Article: Dirac Live 2 Digital Room Correction Software Walkthrough

[mitchco]

View full article

[mitchco]

Thanks @sem115 One will require an application to "host" the Dirac Live Processor (DLP) plugin. An oldy but goody is VSTHost on Windows. I just tried loading the DLP VST plugin and was successful. There are several VST host's available, some free, some $$'s, some standalone and some integrated with other music player software. The trick is routing the audio.

I ran out of time to try DLP in this configuration. But essentially, you will want to route Qobuz or Roon's output through the application hosting the DLP plugin and then out to the DAC and rest of the system. Perhaps other members here have tried this approach and can assist.

[mitchco]

18 hours ago, digitaldufferme said:

Complete noob here and was fascinated by your article! You have corrected your sound without the subs connected I believe. I have a 2.2 system (KEF LS50 & REL T9i) how would I do this as I believe the subs have a massive effect on the room sound? I use ROON ROCK on a NUK8i7.

 

Yes, I did not have the subs connected, but the the little Purifi's have solid output below 30 Hz. But my Rythmik dual F18's go down to 6 Hz in my room. How is your 2.2 system connected? Is it Y cabling or ?

[mitchco]

10 hours ago, blue2 said:

Thanks for another great article! $349 sounds reasonable but what about a trial/evaluation to find out what it can do for my room+system? Do you know if that's available? I can't find anything on the Dirac web site.

Thanks! From: https://www.dirac.com/faq-general it looks the free trial is 14 days. Perhaps @flak can confirm...

[mitchco]

3 hours ago, monteverdi said:

I wonder why the corrected step response shows much more ringing than the uncorrected. As far as I can see it is at around 20kHz. Is it an artifact of Umik with step transitions (I see similar ringing just measuring my Wideband drivers using Umik) or is it introduced by the DSP?

I understand digital speaker correction which is what basically Dirac is doing at higher frequencies but I can not understand how the interaction of bass frequencies with room modes can be corrected by just altering the woofer output (past a single point in space). So I am looking forward to learn more about active bass management!

 

I did not use a UMIK-1 for this particular article, even though I do have one and have not come across any ringing artefacts with it yet. But I am going to test it out extensively in an upcoming article. As the speakers are also gone, I can't retest easily to confirm what the issue was. I could have goofed on my loopback test and resampled somewhere in the chain like 48 kHz signal through a 44.1 kHz filter. When I get a chance, I will confirm using another speaker/test, but given the amount of time I have spent on trying to identify (pre)ringing artefacts previously, it certainly inaudible to these ears. The main point I was trying to get across was the time alignment.

 

Understanding room correction is complicated A separate article is required to get though it all and I am contemplating writing one. In a nutshell, good room correction software will take an acoustic measurement and extract the minimum phase response and since speakers are minimum phase devices, correcting the frequency response will also correct the phase response. However, that is only half the picture as there is the room to deal with, which is what Dirac calls mixed phase or others call it excess phase, which includes room resonances and reflections.

 

And in the time domain, we know what the ideal speaker target is by looking at the timing (i.e. step) response of an "ideal" minimum phase speaker. So... the room correction software corrects the time domain to follow the ideal minimum phase target. You can see that if you have my book or some of the articles here at AS show that as well. But the trick is that this is wavelength dependant, so at low frequencies, the correction time window is long (like 600ms for example) and as frequency increases the correction time window becomes shorter so above the rooms transition frequency we are looking at more of the direct sound plus baffle of the speaker. Often called frequency dependant windowing. In other words, at low frequencies, the room correction software has a large time window that corrects both the direct sound and reflected sound (i.e. room resonance) towards a target response. At high frequencies, it is mostly the direct sound because if we tried to also correct for mid-range and high frequency comb filtering room reflections, the high frequency response of the loudspeaker sounds like a dentists drill. This is how one can hear "over correction" in the high frequencies.

 

Hope that helps somewhat. Like I say, requires a separate article to fully explain.

[mitchco]

16 hours ago, monteverdi said:

Thanks for your response, Toole and also in a similar version Geddes are proposing to use multiple subwoofers spaced over the room to allow wider zones of bass correction (with specific phase corrections). I wonder if that is something Dirac can do and if that leads to a significant improvement compared to standard 2 woofers/subwoofers? Of course that adds a lot of complexity to the setup both computational and room design wise. REW room simulator shows some benefits but I wonder how real these simulations are?

 

Yes, I have seen both Toole's, plus Todd Welti's presentation on multiple subs, and Earl's paper, plus Duke LeJeune's swarm system. All show positive benefit's but one still needs room correction to obtain the smoothest bass response based on experiments I have run with multiple subs and locations.

