Dirac Reviews

[Keith_W]

The biggest variable when it comes to DSP room correction is the user. If you use the software wrongly, or give it bad measurements, you will get rubbish results. Dirac is highly automated which has good and bad points - good in that it produces good results for most people most of the time, bad in that if you want more control, you can't have it. The outcome does not necessarily reflect the efficacy of the software, it just as much reflects the skill of the user. A better review of Dirac would involve feature comparison to the competition (Acourate and Audiolense), ease of use, etc.

 

[HarmonicTHD]

I have hundreds of rew measurements without / with d Dirac and every time in exact same context / parameters (same room, same mic placement, same speakers placement, same crossover) across various situations (speakers, rooms, crossovers).
I did not post any here (away from home) and only summarised the outcome, which is always very positive.

I have the same, many REW measurements , using MultiEQX. I just can’t objectively compare the two as I don’t have measurements of each. (But with a bit of tweaking MQX works well too. I just don’t go on the internet claiming it is better or as good as xyz because, as said above, the objective data isn’t there).

 

[GXAlan]

. They have bought it because they like it so I can see why they would be put off by Dirac if it changes a lot of what they liked in the first place. It can easily be that the corrected response is much better, objectively, but that they just don't like it that way.

That’s true (that the speakers someone buys is likely to have something that they like about the speaker).

I am just saying that the jury is out on transition frequency and EQ. People who don’t like Dirac might be correcting full range and better off correcting only the lowest frequencies.

I also read about people who prefer other room EQ, suggesting that there may be better algorithms for addressing different types of issues.

 

[Deleted member 48726]

The biggest variable when it comes to DSP room correction is the user. If you use the software wrongly, or give it bad measurements, you will get rubbish results. Dirac is highly automated which has good and bad points - good in that it produces good results for most people most of the time, bad in that if you want more control, you can't have it. The outcome does not necessarily reflect the efficacy of the software, it just as much reflects the skill of the user. A better review of Dirac would involve feature comparison to the competition (Acourate and Audiolense), ease of use, etc.

Dirac can only calculate a good result from good data. And maybe you need to experiment a bit to get the result you want. Just because it is automated doesn't mean you just push the "good sound button".

 

[Deleted member 48726]

That’s true (that the speakers someone buys is likely to have something that they like about the speaker).

I am just saying that the jury is out on transition frequency and EQ. People who don’t like Dirac might be correcting full range and better off correcting only the lowest frequencies.

I also read about people who prefer other room EQ, suggesting that there may be better algorithms for addressing different types of issues.

That's just speculation. It may just as well come down to the question of room curve taste..

 

[mr-audio]

Great topic for conversation. I expect a vast range of experiences from trash to can't live without it. The reason for the wide range of responses is due to 1) Dirac's biggest misconception that it's an easy button with a minimal learning curve. (I'm guilty of this) and 2) user error. (guilty) This in some cases is not really the user's fault. I have seen countless people struggle with Dirac (some give up) simply because the current software version doesn't jive with their hardware and they needed to switch versions. The user shouldn't have to dig for this. Why not a pop-up "we see you're using xxx AVR we recommend xxx software with your device." or why even make it possible for people with those devices to update to a software version that's not compatible with their device in the 1st place.

The biggest "gripe" I've seen with Dirac is "it kills my bass".
This is especially true with Nad and Onkyo users. Like the above started, they are probably on the wrong software. I've also seen workarounds in the volume calibration section specifically for some AVR's. Again, not user friendly but the info is out there.

This complaint about bass in most cases where subs are capable and positioned correctly happens because 1) the user doesn't fully grasp the target curve section of Dirac. or 2) The user did not perform the volume calibration correctly. (guilty) An example of #2 would be the user has their subwoofers trim post-Dirac at the maximum allowed for their application. Some are -10 others I think are -15. Example of a 2.2 system that incorrect FL +1 , FR 0, Sub 1 -10, Sub 2 -10. You need to adjust the gain on the back of the subs and run the calibration again. Audy always left me wanting more in the bass department as well, I've always adjusted the subs +4 db or so post eq.

