Update – 4 Dec 2023:
- Audyssey is still an independent company and has not been acquired by D&M –> Sound United –> Masimo, so this has been removed from the intro section. Sincere apologies to Audyssey for not fact-checking this.
- Audyssey didn’t create the Audyssey Mobile App, it was D&M Holdings (or later D+M Group), created from the merger of Denon and Marantz at the time, and later acquired by Sound United, and more recently by Masimo.
If you had told me a few years ago that Masimo, owners of the Denon, Marantz and Audyssey brands, would release Dirac Live on Denon and Marantz AVRs and processors, I would have told you that was not going to happen. However, ultimately, the pressure from Dirac Live, and stagnation with Audyssey’s core technology has meant that this was sort of inevitable.
Dirac versus Audyssey
I have written at length about Dirac’s vs Audyssey’s technology and specifically bass performance. However, the main difference between these two technologies are how the time-domain is handled. While Audyssey uses filters that don’t alter the phase or relative timing of different frequencies (or frequency bands – also called group delay), they also don’t correct for timing or phase issues, at least not beyond phase-aligning speakers.
Good-old PEQ will mess with phase and timing of frequencies: correcting the frequency domain but creating a mess in the time-domain in the process. Even more annoyingly, these timing issues vary based on the PEQ filters used so they aren’t uniform.
This is why Audyssey’s approach at the time was a step above: their filters didn’t affect phase and timing and therefore didn’t alter the system’s phase and timing performance – for better or for worse.
This is in contrast with Dirac Live, which uses mixed-phase filters to correct not only frequency-domain issues, but also phase and timing differences between the different frequency groups. This is the defining difference between the two technologies at their core, and why Dirac Live is a more advanced technology. Now this doesn’t mean that Dirac Live is simply better by default.
D&M (Denon & Marantz) first tried to close the gap by releasing the excellent Audyssey Mobile App. More recently Audyssey released the more advanced and configurable MultEQ-X. MultEQ-X allows Audyssey to do filter-calculation on a PC, as opposed to the AVR, which gives them the ability to do more advanced filter selection. How far Audyssey has taken this is a good question, but a calibration with MultEQ-X can sound better than doing a calibration on the AVR, even without doing any customisation in the app. Additionally, it is possible to do more measurement positions and further customisation of Audyssey to get even better performance out of it.
The question is then: how much has Audyssey caught up with Dirac Live and what is still the gap between the two. But first things first, let’s talk about why a good phase-response is important for sound quality.
A System’s Phase Response
Errors in the time-domain are created by the playback system, including the processor and speakers, and the playback system’s interaction with itself and the room. Timing errors can affect all frequencies equally: jitter is a great example of this, or they can affect individual frequencies differently, therefore introducing phase-errors. This is sometimes also called group-delay, as groups of frequencies will take on different delay / phase characteristics.
In a multi-channel system, we also need to distinguish phase errors between speakers and for each individual speaker. This is important, as most Room Correction technologies will try and phase-align speakers as best as they can, but only more advanced technologies such as Dirac Live, will attempt to correct the phase response of individual speakers by time-aligning frequency bands. This of course will also help with the phase-alignment of speakers in a multi-channel system, as you are no longer aligning them by averaging phase delay that may represent a grossly varying phase response.
Speakers are inherently messy with regards to their time-domain response: the different drivers may sit in space at different distance from the listener, and the crossover network will also introduce time-domain issues of its own. Some manufacturers will include a “phase-focused crossover”, but this only guarantees the drivers to be properly in phase in the crossover region, and doesn’t remove the drivers’ inherent phase response characteristics, which can still be messy.
It’s nearly impossible to create speakers with perfect phase-response with only an analogue crossover network, or it costs so much money that it is impractical for 99.9% of products and installations. This is why digital correction is generally necessary to remove these issues. What’s more, I can guarantee that even the most advanced speaker designs will have some sort of residual phase issues.
When all frequencies arrive at the listeners’ ears in phase, the sound becomes clearer, more intelligible and takes on almost a 3D feel, whether listening in two-channel and multi-channel. In fact, you can have two speakers with exactly the same, even ruler-flat frequency have terrible imaging, exactly because of the differing phase response between the two. This is why focusing on only the frequency response of a system isn’t going to deliver accuracy.
The above is also why looking at simply frequency-response graphs is not only delivering half the story, but isn’t entirely insightful:
- We are sensitive to the overall tonal balance of the sound more than we are to individual frequencies.
