• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Dirac Live, FIR-filter - Yes, No?

Flak

Senior Member
Industry Insider
Joined
Nov 18, 2018
Messages
385
Likes
593
Salve Flavio! :)

Heh.. So much text and yet so little usefull information.

In my opinion it would be very usefull if Dirac would implement control measurement procedure so users can verify how good of a job is filter generated by Dirac actually doing. As long as that doesn't exist users are left on their own and the only thing they can do is either measure the log sweep in same 9 points and average them or they can apply RTA moving microphone method in the same space within those 9 points. In my opinion and in my experience both of these methods would result in a very similar curve and would provide very good insight of Diracs results.

Ciao Krunok :)

my opinion is different as I think instead that the document that I linked provides a wealth of very valuable information.
One can understand that the frequency (or magnitude) response doesn't tell the whole story with reference to perceived listening results,
time and geometry play an important role.
For example, late reflections should not be corrected and Dirac Live's algorithm can do that thanks to the fact that those late reflections are late because they are bounced by several walls and change with position, Dirac will not correct those as this will be shown by the different measurements in different positions.
It also seems to me that Amirm correctly commented about side reflections in another thread, the subject is discussed in the section "Some reflections are good, some bad"
A measurement in the frequency domain completely removes the information in the time domain and filters that provide the same exact frequency response can have significantly better or worse listening results... tonal balance is important but it's not all, a slightly less linear response can be audibly better depending on a number of factors.
In a few words I think that this document provides much more than "little usefull information" and might even give away more info than necessary about the way Dirac Live works : http://diracdocs.com/on_room_correction.pdf

That's anyhow for our coforumers to judge, Flavio
 
Last edited:

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Ciao Krunok :)

my opinion is different as I think instead that the document that I linked provides a wealth of very valuable information.

I didn't express myself precisely - what I meant was that the document you linked has been available for some time and I'm sure most of the folks here interested in DRC already read it. In other words I wanted to say that the document you linked is nothing reall new.

Ciao Krunok :)
.. tonal balance is important but it's not all, a slightly less linear response can be audibly better depending on a number of factors.

That is true and nobody argues with that. It would be nice if you would explain here in more details that "number of factors". :)

However, I still think it would be beneficial for all your users to see what effects Dirac correction had on FR and IR by means of control measurement via before and after comparison graphs. Do you think your marketing department would allow that feature to be available in next version? :)

P.S. pls don't get me wrong, I am not criticising your product, for which I think it is probably one of the best on the market today. I am merely suggesting a feature that I believe would be an improvement to the current functionality.
 

Flak

Senior Member
Industry Insider
Joined
Nov 18, 2018
Messages
385
Likes
593
I didn't express myself precisely - what I meant was that the document you linked has been available for some time and I'm sure most of the folks here interested in DRC already read it. In other words I wanted to say that the document you linked is nothing reall new.



That is true and nobody argues with that. It would be nice if you would explain here in more details that "number of factors". :)

However, I still think it would be beneficial for all your users to see what effects Dirac correction had on FR and IR by means of control measurement via before and after comparison graphs. Do you think your marketing department would allow that feature to be available in next version? :)

P.S. pls don't get me wrong, I am not criticising your product, for which I think it is probably one of the best on the market today. I am merely suggesting a feature that I believe would be an improvement to the current functionality.

Thanks for clarifying and agreed on all points :)
and yes, from a marketing perspective a separate measurement session could be an interesting addition/selling feature even if the measurements will not be significantly different from the current computed ones
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Thanks for clarifying and agreed on all points :)
and yes, from a marketing perspective a separate measurement session could be an interesting addition/selling feature even if the measurements will not be significantly different from the current computed ones

Are you able to share any details about how do you perform IR correction?
 

Flak

Senior Member
Industry Insider
Joined
Nov 18, 2018
Messages
385
Likes
593
I cannot share more info than available in that document but I can add that the latest Dirac Live 2.x algorithm might even slightly compromise on FR response of one of the speakers in favor of the best possible phase coherence between them that is important for imaging.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
I cannot share more info than available in that document but I can add that the latest Dirac Live 2.x algorithm might even slightly compromise on FR response of one of the speakers in favor of the best possible phase coherence between them that is important for imaging.

Interesting. I gues achieving best possible phase coherence is important only above Schroeder frequency, or..?
 

Flak

Senior Member
Industry Insider
Joined
Nov 18, 2018
Messages
385
Likes
593
Good to see you here Flavio. I upgraded your title. Welcome to the forum. :)

Thanks Amirm,
your unbiased testing and the knowledgeable information from you and other forumers are quite informative and interesting.
Keep up the good work!
 

patate91

Active Member
Joined
Apr 14, 2019
Messages
253
Likes
137
How can it be called science if there's no way to confirm results? IMO if you can't replicate the results of an experiment you're going against the scientific method.

For now @Krunok seems right with his questions.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
How can it be called science if there's no way to confirm results? IMO if you can't replicate the results of an experiment you're going against the scientific method.

For now @Krunok seems right with his questions.

IMHO you can easilly check the result you got with Dirac Live: use REW RTA moving microphone method to cover the same area you covered with 9 points Dirac requried for correction and you'll get very good picture about amplitude response of each speaker. Make nearfield sweep measurement at say 1m from the speakers, or even closer, 0.5m or 10cm, apply IR delay correction and some gating or FDW (if necessary) and you'll be able to see how Dirac did with time domain correction (phase, step response and GD).

