How Important is Speaker THD for Dirac Live ART?

[NanoGuidingStar]

Hello everyone,

Since last year I’ve been using Dirac Live ART. When adjusting the F Support Low (FSL) values, I’ve always referred to the specifications provided by the speaker manufacturer along with Dirac’s in-room measurements. As for THD, because most reviewers like amirm and Erin point out that bookshelf speakers tend to have relatively high THD in the bass region, I usually set an 80Hz crossover to the subwoofer to mitigate this issue.

Here’s the problem: Dirac Live ART doesn’t follow the traditional crossover concept. The lower the FSL setting for the supporting speakers, the greater the benefit to the overall system. In other words, within the usable bass range of the speaker, the FSL should be set as low as possible.

Take the Ascilab C6B bookshelf speakers as an example. F3 is around 52Hz. Based on my past experience, I would set their FSL to 60~65Hz. However, if we factor in THD (96dB@1M) from Erin, the speaker’s THD below 100Hz already exceeds 3%. This means that theoretically, setting the FSL below 100Hz could result in “distorted bass” being audible.

I’m not sure if this topic has been discussed before, but I really want to understand how to balance between a speaker’s bass performance and THD in order to determine a reasonable and appropriate FSL setting.

For most budget bookshelf speakers, their bass THD performance is generally poor (higher than 3%). Does this imply that they are all unsuitable for supporting bass below 100Hz? Or am I just scaring myself, overthinking?


By the way, my 7.2.2ch system is entirely composed of bookshelf speakers+subs. Normally, I set the AVR’s Master Volume to -25dB for movie. The distance from each speaker to my listening position is about 1.3–1.7 meters. Using a smartphone SPL meter to observe the playback volume while watching movies, I measured: minimum 45dB(A), average 85dB(A), and maximum 100dB(A).

 

[NTK]

Don't know about Dirac ART, but if it works similarly as Harman's SFM, using multiple low frequency sources will supposedly increase efficiency. So the speakers actually don't need to work as hard, and therefore lower THD. But of course you will have to see how the actual parameters come out after measurements and calibrations, which will be different for different rooms and setups.

 

[NanoGuidingStar]

Don't know about Dirac ART, but if it works similarly as Harman's SFM, using multiple low frequency sources will supposedly increase efficiency. So the speakers actually don't need to work as hard, and therefore lower THD. But of course you will have to see how the actual parameters come out after measurements and calibrations, which will be different for different rooms and setups.

Using this Filter Design diagram as an example:

The light-colored lines circled in red represent the ART Filters calculated by Dirac. These filters are applied from the 'Support Speakers' to the 'Supported Speaker.' Their amplitudes generally hover around -10dB. While the 'F Support High' is consistently set at 150Hz for all, the 'F Support Low' varies depending on the specific speaker configuration.

 

[90sNintendoKid]

I am going to do a Dirac ART calibration with my Orbit 11s.

I was going to try post EQ measurements to check for distortion. I guess I would need to disconnect other speakers and subwoofers to see how a speaker responds to F support low by itself.

 

[NTK]

Using this Filter Design diagram as an example:
View attachment 521215

The light-colored lines circled in red represent the ART Filters calculated by Dirac. These filters are applied from the 'Support Speakers' to the 'Supported Speaker.' Their amplitudes generally hover around -10dB. While the 'F Support High' is consistently set at 150Hz for all, the 'F Support Low' varies depending on the specific speaker configuration.

If a "support speaker" is playing at a level 10 dB below, it isn't doing much. -10 dB in power is 10%. That means, if you are adding a support speaker at -10 dB to the sub at 0 dB, the total power sum becomes 110%, which is just 0.4 dB increase, and is probably unnoticeable.

If you have many support speakers each adding 10% (= -10 dB), the total contribution probably will make a difference. But from the little I can read from the plots, not all are at -10 dB. I strong suspect you can turn off those below -10 dB, and there won't be much of a difference.

 

[JRowland]

I think if you watch this video with Joss Walker and think for yourself about his answers and ignore the interviewer comments it may help you. I interpret him as using whatever room gain is available as long as you are aware of your systems capabilities, think, and listen.

At 26:25 he makes a comment "if the speaker is capable of playing it we want to use that ability"
At 59:30 you can watch that segment and like I said ignore the interviewer and focus on Joss' answers and how he answers.

I use all of the room gain for each speaker in my system and set FS Low at where they hit 0db on the multipoint measurement. When you look at the filter design screen for my left front all my channels are loafing at under 0db, mostly around -10db or less. No individual speaker is being pushed anywhere near where it would be without the combined effect ART provides. If my filters showed boosts above 0db i would be more concerned, but that is not the case and the sound is lovely with no bad noises even with really loud volume.

 

[JRowland]

I strong suspect you can turn off those below -10 dB, and there won't be much of a difference.