 

Even if your sub(s) are part of your two channel system (i.e. not using digital XO), room correction is still a major benefit, which is what Dirac does. However, the new Dirac Bass Management will provide further control, but also requires more than 2 DAC channels.

 

REW Room Simulator works very well and is quite accurate! Room mode calculators are mostly based on the physical dimensions of one's room. Room construction and rooms treatments (to a certain degree as most do little below 100 Hz) have an impact, but below the room's transition frequency, it is all about the room ratio. The reality is that no matter where the speakers/subs are placed in the room relative to the listening positions, there is no escaping room modes. All that happens is that the dips and peaks move in frequency based on these (super)positions. One can get lucky and minimize, so it is worth the effort. However, in the end, still need room eq...

[mitchco]

7 hours ago, Zapuan said:

Thanks mitchco, always very useful your articles and reviews. I have been using Dirac 1.x since 2013 and lately I have been studying Audiolense and Acourate to try to have a (perhaps) better room correction by linearizing the speakers first, and after the room in the sweet spot..
Do you think that a better result can be achieved than Dirac 2.0?

Cheers @Zapuan As mentioned in the review, I did not get a chance to try Dirac 1.x, but according to the manufacturer, there are several improvements upgrading to D2. As I understand it, you are entitled to a free upgrade and given the simplified measurement process, it is worth a shot to try first 🙂

[mitchco]

14 hours ago, digitaldufferme said:

Following the advice of REL, I connect the left speaker and the left sub to one RCA output on the amp and the same for the right channel so that the REL T9i get a full range signal.

 

There is no reason why this would not work with Dirac. When the free trial is available, you might want to give it a try. The only issue is that Roon does not support VST plugin's at this time. So you would need to temporally try it in a different music player that supports VST plugins. Then at least you could ascertain the sonic benefit, which I think would be significant in your setup.

[mitchco]

@ZapuanSorry for misunderstanding. If you know what you are doing, yes, because of additional features and/or exposing functionality for the user to control:

- create multi-way digital crossovers of varying types, slopes, you have complete control.

- linearize individual drivers.

- time align individual drivers.

- user control over frequency dependent windowing. Both low and high frequency window widths can be independently adjusted for both magnitude and excess phase correction.

- the amount of correction applied.

[mitchco]

1 hour ago, Tonygeno said:

A while back you reviewed the Dynaudio Focus 600XD and raved about this speaker. Additionally you discussed its step response and its importance. Would Dirac benefit this type of active speaker and in doing so, would the step response be compromised?

 

Nice speaker, a little bright sounding to my ears if I recall... Yes, they are time aligned. No, Dirac would not compromise the time alignment, if anything optimise further if possible. Yes, Dirac would be a benefit as we still have room modes to deal with. The boundary controls on the speaker are helpful, but rudimentary compared to what Dirac can do. 

 

Here is the 600 XD frequency response in my room. Dirac would indeed smooth out the response below 600 Hz:

 

 

 

 

 

[mitchco]

13 hours ago, ted_b said:

Mitch, great review!  As I contemplate, for the first time in my long audio life, the idea of putting my signal through the twists and turns of DSP in order to correct for room interactions (and I'm building said room as I type this) I have two issues, one long-term that is too OT for this thread (is DSP worth not being able to do DSD512, etc) and one very short term:

* why, if you are working hard to alleviate interactions caused by walls and ceilings, first reflections, etc...why change the room before you start (removing coffee table, sofa, etc) and then change it back? 

Thx

Ted

 

Thanks Ted, appreciate it. That's awesome about your room build! Would love to hear more about that or perhaps it is an article? Wrt objects between the speakers and/or chair/couch in the way... If using a full range correction, we are correcting mainly for room interactions below 600 Hz and the loudspeakers direct sound (and some near reflections like the speaker baffle, stand), so we want to move any objects that are in the direct sound path. Not move them out of the room, but off to the side temporarily during measurements and then replace when finished. So the DSP is correcting for room and loudspeaker, not room/loudspeaker and coffee table, for example.

 

Using this approach yields the best sounding correction as one is dealing with the room and the speaker, with it's natural dispersion pattern, not broken up by objects in the way, or reflections from the couch getting in the microphone if using UMIK-1's little stand to sit on the top of the couch. Myself and others have tried corrections with objects in the way of the direct sound field and/or chair couch inches from the measuring mic and then with the objects moved to the side. The consensus is that with the objects temporarily out of the way for measurement and then returned after, sounds better than the correction with objects in the path and/or mic sitting on the chair or couch. 

