Dirac's New Auto Target Curve Delivers a Simplified Approach to Better Sound

(March 29, 2022) Dirac has announced a new Auto Target Curve functionality for Dirac Live, making it easier for enthusiasts to achieve the best possible sound quality from their reference system. Dirac Live's Auto Target Curve automatically generates a target curve based on a system's unique measurements, radically reducing the amount of in-depth knowledge needed to properly fine tune audio output. This release can be enjoyed by owners of NAD, Onkyo, Pioneer, Integra, and Pioneer Elite gear that's compatible with the Dirac mobile app. Dirac says that Auto Target Curve is being launched on its mobile platform first, with desktop functionality due to launch later this year.

"Many of the first uses of Dirac Live—more than a decade ago—were in cinemas or high-end home theaters, where experts calibrated and tuned sound systems using Dirac Live. But that's no longer the case," said Jakob Ågren, Head of Product Management at Dirac. "With more mass-market AVRs now offering our breakthrough solution, Dirac Live has gone mainstream. However, when addressing this new, larger audience, it became clear that attaining the optimal sound experience should be as easy as possible. Our new auto target curve eliminates the need for manual adjustments, allowing new users to sit back and enjoy superior sound."

The new Auto Target Curve feature is designed to faithfully reflect the inherent character of a speaker system, minus the adverse acoustic effects of the room. It gives enthusiasts access to a balanced response with a more natural slope capable of preserving a speaker's individual flavor, while tightening up performance across the board. Should further tweaking be required to achieve optimal sound, enthusiasts can manually adjust Auto Target Curve through a new simplified interface in the Dirac Live mobile app. Dirac says the interface provides a simple drag feature to increase or decrease output. And for old-school fans of Dirac's classic filter design, that approach will remain available for use.

For more information, visit www.dirac.com.

 

[Mike-48]

If this works well in practice, it seems like a good move for Dirac.

 

[JStewart]

So, what happens with a surround system having speakers with different responses?
@Flak , might you know?

 

[AudiocRaver]

A great offering for newcomers to room correction, and those who have developed their own fave target curves over the years might learn a few new tricks from Dirac's engineers as well!

 

[NBPK402]

I wish it would work with my MiniDSP DDRC88A...

 

[DanDan]

I have researched The Target Curve actively for a couple of years. This was a professional use, i.e. attempting to impose a tonality in a Studio CR which would sound essentially the same in several other rooms in the building, on different playback systems.
It was intriguing how different way points and slopes led to success or failure. Ultimately I ended up with a curve based on the original Bruel and Kjaer research. 0dB at 30Hz +3dB at 100Hz, 0dB at 2KHz, -3dB at 10KHz, -6dB at 20KHz. This concurs nicely with Harman. I don't like the look of that Dirac Curve above, nor the 'recommended' one it has had all along.

 

[JStewart]

I have researched The Target Curve actively for a couple of years. This was a professional use, i.e. attempting to impose a tonality in a Studio CR which would sound essentially the same in several other rooms in the building, on different playback systems.
It was intriguing how different way points and slopes led to success or failure. Ultimately I ended up with a curve based on the original Bruel and Kjaer research. 0dB at 30Hz +3dB at 100Hz, 0dB at 2KHz, -3dB at 10KHz, -6dB at 20KHz. This concurs nicely with Harman. I don't like the look of that Dirac Curve above, nor the 'recommended' one it has had all along.

I guess we need different curves for different rooms and speakers. The B&K doesn't work for me, nor have the Harmon curves with accentuated bass. For me a straight line that fits the high frequency slope works best. For me that's 0.65dB per Octave. Dirac standard is 0.5dB per Octave.

The concern in my above post is if Dirac will be producing different targets for different speaker groups. I expect that a surround sound field would be less cohesive if it did, but I really don't know anything other than the current dogma which has always been same target for all speakers, including the sub.

 

[Mike-48]

I guess we need different curves for different rooms and speakers.

Yes, speakers of different directivity IME need different targets.

