Active Room Treatment (ART) by Dirac

[fzst]

Will be interesting to see where i will end up with my subs.

View attachment 542165


From Data-bass, a baffle wall do Wonders for these subs.
View attachment 542167


*edit*
Spelling

Bass management stays the same for the rest of the speakers and I suppose the LFE channel is left to reproduce whatever content is within that channel to whatever point the sub extends.

Could you instead just use the LPF on the sub itself to prevent it from reaching too high (should there even be an LFE-signal that produces significant output high up)?

 

[KenMasters]

Could you instead just use the LPF on the sub itself to prevent it from reaching too high (should there even be an LFE-signal that produces significant output high up)?

That's exactly what I will do if I find it to be a problem.

It would be a pain since I'd have to redo the measurements, but I don't think it will be an issue, not based on prior experience with an LFE LPF of 200Hz.

@smygolf's system though - that is uncharted territory!

 

[smygolf]

Correction limit at 1Khz where the subs start to fall, measures the same in REW
Must say there is a bit more power, slam or what you call it compared low pass on

I have dsp possibility on every ch in the system so to add a low pass above 250hz is done in 2 sec

I dont have the time to do propper measurements until the weekend but will try with/without
Can not se/hear anything strange or bad atm

 

[smygolf]

Could you instead just use the LPF on the sub itself to prevent it from reaching too high (should there even be an LFE-signal that produces significant output high up)?

LFE limit is around 200-250hz ?
More concerned about hizzing sound due to the high sensitivity
Will check at ref lvl later today

 

[mmares056]

@mmares056 if you see this first take a screen shot of your current filter design screen before you upload the beta.

Sorry I was so excited to try this new feature in beta v3.15 that before I read your message I had already uploaded the beta. Anyways, after switching to LFE LFP "disabled" it was audible to me in a positive way instantly. I had a chance to watch a few scenes from BR 2049, Tron Ares, and Dolby Atmos Demo trailers like Amaze and Horizon. I will say that Dirac did a good job adding this option and I concur with those that have posted their first impressions of this update on the AVSF Dirac Live and ASR Dirac ART threads. I instantly noticed a difference in bass intensity especially in the higher bass region. The bass is more pronounced and it could just be me but it sounds more refined too. So my first impression of the LFE LPF update in beta 3.15 is a positive one. And no sub localization which is always a concern with higher LFE LPF.

I have yet to measure with REW to validate impulse response, phase, and GD. Will probably do this If I get the house to myself this weekend.

For comparison sake, this is the predicted/corrected response of LR/+LFE with LFE LPF Enabled and Disabled.

And this is the Impulse Response comparison.

 

[RenPa]

https://www.audiosciencereview.com/forum/index.php?attachments/ir-lfe-lpf-enabled-jpg.542221/

I am releaved to see this image. The doted lins of subs is at the left side as well in Cinema 30 LFE desabled.
Now I am confused with the following. I thought that sub 1/LFE in Dirac plays only LFE signal, isn't it?.
I measured in REW with LFE channel and
a- sub 2 ( my left sub) to OFF = YELLOW is sub1/lfe alone isn't it?
b- sub1/lfe (my right sub) to OFF = RED is the sub2 playing lfe signal from lfe channel in REW?
Is it expwcted or what I don't understand?

 

[Joaquinín]

View attachment 542121
View attachment 542122

From that post in AVSForum that you linked: red = LPF enabled, green = disabled.

Look at that timing curve and see how much post-ringing there is in the green curve. The red curve looks so much cleaner. Yes, it fixes the 5ms delay, but a 5ms delay is trivial to fix. Post-ringing is much more difficult.

Removing a filter causes ringing in this instance? Is it bad, then?

 

[EngineerNate]

I can’t comment on where the ringing is coming from but I’d guess it’s more “the filter was masking the ringing and you can see/hear it with it turned off”.

 

[azzy_mazzy]

there is a tool made by avsforum member that allows blind comparisons between profiles (like dirac profiles), it needs HT web remote and seems to be rather "involved" method but it might be of use https://www.avsforum.com/posts/64614782/

 

[Keith_W]

Removing a filter causes ringing in this instance? Is it bad, then?

Ringing is bad, yes. But it depends on how loud it is, and how long it goes for. Post-ringing is masked by the signal, and its audibility depends on program material. For example, if you listen to continuous tones (e.g. violin, flute, etc) it will be somewhere between inaudible to subtle. But if you listen to music with a lot of transients (the worst case would be large drums with no accompaniment) it would be more obvious. Having said that, unless the DSP has done a horrible job, most ringing is contributed by the room. And sometimes by speakers, drivers, and ports.

I generally try to avoid as much ringing as I can, and I don't care if it's audible or not - I don't want to see it at all. The ultimate way to make sure something is inaudible is if it's not there in the first place

I am guessing that the ringing is coming from higher up in the frequency range and the LPF simply removes it. As for why there is so much ringing - could be a lot of reasons. It's most likely some kind of high Q DSP. Without knowing exactly what that guy (or Dirac) has done, it is impossible to say.

