AVS Forum | Home Theater Discussions And Reviews (https://www.avsforum.com/forum/)
-   Receivers, Amps, and Processors (https://www.avsforum.com/forum/90-receivers-amps-processors/)
-   -   The official Dirac Live thread (https://www.avsforum.com/forum/90-receivers-amps-processors/3137722-official-dirac-live-thread.html)

markus767 04-17-2020 10:15 AM

The official Dirac Live thread
 
People asked for it, here it is, a dedicated thread all things Dirac Live (DL). This post will be updated as we go along...

Manufacturer website:
https://live.dirac.com

Download for Mac, PC, iOS and Android:
https://live.dirac.com/download/

Quickstart guide – AVR/pre-pro:
https://live.dirac.com/content/uploa...nt-version.pdf

Quickstart guide – Dirac Live Processor for Mac/PC:
https://live.dirac.com/processor-step-by-step/

Bug reporting:
http://helpdesk.dirac.se/

History, target curves, bugs:
https://mehlau.net/audio/dirac-live-2/

Configure microphone and setting speaker levels:
https://mehlau.net/audio/dirac-live-...peaker-levels/

How to set crossovers (without "Bass Control"):
https://mehlau.net/audio/dirac-live-2-avr-crossover/

Dirac Live "Bass Control" add-on:
http://diracdocs.com/Bass%20Control%20in%20Live.pdf

Introduction to digital room equalization
http://diracdocs.com/on_room_correction.pdf

Deep dive into room correction with Dirac Live:

wilbur_the_goose 04-17-2020 12:52 PM

Subscribed - Thanks @markus767 !

One question I have about Dirac Live 2 is in regard to channel trims:
  • Dirac 1 didn't adjust trims - that was something I'd manually do after the calibration
  • Dirac 2 does pay attention the the trims.
  • I've found that my rear channels (10' away from the microphone) are dragging down the "set" trim level for all the other speakers. For example, my FL speaker is -8db, and my RL speaker (the quiet one) is set to 0db.
  • I'd much prefer that Dirac 2 set the loudest speaker to 0db and adjust relative to that, not to the quietest speaker.
  • Is there any way to "normalize" the trims, or to tell Dirac 2 to NOT set the trims?

Processor - HTP-1
Amps - Parasound A21, Emotiva XPA-5, Gen 2
Speakers - Paradigm Studio series
Setup: 7.1.0

wilbur_the_goose 04-17-2020 12:55 PM

Hint for Dirac 1 and Dirac 2 running on a PC:
I found that when running on a Windows 10 PC, you'll need to run Dirac under a user account with Administrator level rights.

In my case, I'll login using the Admin account, run Dirac, upload the filters to my device, and then log out (I never typically use the Admin user unless I really need to).

BigCoolJesus 04-17-2020 01:27 PM

Quote:

Originally Posted by wilbur_the_goose (Post 59509470)
Subscribed - Thanks @markus767 !

One question I have about Dirac Live 2 is in regard to channel trims:
  • Dirac 1 didn't adjust trims - that was something I'd manually do after the calibration
  • Dirac 2 does pay attention the the trims.
  • I've found that my rear channels (10' away from the microphone) are dragging down the "set" trim level for all the other speakers. For example, my FL speaker is -8db, and my RL speaker (the quiet one) is set to 0db.
  • I'd much prefer that Dirac 2 set the loudest speaker to 0db and adjust relative to that, not to the quietest speaker.
  • Is there any way to "normalize" the trims, or to tell Dirac 2 to NOT set the trims?

Processor - HTP-1
Amps - Parasound A21, Emotiva XPA-5, Gen 2
Speakers - Paradigm Studio series
Setup: 7.1.0


I would ask why it would matter if the loudest speaker is set to 0dB on the trim and all others are boosted as needed or if the quietest is set to 0dB and all others are cut as needed?

ss3740 04-17-2020 01:33 PM

I just ordered a Lexicon MC-10 which runs Dirac Live. What is the optional mic brand and part number that is recommended to use and where do you purchase them ? Thanks

AustinJerry 04-17-2020 02:08 PM

Quote:

Originally Posted by BigCoolJesus (Post 59509700)
I would ask why it would matter if the loudest speaker is set to 0dB on the trim and all others are boosted as needed or if the quietest is set to 0dB and all others are cut as needed?

