• 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 sounds really bad - am I doing something wrong?

board

Active Member
Joined
Apr 24, 2019
Messages
217
Likes
154
Before I show measurements I'll lay the basics down:

* This is my first time trying Dirac, and I have used a free 14-day trial license, only as software. My plan was to try it out and then buy hardware later. This free trial processes the entire audible range.
* My microphone is the MiniDSP Umik-1. I use the calibration file as well.
* I am aware that many aspects go into this (personal preference, habits, etc.) when hearing the difference that room correction makes, but I don't think this is the problem here. I sense that something has gone wrong somewhere.
* My speakers are GoldenEar Triton 1.
* I have acoustic panels from Gik Acoustics: Bass traps in all four corners (one corner has two) + one more bass trap behind one speaker, as well as thinner panels on the two side walls, and a thick panel right behind the listening position. When installing these panels I noticed a bit less bass, although what I mostly noticed was that the bass became less boomy and more "firm", and the high frequencies became less shrill, and most of the reverberation went away as well. Music became less "shouty".
* I have enjoyed my speakers, but I sense that certain frequency ranges are a bit too lively in this room. The lower treble can be a bit piercing, and bass is not completely even.
* I have remastered a lot of music, and I often (but not always) remaster it pretty much the same way, and I figured that this pattern was perhaps because I needed room correction. However, I often check my remasters on headphones as well and still prefer them to the originals. Also, I haven't felt any need to tinker with really good recordings.

So, unlike Dirac 2 that I tried to install once without any luck, installing and using Dirac 3 on Windows 10 was a breeze. I had installed VST in Foobar, but it didn't seem necessary. As you will see further down, the room correction was applied to any source I played. I don't know if this is what is going wrong.
I went through all the measurements, and I've tried both using "studio" and "home" settings, with the latter being "tightly focused", as I always sit in the same are of my couch.
No matter what I do, it just doesn't sound "right".
As mentioned, I am aware that preferences and habits play into this, but the stories I hear are usually that once you have listened to room correction for a short while, everything becomes clearer and just sounds "right", and playing music without it then just sounds "wrong". For me it has been the opposite.
To be specific:
The bass is gone, and the treble sounds very thin, "flat", dull, and lackluster. My speakers no longer sound "big". As you can see from the screenshots below, a lot of bass has been removed, as I also expected, but the treble also just sounds lackluster and a bit annoying. Without room correction the music is a bit too "lively", but also much more exciting to listen to. It sounds 3D and lifelike, whereas with room correction my big speakers sound a bit like this one:



url.jpg



I also noticed that with room correction imaging has tilted significantly to the left. Without room correction voices were essentially dead-center.
This could in part be because one side of the room is worse than the other, and also because there's more bass closer to the speakers than there is at the listening position.
Nevertheless, whether I complete all 9 measurement positions, or I only do one measurement for the listening position, it just sounds "wrong".
I think I'm doing something wrong somehow, because not only does it sound "wrong", but it also looks wrong - meaning when I do a measurement after applying the room correction, the measurement looks pretty much exactly the same as before.

So, here's a measurement from the listening position:

Gardiner trukket for.png



Although there are many peaks and dips, and some of them are wider than I had hoped, the general trend is still pretty much what I had hoped for, so no broad areas that go in the wrong direction, expect for the bass.
I then did all nine measurements (and I tried it several times), and then pressed "proceed to filter design" and got this filter:


Filter design, tightly focused imaging.png


As mentioned, there seems to be more bass closer to the speakers, so the lowering around 100 Hz in the filter is probably due to there being spikes in that region in other parts of the room, especially on the right hand side of the room, where the bass is audibly louder than in other parts of the room.
Nevertheless, if I export a filter after only doing one measurement from the listening position, where it should then remove less bass around 100 Hz, according to my first picture, the result still sounds pretty much the same.
So, after exporting the filter and naming it, I then choose Dirac as the output on my computer, and I'm wondering if something goes wrong there, when I make it the standard output. Whether I play music from Windows Media Player, Foobar, or Youtube, the Dirac filter is applied to all of it.

