Dirac Adjustments for “bright” sound

[BigDan79]

Hello,

My calibration seems a little “bright” in my opinion. The treble seems high. What adjustments can be made in Dirac to warm the sound a bit? I think that adjusting the treble goes back to zero when restarting my pre amp so I would like to adjust the Dirac filter. I am currently using the Harman curve.

 

[JStewart]

Probably shooting in the dark without seeing your response and house curve. I’ve seen several different “Harman” curves too, with differing high and low ends. (The do all have some degree of bass boost).

Perhaps a screenshot from Dirac showing the average response and target curve? A REW measurement at your listening position would show something different if the average is not a good representation.

You might also try lowering the right curtain to 250Hz and not correct the response above that. I have found this works well with my JTR210RT which like your Arendals have a narrower and more constant directivity. Especially with 2 channel where the added room reflections can be pleasing.

Another thought is to be sure you’ve followed Dirac’s guidelines for mic spacing. Measurements that are too close to each other can produce filters with dry or harsh sounding results.

Oh, and member @serko70 posted one of the better Harman curves I’ve seen because the high end has a nice roll off where some are flat through mids and highs.

 

[BigDan79]

Thank you for the reply. I did follow the Dirac mic spacing guidelines. I will try your recommendations. I am not the greatest at REW so I need to work on getting some measurements. I will try the Dirac changes and feedback the results.

 

[JStewart]

Thank you for the reply. I did follow the Dirac mic spacing guidelines. I will try your recommendations. I am not the greatest at REW so I need to work on getting some measurements. I will try the Dirac changes and feedback the results.

Good plan. If the Dirac changes work for you then REW would be for knowledge only.

 

[BigDan79]

Thanks again. I will mess with it this weekend when I can find some quite time with the kids out of the house.

 

[Randy L]

Not sure which version of Dirac you're using. But in my PC version we are able to adjust the target curve by "grabbing" the high-end tab (or low-end tab) of the curve and sliding it up or down which creates a nice roll-off (or boost) throughout the high frequencies (or low frequencies) based on your resulting room acoustics. In your case just lower the high-end tab by 2 or 3db, save filter, the listen. Tune to taste... Alternatively you can move each plotted point you see along the curve manually but it's more work.... The 'auto' curve is based on your resulting calibration (your room acoustics) and moves multiple points at the same time with a simple tab.

 

[dsnyder0cnn]

Hello,

My calibration seems a little “bright” in my opinion. The treble seems high. What adjustments can be made in Dirac to warm the sound a bit? I think that adjusting the treble goes back to zero when restarting my pre amp so I would like to adjust the Dirac filter. I am currently using the Harman curve.

I agree with JStewart. Tough to say without seeing your target curve. Here's what I'm using for my Legacy Audio Focus XD speakers in our little family room home theater. Tonality sounds neutral to me. The trick seems to be creating a target curve that does not deviate too far from your speaker+room's native response.

Sorry that I don't have the uncorrected response relative to target to show since I think that would illustrate my point a little better. What you see here is flat from 100 Hz to 1 kHz, -1 dB/octave above 1 kHz, and a bit of a rise in low-end below 60 Hz. I did something similar for the center and surrounds (this is a 5.0 system):

Above, the center is the Legacy Audio Marquis XD. Surrounds (below) are JBL 305P active monitors:

For the surrounds, I chose to roll the highs off a little more steeply since they are just sitting on the floor (temporary arrangement until we move to our new place).

For Dirac, I do prefer to use the point-by-point method for building the curve to manipulating the two shelves. I wish there was a way for me to just enter or edit the coordinates for each point rather than dragging them around with a mouse, but the hassle only has to be done once per room.

Hope this helps with another take. Keep in mind that the broader the frequency range of your adjustments, the more impact they will have on tonal balance. The difference between -1 dB/octave and -0.9 dB/octave above 1 kHz is quite obvious.

If your processor has several banks for you to load filters, I'd encourage you to design two or three sets based on slightly different target curves so that you can compare. You may even find it handy to have different versions that you can load depending on source material.