- If we are sensitive to individual frequencies or bands of frequencies, these would be around 1Khz to 3Khz with a drop-off of sensitivity both below and above this. We will however be sensitive to phase and timing-response well below this and well above this.
- Frequency response needs to be balanced across the whole listening space for best results. This is because over-correcting for one seat can create equalisation artefacts in another seat. This does mean that striving for ultimate flat response in just one seat is not ideal. What’s more important is that all seats get a balanced sound.
- We have two ears so over-correcting high frequencies is a futile exercise as room modes above 500Hz can be just inches apart. Broad corrections are generally better over 500Hz with the least chance for artefacts.
- Phase response is a critical part of our hearing and can make or break imaging and clarity, so it must be balanced with frequency response.
The following equipment was used for this review:
- Marantz Cinema 50 in a 7.2.4 configuration with a Rotel RMB-1077 driving the main channels and the Marantz driving the 4 height channels.
- 2 REL Predator Subs sitting under the screen equidistant from the main listening position and the centre channel
- MiniDSP 2x4HD driving the two subs using MSO-selected filters, and levels (delays were not allowed to be altered due to the symmetrical setup and to avoid introducing time-domain issues)
- Main speakers are custom-made high-efficiency speakers with 10″ Faital Pro and 1″ horn-loaded B&C compression drivers to emulate real theatre sound (we will release the design for these early next year – they kick ass). These speakers can knock you senseless even without a sub, and been really enjoying the clarity and range from these.
- Rest of the speakers are Jamo D500 THX speakers in all other speaker positions. They are wide-dispersion speakers which help surround envelopment.
General Product Comparison
Cost / Value
Let’s start here for our comparison: Audyssey comes free of charge with Marantz and Denon receivers. You can also upgrade the AVR with a Mobile App on either iOS or Android, which costs $19.99. The Audyssey Mobile App lets you tweak and save your calibrations beyond what the AVR allows, including customising the Reference curve and enabling / disabling Midrange Compensation.
MultEQ-X and an Audyssey-calibrated microphone are another upgrade path and also cost extra: $199 and $79.99 respectively. However, these are also optional upgrades and aren’t required to run Audyssey.
Dirac Live costs differ based on whether you want full-range or limited range EQ ($259 or $349) and whether you wanted support for one or multiple subs (DLBC not yet released). You will also need a UMIK-1 microphone which will set you back $79.
Let’s tally the costs then:
(EQ Solution + Mic)
|Included in AVR Price|
|Advanced EQ Solution||$199||$349|
|Multi-Sub Upgrade||$0 |
|$428 for single sub|
$607 for multi sub
(no inclusion of
multiple sub support –
will cost extra)
As we can see, Audyssey’s pricing is both free and incredibly flexible when it comes to optional upgrades. In its most upgraded form, it is still around $150 cheaper than Dirac Live, which is not yet the equivalent of Audyssey until multi-sub support is released and will be around $328 cheaper for multi-sub users. This is obviously a large price difference, considering Audyssey is fully feature complete out of the box with upto 4 subwoofer support and full-bandwidth correction with an included microphone.
Audyssey: 5, Dirac Live: 3
Let’s get this out of the way: Audyssey’s setup routine is miles ahead of Dirac Live in terms of ease of use, choices and options. Audyssey can be run on the AVR, on the excellent Mobile App and using a PC with MultEQ-X for those that want to get the very best performance out of it or like to tinker.
Dirac Live on the other hand can only be run on a PC for Denon and Marantz receivers and the setup routine is anything but fool-proof. For example, other EQ routines have had auto-levelling of speakers for 20 years now, and here we are in 2023 manually adjusting sliders to get speakers ready for calibration. It isn’t really that hard to test microphone sensitivity and then test speakers at increasing volumes. Since Dirac Live now allows the UMIK-1 to be plugged directly into the AVR, as opposed to the PC, this should get around the risk of not knowing mic sensitivity and accidentally damaging speakers.
Additionally, a Dirac Live mobile app does exist for Onkyo receivers, so this wouldn’t be out of the realm of possibility.
For our review here, the two systems were set up using the best we could for each technology and by calibrating the whole listening space across two rows of seats. This does mean that we used MultEQ-X for Audyssey.
The microphone pattern was selected based on numerous tests with both technologies and landed on a common pattern that worked very well across multiple rows and for both EQ solutions. I think that this is the most pragmatic approach we could use: it is real-world and tries not to disadvantage either technology. We couldn’t get a better calibration across the two rows by following Dirac’s Mic Patterns to the letter for example. Even then their auditorium pattern that would be suitable for this type of setup isn’t exactly exact.