Because that is the way Dirac works. Are you questioning why Dirac is designed that way?

markus767 04-17-2020 02:12 PM

Quote:

Originally Posted by ss3740 (Post 59509740)
I just ordered a Lexicon MC-10 which runs Dirac Live. What is the optional mic brand and part number that is recommended to use and where do you purchase them ? Thanks

As long as you're using a calibrated, omnidirectional mic and it is recognized by your computer (or AVR/AVP) the brand doesn't matter much.

markus767 04-17-2020 02:25 PM

Quote:

Originally Posted by wilbur_the_goose (Post 59509470)
Subscribed - Thanks @markus767 !

One question I have about Dirac Live 2 is in regard to channel trims:
  • Dirac 1 didn't adjust trims - that was something I'd manually do after the calibration
  • Dirac 2 does pay attention the the trims.
  • I've found that my rear channels (10' away from the microphone) are dragging down the "set" trim level for all the other speakers. For example, my FL speaker is -8db, and my RL speaker (the quiet one) is set to 0db.
  • I'd much prefer that Dirac 2 set the loudest speaker to 0db and adjust relative to that, not to the quietest speaker.
  • Is there any way to "normalize" the trims, or to tell Dirac 2 to NOT set the trims?

Processor - HTP-1
Amps - Parasound A21, Emotiva XPA-5, Gen 2
Speakers - Paradigm Studio series
Setup: 7.1.0

How speaker trims are set depends on the available internal headroom. Manufacturers usually don't provide good control and monitoring for these parameters. The miniDSP DDRC-88BM being the exception.

BigCoolJesus 04-17-2020 02:28 PM

Quote:

Originally Posted by AustinJerry (Post 59509964)
Because that is the way Dirac works. Are you questioning why Dirac is designed that way?

I think you're reading my reply incorrectly. I'm not questioning how Dirac sets the trims....

mraub 04-17-2020 02:57 PM

Many use this microphone:

https://www.minidsp.com/products/aco...urement/umik-1

I recall seeing a comparison which showed calibrations performed with the umik were more accurate than those done with the cheap microphones manufacturers include with processors and receivers.


Quote:

Originally Posted by ss3740 (Post 59509740)
I just ordered a Lexicon MC-10 which runs Dirac Live. What is the optional mic brand and part number that is recommended to use and where do you purchase them ? Thanks


ss3740 04-17-2020 03:19 PM

Quote:

Originally Posted by mraub (Post 59510152)
Many use this microphone:

https://www.minidsp.com/products/aco...urement/umik-1

I recall seeing a comparison which showed calibrations performed with the umik were more accurate than those done with the cheap microphones manufacturers include with processors and receivers.


Thanks. Can't seem to find any in USA for sale. Everything is sold out

neo_2009 04-17-2020 04:59 PM

@markus767 , you mention "... Some features have even been crippled (e.g. test signal/mic gain adjustment) ...", and later you mention "Version 2.5.2 has been released on 08.04.2020 with usability improvements and bug fixes but generally all of the above still applies".

Following your posts on the NAD T758 thread, i was under the impression you were hopping that version 2.5 would address the signal/gain issue.
Could you please clarify?

Thank you very much for having the time to create this thread!

Kal Rubinson 04-17-2020 05:08 PM

Quote:

Originally Posted by wilbur_the_goose (Post 59509470)
Subscribed - Thanks @markus767 !

One question I have about Dirac Live 2 is in regard to channel trims:[*]Dirac 1 didn't adjust trims - that was something I'd manually do after the calibration[*]Dirac 2 does pay attention the the trims.

That may depend on how you used DL. Dirac 1 did do trims in the PC implementation.

Snowmanick 04-17-2020 05:16 PM

Quote:

Originally Posted by ss3740 (Post 59510244)
Thanks. Can't seem to find any in USA for sale. Everything is sold out

I ordered my Umik-1 directly from MiniDSP and had it shipped to the States.

BigCoolJesus 04-17-2020 05:21 PM

2 Attachment(s)
I've decided to try a Dirac cal only correcting up to my rooms Schroeder frequency but was wondering if a good guestimate of where that would be can be gleaned from the below Dirac results for my Front LR and a pair of my Atmos speakers?

