The Official Dirac Live Thread: Start w/1,2,3 🙋‍♂️

[marcl]

Youthman did an hour and 42 minutes LIVE today with Matthias from Dirac.  I haven't watched all of it but there was some interesting stuff sprinkled through it.




One thing he mentioned - I knew about it but hadn't heard the details - was the idea of using Exclusive Mode in Windows while doing measurements.  he said sometimes processes running in Windows could interfere with measurements and cause lost or inaccurate data.  So I was just about to do a calibration (after moving my surround speakers a bit) so I did what he said ... enabled Exclusive Mode and shut down everything in my PC except Dirac.


Well, I can't say that's what made the difference for sure ... but here are my results measured at the MLP (exact location of my initial Dirac measurement) after doing the 6-point "Domino" measurement pattern.  This is with no level tweaking, and with Variable Smoothing (not Psychoacoustic).

[marcl]

As my friends constantly bust my chops about how much measuring and Dirac calibrating I do ... they seem certain that at any hour of the day or night they will catch me sucking sweeps through my UMIK-1 ... I respond that there are two kinds of people in this world: those who will do a Dirac calibration only once if they can get away with it, and dread every second ... and those of us who will get an idea and squeeze one in while their wife is at the grocery store.

So how long DOES it take? How quiet does it HAVE to be? Yesterday I had a quiet house and was able to turn off HVAC and hadn't opened the refrigerator in a while so the compressor was off. I measured noise floor at the MLP 38dbC and 30dbA ... and that is REALLY quiet and Dirac will certainly work fine at ambient levels 10db higher than that!

My current process is to do the 6-point "domino" pattern around the MLP, and measure that MLP again at the end so the mic stays there for post-cal verification. It took me:

35min - Volume Calibration and 7 measurements (x7.1.4 ... 91 sweeps)
5min - Filter Design and save filters
15min - Quick 7.1 verification with REW
20min - Precise level adjustment (to <1db) using Spears & Munsil disk Panning noise - separate adjustments for Preset 1 and Preset 2
5min - Full 7.1.4 measurements using Spatial Audio Toolkit sweeps with REW - Preset 1 only

Note that Preset 1 and Preset 2 have the exact same filters.  The only difference is Preset 1 has center sub=LFE, and Preset 2 has center sub=None (changed after filters are loaded).

[kcdc]

First post, Have had a RMC1 since last Oct or so... Snaps and pops aside, it's been OK if I ignore the rando freezing, menu issues (android app that the generous fellow made does fix a lot of that).. I did do the 100ms timing delay "bandaid" but wasn't sufficient enough at times, still pops and cracks during handshakes or input switchings, and regardless of what's been noted that it doesn't affect lip-syncing, it does. I did apply the 100ms delay to every possible channel as suggested to get past that issue.. I'm in my 40s, but do have a profession in VFX for the past 25 years, so maybe I am just too sensitive to the slight adjustment... a curse, trust me.. to the difference in sound to screen differences... Those are minimal issues at this point. 90% pleased with the purchase since I am used to nothing being perfect.

I wanted to commend marcl for the 6-pack recommendation.. this did help with the baseline problem I've been having since day-1 learning dirac and it's behaviors. The 6-pack for my setup, however, wasn't giving me the breadth I once had, so I made it a 9-pack. Same idea, just add and extra 3 in front. This may be due to my "gaming" and leaning forward a lot, also a large couch (large but MLP is still focused within the 6pack), so it is never a stationary single seating spot. I also took the same idea and added an extra 3 on either side , ideally a "15-pack" for lack of a better term, for a relaxed sofa situation - but keeping the same idea in focus, no higher or lower, same plane.. if one were to keep the 6-pack, then just adding 2 on either side would be the same idea.

While I mostly listen and play/watch by myself, my significant other brought up that it sounded different from her area.. after a successful 6-pack run.. That was a first ever, so it is why I decided to try these other configs... the 6-pack config definitely works for a single theater seat, but just adding a set in front or two/three on either side did help me spread it to please my constant movement and another discerning listener.

If this was already talked about, then disregard my input, I admit I have been mostly screening this thread, not studying everything.

Great unit past it's little "things"!

