Active Room Treatment (ART) by Dirac

[Miniyouuuu]

And 60dB scale

Here you have the comparison 12-200hz 60db
No EQ no subs, just 2.0 (Mofi SP10 ME)

And 2.2 (Mofi + 2xSB2000) with EQ:

 

[Oddball]

This discussion goes back to something I commented on a while ago which is how much headroom do we need to implement Dirac Art? In my particular setup my channels all support each other (5 channel system), none of the support filters are above -10db (and this is looking at the filter graph, not the 'support level setting'). If I'm looking for a reference level of 105 db per channel that means all the other channels must play 95 db in support. If my math is right each channel must play 105 + 95 + 95 + 95 + 95 db in a worst case scenario where each channel is maxed out. 105 + 95 + 95 + 95 + 95 = 106.4 db. So all we have to know is if we have that amount of headroom at f3 or f6 or f10 or whatever and we can safely set our low range support setting there. right?

Not really following the math but it’s also not that simple. Some speakers might be playing in support of the signal, and some will be playing cancelling signal. So will really depend the specific set of filters.

We discussed the support range settings recently but if looking at listening at reference levels you might go a bit higher than measured F3 and reduce support intensity as well.

I was testing first at -10dB master volume, then -5dB, then reference and for good measure +5dB. These are all pretty high SPLs so would be careful and adjust down support range and intensity if you don’t clearly pass every gate. I had lots of experience with Audy that can be even more demanding with LFE plus main setting and LFE distribution so pretty confident in the system.

 

[speakers4me]

Not really following the math but it’s also not that simple. Some speakers might be playing in support of the signal, and some will be playing cancelling signal. So will really depend the specific set of filters.

We discussed the support range settings recently but if looking at listening at reference levels you might go a bit higher than measured F3 and reduce support intensity as well.

I was testing first at -10dB master volume, then -5dB, then reference and for good measure +5dB. These are all pretty high SPLs so would be careful and adjust down support range and intensity if you don’t clearly pass every gate. I had lots of experience with Audy that can be even more demanding with LFE plus main setting and LFE distribution so pretty confident in the system.

I could be wrong but I think the same signal that is sent to the speaker being supported is also sent to the support speaker, a filter is applied to both speakers (and those filters are shown in Dirac art), and phase shifts are used as necessary to do any cancellation that are needed. BUT I’m pretty sure the support filters shown tell you the magnitude of the support speaker relative to the supported speaker. I think you can use math to estimate kind of a worst case scenario where all speakers are at full blast and supporting each other and get a good idea of the headroom you’d need. That math in the first post is just from a db calculator online, it may look funny but that’s just because db’s are a logarithmic scale.

 

[Oddball]

I could be wrong but I think the same signal that is sent to the speaker being supported is also sent to the support speaker, a filter is applied to both speakers (and those filters are shown in Dirac art), and phase shifts are used as necessary to do any cancellation that are needed. BUT I’m pretty sure the support filters shown tell you the magnitude of the support speaker relative to the supported speaker. I think you can use math to estimate kind of a worst case scenario where all speakers are at full blast and supporting each other and get a good idea of the headroom you’d need. That math in the first post is just from a db calculator online, it may look funny but that’s just because db’s are a logarithmic scale.

I don't think support speakers get the same signal as the supported speaker. They get to play supporting role that as noted can be filling or canceling and will be different SPL across the supported range. At least my understanding.

Are you designing a system so want to make sure you hit the reference?

 

[speakers4me]

I don't think support speakers get the same signal as the supported speaker. They get to play supporting role that as noted can be filling or canceling and will be different SPL across the supported range. At least my understanding.

Are you designing a system so want to make sure you hit the reference?

No but I push my system pretty hard and I just wanted some clue as to headroom requirements, I’ve done some testing where I measure a supported speaker with all others muted, then measure the support speaker with all others muted just to get an idea of how hard the support speakers are getting pushed. In my system the support level is not very high relative to the supported speaker (and I’m talking about actual volume in the room, not the support setting in Dirac). I have a suspicion that the headroom requirements of implementing ART might only be 2-3 db’s in many cases. (If you blow your speakers don’t blame me). If anyone else has input on this I’d be interested, for those of us that listen loud this is pretty important factor and I’m surprised there isn’t more concrete info on headroom requirements.