Calibration and tests were conducted by letting both technologies calibrate the subwoofers by themselves first and also with the help of a MiniDSP calibrated with Multi-Sub Optimiser. While Dirac Live will handle multiple subs even in its current form, it doesn’t do so very well without Dirac Live Bass Control. Hence we wanted to give both technologies the same excellent pre-calibrated subwoofer response.
As an additional note, the two sub-woofers used have one of the best, if not the every best, time-domain performance of any subwoofer system for the home theatre: REL Predators. They use minimal filtering and do so in the analogue domain, not by shaping the sub response using PEQ. To make sure MSO didn’t mess with this, the two subs were equidistant from the main listening position and MSO was not allowed to introduce any delay / timing difference between the two subs. While PEQs will still mess with the time-domain to a degree, this was the best base calibration we could give both technologies – until DLBC is introduced which MIGHT be able to extract even better time-domain performance as the two subs could have their group-delays reduced individually, as opposed to as a whole. How much of a difference this would make with two identical subs equidistant to the listening position is questionable but is possible.
Audyssey: 5, Dirac Live: 3
Audyssey has two massive aces up its sleeve: Dynamic EQ and Dynamic Volume. Once set up correctly, Audyssey’s Dynamic EQ works incredibly well to balance equalisation based on volume and is based on sound science and research. If you are new to loudness compensation and wondering why it is so critical for achieving reference playback at all volume levels, you may want to read this article and this article.
Those that complain about Dynamic EQ generally haven’t set up Audyssey properly or don’t know how to use Dynamic EQ. I have worked with clients who – even after reading Secrets of Audyssey, were implementing misguided forum advice and getting terrible results. But once troubleshooted their setup, things started coming together and Dynamic EQ just worked. So true, Audyssey is also not fool-proof, but if people actually do the work, Dynamic EQ works very well.
In addition to the balancing of the frequency response based on volume, Dynamic EQ also tries to balance surround envelopment by changing the front to back volume balance. This can be a bit more of an issue dependent on your setup and I talk about how to resolve this in Secrets of Audyssey. But ultimately, for setups where this is a necessary element, it can work very well.
The same goes for Dynamic Volume: it brings back impact for low-volume listening and works much better than any other dynamic compression algorithms, and I generally recommend it when listening under -20dB.
In contrast, Dirac Live doesn’t include any of these technologies: it includes only one base target curve and expects the user to either use that for all listening or create new ones based on whatever methodology they subscribe to – most not based on science of loudness but “user preferences”. All well and good, but I am not exactly thrilled about this. This is why I spent 3 months modelling different loudness-based technologies and creating loudness-based curves that can be uploaded to Dirac Live. I described my setup methodology in this article in more detail. For background on these curves, you can check out this article. To sum those two articles up: I am able to emulate Audyssey’s Dynamic EQ with Dirac Live at all volume levels and for all content types by loading different loudness-based curves.
There’s a bit of a silver lining for Dirac Live and it is this: due to the better phase-response, the sound is somewhat clearer and more intelligible at all volume levels even without changing the EQ. Of course, the issue is not only clarity but also correct tonality and impact of the soundtrack, especially in the low-end. This is really the biggest casualty in Dirac Live’s one-size-fits-all approach. As the volume is lowered from Reference, Audyssey retains surround envelopment and impact of the soundtrack a lot better than Dirac Live does.
Ultimately, for the listening tests, both Audyssey and Dirac Live were set up with the same curves: Audyssey was running the stock curves with Dynamic EQ, while Dirac Live was using the loudness-based curves based on volume levels and content, matched to Audyssey.
Audyssey: 4, Dirac Live: 2
Dirac Live vs Audyssey Listening Tests
Neither technology is state of the art with regards to subwoofer calibration just by itself. When left to their own devices with two subwoofers, both created an uneven response within the room and across seats – by upto around 20dB difference in the low-end. Now this might be different once Dirac Live Bass Control is introduced, but that will be at an additional cost.
Audyssey was great at balancing the sound across seats in only one row, but since it only time-aligns and levels subwoofers individually, then EQ’s them together, it cannot achieve a more balanced sound across rows or seats with widely different response. It can only try to find a more balanced medium. This however sucks the impact out of some seats, while creates too much low-end for other seats. Dirac Live without DBLC fairs about the same here.
However, once a MiniDSP with Multi-Sub Optimiser was introduced to better-equalise the full listening space, the playfield was levelled and Dirac Live came out as the clear winner.