I see no need for mistifications as this way you can easilly check effects of every correction procedure, not only Dirac Live. This method simply gives you pretty good picture how your speakers are doing at your LP or wherever you have measured them.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
One more thing: I remember seeing @mitchco demonstrating the efficiency of his filters created with Audiolense. I also remember @RayDunzl doing the same with the filters running on MiniDSP platform but I cannot remeber this moment which tool he used to create them (Ray, pls help - is it Acourate?). I also remember one member demonstrated how well the filters worked generated by DRC-FIR tool based on his measurements manually done with REW.

Apart from these I don't remember seeing any other examples where somebody demonstrated on this forum how well DRC tool worked. And I think we can all agree that untill such results are demonstrated we cannot actually judge the efficiency of particular DRC tools.
 

Olli

Senior Member
Forum Donor
Joined
Mar 6, 2018
Messages
331
Likes
242
Here's a comparison of an Audiolense Multichannel Sub/Mains setup/measurement (Subs MiniDSP SHD, Mains Weiss 202 DAC, both connected to a MiniDSP UDIO-8 - red), a Dirac (MiniDSP SHD) 2 CH setup (blue) and a non REQ'ed measurement (MiniDSP SHD - green). Please note that the Target curves are not identical in Dirac vs Audiolense (need to copy the Audiolense target curve in Dirac when I have some time) and that I could not figure out how to measure L and R channels seperately in REW on short notice.

Mains are TAD CE-1, Subs JL Audio E112, Amp March Audio P252. XO is at 110 Hz (TADs are on top of Subs).
849DE57E-807C-4F08-A32D-8A412721BEC1.jpeg
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Here's a comparison of an Audiolense Multichannel Sub/Mains setup/measurement (Subs MiniDSP SHD, Mains Weiss 202 DAC, both connected to a MiniDSP UDIO-8 - red), a Dirac (MiniDSP SHD) 2 CH setup (blue) and a non REQ'ed measurement (MiniDSP SHD - green). Please note that the Target curves are not identical in Dirac vs Audiolense (need to copy the Audiolense target curve in Dirac when I have some time) and that I could not figure out how to measure L and R channels seperately in REW on short notice.

Mains are TAD CE-1, Subs JL Audio E112, Amp March Audio P252. XO is at 110 Hz (TADs are on top of Subs).
View attachment 24918

This would be extremely interesting experiment if you would manage to make it under the same conditions (same number of channels, same target curve etc.)!
 

Olli

Senior Member
Forum Donor
Joined
Mar 6, 2018
Messages
331
Likes
242
This would be extremely interesting experiment if you would manage to make it under the same conditions (same number of channels, same target curve etc.)!

Sure. It will just take some time, but no problem.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
One more thing: I remember seeing @mitchco demonstrating the efficiency of his filters created with Audiolense. I also remember @RayDunzl doing the same with the filters running on MiniDSP platform but I cannot remeber this moment which tool he used to create them (Ray, pls help - is it Acourate?).

It's a MiniDSP OpenDRC-DI (two channels, digital in and out) with AcourateDRC (think Acourate Lite), a version of Acourate specifically aimed at the OpenDRC series.
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
I also remember one member demonstrated how well the filters worked generated by DRC-FIR tool based on his measurements manually done with REW.

Discussion starts here:

https://www.audiosciencereview.com/...orial-for-dummies-part-2.5/page-9#post-163964

I'd like to see more people trying DRC-FIR. It's a command-line tool, and the configuration is pretty technical, which may be off-putting. I don't really understand the meaning of many of the parameters myself, I've just been using the example configurations that come with it. Also, there's not much of a community around it. The Sourceforge mailing list gets almost no traffic, but that's somewhat the nature of Sourceforge mailing lists.

There is a GUI tool called DRCDesigner that might make it a bit easier to tweek the configuration. It's an old Java program, but it seems to work fine on windows with ASIO4ALL. I prefer REW for measuring the impulse response, though.

It would make sense to create a DRC-FIR plugin for REW that would create filters right within REW and allow examining a test convolution of the result. Is the REW plugin API documented somewhere?

http://drc-fir.sourceforge.net/
http://www.alanjordan.org/DRCDesigner/DrcDesignerHelp.html
(Video that shows some usage of DRCDesigner)
 

DaleMurray

Member
Joined
Mar 19, 2019
Messages
9
Likes
4
Location
Wheaton Illinois
I believe looking at response curve is only a very small part of the equation. Speaking with Dirac engineers are axpona this year they emphasized timing correction is more important, overall, than EQ curve.

The anticipated response is the product of all the measurements+correction. To see if it works in reality you'd need to take REW measurements in those same location then spit out an average of all measurements.

I have used REW to compare before and after measurement at the listening position; while the overall response curve was not wildly impressive - Dirac cannot fix my room - but impulse response was better.

As I write this I realize I should do what I suggested above - multiple measurements averaged.
 

TonioRoffo

Member
Joined
Jul 6, 2020
Messages
84
Likes
41
At the time Roon did not have a DSP function and I was not looking forward to getting yet another piece of software to make the correction in there. I tried Dirac, it worked very nicely and that was that. :) FYI I have it only correcting < 200 Hz.

I was wondering how you have Dirac set up. I have a license for Dirac V2 (and now V3) which only works as a VST plugin on PC.

I'm, for now, stuck in the logitech media server ecosystem, using foobar2000 on an intel NUC as streamer. Foobar can do VST on the incoming music and that works nicely.

However I'd love to add Deezer to the mix (cheapest CD quality streaming option out there) but there's no decent plugins on LMS for that.
 
Top Bottom