 

I have tried this many times, in many rooms, and in every case, the correction sounds better with an unobstructed path to the measurement mic and placing the objects back for listening. Leaving objects in the direct sound path and/or having reflections from the couch back getting into the mic produces an inferior sounding correction. Not so much at low frequencies, but at frequencies above the rooms transition frequency, i.e. 600 Hz and above. It can alter the tone and sound like comb filtering or just unnatural sounding and then folks blame the DSP 🙂 Of course, folks are free to do whatever they want, but myself and others have found this measurement approach works more effectively for achieving the best sounding correction.

 

Best of luck with your room!

Mitch

[mitchco]

It is not so much the direction and cues, which is indeed affected to a certain degree, it is more of the “comb filtering” that colours the frequency response is the main issue. I can’t show it because of copyright issues, but if you happen to have a copy of Bob Katz’s excellent book on Mastering, there are a couple of charts that show the frequency response comb filtering issues when, in his case, a mastering desk is placed between him and the speakers. Bob also adjusted the angle of the desk to minimize the comb filter to almost be the same as without the desk there.

 

When I “critically listen” I move my coffee table out of the way. Aside from the comb filtering colourations, I want to hear the full radiation pattern off the speakers for the best possible imaging. When I place the coffee table back, I can hear the comb filter coloration and the image height is reduced (i.e. the bottom half), by a bit, but since our ears/brains quickly adapt, it is soon forgotten and I am back to background music listening. I also have a “half back” couch as when I had a full back couch, it drove me nuts as I moved further back into the couch, I could easily hear the comb filtering from both channels reflecting off the couch and into my ears, aside from cutting off the ambient sound from behind me.

 

But the most important aspect is that if one corrects for the “comb filtering” off the coffee table for example, then that correction is now embedded in the frequency response of the loudspeaker including it's off axis response. We know from Sean Olive’s and Floyd Toole’s extensive work that a smooth off axis frequency response is just as important as smooth on axis response (think spinorama). So if we correct for the comb filter off the coffee table, we have just coloured the off axis response of the loudspeaker. So now the combination of the direct sound and early reflections in the listening window sounds “coloured” even sitting in the sweet spot and worsens when one moves around the room. This is one of the reasons why folks mistaken that it is the DSP that has coloured the sound when what they have done is coloured the off axis response, so the later reflections after the coffee table are coloured and to our ears does not sound right (and never will!).

 

There are no hard and fast rules, and I encourage folks to experiment. Having been using DSP extensively for a decade, these are some of my experiences.

[mitchco]

@One and a half you need digital loopback capability to route the output of Roon to the VST and then to the DAC. On the PC something like https://www.vb-audio.com/Cable/ or https://vac.muzychenko.net/en/ should work.

 

I see this person got it to work in a basic setup:

https://community.roonlabs.com/t/roon-vst-plugins-hifiberry-digi-pro-possible/73928

There is also a Dirac doc on how to do it as well:

http://diracdocs.com/Windows-Using_Dirac_plugins_with_players_not_supporting_plugins.pdf

 

But, I have no idea about DLNA...

 

Unfortunately, most hardware devices are limited by the processing power required for low frequency control. The lower the frequency, the more "filter taps" required. So most hardware is limited to 8,000 FIR filter taps, whereas on the PC we can easily do 65,536 or even 131,072 taps. https://www.deqx.com/ is about the only hardware solution I would consider, but it is considerably more expensive than Dirac...

[mitchco]

19 hours ago, mwhitak said:

Or Sonarworks, another option.

While Sonarworks is indeed another option, it is not in the same category as Dirac, Acourate or Audiolense, it is amplitude correction only. It does not have any timing correction capabilities.

[mitchco]

Hi @One and a half yes, once you take the measurements, and happy with them, you are good to go, and can put the measurement mic away. Of course if you change speakers, speaker placement, listening position, etc., then you will need to measure again. Now that you have the measurements, you can play with the target response and try a partial correction from 600 Hz on down, or full range correction, different target curves, etc., and easily A/B them in Dirac Live Processor while listening to music. Once you have settled on what sounds best to your ears, it is set and forget and enjoy the music!

[mitchco]

Hi @GSWaul Sorry for the delay. Not sure about your details, I presume the laptop and desktop on the same network? If so, it should be able to see it. You might have to make a few adjustments as described here: https://forum.avast.com/index.php?topic=221321.0 

[mitchco]

@JR4321 

Great question!! I can offer one perspective with some thoughts:

 

1) Room correction is a technically complicated subject area as there are several aspects to it. As alluded to in the article, some aspects are; audio digital signal processing, loudspeaker directivity, Schroeder frequency, standing waves, room resonances, early reflections, late reflections, the ears amplitude and frequency response non-linearity’s to sound, what to correct, what not to correct, etc., is what makes it a technically complicated subject area. I have yet to find a well written “layman’s” article on how room correction works and what the results are supposed to be. I go into gory detail in my book, but it isn’t a summary article. There is such a thing as an ideal loudspeaker in an ideal room that can be modelled and emulated. Unfortunately, most explanations are either too technical or are (wildly) incorrect, which leads me to a 2^nd reason.

