Active Room Treatment (ART) by Dirac

[BJL]

But Dirac does allow it. Storm has it. The question, I believe, is where the gain is provided within the signal flow. What Dirac does not allow are multiple changes to the different channels and setting PEQ etc. At least, this is what the technical support at miniDSP has said in response to questions about why we can't adjust levels to all speakers.

Also, maybe Marantz is adding gain post Dirac processing. That would definitely mess up ART.

 

[JRowland]

I really think the “issue for me” is simply the dynamic range within the dialogue. Many movies people are whispering which I struggle to hear so I turn the volume up and then the explosion blows me out of the room. Perhaps BladeRunner2049. The opening few minutes has 22hz LFE and then around the 6 min mark when Ryan Gosling first speaks to Batista when Batista first walks in the kitchen, there is some whispering at low low levels. Now to hear this bit of dialogue at normal levels the opening eyeball scene with 22hz LFE would blow anyone out of the room

I watched movies when BladeRunner came out. Thanks, should be fun. I bought Mad Max Fury road when I put this system together because I liked Mad Max and Road Warrior, that was fun! I need to try it again with ART.

 

[ASharma]

Also, maybe Marantz is adding gain post Dirac processing. That would definitely mess up ART.

Correct, Marantz “was” adding gain post Dirac. Dirac has shut that down now and I understand why. Now on the Tide16 it looks like the trim adjustments are made pre Dirac. Monolith HTP-1 also does this and then these trims are included in the Dirac calculation (I believe). On the Tide16 If you are able to adjust Center trim real time by movie (let’s say u boost Center by 5db for some movies) and you have other channels supporting the Center how do the anti phase signals know to fire louder given the +5db change you made to the Center, given this is pre Dirac?

Sorry to clarify are you saying you can adjust Center trim real time per any movie scene or are you saying you can adjust Center trim once and then Dirac uses the new trim level in its calculation?

 

[BJL]

Correct, Marantz “was” adding gain post Dirac. Dirac has shut that down now and I understand why. Now on the Tide16 it looks like the trim adjustments are made pre Dirac. Monolith HTP-1 also does this and then these trims are included in the Dirac calculation (I believe). On the Tide16 If you are able to adjust Center trim real time by movie (let’s say u boost Center by 5db for some movies) and you have other channels supporting the Center how do the anti phase signals know to fire louder given the +5db change you made to the Center, given this is pre Dirac?

Sorry to clarify are you saying you can adjust Center trim real time per any movie scene or are you saying you can adjust Center trim once and then Dirac uses the new trim level in its calculation?

On the Tide you can adjust the level of the center channel up to plus or minus 10dB using a level control on the remote, or using the Console app on your computer, in real time, while you are watching the show. I just used it, watching an old TV show encoded Dolby surround, where the original media (a DVD) has the dialog in the center and too low. I adjusted the center until the dialog was sufficiently prominent. Checking the console, I raised it 1.5 dB (the console is an application that accesses information and features for the Tide in great detail.) You can raise or lower the level as much as you want while watching, like a volume control.

It does not have any audible effect on the overall ART room correction. I don't know why, I have no technical expertise about ART under the hood, just practical use. I do know that it is adjusting the input level of the center channel post dts/Dolby decoding, but before Dirac processing, and it is only the level of the center channel input that is being raised (or lowered), not the level of any support or other signals being redirected to the center speaker, it is processed by the Dirac suite at the trim or gain level adjustment, and the designated support speakers (for the center) are providing support signals using the level that you set when making any adjustment.

I only use this feature for TV/cinema where the center channel is carrying the dialog, and only to the extent necessary. It is a precise control. A short tap of the up or down control is .5 dB. I was surprised to find out that small changes can make a big difference.

I'll say it again, the Dirac suite is integral to the design of the TIde. It is not an "add on" feature. I can't compare to any other brand, but my experience is that Dirac processing on the Tide is superb.

 

[mmares056]

Hi folks. I’ve been using ART for a month or so now and have a few questions. I really wish Dirac had easily accessible tech support so I could better understand best practices. If they do, please let me know. Ok a couple starting questions. I’m running a 7.3.6 setup. I’ve attached my REW waterfall which “I think” looks pretty good.

1) I’m simply using Dirac’s default curves. They are all slightly different below 150hz depending on the channel. I’ve also loaded a Harman +4 curve in the past. My question is what’s best practice? I’m assuming Dirac’s default curves are somewhat optimized or they wouldn’t be the default? Is there a downside to running these slightly different below 150hz Dirac default curves?
2) I struggle with clear clean dialog. I know with Audessey there were different tools a person could use such as dialogue enhancer, dynamic volume or simply adjusting the trim level. I don’t see those options with Dirac only the choice of designing a custom center curve. My question is if I use the default Dirac center curve and I raise the entire curve by 3db to simulate bumping the Center trim level by 3db is there any downside or would this break the calibration in any way given the Center is used as the timing reference for LFE?