And recordings. I have several targets with different treble cuts for different recordings. Oh, and one with a bass cut, too. A lot of that can be done with tone controls if available, but sometimes easier with various targets.

 

[Randy L]

I have researched The Target Curve actively for a couple of years. This was a professional use, i.e. attempting to impose a tonality in a Studio CR which would sound essentially the same in several other rooms in the building, on different playback systems.
It was intriguing how different way points and slopes led to success or failure. Ultimately I ended up with a curve based on the original Bruel and Kjaer research. 0dB at 30Hz +3dB at 100Hz, 0dB at 2KHz, -3dB at 10KHz, -6dB at 20KHz. This concurs nicely with Harman. I don't like the look of that Dirac Curve above, nor the 'recommended' one it has had all along.

Hey, just for fun I tried the curve you mentioned on my DIRAC (0dB at 30Hz +3dB at 100Hz, 0dB at 2KHz, -3dB at 10KHz, -6dB at 20KHz) - Wow! This is what I was looking for! Thank you so much.... Great bass, midrange predominance, and relaxed highs. Not too bright. Really follows how the human ear hears. I noticed better fullness and detail without too much harshness, less muddy, can hear more frequency separation throughout... I'm using a Marantz Model 30 amp with Elac UF5 speakers... I think you are really on to something. Thanks for doing years of research! Saved me a lot of time.

 

[Sonnie]

I'm looking forward to seeing what it suggests for my room when it comes out on desktop later this year... could be interesting to experiment with at least.

 

[BenToronto]

I guess we need different curves for different rooms and speakers...

Right. Fallacious to believe the sound of a system in a room is just the freq plot with a given mic pattern. Two rooms with the same plot do not necessarily sound the same unless identical in all acoustically relevant respects.

And so for DanDan's preferred curve, it may indeed be the one that works very well for his environments and for others too. But let's get real here. Except for quick-switch A-B tests, barely perceptible to detect 3 dB change at 100 Hz with any music not hot-to-test for that freq. And if you can detect that difference blind, likely half the time you'd prefer one or the other curve - as Toole sometimes found.

Nice after so many decades of folks touting "flat" FR, that psycho-acoustic realities are now becoming mainstream (at least on this forum).

B.

 

[DanDan]

I recommend we consider the sources here. Bruel and Kjaers study was conducted in Hi Fi Dealers and Hi Fi afficiando's spaces. Let us assume Suspended ceilings, carpets, etc. Probably decent sounding rooms. The playback equipment? I reckon safe to assume decent to very good. The whole chain is measured in each instance. That includes the Room Decay. Now in my case I was Mixing and Mastering in my White Room here at soundsound.ie But I live here so have immediate playback options in 4 domestic rooms, on different systems ranging from iMac to TV Speakers to Spendor BC1s in a pretty reverberant Kitchen Living room. Over years I tweaked my work in progress curve periodically and observed how this influenced what came out in the other rooms. Ultimately I somewhat reluctantly accepted that B&K translates best. This is not necessarily my preferred tonality. But I have gotten to love it's musicality as Randy has picked up on. It is amazing how critical the way points are, as they contain very very broad sloped bands. It is IMO notable that 100Hz and 10KHz are featured. Where did we see that pair of frequencies before? Practically every Hi Fi Amp, Baxandall, the Quad Preamp with its +/-3dB Tilt control. Some time ago a guy set about to create the ultimate listening room. For pleasure not work. Ultimately he ended up having to add some reverb to his highly treated space. But once again the summed response of this system, speakers, room, Haas Kickers, Diffusors, Lexicon Reverb....... look at it...... Familiar? https://gearspace.com/board/studio-building-acoustics/817205-my-listening-room.html

 

[AudiocRaver]

All due respect to DanDan, B&K, Toole, and the respected research of many in the industry, to each his own preferred Target Curve! It never hurts to try one out, and it never hurts to say you prefer something else.

Sonnie, I agree it will be interesting to see the new Recommended Dirac target curves, and to read up on the science behind the derivation of them.