 

[Oddball]

No but I push my system pretty hard and I just wanted some clue as to headroom requirements, I’ve done some testing where I measure a supported speaker with all others muted, then measure the support speaker with all others muted just to get an idea of how hard the support speakers are getting pushed. In my system the support level is not very high relative to the supported speaker (and I’m talking about actual volume in the room, not the support setting in Dirac). I have a suspicion that the headroom requirements of implementing ART might only be 2-3 db’s in many cases. (If you blow your speakers don’t blame me). If anyone else has input on this I’d be interested, for those of us that listen loud this is pretty important factor and I’m surprised there isn’t more concrete info on headroom requirements.

I see now. Agreed that one would not expect support speakers (except subs) to work very hard in that role, especially above 80hz or 100hz as that might start to be audible at high SPL. So would not expect that you need significant headroom for support role.

Don’t think many here go up to reference level so purely lack of interest. Perhaps on AVS forum there will be more. I went step by step to test as have more confidence in this method and my speakers survived worse.

 

[dlaloum]

Because I'm going to do timing, phase and fr up to 300 on genelec GLM.

If you haven't followed, I'm attempting to keep the incredible benefits of GLM extended phase linearity and combine with DiracArt bass management.

Look forward to hearing how you go...

 

[dlaloum]

I may derrail a bit the conversation, but I might as well ask and get straight answers from people who have direct hands-on experience on ART over here.

I have red several times that the closer to full range the speakers are, the better ART will work on the whole system. Is that statement true? On the other hand, is it possible to "cheat" and get a subwoofer near the non-full range channels (sides and rears) to get a similar performance capacity and thus, get similar results?

Basically the more speakers you have capable of support within the ART frequency range (20-150Hz) - the better the results (within the limits of the number of filters your processor can handle)

So each "standard" THX speaker limited to 80Hz - will only help with the 80-150Hz range, whereas a full range speaker, capable of 40Hz to 150Hz - will provide support right through almost all of the bass range.

If you have 6 speakers supporting the bass region - you can achieve remarkable results (as per my results using 4 full range speakers and 2 subs)

Dropping off 1 sub, and using 4 full range +1 sub achieved almost identical results.

You want a sub because full range speakers in most cases, won't extend below 30Hz (and often not below 40Hz) - you can see the rise in decay below 35Hz... this is due to (I theorise!) my 2 subs being "alone" below that - whereas above that I have 5 or 6 contributing speakers.

So from an ART decay control perspective, I theorise that you need more than 2 speakers for any part of the frequency range, but definitely less than 6...

Those who have tried straight stereo with full range speakers - have found ART beneficial too... but it seems like you get 60% of the benefit of ART with 2 speakers, probably 85% with 3 speakers (2 full range + sub?) and 99% by the time you reach 5 speakers (in my case 4 full range + 2 subs)

Definitely diminishing returns with additional speakers... and lots more experience from different people is needed before we can with confidence say where the optimal "sweet spots" are.

If you have 4 subs + 5 speakers capable of 80Hz to 150Hz - your results will probably be just as good as with 2 subs and 4 full range speakers.

With the proviso, that I have not tested SPL extremes... my Subs are the 10" sealed Gallo TR1's... circa 20 years old, very musical, focused on music rather than digging deep - various specs say their F3 is 22Hz or 24Hz - in my setup, I have FSL at 28Hz - but they are not a design known for its loudness!
My mains are Gallo Ref 3.2 - with the exact same woofer as the TR1 sub - but a smaller sealed volume leads to slightly more restrained F3 (spec says 24Hz) - I have FSL at 30Hz
My Surrounds are B&O Beovox Penta with F3 35Hz - I have set FSL at 40Hz

My sub positioning is 1 at the front right (next to R) 1 at rear L (next to SL)

Ideally, from a surround positioning perspective, the LFE should be front and center... but the realities of domesticity tend to be greater deciders in terms of sub positioning.... just as WAF has a substantive impact on the type and size of speakers "permitted".... The 45 year old vintage B&O Beovox Penta, is a very very good option from a WAF perspective! (and happens to be a damn good performer too)

 

[dlaloum]

@Vacceo I don't think it is possible to "cheat" and get a subwoofer near the non-full range channels to get a similar performance as towers. Subs just work different in ART and directional bass is achieved on different basis.