Output.png


If I then re-measure just the listening position after making Dirac my default setting, I get essentially the exact same measurement as before applying the room correction, even though at least music sounds radically different. I have also tried exporting that resulting filter. Then measure again just from the listening position and export that resulting filter, then repeat for six generations. It all sounds (essentially) the same.

I have tried to move the microphone for the 8 other measurement positions and play around with the "curtains" before exporting the filter - avoiding that Dirac touches anything below 80 Hz or anything above 1 kHz, and although moving the curtains changes the sound, and perhaps more to my liking, it still doesn't sound "right".

Also, my speakers have a built-in powered subwoofer with a dial. I have had the dial set at 1 o'clock since installing the acoustic panels, and put it at 12 o'clock for these measurements. If I apply the room correction and then turn the dial all the way to the max it sound better, but the treble still sounds "wrong".
I have tried good recordings like Diana Krall (where the impact was the smallest) to somewhat average recordings to very noisy recordings like Black Label Society and Gojira, where the impact was the greatest.
The image shift to the left applied to all of them.

I'm hoping that I'm just doing something wrong, and all suggestions are very welcome, as I've been going a bit nuts over this these last few days.
 
Last edited:

Kal Rubinson

Master Contributor
Industry Insider
Forum Donor
Joined
Mar 23, 2016
Messages
5,273
Likes
9,790
Location
NYC
So, unlike Dirac 2 that I tried to install once without any luck, installing and using Dirac 3 on Windows 10 was a breeze. I had installed VST in Foobar, but it didn't seem necessary. As you will see further down, the room correction was applied to any source I played. I don't know if this is what is going wrong.
A shot in the dark: It seems you have DLP installed as a "Stand-Alone" and serving as your PC's default audio output. If you have it also installed as a VST in Foobar, is it possible that you are running it twice? Can you check on that and, if warranted, delete one of them?
 

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,356
Location
Budapest
Can you post REW measurements about your system at listening position before/after Dirac?

I have been using it for years on multiple systems (all PC based) and it always did wonders on all of them

One hint: you can play around with the target curve to increase/decrease bass and treble - have you tried that? For me it helps a lot
 
OP
board

board

Active Member
Joined
Apr 24, 2019
Messages
217
Likes
154
A shot in the dark: It seems you have DLP installed as a "Stand-Alone" and serving as your PC's default audio output. If you have it also installed as a VST in Foobar, is it possible that you are running it twice? Can you check on that and, if warranted, delete one of them?
This actually did occur to me, so initially I tried to uninstall the VST, which seemed a bit difficult. But I'll try again.
However, wouldn't this only apply if I play music through Foobar? If I play the same song in Foobar or any other play it sounds the same.
I do really appreciate the suggestion though, and I would be happy to hear any other suggestions if you have any, as I really think I'm doing something wrong :).
 
OP
board

board

Active Member
Joined
Apr 24, 2019
Messages
217
Likes
154
Can you post REW measurements about your system at listening position before/after Dirac?

I have been using it for years on multiple systems (all PC based) and it always did wonders on all of them

One hint: you can play around with the target curve to increase/decrease bass and treble - have you tried that? For me it helps a lot
I'll see if I can make REW work. Around a year ago I tried to use it but with no luck.
I did also try to fiddle a bit with the curves in Dirac, boosting the bass a little bit and lowered the lower treble, and while it certainly made it sound different, it still didn't sound good, and the bass became very bloated and "pumping". I'll report back on how it goes with REW.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,424
Likes
7,940
Location
Brussels, Belgium
I think you just don't like the Dirac Target Curve, there is no apparent correction that would make it sound malevolent.

try the file in the attachment, but remove .txt and add .targetcurve first.
 