These two guides along with the assistance I've received here and AVS were essential in getting me up and running.
All credit to the guides provided by the guys over at AVS
Speaker level calibration guide
Dirac Guide with target curves

Dirac issue 2 imaging. I had the same experience, maybe not quite to the extreme as others but I preferred the imaging in my 2.3 system better with Dirac off. Happy to say that is not even close to my feelings today. The solution? Adjusting how many measurements I was taking, the measurement layout per Dirac, and the distance apart from each measurement.
Previously I always measured the 13 pt measurement, I think it may be called focused. but I switched to the tightly focused 9pt measurement and only moved the mic max 18 inches away from the 1st measurement. I could not believe the difference this made in my system in my room. The image was the best I've heard in my system and am now very happy. The trade-off to this would be a more narrow "sweet spot". I'm a selfish listener and others could care less so I'm good with this. I now run the 9pt layout but only measure 6 of the 9 pts. If someone's imaging issue still persists it is most likely a timing issue. I would check delays, the impulse screen in Dirac, and check rew with a timing reference. I'm not the most knowledgeable with this but feel it should get people headed in the right direction.

My Dirac experience has been a journey but an overall positive one. In my very challenging room, I'm very happy with the results from Dirac. I really enjoy the customization the target curves provide and the numerous presets you can save. I would say no matter how "good" Dirac gets for a lot of users Audy is going to be a better choice with a much more pleasant user experience with less margin for error.

A) for A this wasn't as easy as I had thought but is now outstanding. : )

Great comments and around and good food for thought! Thanks

 

[mr-audio]

Thanks all for comments thus far...I guess my perception so far is:

- The impression I HAD was that Dirac was 1) very consistent in delivering better sounding systems, 2) using a fairly simple process of measurements and 3) built in algorithms. Perhaps not idiot proof but relatively simple and highly probable to provide improvement.

- The impression I am getting NOW is more that it 'takes time to experiment and optimize' using Dirac versions/options/features and perhaps even brand of equipment, to get good results.

- I also agree with the comments made around human preferences vs measured sound. e.g. 'If you like the sound of your speakers when you bought them, you may prefer to leave it at that and not make any changes'.

BTW I would compare Dirac to Volume based EQ. Its easy, there is science behind it, and the benefit was very clear to me without changing the nature of your system's sound:
--- Even the old 'Loudness' button was nice in some (low volume) situations
--- I like the Yamaha approach results although a bit too much manual effort to use
--- I loved my old DBX, and so did everybody who heard it, especially on FM and Tape, but even on LPs (never used on Digital, don't need I don't think)

So I am still interested in the Dirac perceptions, and feeling that it requires far more 'twiddling' than I would want to do, and yet still may results in something SUBJECTIVELY no better than without.

 

[RickyC34]

I have been having imaging and delay issues with the most recent versions (3.6.X and now 3.7.X too)
Version 3.4.4 is supposed to be the latest 'fully stable' version but @Flak might be able to comment on this too
What I do to circumvent it is to perform and save more measurements (up to 5) and then I listen to them and choose the best which is the closest to when Dirac is off
This is definitely something that shall be fixed by Dirac ASAP

With all due respect, don't hold your breath. 3.4.4 came out a YEAR ago.

However there is another issue since version 3.6.X - the 'exclamation mark issue': when performing measurements, you see an exclamation mark instead of a checkmark in the measurement 'bubble'. This is bad since in that case Dirac only performs min phase correction instead of the full mixed phase correction (especially audible in the bass)
I simply could not measure my smaller system since everytime I try it I run into this issue. With my main system I get this issue in 50-60% of the cases (still really bad)
Again, this is another bug that should be fixed ASAP (note: in version 3.4.4 this issue is non-existing but the sound is waaay better with version 3.7.X in every other regard)

What hardware are you using to run Dirac? If AVR, which one? I'm running 3.7.X on Mac with Dirac Room Correction Suite and have no ! when taking measurements.
Since you're having issues with 3.6 and 3.7 why not rollback the software to 3.4.4 where you didn't have any issues? I would expect your results would be more consistent and accurate.

 

[ppataki]

What hardware are you using to run Dirac? If AVR, which one? I'm running 3.7.X on Mac with Dirac Room Correction Suite and have no ! when taking measurements.

On my main system I am using a PC with Windows 11 and on the small system I have a Dell laptop with Windows 11 too
On my main system I am using 3.7.3 since I have the exclamation mark issue 'only' in like 50-60% of the cases (as mentioned above) and on my small system I indeed rolled back to 3.4.4 since I was unable to perform any successful measurements with 3.7.3

I am using a brand new UMIK-1 mic (I had an older one that died after 5 years of service but I also had the exclamation mark issue + the imaging issue with that too)

 

[RickyC34]

On my main system I am using a PC with Windows 11 and on the small system I have a Dell laptop with Windows 11 too
On my main system I am using 3.7.3 since I have the exclamation mark issue 'only' in like 50-60% of the cases (as mentioned above) and on my small system I indeed rolled back to 3.4.4 since I was unable to perform any successful measurements with 3.7.3

I am using a brand new UMIK-1 mic (I had an older one that died after 5 years of service but I also had the exclamation mark issue + the imaging issue with that too)

Interested to see if others are having the ! issues with the newly released software. My curiosity as to why this is happening and worries of an underlying issue would make me contact Dirac. They may at least provide the peace of mind that something in your system is not incorrect.
The only thing that comes to mind you may want to double-check is - During the volume calibration page are you able to achieve a sound floor level close to -50 on the mic? Are the speakers then set to at least 20 dB above the noise floor?
From the sounds of things, you have a 2.0 system. Is this correct? May I ask what Dac you're using?

 

[Steve Dallas]

B- with the default target curve
A+ with my custom target curve

I take dozens of measurements as I set up a Dirac system and hammer the results into what I want them to be. I have posted the results in several other threads.

 

[ppataki]

Interested to see if others are having the ! issues with the newly released software. My curiosity as to why this is happening and worries of an underlying issue would make me contact Dirac. They may at least provide the peace of mind that something in your system is not incorrect.
The only thing that comes to mind you may want to double-check is - During the volume calibration page are you able to achieve a sound floor level close to -50 on the mic? Are the speakers then set to at least 20 dB above the noise floor?
From the sounds of things, you have a 2.0 system. Is this correct? May I ask what Dac you're using?

I have already contacted them; they told me what I wrote above about version 3.4.4 (being the latest fully stable version)
Apparently others have been experiencing this too so they are allegedly working on fixing this bug - fingers crossed!
I reckon it must be a bug with the newer versions since with version 3.4.4 I never ever have this issue on any on my systems - only with versions post that

I am well above the noise floor
I am using a quad system (+see my signature) in my main room and a stereo system in my home office (Sabaj A30A like in my main room but connected directly to the laptop via USB)

 

[Deleted member 48726]

B- with the default target curve
A+ with my custom target curve

I take dozens of measurements as I set up a Dirac system and hammer the results into what I want them to be. I have posted the results in several other threads.

Me too. Measurements.. Typical a one-off affair.. Save the project.
Then tweak four different profiles.. Listening tests, ranking the profiles, tweaking the curve with the best one as baseline, saving three new variants, listening test, ranking, choosing the favourite.
Haven't touched it for six months. Sounds glorious.

 

[bluefuzz]

I have used Dirac with three sets of speakers: a pair of big classic 3-way 70s 'bookshelfs', a pair of MTM mini towers and a pair of Linkwitz LxMinis. All used with a single subwoofer. All DIY. My room is smallish (~23 m2), asymmetrical and full of .. ahem ...'stuff.' In this environment Dirac full-range results are somewhere between A++ and life changing with all three pairs of speakers. There are subtle differences in the results depending on positioning of the 9 or 12 measuring points, which microphone etc. even when using the 'same' room curve. It's also fairly easy to screw up the calibration phase and get garbage results.

That said, I have achieved comparable results using REW and RePhase for measurement and manually implementing FIR/IIR filters with CamillaDSP.

 

[Miniyouuuu]

Dirac affects sound a lot. You have to know what you want.
My first Dirac filter (90º Mic position, 13 point measurement, issues with low volume in some positions, default targetcurve and full range) was a disaster for my 2.0 setup. Only improvement was dialogue clarity for TV. Music sounded dull, no "live experience", and sometimes seemed fake sound, with too much conversion.

After that I tried to lower upper conversion up to 500/800hz. Music sounded less fake. Below 500hz you are trying to solve room modes, above that you are affecting Speaker "personality".

Later I tried, 0º Mic, 5 point measurement, with no low volumen issues, and I implement a Harman targetcurve found in this forum. Night and Day. With that measurements I created 2 filters: fullrange one for TV and up to 800hz filter with same targetcurve. Sometimes, for 60/70s songs I prefer the unaffected upper range caracter of the later.

I also tried REW in parallel. During my first Dirac Test, REW gave to me better results. Nowadays, after some Dirac fixings, I tend to prefer Dirac. PEW filters are perfect, but FIR filters with Rephase are very Tricky. Dirac gets both with easy.

 

[levimax]

I am a long time REW/ Rephase user with DIY near full range speakers which I recently added 2 subs to. I was happy with the DIY REW/Rephase results but was always curious about DIRAC. I tried it with the free trial and liked it enough that I bought it (full version with Bass Control) on the black Friday sales. This review is of the latest version 3.7.3 which seemed to me to be the "recommended" stable version on a Windows 10 machine.

1. RE "!" issue. I am not sure if this is a bug or a feature. Through the course of several measurement sessions I got the "!" several times but I just "re-ran" the measurement and it worked. Several times I think this was caused by "other sounds" during the sweep so maybe the newer version checks more closely for measurement errors? Not sure but for me it was not really an issue outside of a few "re-measures".

2. RE: Results compared to my REW/ Rephase filters. The DIRAC ones "looked better" on a REW "MMM" measurement (smoother response).

3. RE: Listening I preferred the DIRAC filters (Using Harmon Curve) especially for the mid-range.

4. RE: Filter gain. With my REW/ Rephase filter I had a problem with "digital clipping" when running a 0 dB sweep and had to reduce digital volume to prevent it which also reduced the loudness of my system considerably (-8 db). The DIRAC filters don't seem to have the digital clipping issue when running a 0 dB sweep and plays much louder, this is kind of a big positive deal for me.

5. RE: Flexibility to fine tune the filters. DIRAC is way easier and faster to take a set of measurements and apply different house curves to them and to compare them. Good for knowing your preferences, bad because it can encourage OCD endless tweaking behavior.

6. RE: Cost. Seems expensive to me especially since REW and Rephase are "free".

Overall A-

 

[mr-audio]

Quick update on my experience since I started this thread:

I contacted Dirac support and I must say they were very accommodating. They even allowed me to move my full spectrum license from one amp/room to another amp/room.

I made more tests and although they were better, my conclusion was pretty much the same.

I certainly noticed some differences, but it was more to me, like making the sound less stereophonic and more mono sounding. Rooms and amps I could argue that the modified sound was better for background listening because it was smoother and less dynamic perhaps.

But in critical listening and even blind A/B comparisons (using a friend to switch w/o my knowledge), the unmodified sound has been more dynamic with more separation.

In conclusion, it certainly isn’t for everybody, but if it works for you that’s great!

Happy New Year

 

[Rottmannash]

Great topic for conversation. I expect a vast range of experiences from trash to can't live without it. The reason for the wide range of responses is due to 1) Dirac's biggest misconception that it's an easy button with a minimal learning curve. (I'm guilty of this) and 2) user error. (guilty) This in some cases is not really the user's fault. I have seen countless people struggle with Dirac (some give up) simply because the current software version doesn't jive with their hardware and they needed to switch versions. The user shouldn't have to dig for this. Why not a pop-up "we see you're using xxx AVR we recommend xxx software with your device." or why even make it possible for people with those devices to update to a software version that's not compatible with their device in the 1st place.

The biggest "gripe" I've seen with Dirac is "it kills my bass".
This is especially true with Nad and Onkyo users. Like the above started, they are probably on the wrong software. I've also seen workarounds in the volume calibration section specifically for some AVR's. Again, not user friendly but the info is out there.

This complaint about bass in most cases where subs are capable and positioned correctly happens because 1) the user doesn't fully grasp the target curve section of Dirac. or 2) The user did not perform the volume calibration correctly. (guilty) An example of #2 would be the user has their subwoofers trim post-Dirac at the maximum allowed for their application. Some are -10 others I think are -15. Example of a 2.2 system that incorrect FL +1 , FR 0, Sub 1 -10, Sub 2 -10. You need to adjust the gain on the back of the subs and run the calibration again. Audy always left me wanting more in the bass department as well, I've always adjusted the subs +4 db or so post eq.

These two guides along with the assistance I've received here and AVS were essential in getting me up and running.
All credit to the guides provided by the guys over at AVS
Speaker level calibration guide
Dirac Guide with target curves

Dirac issue 2 imaging. I had the same experience, maybe not quite to the extreme as others but I preferred the imaging in my 2.3 system better with Dirac off. Happy to say that is not even close to my feelings today. The solution? Adjusting how many measurements I was taking, the measurement layout per Dirac, and the distance apart from each measurement.
Previously I always measured the 13 pt measurement, I think it may be called focused. but I switched to the tightly focused 9pt measurement and only moved the mic max 18 inches away from the 1st measurement. I could not believe the difference this made in my system in my room. The image was the best I've heard in my system and am now very happy. The trade-off to this would be a more narrow "sweet spot". I'm a selfish listener and others could care less so I'm good with this. I now run the 9pt layout but only measure 6 of the 9 pts. If someone's imaging issue still persists it is most likely a timing issue. I would check delays, the impulse screen in Dirac, and check rew with a timing reference. I'm not the most knowledgeable with this but feel it should get people headed in the right direction.

My Dirac experience has been a journey but an overall positive one. In my very challenging room, I'm very happy with the results from Dirac. I really enjoy the customization the target curves provide and the numerous presets you can save. I would say no matter how "good" Dirac gets for a lot of users Audy is going to be a better choice with a much more pleasant user experience with less margin for error.

A) for A this wasn't as easy as I had thought but is now outstanding. : )

Could you elaborate on your comment re: " the current software version doesn't jive with their hardware and they needed to switch versions. The user shouldn't have to dig for this. Why not a pop-up "we see you're using xxx AVR we recommend xxx software with your device." I have the RZ50 and am using an older version of DL and have heard of others experiencing issues w/ newer versions. Is this what you're referring to?

 

[Johnp]

I certainly noticed some differences, but it was more to me, like making the sound less stereophonic and more mono sounding. Rooms and amps I could argue that the modified sound was better for background listening because it was smoother and less dynamic perhaps.

But in critical listening and even blind A/B comparisons (using a friend to switch w/o my knowledge), the unmodified sound has been more dynamic with more separation.

The difference you describe is clear to me. It’s likely that phase and time alignment are making a perfect match, removing what can be described as “spaciousness”. It may be that it’s more accurate with Dirac full range, but it also sounds somewhat artificial to me.

 

[Miniyouuuu]

The difference you describe is clear to me. It’s likely that phase and time alignment are making a perfect match, removing what can be described as “spaciousness”. It may be that it’s more accurate with Dirac full range, but it also sounds somewhat artificial to me.

I agree. Full range Dirac sounds artificial, but you have to chance to edit curtains, and adjust targetcurve manually. The more difference between target and measures on each channel, the more artificial the sound gets. Any variation in the range from 500hz to 20000Hz will affect perceived sound. Voices and instruments sound pitched.

Dirac is perfect for subwoofer integration and room mode correction. Better than REW. By affecting delays and alignement of speakers clearly improves clarity. AV receivers with Dirac improve the sound experience for films, but take care for Music listening.

I suggest 2 sets: film set with full range Dirac, and upto 500/800hz set for music. In both cases try to avoid any +/-3dB between measurement and manually adjusted targetcurve after 100hz.

Dirac take many spot measurements and generate a “phantom” measurement, the one you will use as room reference. Small listenings rooms, asymetric rooms, or close to backwall listening spots retreive very different measurements from spot to spot. Be carefull with this, as “phantom” image is a “media” of all these spots. In order to make a nice experience on all spots, Dirac could kill the main spot listening experience. In my case I got better results with narrow measurement in up to 5 spots.