All of that under -10db action is what is handling the decay. If you remove those support speakers you will almost always see (corrected spread)/feel/hear the decay increase. If you look at the red line on the filter creation screen from ART that I posted the red red line is what my left front is actually being asked to produce. Anywhere a filter line goes above the red red speaker (the 3 subs do below 50Hz - tan trace, blue trace, and pink trace then the left front also gets help at 65Hz and 85Hz) in DB that other speaker (trace) is supporting the frequency response in that area. The down under messiness is decay control. Almost always when you remove one of those speakers (channels) playing just -10db and less the corrected spread measurement gets bigger (increased decay) and it is noticeable in feel mostly.

 

[NanoGuidingStar]

I think if you watch this video with Joss Walker and think for yourself about his answers and ignore the interviewer comments it may help you. I interpret him as using whatever room gain is available as long as you are aware of your systems capabilities, think, and listen.

At 26:25 he makes a comment "if the speaker is capable of playing it we want to use that ability"
At 59:30 you can watch that segment and like I said ignore the interviewer and focus on Joss' answers and how he answers.

I use all of the room gain for each speaker in my system and set FS Low at where they hit 0db on the multipoint measurement. When you look at the filter design screen for my left front all my channels are loafing at under 0db, mostly around -10db or less. No individual speaker is being pushed anywhere near where it would be without the combined effect ART provides. If my filters showed boosts above 0db i would be more concerned, but that is not the case and the sound is lovely with no bad noises even with really loud volume.

View attachment 521250

To be honest, many people on the forums suggest setting the FSL with at least a 10–20Hz buffer. I think this probably comes from StormAudio’s recommendations and the traditional crossover concept, aiming to avoid overdriving the speakers by forcing the drivers to maintain output beyond their limits.

But from the interviews I’ve seen with Joss and Masimo regarding ART, as well as the recent one with Joss and Joe, it seems Joss doesn’t really worry about having that 10–20Hz buffer. Looking at the ART Filter, you can indeed notice that even if you set the FSL at the -3dB or 0dB point of the speaker’s bass extension, it doesn’t produce very high output in those low ranges. People seem overly concerned about this.

My consideration is more about the impact of THD on FSL. If I set bookshelf speakers’ FSL down to 50Hz, based on the above results, I don’t need to worry about overdriving the woofer (at worst it’s around -15 to -10dB). But on the other hand, the bookshelf speakers would still have to reproduce sound down to 50Hz, which might affect clarity in the mid-bass and midrange. For example, the Ascilab C6B and F6B I plan to buy already show distortion exceeding 3% below 100Hz. So should I instead choose a 3-way bookshelf design (like the KEF Q Concerto) for better performance?

 

[illusonic]

So should I instead choose a 3-way bookshelf design (like the KEF Q Concerto) for better performance?

3-Way will not have IMD in the midrange caused by bass content. When pushed hard it will be cleaner than most 2-ways. THD alone is not an issue in bass.

 

[illusonic]

Dirac ART utilizes multiple bass speakers to even out bass response in the seating area. Your system seems to have only 2 subwoofers. This will limit effectiveness of ART. Below the crossover frequency towards the satellites you basically have a 2-channel system.

To improve this it might be an option to add 2 more subs if supported by your AV-receiver. In this case x-over frequency to the satellites could be higher (e.g. 80Hz) without loosing effective ART support. If you need ART support below satellity x-over frequency also depends on room size. In a small room the lowest room mode may be somewhere around 50Hz.

As ART works best with a multitude of low-frequency sources, a full setup of (small) fullrange 3-way speakers like Orbit 11 may be the best one can do if starting from scratch.

Regarding THD and IMD there is a good writeup from Purify. Basically THD in the bass is only really bad if causing IMD. Purify makes speakers that don't cause much IMD. Going 3-way is another option to achieve the same.

Distortion, The Sound That Dare Not Speak Its Name

purifi-audio.com

 

[90sNintendoKid]

Dirac ART utilizes multiple bass speakers to even out bass response in the seating area. Your system seems to have only 2 subwoofers. This will limit effectiveness of ART. Below the crossover frequency towards the satellites you basically have a 2-channel system.

To improve this it might be an option to add 2 more subs if supported by your AV-receiver. In this case x-over frequency to the satellites could be higher (e.g. 80Hz) without loosing effective ART support. If you need ART support below satellity x-over frequency also depends on room size. In a small room the lowest room mode may be somewhere around 50Hz.

As ART works best with a multitude of low-frequency sources, a full setup of (small) fullrange 3-way speakers like Orbit 11 may be the best one can do if starting from scratch.

Regarding THD and IMD there is a good writeup from Purify. Basically THD in the bass is only really bad if causing IMD. Purify makes speakers that don't cause much IMD. Going 3-way is another option to achieve the same.

Distortion, The Sound That Dare Not Speak Its Name

purifi-audio.com

I agree with this. I want to experiment with Orbit 11s as Atmos height speakers, at the very least.

 

[Oddball]

Hello everyone,

Since last year I’ve been using Dirac Live ART. When adjusting the F Support Low (FSL) values, I’ve always referred to the specifications provided by the speaker manufacturer along with Dirac’s in-room measurements. As for THD, because most reviewers like amirm and Erin point out that bookshelf speakers tend to have relatively high THD in the bass region, I usually set an 80Hz crossover to the subwoofer to mitigate this issue.