My guess is maybe around 500Hz?

llang269 04-17-2020 06:00 PM

Maybe someone can help me here. I have Dirac on a Datasat Ls10 and I can’t get my speakers to make noise to calibrate. It finds the processor, recognizes the mic and when it come to playing the noise back I get nothing, zero sound

Stereojeff 04-17-2020 06:08 PM

Quote:

Originally Posted by BigCoolJesus (Post 59510786)
I've decided to try a Dirac cal only correcting up to my rooms Schroeder frequency but was wondering if a good guestimate of where that would be can be gleaned from the below Dirac results for my Front LR and a pair of my Atmos speakers?

My guess is maybe around 500Hz?


500 Hz is a good estimate. Or you can use this equation to determine the actual Schroeder frequency for your listening room: http://www.acoustic.ua/forms/room-modes-eng.html


Jeff

BigCoolJesus 04-17-2020 06:26 PM

72 Attachment(s)
Quote:

Originally Posted by Stereojeff (Post 59510930)

500 Hz is a good estimate. Or you can use this equation to determine the actual Schroeder frequency for your listening room: http://www.acoustic.ua/forms/room-modes-eng.html

Jeff

Thanks. Just tried the calculator and it says the Schroeder frequency should be around 170Hz based on 22'x14'x8' and default temp and reverb time.

Visually, when looking at frequency response of speakers in the room, is it possible to roughly guess the frequency? Reading up it seems like the visual indicator is when the response goes from large dips/peaks and "narrows" more?

Better yet, is there an easy way to use REW and find out? I'm no REW expert but figure if there is an easy way to find out it's worth a shot.

markus767 04-18-2020 12:04 AM

Quote:

Originally Posted by neo_2009 (Post 59510698)
@markus767 , you mention "... Some features have even been crippled (e.g. test signal/mic gain adjustment) ...", and later you mention "Version 2.5.2 has been released on 08.04.2020 with usability improvements and bug fixes but generally all of the above still applies".

Following your posts on the NAD T758 thread, i was under the impression you were hopping that version 2.5 would address the signal/gain issue.
Could you please clarify?

Thank you very much for having the time to create this thread!

The gain matching algorithm issue went much deeper and isn't related to settings in the "Volume Calibration" tab. It has been fixed beginning version 2.5. So finally after 16 months (or 2.5 years, depending on when you start counting) we finally have a Dirac Live version that provides more or less the same functionality as version 1... Let's hope the upcoming "Bass Control" feature was worth the wait.

markus767 04-18-2020 12:27 AM

Quote:

Originally Posted by BigCoolJesus (Post 59510786)
I've decided to try a Dirac cal only correcting up to my rooms Schroeder frequency but was wondering if a good guestimate of where that would be can be gleaned from the below Dirac results for my Front LR and a pair of my Atmos speakers?

My guess is maybe around 500Hz?

It can be eyeballed looking at multiple spatially distributed measurements. Check the "Measured/Corrected Spread" box.

P.S. The "Measured Spread" checkbox is currently missing due to a bug.

markus767 04-18-2020 12:28 AM

Quote:

Originally Posted by llang269 (Post 59510910)
Maybe someone can help me here. I have Dirac on a Datasat Ls10 and I can’t get my speakers to make noise to calibrate. It finds the processor, recognizes the mic and when it come to playing the noise back I get nothing, zero sound

Get in contact with Dirac. See post 1 for the link.

BigCoolJesus 04-18-2020 12:51 PM

Redid Dirac using the Harman Curve and only EQing up to 500Hz (picked this since the mid-range crossover on the center starts at 600Hz).

Rewatching the same movie scenes I've been using over the last week for testing and this definitely sounds the best. The other times when I let Dirac EQ the entire frequency range it never sounded bad. And I would have never noticed it could have been "better" had I not tried. But having done so it definitely sounds better. There is more "life" it seems. How the speakers were meant to sound after EQing the low end to remove some room interaction.

AustinJerry 04-18-2020 01:53 PM

Quote:

Originally Posted by BigCoolJesus (Post 59514378)
Redid Dirac using the Harman Curve and only EQing up to 500Hz (picked this since the mid-range crossover on the center starts at 600Hz).