That’s it “for now”. Any insights would be appreciated.

Thanks folks

As you have noticed everyone has different tastes when it comes to target curves. Personally I don't use generic target curves because in my opinion you leave way too much on the table and the idea is to get the best sound out of your speakers in your room. I use custom, room/speaker specific curves where below the transition region the same target curve is applied(Unified) to all the speakers and above the transition region a per speaker correction curve is applied. I don't boost or cut channel levels post calibration. I do full range correction because that makes my speakers sound neutral and that's what sounds the best in my system. I use the same preset for movies and music. Dialogue clarity is crystal clear. The bass sounds natural, awesome, all speakers sound cohesive and accurate and object panning is superb.

For illustration sake, here are a few screenshot of my ART project. I have 5.4.2 Setup.

This is the Center predicted response

This is the Center Actual response with center room curve overlayed.

These are all speaker predicted response.

These are all speakers actual response (RTA MMM replicating 12 point ART measurement pattern, 50% weighted towards the MLP.

 

[MouseOnMars]

Can you show the measured response of those 2 subs, dirac screenshot just like above? Wondering if one or both have issues around 40Hz where they are being used to boost the mains. Sub 2 is always boosting in every screen shot at 40Hz so I'd look there first. The boost doesn't look excessive (looks to be around +3db) but if they (it) are (is) coming from say -10db down at 40Hz and being tasked with boosting for the mains that could cause issues I would imagine. Also, your mains have good in room response to 35Hz (ish), you could try moving the FS low on the mains from the default 50Hz to 40Hz and check the new filters to see if that takes some of that load off of the subs, I would think it would. I know it is the opposite of what you tried already but it's worth a quick test. To quickly negate that sub 2 boost if desired lowering the support level on sub 2 from -18 to say -12 (or -6 if needed) on all speakers it supports should bring that filter line down to or below the 0db line. Start with -12 and calculate and then look at the filters. Sub 1 is always boosting a little above 100Hz so it could be there as well, same thing applies as what i suggested for sub 2 but you could also try a curtain for sub 1 at 100Hz (if it is the problem and it goes away with say a -6 support level). Hopefully you are using the dirac "screenshots" and relabeling them to match what you have uploaded into your filter slots, that allows you to quickly flip back and forth to see filter differences.

I had to lower the sub support level to -6dB to get sub filters close to 0dB.

When I lower the mains FS low to 40Hz, the subs filters go up quite a bit. They must have to work harder to support the mains at the lower frequency.

It's too late to see if these filter changes reduce the distortion. I'll have to wait until tomorrow to test.

 

[temeq]

Great idea @temeq . I’m looking more closely at the filters to try and spot a problem. Some of them don’t make any sense to me, so I’m still trying to figure them out. I attached images of them in hopes someone else may spot an issue. It’s a 2.2 system, so there are four filters per main speaker. The measured (uncorrected) response of the main speaker is also in each image. All Dirac settings were left at default values except the target curves were zeroed.

If you are sure that distortion is at 40hz or so , I see a peak in the filters around 42Hz. I would draw a sharp -5dB dip to 42Hz to your fronts target curve and see if that reduces the distortion. There's also a big peak at 100Hz so I would confirm with long frequency sweeps what the distortion frequency is.

 

[smygolf]

A question for those with 4 subs..
Have you noticed any major differences when changing sub placement with ART ?

I have a 3 sub setup, two in baffle wall and one behind mlp. (JBL 4645C)
Before ART i only had the front subs and they played in parallell, and with ART i do loose some bottom end.

Now i might have 2 subs on the way in (DIY) and the plan is to go back with parallell connection on the front subs.
Question is where to place the new ones, they are built for corner placement so in the corners behind mlp sound as the logical thing to do ?

 

[OCA]

An idea:

Atmos tracks have a "dialnorm" setting embedded in them which stands for dialogue normalization. If you do not turn off "Loudness management", processor will follow this setting during playback and it varies from track to track. I always turn it off (it's possible the opposite is useful for dialogue clarity for certain tracks - have not tested that thoroughly). That setting will not show up in the AVR menu unless it's being fed an Atmos signal btw.

 

[JRowland]

View attachment 551551

I had to lower the sub support level to -6dB to get sub filters close to 0dB.