Attachments

  • Speaker natural roll-off.txt
    165 bytes · Views: 403
OP
board

board

Active Member
Joined
Apr 24, 2019
Messages
217
Likes
154
I think you just don't like the Dirac Target Curve, there is no apparent correction that would make it sound malevolent.
That is certainly possible, although it sounds so "wrong" that I think something has gone wrong somewhere, as I have more faith in Dirac than that :). In other words: People usually say that using Dirac everything just sounds "right" all of a sudden, which it doesn't here - quite the opposite. So I'm thinking something must have gone wrong.

try the file in the attachment, but remove .txt and add .targetcurve first.
I downloaded the file and saved as filetype .targetcurve instead of .txt, but in what folder am I supposed to put it, and what do I need to do with it? I couldn't find a folder with any other targetcurves. In the user folder on my computer I did find a folder called "Dirac/targets", but it was empty. I tried putting your file there, but I couldn't open it with Dirac. I don't see any Option in Dirac to open any curves like that (I can save the curves I make by measuring and them assign them slots, but that's it).

As for Ppataki's suggestion of measuring with REW, I haven't installed REW yet, but I will see if I can get that done on Sunday.
And sorry about the late response. I've been overloaded with work this week, so I've had to postone this.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,424
Likes
7,940
Location
Brussels, Belgium
That is certainly possible, although it sounds so "wrong" that I think something has gone wrong somewhere, as I have more faith in Dirac than that :). In other words: People usually say that using Dirac everything just sounds "right" all of a sudden, which it doesn't here - quite the opposite. So I'm thinking something must have gone wrong.


I downloaded the file and saved as filetype .targetcurve instead of .txt, but in what folder am I supposed to put it, and what do I need to do with it? I couldn't find a folder with any other targetcurves. In the user folder on my computer I did find a folder called "Dirac/targets", but it was empty. I tried putting your file there, but I couldn't open it with Dirac. I don't see any Option in Dirac to open any curves like that (I can save the curves I make by measuring and them assign them slots, but that's it).

As for Ppataki's suggestion of measuring with REW, I haven't installed REW yet, but I will see if I can get that done on Sunday.
And sorry about the late response. I've been overloaded with work this week, so I've had to postone this.

when you have a Dirac project open in the filter making stage, just press ‘file’ or the menu icon, then ‘load target curve’ then ‘all groups’.

indeed then it will open the directory you copied to file to and you will be able to import it.

don’t forget to adjust the curtains to your own speaker roll-offs.
 

puppet

Senior Member
Joined
Dec 23, 2020
Messages
446
Likes
284
This may come off as a dumb question ... with apps like this, given that the OP has applied various forms of physical room correction, what is the app measuring? Is the app, perhaps, ignoring the physical correction measures in place by windowing out reflections and therefore "doubling up" on the physical compensation(s) the user has applied?
 

FeddyLost

Addicted to Fun and Learning
Joined
May 24, 2020
Messages
752
Likes
542
In other words: People usually say that using Dirac everything just sounds "right" all of a sudden, which it doesn't here - quite the opposite. So I'm thinking something must have gone wrong.
You have bass boost below 100 hz up to + 20 db. It's a lot if you don't listen at whisper levels.
If you was listening to this FR for long time, you just feel this "correct".
Almost flat target response will be thin and light until your earbrain system adapts to less bass.
And without measurements "before" and "after" all theories are useless. We don't know what exactly is going on at your LP.
 

txbdan

Active Member
Joined
Apr 21, 2020
Messages
213
Likes
198
The answer is in your plots. You had almost 20db of boost at 35hz and a good 10db from 150Hz and down and you got used to that "incorrect" sound. Dirac cut that all out so it now sounds relatively thin. In addition, the default Dirac curve is pretty bright. I use some B&K and OliveToole targets that I like much better. Try these (just remove the .txt and load them as targets):
 

Attachments

  • OliveToole.targetcurve.txt
    233 bytes · Views: 404
  • B&K.targetcurve.txt
    252 bytes · Views: 370

f1shb0n3

Senior Member
Forum Donor
Joined
Aug 29, 2020
Messages
322
Likes
498
Location
Seattle Area
@board - did you use Dirac Live 3.1.1 on Mac by any chance? There's a known bug in that version that I also experienced that makes the speakers out of phase and generally sounding horrible, see the discussion at https://www.audiosciencereview.com/forum/index.php?threads/dirac-live-3-1-1-bugs.26027/#post-912429
The fixed version 3.1.2 is out - check it out.