Rewatching the same movie scenes I've been using over the last week for testing and this definitely sounds the best. The other times when I let Dirac EQ the entire frequency range it never sounded bad. And I would have never noticed it could have been "better" had I not tried. But having done so it definitely sounds better. There is more "life" it seems. How the speakers were meant to sound after EQing the low end to remove some room interaction.

Not sure whether when you say “redid Dirac” if you meant that you ran the measurements again. To accomplish what you wanted to try, you can simply load your previous project, lower the right “curtain” to the desired cut-off frequency for correction, click Optimize, and save the results in a different preset. If this is exactly what you did, then never mind.

blake 04-19-2020 01:55 AM

Why doesn’t Dirac do real measurements again after it shows you the predicted target curve ?

markus767 04-19-2020 04:08 AM

Quote:

Originally Posted by blake (Post 59516552)
Why doesn’t Dirac do real measurements again after it shows you the predicted target curve ?

No need because the measured room behaves largely as a linear time-invariant system (LTI system), and so does the whole audio component chain. There is some non-linear behavior but it is insignificant. So taking "after" measurements doesn't fulfill any other purpose than satisfying OCD behavior and is a complete waste of time.

steakhouse_ 04-19-2020 04:47 AM

The predicted curve is accurate. One just has to keep in mind that this is the average of all measured positions. It is also possible to see the familiy of all predicted curves.

AustinJerry 04-19-2020 07:19 AM

Quote:

Originally Posted by blake (Post 59516552)
Why doesn’t Dirac do real measurements again after it shows you the predicted target curve ?

Quote:

Originally Posted by markus767 (Post 59516708)
No need because the measured room behaves largely as a linear time-invariant system (LTI system), and so does the whole audio component chain. There is some non-linear behavior but it is insignificant. So taking "after" measurements doesn't fulfill any other purpose than satisfying OCD behavior and is a complete waste of time.

Quote:

Originally Posted by steakhouse_ (Post 59516776)
The predicted curve is accurate. One just has to keep in mind that this is the average of all measured positions. It is also possible to see the familiy of all predicted curves.

Another reason not to do this is the mic placement. Placement affects the measurements, and it would be virtually impossible to reproduce the exact mic placements for a full calibration.

And as for the accuracy of the predicted curve, I agree that it is mathematically accurate. However, in my experience post-calibration REW measurements rarely show results precisely matching the predictions. Anyone care to hypothesize why this would be the case?

markus767 04-19-2020 07:24 AM

Quote:

Originally Posted by AustinJerry (Post 59517176)
Another reason not to do this is the mic placement. Placement affects the measurements, and it would be virtually impossible to reproduce the exact mic placements for a full calibration.

And as for the accuracy of the predicted curve, I agree that it is mathematically accurate. However, in my experience post-calibration REW measurements rarely show results precisely matching the predictions. Anyone care to hypothesize why this would be the case?

They don't match for the very same reason you just mentioned, mic placement. When I run a single point optimization and remeasure without the mic being moving, the measured results are virtually identical.

AustinJerry 04-19-2020 07:30 AM

Quote:

Originally Posted by markus767 (Post 59517186)
They don't match for the very same reason you just mentioned, mic placement. When I run a single point optimization and remeasure without the mic being moving, the measured results are virtually identical.

I’m sure you are correct. Easy to test.

steakhouse_ 04-19-2020 09:25 AM

Quote:

Originally Posted by AustinJerry (Post 59517176)
Another reason not to do this is the mic placement. Placement affects the measurements, and it would be virtually impossible to reproduce the exact mic placements for a full calibration.

And as for the accuracy of the predicted curve, I agree that it is mathematically accurate. However, in my experience post-calibration REW measurements rarely show results precisely matching the predictions. Anyone care to hypothesize why this would be the case?

Hmmm, maybe dirac sometimes fails to detect wether the measured peaks and dips are really 100% minimum phase or at least only partly. In that case a correction would not enterely behave as predicted.


All times are GMT -7. The time now is 09:15 AM.

Powered by vBulletin® Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.
vBulletin Security provided by vBSecurity (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
User Alert System provided by Advanced User Tagging (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.