Here’s the problem: Dirac Live ART doesn’t follow the traditional crossover concept. The lower the FSL setting for the supporting speakers, the greater the benefit to the overall system. In other words, within the usable bass range of the speaker, the FSL should be set as low as possible.

Take the Ascilab C6B bookshelf speakers as an example. F3 is around 52Hz. Based on my past experience, I would set their FSL to 60~65Hz. However, if we factor in THD (96dB@1M) from Erin, the speaker’s THD below 100Hz already exceeds 3%. This means that theoretically, setting the FSL below 100Hz could result in “distorted bass” being audible.
View attachment 521211
View attachment 521212

I’m not sure if this topic has been discussed before, but I really want to understand how to balance between a speaker’s bass performance and THD in order to determine a reasonable and appropriate FSL setting.

For most budget bookshelf speakers, their bass THD performance is generally poor (higher than 3%). Does this imply that they are all unsuitable for supporting bass below 100Hz? Or am I just scaring myself, overthinking?


By the way, my 7.2.2ch system is entirely composed of bookshelf speakers+subs. Normally, I set the AVR’s Master Volume to -25dB for movie. The distance from each speaker to my listening position is about 1.3–1.7 meters. Using a smartphone SPL meter to observe the playback volume while watching movies, I measured: minimum 45dB(A), average 85dB(A), and maximum 100dB(A).

You already got some good advice how ART filters work. Couple of points on my part.

-25dB master volume and 100dB peaks don't seem right. Perhaps the reference was some streaming service where the overall volume is boosted. In reference setup you should be at 90dB peaks (115dB sub peak -25 = 90db). So you need to be aware of that if you crank up the volume. Can easily get uncomfortable if you crank up such soundtrack to reference.

For this and many other reasons, it is not so easy to conclude what is reasonable low end support for speakers in ART. It is really very case dependent and depends on the experience of the one operating the gear. While I did not hear yet that anyone blew their speakers away due to ART, perhaps it is just that people did not (yet) get the courage as widely available for only 6 mos. Best is to listen and have a remote close in case you need to tone things down.

How much of THD you want to hear is really individual. I would not push the bed/atmos too far. If needed probably better to buy another sub (assuming current ones are overloaded) to provide more headroom and enable cleaner response. I have fairly large speakers all around, but they never sounded better than with ART when 4 relatively big subs are doing a lot of hard lifting in the ART range.

 

[NanoGuidingStar]

You already got some good advice how ART filters work. Couple of points on my part.

-25dB master volume and 100dB peaks don't seem right. Perhaps the reference was some streaming service where the overall volume is boosted. In reference setup you should be at 90dB peaks (115dB sub peak -25 = 90db). So you need to be aware of that if you crank up the volume. Can easily get uncomfortable if you crank up such soundtrack to reference.

For this and many other reasons, it is not so easy to conclude what is reasonable low end support for speakers in ART. It is really very case dependent and depends on the experience of the one operating the gear. While I did not hear yet that anyone blew their speakers away due to ART, perhaps it is just that people did not (yet) get the courage as widely available for only 6 mos. Best is to listen and have a remote close in case you need to tone things down.

How much of THD you want to hear is really individual. I would not push the bed/atmos too far. If needed probably better to buy another sub (assuming current ones are overloaded) to provide more headroom and enable cleaner response. I have fairly large speakers all around, but they never sounded better than with ART when 4 relatively big subs are doing a lot of hard lifting in the ART range.

At -25 dB MV, I measured a 100 dB peak at my seat during a tense and thrilling scene in a Netflix surround sound movie, using my phone with an SPL Meter app (likely A-weighted, not C-weighted). This was not a rigorous test of a single speaker; I’m just providing these numbers to roughly indicate my usual listening volume when watching movies.

Both of my subwoofers were not overloaded. From what I recall, the SVS volume range is -60 to 0 dB. I set both to -15 dB (after Dirac volume calibration), and afterward, the AVR further reduced the speaker level of both subs by another 5 dB.

 

[jhaider]

IMO for ART compression/limiting (i.e., changes to the frequency response of the speaker compared to the response as calibrated) are far more important to consider than distortion. Unless you're actually hearing overload. Then get something bigger.

 

[Oddball]

At -25 dB MV, I measured a 100 dB peak at my seat during a tense and thrilling scene in a Netflix surround sound movie, using my phone with an SPL Meter app (likely A-weighted, not C-weighted). This was not a rigorous test of a single speaker; I’m just providing these numbers to roughly indicate my usual listening volume when watching movies.

Both of my subwoofers were not overloaded. From what I recall, the SVS volume range is -60 to 0 dB. I set both to -15 dB (after Dirac volume calibration), and afterward, the AVR further reduced the speaker level of both subs by another 5 dB.

While obviously iPhone SPL meter is not a very accurate tool, it might have been in the ballpark. Netflix and Amazon do run some titles quite hot. My usual listening volume is -10dB for e.g. streaming from my Apple library that tend to reproduce master as is (albeit heavily compressed), but on Netflix and Amazon I sometimes back that down to -15/20dB.