When I lower the mains FS low to 40Hz, the subs filters go up quite a bit. They must have to work harder to support the mains at the lower frequency.

It's too late to see if these filter changes reduce the distortion. I'll have to wait until tomorrow to test.

Sure seems strange, both subs and both mains don't appear from the measurements to have any issues at 40Hz. Are you sure it's not a setting issue like the loudness setting OCA mentioned, having Directional Bass turned on, or the crossover in 2Ch advanced settings being set to anything other than auto (those settings are only available with Dirac set to off)?

 

[90sNintendoKid]

As you have noticed everyone has different tastes when it comes to target curves. Personally I don't use generic target curves because in my opinion you leave way too much on the table and the idea is to get the best sound out of your speakers in your room. I use custom, room/speaker specific curves where below the transition region the same target curve is applied(Unified) to all the speakers and above the transition region a per speaker correction curve is applied. I don't boost or cut channel levels post calibration. I do full range correction because that makes my speakers sound neutral and that's what sounds the best in my system. I use the same preset for movies and music. Dialogue clarity is crystal clear. The bass sounds natural, awesome, all speakers sound cohesive and accurate and object panning is superb.

For illustration sake, here are a few screenshot of my ART project. I have 5.4.2 Setup.

This is the Center predicted response
View attachment 551550

This is the Center Actual response with center room curve overlayed.
View attachment 551552

These are all speaker predicted response.
View attachment 551553

These are all speakers actual response (RTA MMM replicating 12 point ART measurement pattern, 50% weighted towards the MLP.
View attachment 551554

I use the same approach as @mmares056 and I'd like to say how awesome your post EQ measurements are......................I think I can guess how great your system sounds.

 

[90sNintendoKid]

An idea:

Atmos tracks have a "dialnorm" setting embedded in them which stands for dialogue normalization. If you do not turn off "Loudness management", processor will follow this setting during playback and it varies from track to track. I always turn it off (it's possible the opposite is useful for dialogue clarity for certain tracks - have not tested that thoroughly). That setting will not show up in the AVR menu unless it's being fed an Atmos signal btw.

I always turn off Loudness Management.

 

[ASharma]

I use the same approach as @mmares056 and I'd like to say how awesome your post EQ measurements are......................I think I can guess how great your system sounds.

Thanks man…To clarify, are you using the Dirac default curves above 150hz and then the Dirac LFE curve for all curves below 150hz?

 

[90sNintendoKid]

Thanks man…To clarify, are you using the Dirac default curves above 150hz and then the Dirac LFE curve for all curves below 150hz?

I use custom per speaker target curve that uses the same target curve below the transition region for the bass and per speaker curves to correct the nearfield to flat using nearfield and mlp moving mic method measurements.

I do this with Magic Beans True Target.

 

[mmares056]

I always turn off Loudness Management.

Same here.

 

[patoulol]

ggood point on dialnorm. Technical side, it's per title metadata for avg dialogue level, with -31 dBFS Leq(A) as the Dolby ref point (where no attenuation gets applied by the decoder)
most encoders default around -27 dBFS tho, which is prob why ppl notice such different levels title to title when Loudness Mgmt is on


no idea where that gain lands relative to the ART chain (pre vs post correction) tho, never seen dirac document that publicly
anyone gotten a straight answer from their support on this?

 

[Flak]

merci pour tout!

Thank you, RenPa

 

[smygolf]

I must confess, i am back at my old ways, more is better !
So the plan was to pick up two subs, measured a bit and the Mrs Tiguan should be enough for both
Yeah, right
So i will pick up the other one next weekend

2 x Beyma 15LEX1600ND with roughly 308L and tuning at 26hz

Simulation with 600w input (fraction of what they can take), F3 is nearly at the same point as my current subs
2 of these in each rear corner will prob be enough

*edit*
Max output

 

[ASharma]

On the advice of a very knowledgeable forum member I lowered all my curves by 2.5db to minimize boosting for infrasonic bass…Good extension now with +/-3db at around 16hz…

 

[kawauso]

I must confess, i am back at my old ways, more is better !
So the plan was to pick up two subs, measured a bit and the Mrs Tiguan should be enough for both
Yeah, right
So i will pick up the other one next weekend

2 x Beyma 15LEX1600ND with roughly 308L and tuning at 26hz
View attachment 553808



Simulation with 600w input (fraction of what they can take), F3 is nearly at the same point as my current subs
2 of these in each rear corner will prob be enough
View attachment 553809

*edit*
Max output
View attachment 553822

Bigger is better! But there’s absolutely no way two of those are fitting in a Tiguan.