KC

[marcl]

May 25, 2024 18:12:17 GMT -5 kcdc said:

First post, Have had a RMC1 since last Oct or so... Snaps and pops aside, it's been OK if I ignore the rando freezing, menu issues (android app that the generous fellow made does fix a lot of that).. I did do the 100ms timing delay "bandaid" but wasn't sufficient enough at times, still pops and cracks during handshakes or input switchings, and regardless of what's been noted that it doesn't affect lip-syncing, it does. I did apply the 100ms delay to every possible channel as suggested to get past that issue.. I'm in my 40s, but do have a profession in VFX for the past 25 years, so maybe I am just too sensitive to the slight adjustment... a curse, trust me.. to the difference in sound to screen differences... Those are minimal issues at this point. 90% pleased with the purchase since I am used to nothing being perfect.

I wanted to commend marcl for the 6-pack recommendation.. this did help with the baseline problem I've been having since day-1 learning dirac and it's behaviors. The 6-pack for my setup, however, wasn't giving me the breadth I once had, so I made it a 9-pack. Same idea, just add and extra 3 in front. This may be due to my "gaming" and leaning forward a lot, also a large couch (large but MLP is still focused within the 6pack), so it is never a stationary single seating spot. I also took the same idea and added an extra 3 on either side , ideally a "15-pack" for lack of a better term, for a relaxed sofa situation - but keeping the same idea in focus, no higher or lower, same plane.. if one were to keep the 6-pack, then just adding 2 on either side would be the same idea.

While I mostly listen and play/watch by myself, my significant other brought up that it sounded different from her area.. after a successful 6-pack run.. That was a first ever, so it is why I decided to try these other configs... the 6-pack config definitely works for a single theater seat, but just adding a set in front or two/three on either side did help me spread it to please my constant movement and another discerning listener.

If this was already talked about, then disregard my input, I admit I have been mostly screening this thread, not studying everything.

Great unit past it's little "things"!

KC

Good logic starting with 6 and expanding.  The guy who came up with the 6 spent a week just trying to optimize his seat.  If you take that as the basis and experiment with expanding around it until the other seats sound pretty good without too much compromise  at the MLP ... that makes good sense.   And you can always save two or more configurations.

[marcl]

Dirac Live Beta 3.10.1 Software Changelog


helpdesk.dirac.com/en


Dirac Live Beta 3.10.1
2024-05-22

Changes listed are relative to the latest public stable release.

New features

Improved the measurement qualification algorithm to more reliably detect occasional data loss in the recording.

Logging Operating System version number information to correlate with measurement problems.

To improve the quality of measurements with microphones connected to Windows PCs, "WASAPI Exclusive Mode" will be activated by default in some cases.

To control whether "Exclusive Mode" is used or not in your case, and thereby override the default behavior, please follow these steps:

Click the Windows button, type “env”, and hit Enter.

Click the "Environment Variables..." button.

Click the "New..." button under "User variables for <user>" (or select DAUDIO_WASAPI_NON_EXCLUSIVE and click “Edit…” if it already exists).

Enter the following:

Variable name: DAUDIO_WASAPI_NON_EXCLUSIVE

Variable value:
“OFF” to enable Exclusive Mode
“ON” to disable Exclusive Mode (the old behavior)

Click "OK".

Restart the Dirac Live application, and verify that it is possible to complete a calibration.

Bug fixes

Fixed a system error when modifying large ART projects.

Fixed a crash on Windows and Mac if microphone access has not been granted by the user when entering the Select Microphone view.

Fixed a problem where the WAV file dump of the last recording was clipped due to conversion to fixed point - it now supports floating point data.

Several minor UI glitches were fixed on both desktop and mobile app versions.

Known issues

Filter export to devices from Onkyo/Integra/Pioneer take longer time than usual.

Download
Mac OS: artifacts.connect.dirac.com/public/accord/release-candidate/macos/diraclive-3.10.1-setup.zip

Win OS: artifacts.connect.dirac.com/public/accord/release-candidate/win64/diraclive-3.10.1-setup.exe

[marcl]

Testing Dirac Live 3.10.1

I uninstalled 3.9.7 and installed 3.10.1.   I loaded a previously saved project and it took some time and recalculated filters as expected ... and it loaded okay.

But ...

It is no longer possible to load previously saved measurements and add, delete, or repeat a measurement. It goes directly to the default Filter Design screen and if you go back to Measurements the ability to measure is disabled. If you back up to Volume Calibration it will not let you do Volume Calibration.

So I exited Dirac and went back in. I first did Volume Calibration, then proceeded to the Measurements screen and loaded the previously saved measurements. Same behavior. It loads the measurements and goes directly to default Filter Design and if you go back to Measurements you can’t add, delete or repeat a measurement. It’s all disabled.

