Dirac for stereo noob questions.

Timmy C

Distinguished Member
Joined
Oct 10, 2003
Messages
7,972
Reaction score
1,822
Points
1,604
Age
51
EDIT: I've changed the thread title to save me starting a new thread. I figured anyone with future questions could post in here too.

I've been playing around with Dirac for a week or so and haven't really made much progress. So I decided to try and adjust a target curve to sound pretty similar to how it sounds with EQ off and then adapt from there but have struggled to do so.

So as an experiment I decided to close the Dirac curtains as much as was possible leaving a gap from about 5k upwards. Now please correct me if I am wrong but I was under the impression this would mean Dirac would only work it's magic on that small high frequency area and leave everything else untouched. So I was expecting to hear no real difference to the overall sound. I also assumed it wouldn't matter how badly I had taken quick room measurements as it shouldn't be changing much of it anyway.

So why the hell with it loaded as it is (see pic) does it sound completely different with, amongst other things, very obviously more bass when switched off. Have I misunderstood how these curtains work or does Dirac do things other than try and match a target curve, that we have no control over? I'm really hoping someone can suggest an obvious mistake I may have made but despite checking everything I can think of, I can't figure out what's going on.

Second image is zoom in of only area Dirac is supposed to have targeted, assuming I have understood how the curtains work.

Please help!
 

Attachments

  • 2022-03-11 (2).png
    2022-03-11 (2).png
    358.4 KB · Views: 137
  • 2022-03-11 (3).png
    2022-03-11 (3).png
    346.2 KB · Views: 138
Last edited:
I don't have dirac, I'd guess you getting hi/lo pass back to front.
 
Dirac also adjusts for time delay - so if you are hearing a lot more bass that is probably because your speakers are relatively close to the wall and you are getting a lot of extra room gain. This is more bass and louder, but boomy and less tight - you can really hear this if you concentrate on drums. Boomy v tight.

Another comment I have is that you have the most bizarre target curve in the left graphic that I've ever seen. With that target curve I'm not surprised you are not getting much of an improvement in sound.

It doesn't really look like you have any target curve loaded at all. Re-reading I see that you have done that on purpose because you can't hear a difference? What tracks are you using as a reference - your in-room FR is all over the place, so I would have thought the difference would be incredibly easy to hear.

From the menu - can you try clicking "Set default target curve -> All groups. Then take a snip of what it looks like? Export the Filter and have a listen after that.

1647068297534.png



You should end up with something that looks a bit like this...

1647068346179.png



You can then adjust the shape of the target curve to taste, and have multiple target curves for different purposes; my standard curve has more bass than the Default Dirac one, and I have one for late night listening at low volume to emulate a Loudness knob - with a lot more bass and elevated treble.

You can download the Harman target curves for Dirac Live from here:


As the link says - the default Dirac Live curve is a bit bass-light for some people. The NAD target curves look to be very similar to the Harman -6dB curve.

1647069763072.png
 
Last edited:
I like to eq up to 300hz and cut only to preserve headroom.

To cut less do you raise the target curve, balancing weaker bass due to dips?

Can you disable boost other than keeping the target curve under dips?
 
You can set "curtains" to choose the upper and lower limits of where you want to EQ. If I set my curtains to leave things alone above 300Hz the system response would look like this. This is actually fine in my room and with my speakers. It's flat enough in the upper registers and retains the special character of my lovely C300s and their amazing soundstage. But of course you have the option to set multiple filters and adjust according to need. For example in an action movie you might want a -12dB Harman curve and late at night you might want a "loudness" curve with elevated bass and treble. For normal listening you could leave it with 300Hz curtains.


1647070243126.png



Actually as you can see on the curve below I would usually set the curtains at around 500Hz to smooth out the slightly lumpy early midrange.


1647070561985.png
 
Dirac is definitely for the technically minded ….. it looks hugely complicated for a Luddite such as myself.
 
Dirac adjusts impulse response across the audio band no matter what, but floyd Toole says this is inaudible, so who knows, it should sound the same.
 
Interested to know what speakers you have and what your room looks like too.
 
Dirac also adjusts for time delay - so if you are hearing a lot more bass that is probably because your speakers are relatively close to the wall and you are getting a lot of extra room gain. This is more bass and louder, but boomy and less tight - you can really hear this if you concentrate on drums. Boomy v tight.

Another comment I have is that you have the most bizarre target curve in the left graphic that I've ever seen. With that target curve I'm not surprised you are not getting much of an improvement in sound.

It doesn't really look like you have any target curve loaded at all. Re-reading I see that you have done that on purpose because you can't hear a difference? What tracks are you using as a reference - your in-room FR is all over the place, so I would have thought the difference would be incredibly easy to hear.

From the menu - can you try clicking "Set default target curve -> All groups. Then take a snip of what it looks like? Export the Filter and have a listen after that.

View attachment 1666186


You should end up with something that looks a bit like this...

View attachment 1666187


You can then adjust the shape of the target curve to taste, and have multiple target curves for different purposes; my standard curve has more bass than the Default Dirac one, and I have one for late night listening at low volume to emulate a Loudness knob - with a lot more bass and elevated treble.

You can download the Harman target curves for Dirac Live from here:


As the link says - the default Dirac Live curve is a bit bass-light for some people. The NAD target curves look to be very similar to the Harman -6dB curve.