 

2) Because of the technical complexity, most folks don’t have the technical skills to understand what they are seeing on a chart and how it correlates to what we hear. So it makes it difficult to know what the end result should be or sound like, and most importantly, how to obtain it. There have been several scientific studies that correlates what makes for a good sounding loudspeaker in a room with objective measurements. Floyd Toole and most notably, Sean Olive have undertaken several scientific studies in this area correlating peoples listening preferences with objective measurements with loudspeakers in rooms and room correction systems. I have linked to a few studies in this article, plus other room correction articles I have written. In fact, there are literally dozens of repeatable scientific studies in this area, but again requires some level of technical acumen to spend the time reading and understanding.

 

3) This ain’t your granddaddies 31 band eq 🙂 Modern Digital Signal Processing (DSP) software is very sophisticated and coupled with powerful computers one can alter the frequency and timing response of a loudspeaker in a room to just about anything you want. Put another way, not all DSP room correction systems/techniques are the same and some are orders of magnitude better than others. Some work in the time domain, others do not. Some offer tailoring of every possible parameter, others do not. The list goes on. But perhaps the most important point is that most folks don't know how powerful software room correction has become.

 

4) And the elephant in the room, double pun, is that loudspeaker manufacturers (mostly) do not want folks to use room eq with their loudspeakers for a whole pile of reasons (e.g. don’t mess with my voicing, mics aren’t ears, yet those are used in the design of the loudspeaker in the first place, and a laundry list of other myths built up over the years - some justified with very poor eq systems from yesteryear). The biggest reason being that if one were to use state of the art DSP room correction system on Speaker A and then use the same target frequency response for a similar Speaker B and then A/B them (which is difficult to do) one may be hard pressed to tell the two speakers apart. Kudos to some speaker manufacturers, who have the confidence in their products to actually make a statement like that. For example, Martin Mensink, designer of the Dutch and Dutch 8c, “I've had the Kii's and the 8c's side by side in my living room for a while. The Kii's too are remarkably good speakers. With just some subtle EQ the two could be made to sound very similar on most program material - to the extent that I might not be able to distinguish them in a proper blind test. I'm still amazed sometimes by the extent to which differences in sound can be explained by frequency response.” Having performed the same experiment above myself with the speakers mentioned and others, there is considerable truth to this.

 

The irony is that for around $500, which includes the calibrated measurement microphone and state of the art room correction software, will make the biggest sonic optimization/improvement that one can make to your existing sound system that is both audible and measurable for the $’s spent. The caveat is one must know what one is doing to achieve a successful result. Companies like Dirac are making it easier to obtain a good result by limiting the amount of variables the user can play with. But as I mentioned in the article, that in itself is also a trade-off.

 

At this stage in the progression of software based room correction adoption, most folks that have experienced good room correction, even just for levelling out the bass frequencies, would not go without it. The audible difference in smooth sounding bass versus uneven bass response, virtually everyone can hear the difference as the difference is considerable. We are talking going from +20 dB peak to peak ripple response in the low end to +- 3 dB envelope. Everybody that has a pair of ears can hear the difference 😉 Again, part of this is the education to know that below the room’s transition frequency into standing waves/room modes, is to know that the room is in control of the bass response, not the loudspeaker. It takes a bit of understanding/time to wrap ones head around that.

 

Hope some of that is useful. I have considered writing a layman’s article on the subject area to help folks understand what is being measured, what is being corrected and why. Maybe it is time...

 

Have a great weekend!

Mitch

[mitchco]

Hello @cpdk welcome to Audiophile Style and thanks for your comment.

 

Yes, the answer is in the 5th paragraph of the conclusion. 

 

Kind regards,

Mitch

[mitchco]

Hi @hulkss good to hear from you! Thanks for your comments. I have used MSO as well. Excellent software!  Would be curious to see the corresponding phase response that goes with the frequency response plots in your MSO post...

Hope you are well.

Kind regards,

Mitch

[mitchco]

Bill, great to hear from you! I hope all is well.

Dirac suggests: http://diracdocs.com/Roon_&_Dirac_Live_2.0.pdf to work with Roon. 

