t set the crossover in the plugin.