Although.... many (most?) subs come with inbuilt crossover options, so they can be used in straight analogue systems, if you use that option, you can take 2 subs, use their internal XO to mate them to the main speakers, turning them into "full range" speakers. (and with external subs they might be higher performance full range than most!)

If you did that for the 4 primary channels (L/R/SL/SR) then set up your decoder/mixer without a sub (so it mixes LFE back into the mains) - you could absolutely run it as a "full range speakers only" rig. I would think ART would handle that just fine, and you would have full directional bass as a bonus.

 

[dlaloum]

Start at the 1 hour and 2 minute mark.

At 46 minutes in there is talk of some things brought up recently in this thread as well.

The talk about maximum SPL's and distortion artifacts brings up an interesting thing that I observed...

When measuring and configuring at my preferred listening level (75db @ MLP) - I noticed in the REW measurements that the THD for the main supported channels (L & R) was LOWER using ART (in the ART frequencies) than it was at the same SPL running straight through.

To me that implies that the speaker is working less hard, due to support from other speakers - so everyone is cruising along at lower THD levels for the same SPL...

It also implies that the maximum SPL levels - the levels at which THD will really kick in audibly, are probably higher in than they were without ART (at least in my configuration with full range mains)

So contrary to many others discussions, (and probably to some degree system dependent), the ART based setup may have higher dynamic headroom (max SPL) than the more basic Dirac Live or unprocessed system....

It might be useful if those taking REW measurements also looked at the THD for their subs and full range mains, to see what impact ART has on the THD in their configuration!

 

[dlaloum]

This discussion goes back to something I commented on a while ago which is how much headroom do we need to implement Dirac Art? In my particular setup my channels all support each other (5 channel system), none of the support filters are above -10db (and this is looking at the filter graph, not the 'support level setting'). If I'm looking for a reference level of 105 db per channel that means all the other channels must play 95 db in support. If my math is right each channel must play 105 + 95 + 95 + 95 + 95 db in a worst case scenario where each channel is maxed out. 105 + 95 + 95 + 95 + 95 = 106.4 db. So all we have to know is if we have that amount of headroom at f3 or f6 or f10 or whatever and we can safely set our low range support setting there. right?

See my earlier posting about SPL and THD.... #3410

With all the speakers supporting one another we have a situation which may potentially be similar to the "peak power" measurements for amplifiers...

What I mean by that is, if you try and drive everything simultaneously to its spec maximum - your system may well get into trouble.

But in actual use, that does not happen, and you may end up getting much cleaner sound, with peak capabilities higher than the "spec" value, but only for individual (or a couple) of speakers at a time... as in reality when L is driven to 105db - it will actually be producing perhaps 100db and 5db of the end result would be from the multiple support speakers assisting... - but if they themselve are being driven to the max, and then adding the assistance on top of that, then the whole structure falls appart.

So we may find that with the ART setup, our 105db max spl mains are now capable of 110db SPL - individually.... but if all channels are being driven simultaneously they might each be only capable of 100db... (Hence my comparing it to amp peak power measurements...)

This would also imply a seperation between "real life" behaviour, and "lab testing" behaviour - and the relevant lab tests might need to be chosen judiciously!

Isn't ART/Unison interesting!

 

[dlaloum]

Not really following the math but it’s also not that simple. Some speakers might be playing in support of the signal, and some will be playing cancelling signal. So will really depend the specific set of filters.

I wonder about that... I am not sure that Dirac ART actually tries to do any cancellation... (although I used to think that cancellation was involved... )
It seems to me that the magic is achieved by summing of the differently positioned speakers, and using standard EQ filtering to control peaks generated by individual speaker positions.

 

[Frank207be]

I tried -24dB, but as @middlemarch says, could not really hear any difference from -18dB even at reference level. It also measured pretty consistently with -18dB. My best guess is that this is room/system dependant, so might not work quite the same for all.

