Active Room Treatment (ART) by Dirac

[IngMishael]

Hey Now and Bad Guy

oh these two will work out a driver

 

[JRowland]

oh these two will work out a driver

Most Future songs as well.

 

[umamiaudio]

Would anyone here be kind enough to help compare the performance of my MIMO speaker calibration software to Dirac ART? Mine is fully automatic and also takes your distortion measurements into consideration as well.

I know it probably seems a little bit outrageous and silly to think that a solo developer could have any chance of edging out Dirac, but we all know what happened when David met Goliath.

Currently it only runs as VST / AU. https://umamiaudio.com

 

[RenPa]

I begin with Art and would appreciate advise on FSL settings for my surrounds.
They are B&W 705
Data sheet F3 is 46Hz.
In room response has a fairly wide valley between 70Hz and 105Hz and returns up between 50Hz and 70Hz.

A good practice is to add around 15Hz to the F3 so 61Hz or up. At this setting FSL Art set a filter with + 18dB. Is it excessive? Should a better practice be to set FSL above 105Hz?

Spread at FSL 65Hz

Spread FSL at 105Hz (deeper above 105)

 

[JRowland]

I begin with Art and would appreciate advise on FSL settings for my surrounds.
They are B&W 705
Data sheet F3 is 46Hz.
In room response has a fairly wide valley between 70Hz and 105Hz and returns up between 50Hz and 70Hz.
View attachment 526765


A good practice is to add around 15Hz to the F3 so 61Hz or up. At this setting FSL Art set a filter with + 18dB. Is it excessive? Should a better practice be to set FSL above 105Hz?

View attachment 526766

Spread at FSL 65Hz

View attachment 526770

Spread FSL at 105Hz (deeper above 105)
View attachment 526771

I'd try FS Low of 50Hz but I'm unconventional and think the "good practice" is not.

 

[Keith_W]

If you see a wide valley like that in your FR, it's usually due to the confluence of two or more causes of dips in the room modes. Maybe two SBIR cancellations at slightly different frequencies (especially if the distance between the speakers and front wall is equal to distance to the side wall). Maybe two room modes. The solution is to find better placement. 70-100Hz is almost low enough to DSP your way out of it if you have multiple subwoofers, but it's better if the dip didn't exist in the first place.

 

[RenPa]

If you see a wide valley like that in your FR, it's usually due to the confluence of two or more causes of dips in the room modes. Maybe two SBIR cancellations at slightly different frequencies (especially if the distance between the speakers and front wall is equal to distance to the side wall). Maybe two room modes. The solution is to find better placement. 70-100Hz is almost low enough to DSP your way out of it if you have multiple subwoofers, but it's better if the dip didn't exist in the first place.

I think that it is 2 or more room modes. I can't change the placement. I have 2 subs. So FSL to 100Hz then? I could unsealed the ports which would help a little around 85Hz with a little 3dB see REW RTA :

 

[Keith_W]

I am sorry, I don't use Dirac. I don't know what "FSL" means. I know what I would do, but I don't know what an ART user would do.

 

[quattro98]

The Fsl will change what each group contributes to support. In your case, it looks like other groups are supporting your surrounds and filling in the null which is what you see in the corrected response (your spreads don't change since other speakers are supporting group 4, so this isn't where a change in Fsl for group 4 would be seen).

To assess the effects of changes to Fsl for this group (looks like you have surround right as group 4), you need to see the impact on the corrected response of the groups which are supported by group 4. The null in your surround right isn't what determines Fsl for that speaker, it's the response and capabilities of the speaker. You have essentially flat response down to at least 50 Hz, so can you use that or a little higher for Fsl. If the output capabilities of the speaker are limited, you can increase the Fsl or reduce the support (less negative numbers) provided by group 4 to other groups.

When making changes, having similar speakers in a single group makes things easier unless you have a good reason to have each speaker be it's own group.

I begin with Art and would appreciate advise on FSL settings for my surrounds.
They are B&W 705
Data sheet F3 is 46Hz.
In room response has a fairly wide valley between 70Hz and 105Hz and returns up between 50Hz and 70Hz.

A good practice is to add around 15Hz to the F3 so 61Hz or up. At this setting FSL Art set a filter with + 18dB. Is it excessive? Should a better practice be to set FSL above 105Hz?

Spread at FSL 65Hz

Spread FSL at 105Hz (deeper above 105)

 

[quattro98]

I am sorry, I don't use Dirac. I don't know what "FSL" means. I know what I would do, but I don't know what an ART user would do.

Frequency support low

In Dirac ART, this is the lower bound for the range in which a speaker will support other speakers. For the surround pictured, it plays down to about 50 Hz so something like 60 Hz is a reasonable lower limit for that speaker to support other speakers. ART will use other speakers (including subs) that can contribute to fill the null for this speaker. This process works up to 150 Hz, the upper limit for ART.

 

[RenPa]

(looks like you have surround right as group 4),

Indeed. I was testing a group for each speaker . The fronts are supporting it and of course the 2 subs.

