Article: Integrating Subwoofers with Stereo Mains using Audiolense

[mitchco]

Hi @3ll3d00d Cool. I will have to look into MSO more, as it seems like a an interesting approach. Thanks for the info Matt. Wrt Audiolense and prefilter, no such feature at this time.

 

@R1200CL Thanks. Wrt Roon, I believe the answer is yes: https://community.roonlabs.com/t/audiolense-convolution-filters-in-roon-resolved-build-298/35603 @dallasjustice may be able to say more.

[dziemian]

Regarding multisub approach. This is Bernt's answer to my early questions about dealing with multisubs. 

 

Hi Tomek,

 

I guess you already have the answer. It is one speaker, one channel at a time with Audiolense.

 

https://groups.google.com/forum/#!searchin/audiolense/upgrade|sort:date/audiolense/gvRMwB4reTM/8PkmYZ94AwAJ

[dziemian]

double post, sorry

[ronkuper]

On 5/13/2018 at 8:48 PM, mitchco said:

Hi @ronkuper this stuff makes my head spin ?

 

That's an old JRiver thread... I sent a note to Bernt and his view is slightly different than JRiver's on the summing. Here is the note I got back: Imagine the following simple scenario: Playing old Beatles over a 2.1 rig. All bass in the source signal is in the left channel. You have to use the factor 1.0 in the bass offloading to make that sound right. If you use two subs, the factor has to be 0.5 since two subs will play 6 dB louder than one.

 

Imagine another example: Mono played through stereo speakers: The bass from the two speakers combined will have 6dB higher SPL than each alone. In the bass, the signal will be in phase unless a very bad setup. To get the same spl out of a mono sub you need to use the factor 1.0 as above.

 

Another case that may be relevant: Playing a center channel through left and right speaker. This is where they use 0.71 in AC3 filter. To get a correct phantom center in the sweet spot you need to redirect the center channel with a factor of 0.5 to both speakers. It will be just like the mono scenario above. The sum will be 6dB louder than the contribution from just one speaker. Note that I am still talking about a corrected pair of speakers that are practically in phase for the whole bandwidth – in the sweet spot. Outside the sweet spot above some frequency the signals from  two speakers will have a random phase difference. The combined output for that region will as a theoretical rule of thumb be approx. 3dB higher overall than each of the speakers. But there will be plenty of frequencies where the figure is 6 db, and also plenty of frequencies where the two speakers cancel each other out. If the listening seats are spread out from  left to right, the best compromise might be to use a factor that is higher than 0.5, but it will be substantially lower than 0.75 (sqrt of 0.5). But I wouldn’t bet much money against using 5.0 here too. Those on the “left wing” will get extra spl from the left speaker and vice versa on the right wing…

 

The errors in AC3 filter will amplify the center channel and the bass above what’s correct and neutral This will probably sound sweet to a lot of listeners, and that may explain why the error prevails (if it does).

 

My own experience mirrors the last paragraph above. If I use 5.1 with JRSS surround processing, the center channel and bass is a bit above what I would normally expect. If I use 2 channels (inside a 5.1 channel container) the output does not have the center or bass channels amplified. At least that is how I remember it, and watching movies, I do like the former, but can switch to the latter for a more neutral sound.

 

You can choose either way and your ears can be the judge of which one you like better. 

 

If you want to drill down further, there is a section in the Audiolense help file on bass management. Also, it may be good to post to the Audiolense support forum to get other user experiences as well.

 

Kind regards, Mitch

 

Thanks a lot Mitch!

 

Before summing with factor 1 (1L+1R) should the whole signal be reduced by 6dB to avoid digital clipping? 

 

 

Ron

 

[mitchco]

Hi @ronkuper I would let Audiolense generate the FIR filters and use them in JRiver with no manual intervention. I have not run into any digital clipping. Cheers, Mitch

[james45974]

The Lynx Hilo looks pretty flexible but I have a question that I haven't been able to locate an answer too.  Can you output both an analog and digital signal at once?  The scenario I am thinking about is having a multi-channel digital signal sent to the Hilo, the subwoofer R&L signals converted to analog and sent to the subs, the mid-high R&L digital signals sent to an outboard DAC of my choice and converted there.  Is this a possible scenario?

 

 

[dallasjustice]

Yes. That’s what I do. But you need to make sure the propagation delay through your R/L DAC is exactly the same every time. This is a big problem with many DACs. You should check with the manufacturer. Many DACs have goofy jitter attenuation that creates a variable propagation delay. These DACs will never work with this setup. My Benchmark DAC works perfectly. You can check the group delay in REW loopback. You need to take multiple measurements to make sure the R/L DAC isn’t drifting. That’s the only way to make sure. 

