Dirac Live ART vs. Custom DAW-based MIMO Calibration Objective and Subjective Comparison

[Rmjesty]

It's a great pity that the X4800H used for this comparison test was borrowed from a friend, and we only had half a day for measurement and comparison listening. I am very interested in conducting in-depth research on various calibration methods. However, due to economic and some technical reasons, I don't have an AVR or AVP myself that can be used to run Dirac ART. Perhaps in the future, after they release the PC and plugin versions, I will have the opportunity to continue making more refined adjustments and comparisons.

Moreover, these "economic and technical reasons" are also the driving force that urges me to study more complex and in-depth manual calibration schemes. The Dirac Live ART alone costs at least $799, not including the hardware needed to run it. And $1,100 is basically sufficient to directly build the complete DSP system I'm using, and it has better analog interface performance.

 

[Oddball]

Regarding this point, I have already used this method for calibration in many scenarios outside of this studio, such as my own mixing studio and the mixing/home theater rooms of other friends. As you can see, these rooms all have very distinct features. They vary in configuration and size, but all can achieve good results by the same method. If you are interested, I can contact the users in these places to see if they can provide some complete measurement data. Since I use well-designed coaxial speakers, I think slightly blocking one of the bass units won't cause major problems, and this is indeed the case in the measurement.View attachment 533933View attachment 533934View attachment 533935

They are very small rooms and very nearfield. What you are getting is what you put into it. Lots of folks, me included have very large multi purpose rooms that have completely different acoustics. Again, aplause for the effort.

 

[Rmjesty]

They are very small rooms and very nearfield. What you are getting is what you put into it. Lots of folks, me included have very large multi purpose rooms that have completely different acoustics. Again, aplause for the effort.

I indeed failed to make attempts in a larger space. Part of the reason is that for Atmos mixing and mastering work, overly large Spaces are not common, and in most cases, the listening distance does not exceed 2 meters. My subsequent attempts in more types of rooms will also be updated in this post. Thank you for your understanding.

 

[webmst007]

Please understand any implied critique of your work is not anything other than respect for your endeavours
I've only been able to work with Dirac Live and performed many measurements with that system, but it has always frustrated me not being able to work with ART ( a processor limitation).
But having watched ART from the outset, a few aspects seem to hold true.
Considering its overall complexity, once ART was released on multiple platforms it was inevitable the experiences and results would vary from the very good to the terrible. I would argue this result reflects the huge variety and quality of rooms and systems users are deploying this technology on.
There is also the practical aspects of ART ( or any dsp- room correction system) - how invested is the owner in working with the systems?
Certainly ART can be considered the most sophisticated audio tool to be unleashed on the general audio population to dare.

So how extensive is their understanding of what the system does and how to refine the outcomes based on an initial set of results?

Well the level of interest and confusion could be gauged I would suggest by the the huge attendance on the ART forums here and on AVS.

Perhaps it can be also be concluded that, based on so many posters experiences, their initial results are substantially improved apon with persistent refining.

Of course, this accepts the user needs to invest in a more than a basic understanding of the Dirac environment and how the system interacts with speakers combined with room acoustics.

And despite the negativity we see users enthusiastically continue to take up these challenges.

Isn't this an excellent case study whereby the specific forum environment of shared user experiences has provided practical insights and solutions for new users which aid people's understanding and enjoyment well beyond vendor help videos or user manual s?

Certainly these collective forums promote users to explore beyond the simple default Dirac ART filters and refine their own results.

This alone promotes people to expand their own system and general acoustic knowledge which is ultimately a great outcome for our hobby.

I often here the argument people are moving away from "old fashioned" hifi gear and want new generation "simple" all-in-one everything nowadays. Well here is an example of how new generation is interesting, complex, and all-in-one, and requires learning and skill to achieve better outcomes.

Skills that you have acquired over years of work and put into practice in your use of DAW systems every day.

Asking everyday hifi buyers of home theatre AVP's to learn some of those skills can only be a good thing.