View attachment 1666193

The target curve is all over the place because the Dirac curtains are almost completely closed and my original uncorrected room measurements were all over the place . You can see if you look again that most of my graph is greyed out This was all for the purpose of an experiment and is done with an unaltered default Dirac curve loaded. If I were to open the curtains then the target curve would completely flatten out.

Does everyone agree, looking at my graph that with those curtains closed over almost everything as shown, there shouldn't be any big noticeable difference between EQ on and off? This is what I can't get my head around as there is still very obvious differences and that includes lower frequencies that Dirac shouldn't be touching in this example.

Can you tell me more about the time delay you mention? My speakers are very poorly placed in corners. No choice in this particular room. Nothing can be done about it. Lets just say Dirac allowed you to close the curtains completely over the entire frequency range, are you saying if I did that then the time delay could still make an obvious effect on the bass sound?

If anyone can be bothered to try, I'd love to know if you get similar results with closed curtains.
 
Last edited:
Interested to know what speakers you have and what your room looks like too.
Sorry I replied just as you did. Room is awful, speaker positioning awful and can't be helped. Speakers are MA GR10's. The purpose of this thread is not to understand how to get the best out of Dirac with my room though, at least not at the moment. I just wanted to understand more about what it might be doing that I'm not aware of.
 
Is it not automatically lowering overall volume in anticipation of a need of headroom for boosting.

ie. If you level match the volume do they sound the same?
 
Is it not automatically lowering overall volume in anticipation of a need of headroom for boosting.

ie. If you level match the volume do they sound the same?

I wondered that but that's not what's causing it I don't think. Dirac auto sets the input gain on my Arcam to 10 (max). There are independent settings for EQ and non EQ so I have had to raise the level for non EQ to roughly match the output for comparison.

I had a play about with the volume control too when comparing EQ and no EQ just to be certain it wasn't just the EQ'd output being quieter overall. There's definitely something else going on but I have no idea what!
 
I've just discovered a potential mistake I may have made but I'm not sure it should have made a difference to the unexpected result of this experiment. I think I may have used my laptops mic when taking measurements rather the the Arcam one I had planned to. So everything would be way out.

Does anyone know if there's a way to tell, when loading an old project, which mic was used for measurement?

Update: Just tried the same experiment only using new measurements taken today with a UMIK. Seems everything is behaving pretty much as expected this time. Still doesn't really explain what happened but I no longer care as I'm back on track now. Thanks everyone for your input.
 
Last edited:
OK so I have a new questions and interested to hear your thoughts. Having had success in one room, I'm now trying Dirac in a room with very badly placed furniture/speakers and nothing can be done about that. It's causing big differences between left and right speaker measurements.

It seems Dirac will boost by up to 10 db but I've seen it recommended you manually ensure no more than a 4db boost. What are peoples views on this? I'm trying to play it safe but am I being over cautious?

As far as my bad room goes, what would be best practice where one speaker has a big dip? As you can only have one target curve then surely if I manually add a dip to the curve to prevent a big boost, it's just going to bring the other speaker down with it and I end up with the same differences. Or have I got this wrong?
 
Ideally you will be able to measure with REW and dial in some PEQ adjustments before running DL.

If you use a miniDSP DDRC24 or better then you can use that with REW to make much larger adjustments, best advice when using miniDSP is to get things as close to perfect as possible using REW & the miniDSP PEQ adjustments then after that fine tune with DL.

Some of the miniDSP products have DL and some do not, but for problem rooms they are recommended.

If you simply have a channel imbalance because of speaker placement then you could try a simple balance adjustment to try to even things out before running DL - that might get you close enough.

I don't know whether Arcam has any DSP functions apart from DL - I presume not.
 
Place the Mike centrally left to right to balance left and right.

Ear height to be accurate for height modes.

MLP depth to be accurate for depth modes.
 
Change as little as possible, follow dips with curve and cut less if bass is too quiet.
 
Ideally you will be able to measure with REW and dial in some PEQ adjustments before running DL.

If you use a miniDSP DDRC24 or better then you can use that with REW to make much larger adjustments, best advice when using miniDSP is to get things as close to perfect as possible using REW & the miniDSP PEQ adjustments then after that fine tune with DL.

Some of the miniDSP products have DL and some do not, but for problem rooms they are recommended.

If you simply have a channel imbalance because of speaker placement then you could try a simple balance adjustment to try to even things out before running DL - that might get you close enough.

I don't know whether Arcam has any DSP functions apart from DL - I presume not.

REW is something I briefly looked into a few years ago but I didn't get far. I will try again in future but for now I'm trying to make the best of Dirac as is.

These measurements were taken with room as is and then again with the enormous sofa that partially blocks the right speaker repositioned. Assuming there was little difference in measuring technique, it would seem the sofa is causing that crater between 100 and 200 hz. That's the spot that I'm not sure what to do about.

Currently I'm getting best results using the measurements with sofa repositioned for Dirac even though it's back in place for listening. Not sure if there's a better way to go about it though. I'll start with some new measurements now the sofa is back just to make sure I didn't do something daft first time around.
 

Attachments

  • UMIK Wide Sofa Full Living room.png
    UMIK Wide Sofa Full Living room.png
    313.5 KB · Views: 91
  • UMIK-1 Wide Sofa (Moved) Living Room.png
    UMIK-1 Wide Sofa (Moved) Living Room.png
    356.3 KB · Views: 112

The latest video from AVForums

Is 4K Blu-ray Worth It?
Subscribe to our YouTube channel
Back
Top Bottom