 

[Randy L]

I agree with JStewart. Tough to say without seeing your target curve. Here's what I'm using for my Legacy Audio Focus XD speakers in our little family room home theater. Tonality sounds neutral to me. The trick seems to be creating a target curve that does not deviate too far from your speaker+room's native response.

View attachment 58988

Sorry that I don't have the uncorrected response relative to target to show since I think that would illustrate my point a little better. What you see here is flat from 100 Hz to 1 kHz, -1 dB/octave above 1 kHz, and a bit of a rise in low-end below 60 Hz. I did something similar for the center and surrounds (this is a 5.0 system):

View attachment 58989

Above, the center is the Legacy Audio Marquis XD. Surrounds (below) are JBL 305P active monitors:

View attachment 58990

For the surrounds, I chose to roll the highs off a little more steeply since they are just sitting on the floor (temporary arrangement until we move to our new place).

For Dirac, I do prefer to use the point-by-point method for building the curve to manipulating the two shelves. I wish there was a way for me to just enter or edit the coordinates for each point rather than dragging them around with a mouse, but the hassle only has to be done once per room.

Hope this helps with another take. Keep in mind that the broader the frequency range of your adjustments, the more impact they will have on tonal balance. The difference between -1 dB/octave and -0.9 dB/octave above 1 kHz is quite obvious.

If your processor has several banks for you to load filters, I'd encourage you to design two or three sets based on slightly different target curves so that you can compare. You may even find it handy to have different versions that you can load depending on source material.

Wow, great answer! - Complete with graphs! - Ya, we really need to see his calibration info.

 

[BigDan79]

Please see the screenshots. Not sure why my line is so thick.

 

[BigDan79]

Looks like after 10k the speakers roll off but the program boosts them. Should I just bring the curtain over to 10k to reduce this end? I am a novice at this so please be patient.

 

[dsnyder0cnn]

Sliding the curtain left will help. It looks like your target only drops about 3 dB from 1 kHz to 20 kHz. That seems a little too gradual to me. I generally prefer a target that's closer to -5 dB over that range. This will more closely match the natural response of your loudspeakers + room.

 

[Randy L]

Looks like after 10k the speakers roll off but the program boosts them. Should I just bring the curtain over to 10k to reduce this end? I am a novice at this so please be patient.
View attachment 59019

I downloaded that Harmon curve too and got the same 'thick line'. Lol, not sure why either. Maybe it has something to do with a zoom-scale mismatch between the original program and Dirac. - I have Elac speakers which benefit from the Dirac 'boost' at the tweeter level. When the mids and bass are increased the trebble is preceived to sound softer by comparison so sometimes I prefer to increase trebble to create the perception of balance (flatter) response across the spectrum... But of course sound is a subjective thing. And different rooms create different perceptions.

 

[BigDan79]

I slid the curtains. I will give it a listen and see if its calmed down. After I address this issue I need to fix my bass. I have a big dip at 30hz. Thats a project for when I have a lot if time. Thanks for all the comments.

 

[JStewart]

Not sure why my line is so thick.

I downloaded that Harmon curve too and got the same 'thick line'. Lol, not sure why either.

It's because of the large number of closely spaced points in the file.

Attached is the same curve with less points so there's no thick line and a 4dB bass boost Version.

Also a version with 0.9dB/Octave roll off starting at 1kHz as @dsnyder0cnn suggested. It's has Big Dan in the title. Here's a comparison of the two in REW. Note that absolute SPL is arbitrary for illustration.

If you where used to listening with the bass between 100hz and 200hz I think its going to sound a bit brighter with it removed. It's not uncommon for new dirac users to have that perception when peaks are removed. That's not to say you shouldn't have it sound the way you like, just offering a possible explanation for your perception.

I slid the curtains. I will give it a listen and see if its calmed down. After I address this issue I need to fix my bass. I have a big dip at 30hz. Thats a project for when I have a lot if time. Thanks for all the comments.

Re: the 30Hz dip. If the response from the subs (still multiple on a MiniDSP?) is good, an 80Hz XO will be make this a minimal concern. See this excellent guide from Markus767 on AVS for determining crossover levels in the processor/avr with Dirac.

Setting crossovers in an AVR/pre-pro with Dirac Live 2 room correction