So bravo - I look forward to your next series of experiments with great interest

 

[Stingber09]

Great Work! I bet it took a lot of time to calibrate it with the custom DAW mode.

I have a couple questions:

1)What plugin is "Time Adjustment De.... "? Is this Eventide Precision Time Align plugin?

2)Did you use kHs Convolver for each group support correction?

3)What process did you used to capture and load each impulse response?

4)Can you share your Reaper routing layout?

5)This was calibrated at a sample rate of 48KHz or 96KHz?

I had a similar setup but I used Waves TRACT and Pro-Q4

 

[Stingber09]

Try this and let us know how it compare to Dirac ART and Custom DAW Mode.

Umami Audio - Perfect Soup MIMO Speaker Calibration

Professional MIMO speaker and room calibration software. Measure, correct, and perfect your listening environment with GPU-accelerated multi-input multi-output optimization, room mode correction, and reflection analysis.

umamiaudio.com

 

[Rmjesty]

Great Work! I bet it took a lot of time to calibrate it with the custom DAW mode.

I have a couple questions:

1)What plugin is "Time Adjustment De.... "? Is this Eventide Precision Time Align plugin?

2)Did you use kHs Convolver for each group support correction?

3)What process did you used to capture and load each impulse response?

4)Can you share your Reaper routing layout?

5)This was calibrated at a sample rate of 48KHz or 96KHz?

I had a similar setup but I used Waves TRACT and Pro-Q4

"Time Adjustment Delay" is the built-in delay adjustment tool of Reaper.
I did apply FIR for frequency response calibration to the input of each group of tracks.
The generation method of FIR is to measure with REW and then perform curve operations based on the target curve to obtain the calibration IR.
The internal routing of Reaper is signal input → crossover → respectively enter the main speaker, subwoofer, and support speaker output.
Operate at 48khz. It's great to see that someone has once realized the same idea!

 

[Rmjesty]

Try this and let us know how it compare to Dirac ART and Custom DAW Mode.

Umami Audio - Perfect Soup MIMO Speaker Calibration

Professional MIMO speaker and room calibration software. Measure, correct, and perfect your listening environment with GPU-accelerated multi-input multi-output optimization, room mode correction, and reflection analysis.

umamiaudio.com

View attachment 533956
View attachment 533957

I'll try it later.

 

[ace_xp2]

As usual, Dirac may be as good as custom in a given room. It may also fall short in the wrong (right?) room/setup. But as always, a typical user will turn it on and decide it's the best, in direct relationship to current cost/newness.

 

[Rmjesty]

Try this and let us know how it compare to Dirac ART and Custom DAW Mode.

Umami Audio - Perfect Soup MIMO Speaker Calibration

Professional MIMO speaker and room calibration software. Measure, correct, and perfect your listening environment with GPU-accelerated multi-input multi-output optimization, room mode correction, and reflection analysis.

umamiaudio.com

View attachment 533956
View attachment 533957

It's possible there's something wrong with my operation or that it's not compatible with my audio chain, but based on my current experience after performing 30 measurements and generating a calibration file as instructed, it's clearly still under development and has many bugs to be fixed. I mounted the plugin in my DAW and was surprised to find that the channel volumes were completely mismatched; even after bypassing it, it still affected the volume of every channel, resulting in the same mismatch. Due to such a severe error and the need for 30 measurements per retrieval, and the software's lack of support for separate input and output device selection, which necessitates a more complex signal chain, I haven't conducted further testing for now. I hope the software developers can reduce the number of required measurement points and support simultaneous selection of different audio devices.

 

[Stingber09]

That was quick! I try it with a 9.8.6 system and Reaper as DAW. I didn't found any imbalance. I did a loopback test to verify this. It was on Windows 11 with Blu Link audio network protocol and latest version of Reaper. What OS you where using? mac or Pc? What Daw did you used Studio One or Reaper? Did you verify Dante digital Trims levels?