This morning I had the home for myself so I uploaded a medium boost curve with -18dB and one with -24dB support and the last one sounded drier especially with live recordings.

So I rearranged my ART slots going from a low boost curve with -24dB support on 1 going to a medium boost curve with -18dB support on 2 and a high boost curve with -12dB support on 3. Curious how that last "wet" one sounds with high octane movies.

 

[Oddball]

I wonder about that... I am not sure that Dirac ART actually tries to do any cancellation... (although I used to think that cancellation was involved... )
It seems to me that the magic is achieved by summing of the differently positioned speakers, and using standard EQ filtering to control peaks generated by individual speaker positions.

If you watch the video posted previously, Joss is talking about cancellations. The only comparable solution is Trinnov wave-forming which also works on cancelation principle as advanced back-front bass array.

 

[smygolf]

Here you have the comparison 12-200hz 60db
No EQ no subs, just 2.0 (Mofi SP10 ME)
View attachment 507484

And 2.2 (Mofi + 2xSB2000) with EQ:

View attachment 507483

Almost right, lower end @ 40 - 45Db and higher end @100 - 110dB reason is the noice floor is around 40-45dB
With correct scale it is much easier to see changes and faults.

Ex

 

[TimoJ]

Joss is talking about cancellations.

Is he? Approx. at what time of the video? I only remember him talking about how ART collects a puzzle of sounds from different speakers. Also those graphics that Dirac ART website earlier had, that showed cancellation waves, seem to have disappeared.

 

[Oddball]

Is he? Approx. at what time of the video? I only remember him talking about how ART collects a puzzle of sounds from different speakers. Also those graphics that Dirac ART website earlier had, that showed cancellation waves, seem to have disappeared.

Don't have time to re-watch but from recollection in last 30 min of the video.

EDIT: Additional tip that was very useful was that subs should be at front and back for max effect (decay reduction) and that front only will not be so effective. Again probably in the last 30 min of the video.

 

[JRowland]

EDIT: Additional tip that was very useful was that subs should be at front and back for max effect (decay reduction) and that front only will not be so effective.

As far as the rear sub that is only if your back wall is waaay behind the listening position. At 1:17 he speaks on this and @TimoJ he uses the wording of a speaker being able to reduce rt60 and also if a speaker is "only additive". By separating those two things it seems to me there is something else going on to reduce RT60 although he did not use the term cancellation.

 

[Oddball]

As far as the rear sub that is only if your back wall is waaay behind the listening position. At 1:17 he speaks on this and @TimoJ he uses the wording of a speaker being able to reduce rt60 and also if a speaker is "only additive". By separating those two things it seems to me there is something else going on to reduce RT60 although he did not use the term cancellation.

Your MLP should always be away the wall, so this is quite normal. You can have closefield subs behind, will work well - have it in my setup. Think it was clear enough that this positioning is prefered to loading the front with subs.

Sometimes you have to read between the lines. If you guys doubt there is cancelation, I don't, but we are all allowed to our own speculations and each of them is probably worth a cent on a Euro anyway.

I continue to stand on my position that tweaks in ART result in very little difference in measurements in my room and system. I decide by ear as all of them measure decently. Not saying that is the case for all. So if people are looking for universal tips, that is probably difficult to provide.

What I feel strong about is that bookshelves with F3 of 60hz in the room will do just fine with 4 subs - 2 front and 2 back. What you want out of your subs is different. Those 4 could be SVS 1000 Pro or SVS super ultra, just to stay within the same brand that has many models available. You could probably pull it out with 2, but then since I have 4 can't really comment on it.

 

[JRowland]

Think it was clear enough that this positioning is prefered to loading the front with subs.

Sometimes you have to read between the lines.

Respectfully I do not need to read between any lines. To quote Joss (1:17) "if you've got a big open space and it goes far behind you..." That is different than some space between the mlp and rear wall. I'm not arguing about if subs behind the mlp help/hurt/make no difference. I am just pointing out that your post of "EDIT: Additional tip that was very useful was that subs should be at front and back for max effect (decay reduction) and that front only will not be so effective. Again probably in the last 30 min of the video" is not what Joss stated in the video you referenced. If the way you worded it was indeed the tip he was giving I wouldn't be typing.