I also don't like the default target curve, it's too bass-lean and bright. Try some of the ones suggested, attached mine too if you'd like to try it out. I'd focus my experimentation with target curves to how many dB the bass is boosted and how many dB the treble is rolled-off at 20kHz, these are the essential personal preference customizations that matter most.
 

Attachments

  • f1shb0n3's target curve +4dB 27Hz -3.0dB 20kHz.targetcurve.txt
    247 bytes · Views: 242

GimeDsp

Senior Member
Joined
Oct 30, 2020
Messages
418
Likes
362
Location
Earth
If you are used to a huge bass boost it will take a while to get used to corrected sound.
However i will say this. dirac has many flaws and it had worst sound of all the auto call I have used.
 

Stefan

Member
Joined
May 29, 2021
Messages
67
Likes
13
The fixed version 3.1.2 is out - check it out.
And please report back, because on my side DL 3.1.2 is not fixing it (refering to the same thread as stated above).
 

GimeDsp

Senior Member
Joined
Oct 30, 2020
Messages
418
Likes
362
Location
Earth
Dirac has serious issues.
When I get time I would like to email a list of changes.

1. Add option to not boost any levels
2. Add max boost option

Dirac will boost nulls up 20db, not good.

EDIT, I should be more clear, Dirac is missing many features and does things I consider "bad". Dirac should have AT LEAST the options REW has in filter design. The fact it doesn't makes Dirac, IMO, a gimmick and not a fully functional solution.

However, with these features added I would recommend it to others, for now I can't.
 
Last edited:

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
If you are used to a huge bass boost it will take a while to get used to corrected sound.
However i will say this. dirac has many flaws and it had worst sound of all the auto call I have used.
If you are used to a huge bass boost it will take a while to get used to corrected sound.
However i will say this. dirac has many flaws and it had worst sound of all the auto call I have used.
Correcting reflections higher than 80 Hz from listening position always makes the sound worse in some ways . And thats unfortunately regardless of the room correction programe/hardware.
 

GimeDsp

Senior Member
Joined
Oct 30, 2020
Messages
418
Likes
362
Location
Earth
Correcting reflections higher than 80 Hz from listening position always makes the sound worse in some ways . And thats unfortunately regardless of the room correction programe/hardware.

That's another issue, With Dirac if you want to create a full range "tilt" or house curve it will boost and cut everything to get there, unlike a global tone control that will tilt at specific place and not boost or cut to correct things.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,687
Likes
4,068
Just tame the 32Hz peak to begin with, with a -15dB, Q=6, and see if it improves the sound.
Then try adding -10dB, Q=6, at 65Hz.
In any case, for the moment, don't touch over Schröder frequency (around 250Hz).
 

GimeDsp

Senior Member
Joined
Oct 30, 2020
Messages
418
Likes
362
Location
Earth
Here is the problem with Dirac, for it to do things that don't completely screw up the natural and proper sound you need a FULLY treated room.

Once you have a fully treated room you are only correcting below 400hz with DSP.

So really Dirac should take measurements of all speakers and see itself/market it as "speaker correction" and not room correction.

I have used Dirac in many set ups and above 400hz everything it has every done was bad.
GLM on the other hand seems to get it.

I'm glad I bought the miniDSP ddrc-24 with Dirac because it was a minor investment to figure out I don't like anything Dirac does.
 
Top Bottom