 

[DanDan]

I am trying....unsuccessfully it seems! to say my research was not a search for my preferred curve, but rather one which translated best to multiple playback scenarios. I suggest than any other engineer would experience the same overwhelming consistent similarity.

 

[witwald]

I guess we need different curves for different rooms and speakers.

From this evidence, it would strongly appear that applying DSP room correction above the Schroeder frequency of the room is quite a futile pursuit, irrespective of the marketing stance associated with any given product. It's a technology that is being incessantly tweaked and upgraded, but in its actual application, it does not appear to withstand closer scrutiny. It should really only be applied to flattening the on-axis frequency response curve of a loudspeaker, to achieve reasonable results that actually benefit all listeners independent of the room characteristics. The rest is up to the design of the loudspeaker and the actual listening room, with or without some room treatment and/or any applicable low-frequency room correction.

The B&K doesn't work for me, nor have the Harman curves with accentuated bass. For me, a straight line that fits the high-frequency slope works best. For me, that's 0.65dB per Octave. Dirac standard is 0.5dB per Octave.

The B&K curve is a little nonlinear. It is relatively flat in the decade between 20Hz and 200Hz, but between 500Hz and 20kHz it drops off at about 0.9dB per octave. This is quite a bit more than the other curves mentioned above. B&K noted that their curve represents the corrections necessary when using a C-weighting filter, but I don't know how this might affect their curve when it is compared to other target curves.

 

[witwald]

Nice after so many decades of folks touting "flat" FR, that psycho-acoustic realities are now becoming mainstream...

When you refer to a "flat" frequency response, are you referring to a flat on-axis loudspeaker anechoic frequency response or a flat in-room frequency response as measured at the listening position? Isn't a flat on-axis anechoic frequency response a desirable characteristic of a loudspeaker? If the loudspeaker doesn't possess that particular characteristic, would it not reasonably be regarded as being a flawed transducer? By way of analogy, who would buy an amplifier with pre-set immovable tone controls that attenuate the bass and boost the highs? Should loudspeakers be held to a lower standard?

 

[witwald]

All due respect to DanDan, B&K, Toole, and the respected research of many in the industry, to each his own preferred Target Curve! It never hurts to try one out, and it never hurts to say you prefer something else.

The emphasis with the Dirac DSP is the ability to try things out, much like using the simple tone controls that are present on many amplifiers.

This is all part of the Dirac approach, which was presented in an AXPONA 2019 presentation, Ins and Outs of Room Correction with Dirac, by Jakob Ågren, Head of Product at Dirac Research. Some takeaways were that "Good sound is subjective" and that "A flat system will sound really dull and really boring." Do we believe the latter to be true?

The presenter also admitted to buying hi-fi speakers that had a smile-shaped frequency response. That's what he liked, and he wants the room correction system to be able to maintain that inherent characteristic.

Sonnie, I agree it will be interesting to see the new Recommended Dirac target curves, and to read up on the science behind the derivation of them.

The abovementioned presenter seemed to be suggesting, in a somewhat non-scientific manner, that the premise of Dirac room correction software is that it is a tool to get the sound that you like, because in the end it is subjective and there isn't a one-stop solution here. So the end-user should look for a system that allows them to get to where they want to be for their own particular use case. This seems very much like using simple things such as amplifier tone controls or a frequency response equalizer in one's audio system.

 

[BenToronto]

...that the premise of Dirac room correction software is that it is a tool to get the sound that you like, because in the end it is subjective and there isn't a one-stop solution here. So the end-user should look for a system that allows them to get to where they want to be for their own particular use case. This seems very much like using simple things such as amplifier tone controls or a frequency response equalizer in one's audio system.

Slow down everybody.

The room analysis and auto-correction schemes are just the first cut at fixing your sound. Old-timers with hands-on experience know that your first room exam will reveal all kinds of hideous problems. So the initial benefit of the auto-correction is to address the hideous problems and aim for a favourite "house curve" (such as the evolving Dirac one.... previously aimed at movie houses and now aimed at what they in their corporate wisdom thinks suits homes).