So no point saving measurements from now on with 3.10.1.  And I don't think this is a bug.  I'll bet it is intentional.  I'll try putting in a Support ticket and see what they say.

[ttocs]

Jun 2, 2024 11:19:37 GMT -5 marcl said:

Testing Dirac Live 3.10.1

I uninstalled 3.9.7 and installed 3.10.1.   I loaded a previously saved project and it took some time and recalculated filters as expected ... and it loaded okay.

But ...

It is no longer possible to load previously saved measurements and add, delete, or repeat a measurement. It goes directly to the default Filter Design screen and if you go back to Measurements the ability to measure is disabled. If you back up to Volume Calibration it will not let you do Volume Calibration.

So I exited Dirac and went back in. I first did Volume Calibration, then proceeded to the Measurements screen and loaded the previously saved measurements. Same behavior. It loads the measurements and goes directly to default Filter Design and if you go back to Measurements you can’t add, delete or repeat a measurement. It’s all disabled.

So no point saving measurements from now on with 3.10.1.  And I don't think this is a bug.  I'll bet it is intentional.  I'll try putting in a Support ticket and see what they say.

It'll be interesting to find out if after running 3.10.1 and saving that session after Measurement if one can then load that project later and add more or delete Measurement locations like we've been able to do prior to this version? It might be, like you say - intentional, because it might not be completely backwards compatible with previous DL versions with respect to saved projects, but, might be ok to edit saved projects made with 3.10.1.

[KeithL]

So... after you loaded your previous measurements... you were unable to add or repeat new measurements.
This doesn't exactly surprise me... but I am curious about a few of the details...

You were unable to add measurements after loading measurements that were taken with a different version of the software.

1. Was the same thing true when you loaded a previous project whose measurements were taken with the NEW software?
    (Does it not let you add measurements to an existing project in general... or is that only the case if it involves a project done with a previous version of the software?)
    (I would assume that this probably means that they've changed how the measurements are taken or calculated so "they don't mix".)

2. Did it let you assign and try different / new Target Curves with the old measurements?

Jun 2, 2024 11:19:37 GMT -5 marcl said:

Testing Dirac Live 3.10.1

I uninstalled 3.9.7 and installed 3.10.1.   I loaded a previously saved project and it took some time and recalculated filters as expected ... and it loaded okay.

But ...

It is no longer possible to load previously saved measurements and add, delete, or repeat a measurement. It goes directly to the default Filter Design screen and if you go back to Measurements the ability to measure is disabled. If you back up to Volume Calibration it will not let you do Volume Calibration.

So I exited Dirac and went back in. I first did Volume Calibration, then proceeded to the Measurements screen and loaded the previously saved measurements. Same behavior. It loads the measurements and goes directly to default Filter Design and if you go back to Measurements you can’t add, delete or repeat a measurement. It’s all disabled.

So no point saving measurements from now on with 3.10.1.  And I don't think this is a bug.  I'll bet it is intentional.  I'll try putting in a Support ticket and see what they say.

[marcl]

Jun 3, 2024 8:47:03 GMT -5 KeithL said:

So... after you loaded your previous measurements... you were unable to add or repeat new measurements.
This doesn't exactly surprise me... but I am curious about a few of the details...

You were unable to add measurements after loading measurements that were taken with a different version of the software.

1. Was the same thing true when you loaded a previous project whose measurements were taken with the NEW software?
    (Does it not let you add measurements to an existing project in general... or is that only the case if it involves a project done with a previous version of the software?)
    (I would assume that this probably means that they've changed how the measurements are taken or calculated so "they don't mix".)

2. Did it let you assign and try different / new Target Curves with the old measurements?

Jun 2, 2024 11:19:37 GMT -5 marcl said:

Testing Dirac Live 3.10.1

I uninstalled 3.9.7 and installed 3.10.1.   I loaded a previously saved project and it took some time and recalculated filters as expected ... and it loaded okay.

But ...

It is no longer possible to load previously saved measurements and add, delete, or repeat a measurement. It goes directly to the default Filter Design screen and if you go back to Measurements the ability to measure is disabled. If you back up to Volume Calibration it will not let you do Volume Calibration.

So I exited Dirac and went back in. I first did Volume Calibration, then proceeded to the Measurements screen and loaded the previously saved measurements. Same behavior. It loads the measurements and goes directly to default Filter Design and if you go back to Measurements you can’t add, delete or repeat a measurement. It’s all disabled.

So no point saving measurements from now on with 3.10.1.  And I don't think this is a bug.  I'll bet it is intentional.  I'll try putting in a Support ticket and see what they say.