I don't know how your other sources are hooked up, but the idea is that the input signal would be routed through Audio Hijack to Dirac Live Processor and then out to your LS50W.

Don't have a Mac or LS50W, so could not say for say for sure...

Hope that helps.

Mitch

 

[mitchco]

Bill,

 

The miniDSP DDRC-24 should work no doubt. I have never used one, but I know it has a limited amount of filter taps as compared to the processing on a PC. Less taps generally means less low frequency resolution. However, I can't speak about how Dirac 2 is implemented in the miniDSP unit and may or may not be an issue. Perhaps @flak can comment or a post on the miniDSP forum or maybe another member has experience with this combo and can comment...

 

Wrt 2nd question, hard to say without seeing any before and after graphs. It could be room modes, but could also be the target is a bit too bright that makes the bass sound thin. If you could post some charts that would assist. Also, feel free to add a bit of bass boost like in this target from the miniDSP Dirac 2 user manual: https://www.minidsp.com/images/documents/miniDSP Dirac Live 2.0 User Manual.pdf  See page 36.

 

Kind regards,

Mitch 

[mitchco]

Hi Bill, Dirac I believe is cut only filters, so if you lower the peaks to be even with the majority and not worry about the narrow dips. Narrows dips we don't hear.

 

Sure. just use REW's default settings and sweep 10 Hz to 24 kHz using a 48 kHz sample rate of the left and right speakers at the listening position. Just PM me the .mdat or send to [email protected] and once we sort it out, you can post the results here if you wish.

 

Mitch

 

 

[mitchco]

I agree with @Kal Rubinson. My concern for any hardware DSP is the limited processing power available. Typically they are limited to 4096 FIR filter taps, which is not enough for low frequency control. For example, the miniDSP 2x4 HD in the datasheet talks about 4096 taps. But this is the total number. For 4 channels you have 1024 taps each channel available. The frequency resolution of a 1024 taps filter @ 48 kHz samplerate is 48000/1024 = 46.875 Hz. So below 100 Hz there are just 2 frequency bins at 46.875 and 93.75 Hz. This clearly means that you have no control over the lower frequency range.

 

I don't know how Dirac is implemented in these h/w devices, but on a PC you typically use 65,536 taps per channel which gives full low frequency control. Remember the primary purpose of DRC is to correct the low frequencies below Schroeder. So my review of Dirac 2 is only in the context of using Dirac on a PC. I make no claims on how it works/effectiveness with miniDSP gear or any other hardware.

[mitchco]

Hi @Tp no need to remove the carpet or other objects on walls, etc. What we are trying to avoid is anything in the direct path between the speakers and microphone like plants coffee table, etc. Also, anything around the microphone can cause odd sounding corrections, hence moving the chair or sofa from around the mic.

 

Hi @mlknez Do you mean DiracLiveProcessor.dll? On my Win10 computer it is located at: C:\Program Files\Common Files\VST2

[mitchco]

Cheers @Andyman

 

Wrt questions about Audirvana, AU plugin and sample rates, are product questions that perhaps Flavio @flak can answer as I don't have a Mac.

 

While one could use Dirac to measure the correction result, it is typical to use a 3rd party application for verification measurements. In my case, I have been using REW for +ten years and as a general acoustic measurement software, it has several additional acoustical analysis displays that Dirac does not have. For example, one typically uses a Step response as one of the verification measurements to view the timing response of the loudspeaker, which I show in the article. Also REW allows us to share measurements in a portable format and compare measurements in overlays.

 

I usually put "ideally" in front of moving objects out of the path between the speakers the measurement mic. Of course, it may not be possible. The issue with the couch depends on how close the back of the couch is to the microphone and how reflective the couch surface is. For example, if the mic is 6" or 12" away from a reflective surface is right in the midrange and upper midrange frequency range, where our ears are the most sensitive. Further, because of the close proximity and how reflective the couch is, the reflected sound may be as high in amplitude as the direct sound. Producing a correction in this situation will not sound very good. If there is no choice, try and throw thick wool blankets over the couch area that the mic is moving around which will limit the comb filtering in the analysis and produce a much better sounding correction filter.  I encourage folks to try it for themselves and hear what I am talking about.

 

Re: B&W floorstanders. Take a measurement with the bungs in, then take another measurement with the bungs out Which one produces the flattest response? Go with that. As a general guide, we want to try and get the best sound possible before room correction. It is not so much about less work for the room correction software but the better the setup before hand the better the correction, meaning better sound quality.

 

Thanks Andy, I am looking forward to putting Dirac Live Bass Management through its paces!

 

Kind regards,

Mitch