After that first run at fixing major issues, THEN you can (and you should) tinker with the tone colour.

BTW, whenever I hear, ".... this curve is what I like....", I say to myself, "Golly, I wonder if this person has actually had their hearing tested by a qualified audiologist this decade?"

B.

 

[BenToronto]

When you refer to a "flat" frequency response, are you referring to a flat on-axis loudspeaker anechoic frequency response or a flat in-room frequency response as measured at the listening position? Isn't a flat on-axis anechoic frequency response a desirable characteristic of a loudspeaker? If the loudspeaker doesn't possess that particular characteristic, would it not reasonably be regarded as being a flawed transducer? By way of analogy, who would buy an amplifier with pre-set immovable tone controls that attenuate the bass and boost the highs? Should loudspeakers be held to a lower standard?

The reference and target goal should be sound you like when playing recorded music at your chair.

As Toole's extensive Spinorama* studies show, when buying a speaker, on-axis flat is a weak criterion - what you need is the whole Spinorama profile to identify a good candidate before hauling the speaker home. And THEN you bring the speaker home and do your own listening tests to see if the speakers which seemed good on the lab Spinorama also seem good in your specific room. (And then then you buy large dipoles - esp planar or electrostatic - which are likely to sound better than anything else, at least for quality acoustic recordings.)

B.
* Toole/Harman built a gizmo that takes 360-degree measurements all around. Hence a speaker that shows nice properties when examined that way, will be a good choice for playing-nice in your room at home. Until you open the shipping case and hear it at home.

 

[JStewart]

In the real world we have users that haven't read Toole, don't have have perfect rooms and don't have perfect loudspeakers.

Dirac makes a product for the real world and was among, if not the first, to give their customer's an easy to use GUI to enable target changes to get a better sonic result. Including the ability not to apply correction above and below certain points along the frequency spectrum.

I applaud Dirac's efforts with this new feature to get a more pleasing result for their non forum reading geeky customers (like me) with the initial calibration who find Dirac installed on their shiny new audio device.

For those of us who want to go beyond what Dirac can do, there's nothing stopping us. We can even get professional help. No, I'm not talking about a shrink...

https://accuratesound.ca/services.html

 

[witwald]

As Toole's extensive Spinorama* studies show, when buying a speaker, on-axis flat is a weak criterion...

I'm not sure what is meant by the above. The following two quotes are taken from some of Toole's writings:

QUOTE 1: "Almost 50 years of double-blind listening tests have shown persuasively that listeners like loudspeakers with flat, smooth, anechoic on-axis and listening-window frequency responses. Those with smoothly changing or relatively constant directivity do best. When such loudspeakers are measured in typically reflective listening rooms the resulting steady-state room curves exhibit a smooth downward tilt. It is caused by the frequency dependent directivity of standard loudspeakers - they are omnidirectional at low bass frequencies, becoming progressively more directional as frequency rises. More energy is radiated at low than at high frequencies. Cone/dome loudspeakers tend to show a gently rising directivity index (DI) with frequency, and well designed horn loudspeakers (like the M2) exhibit quite constant DI over their operating frequency range. There is no evidence that either is advantageous - both are highly rated by listeners."

QUOTE 2: "What has been found over several decades of conscientious investigation and publication is:

(a) in double blind tests in normally reflective rooms (different ones over the years) listeners give the highest ratings to loudspeakers that measure essentially flat and smooth on axis, and at least smooth off axis in an anechoic chamber or functional equivalent. What they are recognizing and responding favorably to is the absence of resonances - i.e. neutrality.

(b) Room curves do not correlate as well with listener preferences, except at bass frequencies, below the about 300 400 Hz transition frequency. Adjusting loudspeakers having different flaws to match full-bandwidth room curves of highly rated loudspeakers cannot yield the same high quality sound. This is especially true if narrow-band equalization is used above the transition frequency. This fact is not to be found in the advertising literature of "room EQ" products. Guess why?"