I have not yet taken new measurements with 3.10.1 .... I will do that today and see if the behavior changes.  It is possible that with measurements taken with 3.10.1 the ability to reload and alter them will be back.

Loading measurements and the filter project done with 3.9.2, I could change filter settings and save new filters.  The only thing I could not do - either from loading the measurements alone or the filter project - was go back and alter the measurements.  

In all previous versions, it was possible to load a measurements file and it would take you to the Measurements screen and it would be as if you never left ... all options were available and you could add, delete, replace.  If you wanted to take new measurements you could back up to the Volume screen and do the volume calibration, then go forward and add or replace measurements.  Starting with this 3.10.1 version you can't do that. 

Note also that the reason we started saving measurements on the Measurements screen before proceeding to Fiter Design, was that we found going backward from Filter Design back to Measurements did not work well.  I don't recall the errors exactly but I had an exchange with Flavio about it and he could not explain it ... but was not surprised.

[marcl]

As ttocs suggested to me the other day, I did a new set of measurements using 3.10.1 and the behavior when reloading measurements is as it was before. You CAN reload measurements to add, delete or replace ... IF the measurements were taken and saved with DL 3.10.1.

And otherwise the calibration run went fine ... on to level checking and response verification.

[lhracing]

I know there are individuals on this site who know much more about Dirac than me. I have a question about "Curtains" and limiting Dirac's correction area.

I have read so many opinions about leavening the curtains where Dirac puts them, adjusting the high end down to about 200hz and only correcting below the schroeder frequency of the room (which in my case is about 100hz).

Is there a consensus here as to where to set them?

Below is my mains (L+R) set to Large, no subs, the RED trace is without Dirac and the GREEN is with Dirac. Looks like at about 300hz and up Dirac doesn't do much.

[marcl]

Jun 4, 2024 17:00:03 GMT -5 lhracing said:

I know there are individuals on this site who know much more about Dirac than me. I have a question about "Curtains" and limiting Dirac's correction area.

I have read so many opinions about leavening the curtains where Dirac puts them, adjusting the high end down to about 200hz and only correcting below the schroeder frequency of the room (which in my case is about 100hz).

Is there a consensus here as to where to set them?

Below is my mains (L+R) set to Large, no subs, the RED trace is without Dirac and the GREEN is with Dirac. Looks like at about 300hz and up Dirac doesn't do much.

View Attachment

There's a very important thing to know about how Dirac corrects the response.  Unlike other products, if you allow Dirac to correct across the entire audible range it will NOT try to correct every anomaly to flat.  It will not.  Dirac folks have said many times that they analyze the multiple measurements to determine which anomalies are minimum phase or non-minimum phase.  Without getting into the deep definition, it simply means that only minimum phase anomalies should be corrected with PEQ.  

When people say not to correct above the Schroeder Frequency, they are assuming the application will try to flatten everything.   But that is not the case with Dirac.  So it's okay to run Dirac all the way up.  If you happen to have a peak or dip that is narrow and only occurs at some of the measurement positions ... Dirac will ignore them.

Even below the Schroeder Frequency, I have a couple dips that Dirac ignores.

Regarding your measurement ... L+R is not really a useful measurement.  It would be more useful to see each channel separately because very often they are different.  Also, please be sure to use Variable Smoothing so we can see what's going on in the low end.  That would give us an idea where the Schroeder Frequency is ... which is likely at 200 unless you have a very large room.

[PaulBe]

Jun 4, 2024 17:00:03 GMT -5 lhracing said:

I know there are individuals on this site who know much more about Dirac than me. I have a question about "Curtains" and limiting Dirac's correction area.

I have read so many opinions about leavening the curtains where Dirac puts them, adjusting the high end down to about 200hz and only correcting below the schroeder frequency of the room (which in my case is about 100hz).

Is there a consensus here as to where to set them?

Below is my mains (L+R) set to Large, no subs, the RED trace is without Dirac and the GREEN is with Dirac. Looks like at about 300hz and up Dirac doesn't do much.

View Attachment

Disclosure: I use room correction but not Dirac

I used to think that a small multiple of the Schroeder Frequency was a reasonable limit. After using room EQ, and, seeing and hearing the changes made by the room EQ, I now think it best to run room correction full range. There are significant frequency response differences between similar speakers at higher frequencies too.

I use the same speakers and crossovers for all 7 base channels. the tweeters are matched sets for L&R, Surr, and Back. The speakers are at slightly different distances from the floor between L&R, C, and Surrounds. I raise the surrounds about 5 degrees. The center is lowered about 5 degrees to fit under the TV.