 

[umamiaudio]

Hi! I'm the developer. Can you join us on discord? I'd like to figure out what's going on. How you took the measurements etc.

Join the Umami Audio Discord Server!

Check out the Umami Audio community on Discord - hang out with 15 other members and enjoy free voice and text chat.

discord.gg

Happy to provide support here as well, but I think it's just faster to be able to chat about it.

 

[umamiaudio]

Just had a look at your plots. I think you didn't pick the correct mic input or forgot to turn phantom power on. The first peak being at 120hz strongly suggests that you recorded hum.

Also I noticed that you ran it with Dante Via, which is not a supported configuration. We need the input and output clock to be hardware locked which usually means using a single "real" device (as opposed to a composite logical device). You have a Motu 16A, and should use that directly.

The in-plugin bypass button is a gain matched bypass so that when it's calibrated properly, if you turn the calibration off, you don't get a volume jump. But when the measurement is incorrect, that makes the calibration incorrect, which makes the bypass incorrect as well.

 

[BestGreed]

View attachment 534107
Just had a look at your plots. I think you didn't pick the correct mic input or forgot to turn phantom power on. The first peak being at 120hz strongly suggests that you recorded hum.

Also I noticed that you ran it with Dante Via, which is not a supported configuration. We need the input and output clock to be hardware locked which usually means using a single "real" device (as opposed to a composite logical device). You have a Motu 16A, and should use that directly.

The in-plugin bypass button is a gain matched bypass so that when it's calibrated properly, if you turn the calibration off, you don't get a volume jump. But when the measurement is incorrect, that makes the calibration incorrect, which makes the bypass incorrect as well.

Actually getting multiple independent internal routing devices to play nice together is surprisingly tricky. In my setup, I have two separate hardware Dante output interfaces connected to DVS, plus a Dante-connected microphone. I ran into the exact same issue (and it was even worse) with SoundID Reference, where some measurement points would fail to register entirely. This is almost certainly caused by clock synchronization problems. I suspect rmj is experiencing the same thing, but interestingly enough, Dirac (VST) does not have this issue at all.

 

[umamiaudio]

He just picked the wrong input I think. He recorded again and this time it recorded real signals

 

[Rmjesty]

Update: Remote guidance for an friend to perform calibration. He has a pair of stereo speakers and three subwoofers.

 

[test1223]

Subwoofers: TAULSSEN T12 x2 (placed at the midpoints of the front and rear walls)

I would strongly recommend to place both subwoofer somewhere left and right from your listening position about 90°. The phase relationship of deep bass carries information of the room acoustics of the recording and can significantly increase envelopment. If you want to learn more please read the scientific work of David Griesinger and others.

 

[jhaider]

Moreover, these "economic and technical reasons" are also the driving force that urges me to study more complex and in-depth manual calibration schemes. The Dirac Live ART alone costs at least $799, not including the hardware needed to run it. And $1,100 is basically sufficient to directly build the complete DSP system I'm using, and it has better analog interface performance.

OTOH, what's the time commitment for the alternative approach? Depending on that, $799 for a done-in-an-hour solution that works could be considered extremely high value-for-money.

 

[Rmjesty]

OTOH, what's the time commitment for the alternative approach? Depending on that, $799 for a done-in-an-hour solution that works could be considered extremely high value-for-money.

If I operate this system myself or directly guide users to operate it, it usually takes two or three hours. If users want to learn the technology from scratch, it is impossible to estimate. However, learning technology is clearly not a bad thing

 

[Rmjesty]

I would strongly recommend to place both subwoofer somewhere left and right from your listening position about 90°. The phase relationship of deep bass carries information of the room acoustics of the recording and can significantly increase envelopment. If you want to learn more please read the scientific work of David Griesinger and others.

In fact, during the calibration process, the subwoofer was completely in the "mono bass" mode and the crossover point was relatively low. Whether from the perspective of music production or room acoustics, a correlation of strictly 1 will not have any "phase relationship" issues. They do indeed trigger different room modes, so I chose a position with better effects.