 

[DanDan]

The Subjective and Objective Evaluation of Room Correction Products

In a recent article, I discussed audio’s circle of confusion that exists within the audio industry due to the lack of performance standards...

seanolive.blogspot.com

Over time the curves chosen have one shared notable aspect. They are remarkably similar.

 

[BenToronto]

I'm not sure what is meant by the above. The following two quotes are taken from some of Toole's writings:

QUOTE 1: "Almost 50 years of double-blind listening tests have shown persuasively that listeners like loudspeakers with flat, smooth, anechoic on-axis and listening-window frequency responses. Those with smoothly changing or relatively constant directivity do best. When such loudspeakers are measured in typically reflective listening rooms the resulting steady-state room curves exhibit a smooth downward tilt. It is caused by the frequency dependent directivity of standard loudspeakers - they are omnidirectional at low bass frequencies, becoming progressively more directional as frequency rises. More energy is radiated at low than at high frequencies. Cone/dome loudspeakers tend to show a gently rising directivity index (DI) with frequency, and well designed horn loudspeakers (like the M2) exhibit quite constant DI over their operating frequency range. There is no evidence that either is advantageous - both are highly rated by listeners."

QUOTE 2: "What has been found over several decades of conscientious investigation and publication is:

(a) in double blind tests in normally reflective rooms (different ones over the years) listeners give the highest ratings to loudspeakers that measure essentially flat and smooth on axis, and at least smooth off axis in an anechoic chamber or functional equivalent. What they are recognizing and responding favorably to is the absence of resonances - i.e. neutrality.

(b) Room curves do not correlate as well with listener preferences, except at bass frequencies, below the about 300 400 Hz transition frequency. Adjusting loudspeakers having different flaws to match full-bandwidth room curves of highly rated loudspeakers cannot yield the same high quality sound. This is especially true if narrow-band equalization is used above the transition frequency. This fact is not to be found in the advertising literature of "room EQ" products. Guess why?"

Those are helpful quotes from Toole. He is saying on-axis response needs to be supplemented by examining off-axis response. In particular, as he says, suitably crafted off-axis response can be accomplished using quality dome tweeters or quality horns and presumably also done poorly with some dome tweeters and some horns. (I won't intrude here with my reasons for thinking dipoles make home music wonderful, except to say it is for reasons other than Spinorama.)

Interesting, in this thread, that he diss'es room EQ when you have poor speakers. In recent years, he's started a major rethinking of room EQ. Maybe googling "Toole's Critique" would bring up that debate.

B.

 

[DanDan]

IMO separating the Direct Field Response and the Room Decay Spectrum and the Loudspeakers is a fools errand. Or perhaps an exercise for the Lab, Speaker designer etc.
What we hear, as measured by B&K, Harman, Sonarworks more recently, is a response combining all three aspects. It is unfortunate to see the term Room Eq...... These systems are speaker controllers.
My experience of Live Sound over decades says that speaker controllers are invaluable. The poorer the speaker the more useful Eq is. The poorer the hall acoustic the more useful speaker preconditioning is.

 

[DanDan]

The B&K curve is a little nonlinear. It is relatively flat in the decade between 20Hz and 200Hz, but between 500Hz and 20kHzi

Not sure which meaning of Flat is intended here. To my eye. I see +3dB peaking at round 70-80Hz with a very gradual slope all the way up to 2KHz. This is a curved tilt encompassing by far the majority of significant spectrum. AM Radio only covers to 4KHz. Again, not even a shelf, but a CURVED TILT.

To note, my apprehension regarding Direct Auto Curve is based on a few factors. The image posted is not like any favoured curve I have seen or heard. But perhaps that is just the PR people getting it wrong.
There are some incorrect statements in the Manual too. DL has always presented a Suggested Curve, but it has been way too mild. I wonder what the new one will be like.

Just as everything seemed settled (in my mind) ........ Wow very aggressive HF Roll Off. But aaaah, that old familiar 5-6dB Tilt between 100Hz and 10KHz Same as it ever was.