Woofer response is different above and below the Schroeder frequency for each woofer, though there is a consistent loose pattern in woofer response deviation.  There is still differences in the room curve at higher frequencies well above the 800Hz crossover frequency, for each speaker.

Distance to MLP; speaker distance from the adjacent walls; reflecting objects near the speaker; will all change the frequency response of the same type of speaker, well above the Schroeder Frequency at MLP. I can imagine the issues get worse when speaker types are different from front to surrounds.

It's full range correction for me. My mind changed when I saw the graphical results. My ears hear the improved evenness of the treble.

[runner]

I would like to ask for the link to the latest version of the guide for using dirac, unfortunately the tread is too large and the information is not easy to find. Thank you

[webmst007]

Not sure where the guide you're after is but this white paper just got released by Dirac....

www.dirac.com/live/dirac-live-a-technical-overview/?submissionGuid=13cadc6d-fc67-4fc9-a69c-f6ea16b09d8c

[PaulBe]

Jun 8, 2024 0:31:37 GMT -5 runner said:

I would like to ask for the link to the latest version of the guide for using dirac, unfortunately the tread is too large and the information is not easy to find. Thank you

The second post in this thread contains many links.

[lhracing]

Every time I do a Dirac calibration my subwoofers are around 10bd too hot. I have a question about Dirac, target curves and base management subwoofer levels.

Typically most target curves add some bass boost. When looking at the different speaker groups (I am going to limit my discussion to the Mains and Subs) this boost is usually applied to the both the mains and subs. As an example, the mains are crossed over at 80hz, at 50hz both the mains and subs EQ target curves have about a 6db boost.

With base content at around 50hz (below the crossover point), do the subs see a 12db boost of this content because of the additive effect of both the mains and subs having the 6db EQ boost?

IF so, would I be better off not applying and boost (flat) EQ to the subwoofers?

[ttocs]

Jun 14, 2024 12:13:09 GMT -5 lhracing said:

Every time I do a Dirac calibration my subwoofers are around 10bd too hot. I have a question about Dirac, target curves and base management subwoofer levels.

I noticed this too, I can't recall which version it started with though. There have been times when I temporarily cut the gain of the subs in the miniDSPs by -10dB just before running Dirac, then put the gain back afterwards. But this week I've been testing some new subwoofery and haven't needed to adjust the gain of the subs before/after, and I haven't changed Dirac versions. Don't know what to say about this one.

Jun 14, 2024 12:13:09 GMT -5 lhracing said:

Typically most target curves add some bass boost. When looking at the different speaker groups (I am going to limit my discussion to the Mains and Subs) this boost is usually applied to the both the mains and subs. As an example, the mains are crossed over at 80hz, at 50hz both the mains and subs EQ target curves have about a 6db boost.

With base content at around 50hz (below the crossover point), do the subs see a 12db boost of this content because of the additive effect of both the mains and subs having the 6db EQ boost?

IF so, would I be better off not applying and boost (flat) EQ to the subwoofers?

Interesting, hmmm . . . . , I'll do a test to find out.

I usually adjust my mains with just a gentle tilt of maybe a dB or two, and slightly more tilt for the subs, but only a few dB

[ttocs]

Jun 14, 2024 12:13:09 GMT -5 lhracing said:

Typically most target curves add some bass boost. When looking at the different speaker groups (I am going to limit my discussion to the Mains and Subs) this boost is usually applied to the both the mains and subs. As an example, the mains are crossed over at 80hz, at 50hz both the mains and subs EQ target curves have about a 6db boost.

With base content at around 50hz (below the crossover point), do the subs see a 12db boost of this content because of the additive effect of both the mains and subs having the 6db EQ boost?

The answer is no.

Here's a plot showing my result from adding +6dB boost @50hz in Dirac to both the Fronts and Sub. It does result in about a +6dB higher level @50hz, but it's not an additive +12dB result.

One trace is sub only, one is L with sub on, and the third trace is L with sub off just to show how the Left Front is on its own while setup as Small. Disregard the fact that my main speakers don't like an 80Hz crossover. Aslo, this was a quick and dirty Dirac run with just L, R, and Sub, so I just used the "paddles" instead of the points like I usually do, and I don't ever boost anything so the dip from 300Hz to about 900Hz is something I don't see in my normal results.

[lhracing]

Thanks for taking the time to do your evaluation, good to see actual results.