The subwoofer frequencies were clearly time-aligned much better by Dirac Live which created more impact and cohesiveness between the main speakers and the subwoofers. This was especially obvious in the crossover region where Audyssey left a bit of a gap, while Dirac Live filled it out very well. This was also the case across both rows of seats: while Audyssey was a bit less consistent seat to seat especially in the crossover region, Dirac Live had a consistent clarity and impact in the bass region across all seats.
Dirac Live sounds like all frequencies are hitting together at once and the subwoofer bass has both excellent timing and clarity without phase and cancellation issues.
I would hazard a guess that since Audyssey isn’t able to remove phase issues, there could be some cancellation issues especially in the crossover region between the main speakers and the subwoofers. This was the case at all sensible crossovers: 40Hz, 60Hz, 80Hz and 100Hz. Audyssey simply lacked impact compared to Dirac Live in my setup. Again, it wasn’t bad by any stretch of the imagination, and I would be very happy with its performance without having heard Dirac Live on the same setup, but Dirac Live was a step above.
Audyssey: 3, Dirac Live: 4
Clarity and Intelligibility
While they were relatively close, Dirac Live won out out with regards to clarity and intelligibility, especially obvious on centre channel dialogue. Dialogue was clearer, had more body – likely due to better phase alignment in the crossover region and sounded more grounded.
Dialogue also sounded a bit more natural with Dirac Live than with Audyssey. This is likely due to Audyssey’s insistence on using detailed frequency correction above 500Hz. This is one area of improvement I could suggest to Audyssey: they need to introduce broader filters above this region, not simply allow the user to turn off correction, as that also isn’t ideal. If there was a way in MultEQ-X where you could set the “resolution” of correction per frequency region, that might allow further fine-turning and getting closer to the results Dirac Live is delivering here – at least it would remove some of the slight artefacts with dialogue, and return a more natural presentation.
Movie score and music in general had better clarity with Dirac Live, but Audyssey was not too far behind.
Audyssey: 4, Dirac Live: 5
Audyssey did very well with regards to impact, especially with Dynamic EQ (and Dynamic Volume) on. Dirac Live is lacking with regards to impact with the default curve especially at lower volume levels. However, once loading up the loudness-adjusted curves into Dirac Live and adjusting surround volume levels, it bests Audyssey on impact because the subwoofers hit harder and with more control. It feels like all the frequencies hit at the same time and actually has an in-your-chest impact even more so than Audyssey does.
The same conclusion is true at reference volume with the default curves, as no loudness-compensation is necessary when listening at or close to reference. However, this is way too loud for most rooms to be enjoyable.
This is difficult with scoring but I will give Dirac Live a leg up, in spite of the defaults being a bit lacklustre here for 95% of people. The technology can deliver it, what lacks is the ease of use here, which we have already scored.
Audyssey: 4, Dirac Live: 5
This was a bit of a surprise. I feel that Audyssey is better at this overall, especially if you are using the stock curves on Dirac Live and don’t adjust surround volumes. However, at higher volumes, and if you tweak the curve and surround volumes, Dirac Live sounds awesome as well, but still somewhat different. I feel like Audyssey brought out a bit more surround detail that had more layers, at the expense of some clarity.
But on another level, I could have given this a tie too. I will score Audyssey a little higher, but ultimately there isn’t a huge difference once both are set up well.
Audyssey: 5, Dirac Live: 4
I wasn’t sure which solution would win overall before I tallied the scores, but here we are. Ultimately, when set up well, Dirac Live sounds better, so it got the better sound quality score total (18 versus 16 for Audyssey).
|Clarity and Intelligibility||4||5|
|TOTAL (Sound Quality)||16||18|
However, Dirac Live is let down by its setup routine, lack of loudness compensation curves that adjust the curve based on volume and content, and lack of a sensible and compatible dynamic range compression technology for low-volume and night-time listening. Audyssey clearly has the leg up in both of those areas, so it has the much better value and usability score of 14, compared to Dirac Live’s 8.
|Cost / Value||5||3|
|TOTAL (Value and Usability)||14||8|
With loudness compensation curves in place using the work done for Dirac Live Perfection, Dirac Live actually has better clarity and intelligibility, better impact and overall excellent surround envelopment on par with Audyssey so it gets the higher sound quality score. However, to get these requires quite a bit of work on the user’s part unless the user is willing to listen at reference volume. So when looking at the overall picture, Audyssey still comes out on top.
|Value and Usability||14||8|
Looking at the defaults in Dirac Live, cost and pain of setup as well as resulting usability issues, I feel that Dirac Live still has some way to go before it is ready for prime-time.