 

[Keith_W]

I'll put it this way: if it's two room modes causing your wide and deep valley, your speakers are fighting each other instead of working together. DSP CAN address this, but it would have to use pretty aggressive phase shifts at those problematic frequencies to address the problem, and doing aggressive DSP like that may have side effects in the time domain. Furthermore, it's not clear if it would be entirely successful.

The root cause is speaker, subwoofer, and listener placement. No matter how sophisticated your DSP is, or how smart you are (if you are doing this manually), it is always better to tackle the root cause by repositioning. Yes, I know you said that you can't reposition your speakers. I urge you to look at your room and rethink it more seriously. That dip is in the 70-100Hz range, the wavelength of 80Hz is 4.3m. If you can shift your position by a 1m, that would be a quarter wavelength / 90deg phase shift and it will likely remove the dip. Given the nature of reflections in rooms, it is likely you would need even less.

 

[janbth]

I think that it is 2 or more room modes. I can't change the placement. I have 2 subs. So FSL to 100Hz then? I could unsealed the ports which would help a little around 85Hz with a little 3dB see REW RTA :
View attachment 526981

It's not clear to me what you show on this plot? Is this a measurement of one of the surround channels? Is it with or without Dirac ART?

 

[RenPa]

It's not clear to me what you show on this plot? Is this a measurement of one of the surround channels? Is it with or without Dirac ART?

You are right. It is the Dirac measurement and in the second post the MMM RTA measurement around the MLP of the surround right speaker with and without the foam plug I can use to seal the port.

That dip is in the 70-100Hz range, the wavelength of 80Hz is 4.3m. If you can shift your position by a 1m, that would be a quarter wavelength / 90deg phase shift and it will likely remove the dip. Given the nature of reflections in rooms, it is likely you would need even less.

Believe me I've tried to find a way to move either the mlp or the surrounds, WAF + WAF.2 ! is the first most important parameter in room acoustic after which is speaker and mlp placement!!!!!!
In the 2nd phto, you see the surround left. The surround right is quiete exactly at the opposite against the right wall. 4.3m = the distance between the surround right speaker and the end of the white wall (next the frig). More likely is half of 8.6m which is the diagonale of the floor corner of the right wall and the far top corner of the apex of the cathedral ceiling and the left wall at the top of the mezzanine. Thanks!

 

[Oddball]

You are right. It is the Dirac measurement and in the second post the MMM RTA measurement around the MLP of the surround right speaker with and without the foam plug I can use to seal the port.

Believe me I've tried to find a way to move either the mlp or the surrounds, WAF + WAF.2 ! is the first most important parameter in room acoustic after which is speaker and mlp placement!!!!!!
In the 2nd phto, you see the surround left. The surround right is quiete exactly at the opposite against the right wall. 4.3m = the distance between the surround right speaker and the end of the white wall (next the frig). More likely is half of 8.6m which is the diagonale of the floor corner of the right wall and the far top corner of the apex of the cathedral ceiling and the left wall at the top of the mezzanine. Thanks!

Based on your graphs it looks like you would have 5dB or so spread in ART range which is good. If you can verify that with REW would be even better, as a bit difficult room. Based on my experience in a difficult room, ART prediction is usually very relevant if not accurate.

If you show us the full ART filter graph, my best bet is that subs are doing heavy lifting in up to 100hz range as @Keith_W it might be difficult to align all the supporting speakers so ART probably (just a guess) picks subs as easier to phase and time align.

 

[RenPa]

If you can verify that with REW would be even better,

For the moment I only have Art results for MLP only and for the speakers with ports sealed.
I measured yesterday with ports openned. Art sounds very different now.

If you show us the full ART filter graph

Is this what you asked for? Surrounds in one group and subs in one group

 

[Oddball]

Yes, this helps but does not show the whole thing, or at least not in format I am used to. The full graph would be something like this (kind of check all boxes if in doubt).

This shows the spreads and corrected response as well for my front L.

You seem to also have a narrow dip for your Sub 1 at around 100hz, but not on your Sub 2, so I would expect ART to fix that dip for all your speakers. If you look at my spreads, you can see that it is tightening up at some 50-60hz when more support speakers are kicking in. That is not really how it measures in REW FQ response, but not really relevant as having like 3dB spread is not something you can actually hear.

 

[janbth]

… the second post the MMM RTA measurement around the MLP of the surround right speaker with and without the foam plug I can use to seal the port.

Thanks, but I’m still a bit confused about your second plot from the REW RTA. Do you really mean ‘speaker’ and not ‘channel’? If you measure the channel, then I would expect the supporting subs to fill in the dipped region, as pointed out by @Oddball.

 

[RenPa]

Thanks, but I’m still a bit confused about your second plot from the REW RTA. Do you really mean ‘speaker’ and not ‘channel’? If you measure the channel, then I would expect the supporting subs to fill in the dipped region, as pointed out by @Oddball.

I mean channel and it is RTA REW surround full sweep 0 to 20 000Hz. Sorry for the confusion. no EQ no Dirac.

 

[RenPa]

This shows the spreads and corrected response as well for my front L.

I misunderstood sorry for my confusion again!. I thought that you wanted me to measure Art filter on and MMM in real with REW.

In Dirac I knew what is spread. Here it is surround right spread: