Yes, particularly NAD at the time of my post. 3.4.4 was the recommendation for NAD AVRs for a very long time, I believe a recent software upgrade "fixes" this. You can go to the Dirace knowledge base center and then Dirac software change log to check. If you're not having issues and there are no new "features" that have been released in the software changelog you desire, I wouldn't upgrade your software. If it's not broken don't fix it.
-
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 Reviews
- Thread starter mr-audio
- Start date
-
- Tags
- dirac dirac live
Steve Dallas
Major Contributor
For your information...
Be careful of the SPL differences when comparing Dirac correction to uncorrected sound. Dirac pulls down most of the FR, including midrange and high frequency peaks, which makes the overall SPL lower, which makes Dirac correction sound less "exciting." I usually have to bump the volume few dB when switching to the Dirac-corrected slot.
Also...
My first experience with Dirac was the standalone product for PC. After buying a DLBC license for my Denon 4800, I understand its performance is more reliant on the DSP horsepower than I realized. It works fine on my AVR, but I am not able to achieve results as good with the AVR in the same room driving the same speakers.
Be careful of the SPL differences when comparing Dirac correction to uncorrected sound. Dirac pulls down most of the FR, including midrange and high frequency peaks, which makes the overall SPL lower, which makes Dirac correction sound less "exciting." I usually have to bump the volume few dB when switching to the Dirac-corrected slot.
Also...
My first experience with Dirac was the standalone product for PC. After buying a DLBC license for my Denon 4800, I understand its performance is more reliant on the DSP horsepower than I realized. It works fine on my AVR, but I am not able to achieve results as good with the AVR in the same room driving the same speakers.
A+ for Dirac full range on a Nad T778.
Sure, at first I am missing this thick bass - which were just room modes that made me think I had a good bass reproduction. Above 500 kHz live and death is the room curve. Trying to make the system flat by believing high frequencies should reach the ear at even level as the lower octaves results in a dull sound. That was a huge surprise for as I believed in the old style flat idea.
Sure, at first I am missing this thick bass - which were just room modes that made me think I had a good bass reproduction. Above 500 kHz live and death is the room curve. Trying to make the system flat by believing high frequencies should reach the ear at even level as the lower octaves results in a dull sound. That was a huge surprise for as I believed in the old style flat idea.
I'm having an issue with deep bass with DBLC (two subs). How do I set the software volume vs hardware volume on the sub? Do I set the hardware volume at max and let the processor/dirac set the actual volume, or do I play with them? ie. do I let the subs have full gain of a small signal or try to balance them with lower gain but higher input?
ZolaIII
Major Contributor
- Joined
- Jul 28, 2019
- Messages
- 4,195
- Likes
- 2,475
This is ridiculous! You are comparing automatised room correction system in this case Dirac with nothing. It's much better than nothing that much is certain but compared to fine tuning by hand with it or all the way on your own and to the extent which Dirac doesn't support (just assume at best) it doesn't get even near. Many buy such for simplicity and the fact that it does work and gives good results (especially compared to nothing) and defend their purchases based on personal bias. Not even ARC which is much more complex (mathematical) can cope with fine tuned by hand DSP set of filters and effects but then again almost no one is really ready to deel with it all by hand (honor to the exceptions).
Keith_W
Major Contributor
After reading @ZolaIII's message I am pretty sure he is responding to another post, and not to you.
Anyway, to answer your question in post #44 re: hardware vs. software volume control, what you should do is set the VC on your sub to be high enough so that Dirac can cut the volume as required. This may/may not be at max depending on your situation. In my system, if my subs are at max, the whole house rattles and I get unreliable measurements. So I set my sub to be 15-20dB higher than my mains, which gives plenty of headroom for the software to cut volume to fill in dips.
ZolaIII
Major Contributor
- Joined
- Jul 28, 2019
- Messages
- 4,195
- Likes
- 2,475
To answer both I whose talking generally not addressed to anyone individually.
I do it by hand and differently. First I EQ the room mode on mid SPL levels (so that modes are present but not going wield) for sub chenel/chenels, then model the slape (for crossover) and frequency response, and I do the same for mains (as it's 2.2 system) per channel of course and only then match them (physical or digital gain), do the white noise calibration and enable loudness and that's about it, from there I use digital volume.
I recently got DIRAC DLBC for my 2.2 system and it managed to beat my best "manual" attempts at sub integration. I am sure it is possible to do a better job manually but it takes a lot of experience to know what to do and what not to do and how what you measure sounds to you. Measuring what DIRAC actually does it seems like it makes reasonable choices like knocking down peaks, leaving dips caused by room modes and filling in ones that respond to EQ. It is also nice in that you can use one set of measurements to quickly and easily create and compare different filters to fine tune things to your tastes and to educate yourself on what different measurements sound like to you. If there is any "secret sauce" to their algorithm it seems like it does some work on the midrange to make vocals clear and easy to understand, probably because of its use for Home Theater.
In case I can save someone some time I first added DLBC to my 3 way + 2 subs system. The mains are DIY and used analog active crossovers. Getting DIRAC to work with this setup was easy and straightforward. Since I bought a Motu Ultralite Mk5 with 10 channels I decided to ditch the analog crossovers on the 3-way mains and do all digital crossovers. I tried EKIO, Camilla DSP, Equalizer APO, and deFonica to create the crossovers. For me deFonica was by far the easiest for crossovers and it is free but some of the other programs had more flexibility but were less "dedicated" crossover software. All in all it was easier than I thought to transition to the digital crossovers.
Where things get harder is trying to use DLBC with the digital crossovers on the mains and then separately having DIRAC DLBC control the sub. Using JRiver or Foobar for the crossover won't work because DIRAC needs to be "upstream" of the digital crossovers. In my case I use Foobar to DIRAC and then sent the "main" output to deFonica (which sent 6 channels to the Motu) and then tried to rout the DIRAC sub channels to the MOTU but DIRAC sees this as a different "device" so it won't work. I then came across VBMatrix software https://vb-audio.com/Matrix/ which is made by the Voicemeeter people, this software, unlike the other Voicemeeter software which has tons of other features I don't need or understand, it is dedicated to "sound routing" and virtual device creation. I managed to combine the deFonica ASIO sink driver with the Motu Ultralite ASIO driver into a virtual ASIO device which routed the channels where they needed to go on the Motu and then DIRAC saw this as one device and allowed me select it and surprisingly it worked. One other issue I came across was originally I set up the crossover filters in deFonica as FIR but when I took measurements in DIRAC the delay was way too much and it wouldn't work right, changing the crossover to IIR solved this.
Now it is all working and I got rid of a whole bunch of boxes and all my drivers are "time aligned" and the crossover have acoustic and electrical slopes taken into account and everything is dead quiet. I wish I could say that my system now sounds "night and day" better but I can't. It does sound subtly better and I am enjoying listening to my favorite songs and my "rack" is now down to a PC, the Motu, and the amps.
Where things get harder is trying to use DLBC with the digital crossovers on the mains and then separately having DIRAC DLBC control the sub. Using JRiver or Foobar for the crossover won't work because DIRAC needs to be "upstream" of the digital crossovers. In my case I use Foobar to DIRAC and then sent the "main" output to deFonica (which sent 6 channels to the Motu) and then tried to rout the DIRAC sub channels to the MOTU but DIRAC sees this as a different "device" so it won't work. I then came across VBMatrix software https://vb-audio.com/Matrix/ which is made by the Voicemeeter people, this software, unlike the other Voicemeeter software which has tons of other features I don't need or understand, it is dedicated to "sound routing" and virtual device creation. I managed to combine the deFonica ASIO sink driver with the Motu Ultralite ASIO driver into a virtual ASIO device which routed the channels where they needed to go on the Motu and then DIRAC saw this as one device and allowed me select it and surprisingly it worked. One other issue I came across was originally I set up the crossover filters in deFonica as FIR but when I took measurements in DIRAC the delay was way too much and it wouldn't work right, changing the crossover to IIR solved this.
Now it is all working and I got rid of a whole bunch of boxes and all my drivers are "time aligned" and the crossover have acoustic and electrical slopes taken into account and everything is dead quiet. I wish I could say that my system now sounds "night and day" better but I can't. It does sound subtly better and I am enjoying listening to my favorite songs and my "rack" is now down to a PC, the Motu, and the amps.
This is the best guide I have found. On step 7 I turn the sub gain up on the back of the sub until it is +5db louder than the softest speaker. I had good results with the +8 Harman curve. Apply the curve to all speaker groups and subs. Before saving I would suggest using the dotted vs shelf editor. dirac-live-2-mic-speaker-level
A+ now, but it took some laps.
All measurements done with a Onkyo TX-NR7100's built-in Dirac Live and 5.0 layout.
Main EQ reason are some large room modes, over +15dB.
- First try with the accessory mic. Although advertised as calibrated, it had an obvious bad frequency response.
- Measurements with the UMIK-1 looked more reasonable. Expect a wiggled tilted line.
My listening area is quite asymmetrical, the main place almost on one edge.
Additionally, one side is a large living room closet, the other side open to the rest of the room.
The room modes were gone, but imaging worse.
- Tried partial correction up to 300Hz. Lively, but also a bit "phasey".
The transition to non-corrected looks good on the graph though.
- Re-did the 13-point measurements on a more focused and symmetrical area.
Loudspeaker levels corrected, Dolby Surround, Center spread on.
Big change, balance and imaging very natural now for 2.0 and all other material.
And less of a disaster for the other places as thought.
Target curve is, as always, a question of taste. I use some cues for credibility:
- LFE, low bass: effects/umph. Depends.
- mid bass: male and female voices, outdoor (too thin?) and in a close room (too boomy?)
- tilt: besides general too bright or dark, does dialog sound recessed or leaning forward?
Usually sound is most natural if you preserve the measured slope in the mid/upper
frequency range.
I was not able to jointly optimize the last two points using the DL GUI (bass-flat-treble) controls.
So far happy with Bruel&Kjaer / Bob Katz / Quad amp family of tilted curves.
All measurements done with a Onkyo TX-NR7100's built-in Dirac Live and 5.0 layout.
Main EQ reason are some large room modes, over +15dB.
- First try with the accessory mic. Although advertised as calibrated, it had an obvious bad frequency response.
- Measurements with the UMIK-1 looked more reasonable. Expect a wiggled tilted line.
My listening area is quite asymmetrical, the main place almost on one edge.
Additionally, one side is a large living room closet, the other side open to the rest of the room.
The room modes were gone, but imaging worse.
- Tried partial correction up to 300Hz. Lively, but also a bit "phasey".
The transition to non-corrected looks good on the graph though.
- Re-did the 13-point measurements on a more focused and symmetrical area.
Loudspeaker levels corrected, Dolby Surround, Center spread on.
Big change, balance and imaging very natural now for 2.0 and all other material.
And less of a disaster for the other places as thought.
Target curve is, as always, a question of taste. I use some cues for credibility:
- LFE, low bass: effects/umph. Depends.
- mid bass: male and female voices, outdoor (too thin?) and in a close room (too boomy?)
- tilt: besides general too bright or dark, does dialog sound recessed or leaning forward?
Usually sound is most natural if you preserve the measured slope in the mid/upper
frequency range.
I was not able to jointly optimize the last two points using the DL GUI (bass-flat-treble) controls.
So far happy with Bruel&Kjaer / Bob Katz / Quad amp family of tilted curves.
Last edited:
I have Dirac Live in a SHD (MiniDSP) 4 channels out to poweramp/speakers and two subs. I find it outstanding compared to Audessy on a Denon receiver.
This is an "apples to apples" comparison as the speakers/subs are the front of my Atmos system. I switch line out to change system type, as other people here do.
Dirac Live plus adding a second sub solved my base management problems. I now have big appropriate bass without room reflection issues.
Another improvement I particularly hear is more instrument separation with Dirac. But I also hear problems in recording I know well that were as much revealed in my Denon/Audessy system. In particular the sound stage can be unrealistically big with Dirac. I assume this comes from the recording being enginnered to sound big on lesser systems.
I also think the SHD may contribute by being an excellent crossover between my stereo bookshelf main speakers and the subs. While it is now pricey for four channels I don't regret the purchase.
This is an "apples to apples" comparison as the speakers/subs are the front of my Atmos system. I switch line out to change system type, as other people here do.
Dirac Live plus adding a second sub solved my base management problems. I now have big appropriate bass without room reflection issues.
Another improvement I particularly hear is more instrument separation with Dirac. But I also hear problems in recording I know well that were as much revealed in my Denon/Audessy system. In particular the sound stage can be unrealistically big with Dirac. I assume this comes from the recording being enginnered to sound big on lesser systems.
I also think the SHD may contribute by being an excellent crossover between my stereo bookshelf main speakers and the subs. While it is now pricey for four channels I don't regret the purchase.
I am not sure what DIRAC is doing with the soundstage but I was playing around with "Pink Noise" and turning the DIRAC filter on and off and was surprised to hear the "soundstage" of Pink Noise become much wider with DIRAC on. To me it sounds like there is some "out of phase" signals being added in to "widen" the soundstage "artificially", this is similar to what happens with LP's which have added "out of phase" noise which makes the soundstage sound wider. If DIRAC is adding in out of phase signals I do like the effect but I haven't seen it documented and it is not making things more accurate to the source.
- Joined
- Mar 18, 2023
- Messages
- 669
- Likes
- 1,039
Should be exactly the opposite, probably the measurements were wrong or it assumed the distance of speakers innacurately, you can turn off the delay it introduces to hear if it's better
Nice idea, this poll. My experience:
- Amplifier doesn't matter (hopefully we can all agree on that).
- Speakers play a major role.
- Room plays an even bigger role.
- Home cinema (somewhat acoustically optimised): NAD T778, Canton Reference K speakers, 5.1.4 setup: Result: significantly better, better than Audyssey, which I used before, big improvement (B+).
- Living room (a room with lots of books etc.): NAD M10 V2, Dynaudio Heritage Special. Rather small improvement, worth a little, but not much (B-).
- Home Office (acoustically poor, sloping roof, desk in the corner): Elac 310.2, FosiAudio V3, MacBook Air M2 with DiracLiveProcessor, headphone output: huge improvement, a real ear-opener (A++).
- the less favourable the room, the more useful Dirac is.
- I wouldn't run a system without Dirac any more, it's the one part of the overall system that makes the most difference compared to other components.
There is no delay set and turning it on and off made no difference. I tried the same thing with a 1 Khz sine wave tone (from REW direct to DIRAC Live Processor on PC) and got the same thing. With no filter tone image was locked in dead center, with Dirac more "diffuse" and "wide" image. I don't think there are any issues with the measurements and I am happy how it sounds and testing with REW RTA I get a much smoother response with the filter on than off. Something could be "wrong" or a little out of phase signal could be part of DIRAC's secret sauce. I would be curious if others notice similar.
- Joined
- Mar 18, 2023
- Messages
- 669
- Likes
- 1,039
there's no secret sauce and it should be dead center
Could you share the measured+corrected curves?
If you have a quite ondulated mid/hi frequency response, Dirac can do more harm trying to linearize it.
Same if there is sufficient difference between L and R in this range.
A wider soundstange from the (mono signal I suppose) pink noise points to more differences between left and right.
Cause could be a smaller room having more reflections, bringing more off-axis signals to what the calibration mic hears.
Correcting these will deform an otherwise flat direct response which makes up the perceived sound.
Something similar happened to me.
Similar threads