The issue here is this: while people like me understand these issues and are willing to work around them, most users will not. Even more annoyingly, Dirac Live is NOT compatible with Dynamic EQ or Dynamic Volume, or even THX Loudness Plus on other receivers to make nighttime or lower-volume listening possible for users.
This is NOT a technical issue per se with incompatible algorithms, as fundamentally, Dynamic EQ and Dynamic Volume should not care which equalisation solution is at play. However, it might be a technical issue insofar as not enough processing on the chips to run both at the same time, as it is likely that currently Dynamic EQ and Dynamic Volume are using Audyssey filters to create the EQ curves, which would make sense. But this isn’t the case for THX Loudness Plus so Dirac – or the AVR manufacturer – decided against implementing the feature on Onkyo receivers also.
This is an issue, as the majority of users will gravitate towards a solution that technology enthusiasts like myself recommend, only to be disappointed due to the omission of such important features and migrate back to solutions that offer them. Proper dynamic range compression for night-time listening is not a nice to have, it is a requirement to be successful with the rest of those 95% of users. Listening to technology enthusiasts rave about Dirac Live and think that will fly with the rest of the users is ill-advised.
I think Denon and Marantz are smart not to drop Audyssey as such, as Dirac Live – in spite of its ability for higher-quality sound than Audyssey – still falls short on usability and features. It’s quite the tour-de-force for technology enthusiasts like myself, and frankly I love listening to both music and movies with Dirac Live, I just feel it falls short of what day-to-day users expect. Dirac still has some work to do then if it wants to become mainstream in the Home Cinema space.
Audyssey is STILL the most rounded Room Correction solution on the market today for Home Cinema. The technology base is sound, setup is incredibly versatile and can be scaled to the user’s skills and willingness to tinker and the sound quality can be exceptional when set up correctly.
However, it is bested by Dirac Live on sound quality, especially above 500Hz, which does overlap with the very critical voice region.
As I said earlier, one immediate thing Audyssey could do would be to allow the user to choose the resolution of filtering per frequency band. I would recommend at least 3 bands:
- Under 500Hz
- Between 500Hz and 5kHz
- Over 5kHz
The filter resolution should be adjustable based on this.
Alternatively, have an option whereby filtering resolution above 500Hz is gradually reduced / broadened by using frequency smoothing and only correcting for general tonal issues above this band.
The above should be possible to implement with MultEQ-X and would aid in removing some of the unnatural character Audyssey filtering has in the critical voice region – and above but this isn’t as obvious to most listeners due to less sensitivity in our hearing.
An additional thing Audyssey could focus on is further research in digital filtering and aim to provide a solution for phase / group-delay correction, not only per speaker but also for groups of frequencies. This might mean introducing filters above 500Hz that take less processing power than the current filters and reallocating the remaining processing to phase response correction rather than purely frequency response correction across the whole frequency range.
For End Users / Playback
Dirac Live comes highly recommended, but only if you are willing to work through the above issues and you want the absolute best in sound quality – money and time no object.
Audyssey is recommended for everyone else, and if low-volume (under -20dB) and night-time listening is critical. It is the more well-rounded solution, but falls somewhat short on sound quality compared to Dirac Live. Again, check out our article on Configuring Audyssey – The Right Way to get the most out of Audyssey. Alternatively, our premium guide, Secrets of Audyssey, has helped many users achieve excellent sound quality approaching that of Dirac Live. I am not going to lie: there is an uplift in sound quality when it comes to Dirac Live, but I didn’t think it was going to be this close in the end.
For mastering professionals, I would be inclined to recommend Dirac Live, but NOT with the default curve. It is not suitable for mastering work, which is a whole another article in itself. Dirac needs to offer both a flat and a curve similar to Audyssey Reference with adjustable high-frequency roll-off based on room size.
This is because mastering needs to be done at the correct volume and the frequency response adjusted to flat for music. For movies, a high-frequency roll-off needs to be present dependent on room size and the soundtrack mastered at reference volume. I feel this is one area where Dirac Live is just not on the ball. So if this is something that is hard to implement and tinker with for you, go with Audyssey.
Partly why I think Dirac would be best served to look for previous THX engineers on the job market – and pronto, along with someone trained in equal-loudness contours (ISO:226: 2003), as these are areas that are still clearly lacking in Dirac Live’s current iteration.