 

[mmares056]

I ran single static sweeps, 80dB SPL, 256k long, using acoustic timing reference and this is what the actual response looks like with LPF enabled and disabled. @Keith_W do you see anything on these plots that you think could negatively impact bass quality? So far I like how the bass sounds with the LPF disabled.

Impulse

Step

LFE Actual Response, Bass target curve applied

 

[OCA]

A 120Hz Linkwitz-Riley 4th order (24dB/octave) min phase filter (identical to what the AVR does) will shift the IR peak about 4ms and that's normal. See REW default result below:

Since LFE is a separate track and already lowpassed at 120Hz at source (lossy Atmos - 180Hz otherwise), it's ok to disable it if that 4ms delay is causing extra peaks when played with the rest of the speakers. You can also dim down its +10dB default boost setting (again Dolby spec).

 

[Keith_W]

I ran single static sweeps, 80dB SPL, 256k long, using acoustic timing reference and this is what the actual response looks like with LPF enabled and disabled. @Keith_W do you see anything on these plots that you think could negatively impact bass quality? So far I like how the bass sounds with the LPF disabled.

Firstly you posted rather zoomed out screen caps rather than MDAT's. Assuming that the main speaker impulse is at 0ms, the part I highlighted is the subwoofer / bass impulse starting about 25ms before the main impulse. I see the same thing in EVERY ART measurement I have examined. The first time I saw it, I thought it was a user mistake, but given that all ART measurements show the same thing, I can only assume it's a deliberate choice by Dirac.

The big question is "is it audible". According to theory, yes it should be. It is greater than the 20ms psychoacoustic pre-masking threshold. I have to admit that my knowledge of pre-masking isn't as thorough as it should be. I am aware that pre-masking is frequency-dependent, and I don't know what the threshold is for low frequencies. Whether you can hear it also depends on program material. It is less audible if you listen to mostly smooth tones (violin, flutes) but should be more obvious if you are listening to music with a lot of transients, especially full range transients with silence preceding the transient. Like a Japanese Taiko drum, or cannon fire, etc.

I haven't heard many complaints from Dirac ART users about this pre-ringing artefact that ART introduces, so either everyone has collectively failed to notice, or it is indeed inaudible and I need to upgrade my knowledge. I think the latter is more likely. I am not going to tell you that it is audible when it may not be, and I am not sure.

FYI this is what an ideal step response should look like. Blue is a simulation, red is an actual measurement from my system. Silence before the main impulse, followed by a sharp and steep main impulse, a "compact" head, and decaying/minimal post-ringing. You can make the simulation with REW.

 

[RenPa]

Assuming that the main speaker impulse is at 0ms, the part I highlighted is the subwoofer / bass impulse starting about 25ms before the main impulse. I see the same thing in EVERY ART measurement I have examined. The first time I saw it, I thought it was a user mistake, but given that all ART measurements show the same thing, I can only assume it's a deliberate choice by Dirac.

Is this could help your analysis?
In RED is the step of the front left channel AND the step of only the sub1/lfe channel with the new LPF/LFE 120Hz v3.15 desabled. In pale pink the front left channel step with the step of the only sub1/lfe channel with 120Hz LFE enabled as in v3.14.3
RED lfe channel and front left channel are together.
Edited wooups!

And phase

 

[kawauso]

So, the PC version of Dirac Live ART is finally here.

For a stereo + subwoofer setup, if you do not need multichannel support, the license appears to be $799.
One license can be used on two PCs at the same time.

 

[smygolf]

So, the PC version of Dirac Live ART is finally here.

For a stereo + subwoofer setup, if you do not need multichannel support, the license appears to be $799.
One license can be used on two PCs at the same time.
View attachment 542620View attachment 542621View attachment 542622

More interesting is beta 3,15 that removes the 5ms delay

 

[welwynnick]

And phase

That's fascinating. My ideal target for phase linearity is roughly +/-50 deg above 200Hz (possibly 100Hz), and it looks like you're getting better than that.

Do you know the before & after phase response across the whole frequency range?

 

[TheKoC]

Now that ART is coming to the PC/Mac, can you experienced users recommend some Tutorials / FAQs / Best Practices videos or documents to start out with?
Thanks!

To answer my own question , I found the Youtube Videos of Joss from Dirac with JoeNTell and Marantz/Sound United quite usefull (they overlap on most topics).

 

[kawauso]

More interesting is beta 3,15 that removes the 5ms delay

Yes, it seems so. I’m not sure I’ll be able to notice the difference myself, but I’ll give it a try later.

Does this difference only show up with sources that contain an LFE signal?

 

[RenPa]

That's fascinating. My ideal target for phase linearity is roughly +/-50 deg above 200Hz (possibly 100Hz), and it looks like you're getting better than that.

Do you know the before & after phase response across the whole frequency range?

Unfortunatly no, I forgot to change REW settings for the lfe channel measurement so it stops at 300Hz. I show together with the same color code the front left and the LFE enable- desable and in grey the lfe channel measurement. We see that LPF/LFE desable made a vertical rotation at 293Hz instead of more «erratic»? shift with LPF/LFE enable