[james45974]

8 hours ago, dallasjustice said:

Yes. That’s what I do. But you need to make sure the propagation delay through your R/L DAC is exactly the same every time. This is a big problem with many DACs. You should check with the manufacturer. Many DACs have goofy jitter attenuation that creates a variable propagation delay. These DACs will never work with this setup. My Benchmark DAC works perfectly. You can check the group delay in REW loopback. You need to take multiple measurements to make sure the R/L DAC isn’t drifting. That’s the only way to make sure. 

thanks for the response!  Not sure if this is the scenario I would use but it is an option!

[RobertinMn]

I stopped reading your article when I saw your crossover sweeps and delays/speaker. I'm also wondering how your tri-amping because I just can't see enough hardware in the pic. When truly actively bi or tri amping, one bypasses all of the internal passive crossover networks --going straight from the amp to the speaker--nothing in between. Maybe that's the problem with your 15" bass response. I have a set with dual 12" woofers and when I employ them, they will shake the house and then some--2600 sq ft above ground. I only use the subs when I have my dual 8' woofer speakers set up. I like to swap them in and out. I'm not putting the author down, I just don't see the hardware, speaker wire that one would normally see with actively bi or tri amped speakers. 

 I have never used a digital crossover, so maybe I'm wrong here. But I have been truly tri-amping speakers with electronic crossovers for decades. I vertically bi amp the midrange and tweets and horizontally bi-amp the woofers. Sometimes, i switch that and vertically bi-amp the woofers and mids and then horizontally bi amp the tweets with a smaller amp.

 

And then, I add a fourth class A/B amp (380 watts RMS @ 4ohms) per channel to drive my 12" passive subs (JBLs in 1" thick MDF with a 2" thick baffle). You'll never get good, tight bass from a class D inboard sub amp--never. I also elevate the subs so they aren't picking up shit from the floor.

 

I don't think you can ever get this truly right without using a real electronic xover which sums for the subs and outputs to each amp individually. But, my listening room is a lot larger than yours.

 

So, I use a real xover and SPL meter and it takes about two hours to do using a tone generator. Then again, I only employ fully discrete analog amps. the way one goes about doing it is to set each one individually to hit say 98db from ten feet away. Everything has to be equal to make it sound right--but you should know this given career.

[mitchco]

Hello, @RobertinMn, re: ... I maybe wrong here... Yes, you are incorrect. This is an active triamp setup. You can read the gear list here:

 

The added subs make it a triamp setup from the biamp article above. Both these articles on Audiolense, and if you search on CA, two articles on Acourate, show you what digital XO is all about and what can be achieved. I have used electronic XO's for decades as well, including being a FOH sound engineer touring. Digital XO using software and a computer is the next evolution of active cross over technology. One can also time align and linearize the speaker drivers, correct for excess phase and precision eq the system. There is no passive xo in any of these systems.

 

Do you have a measurement mic? A computer? What about an AD/DA converter? If so, download REW acoustic measurement software. Place the measurement mic at the listening position, calibrate REW for 83 dB SPL C weighting, slow integration. Then take some measurements. My book has all of the details on how to do this... Once you have done this, then we can share our measurements in a common format and compare...

 

Kind regards Mitch

[RobertinMn]

19 hours ago, mitchco said:

Hello, @RobertinMn, re: ... I maybe wrong here... Yes, you are incorrect. This is an active triamp setup. You can read the gear list here:

 

The added subs make it a triamp setup from the biamp article above. Both these articles on Audiolense, and if you search on CA, two articles on Acourate, show you what digital XO is all about and what can be achieved. I have used electronic XO's for decades as well, including being a FOH sound engineer touring. Digital XO using software and a computer is the next evolution of active cross over technology. One can also time align and linearize the speaker drivers, correct for excess phase and precision eq the system. There is no passive xo in any of these systems.

 

Do you have a measurement mic? A computer? What about an AD/DA converter? If so, download REW acoustic measurement software. Place the measurement mic at the listening position, calibrate REW for 83 dB SPL C weighting, slow integration. Then take some measurements. My book has all of the details on how to do this... Once you have done this, then we can share our measurements in a common format and compare...

 

Kind regards Mitch

Yes--I have all of those things and my Rane Xovers--including the "howto" for time alignment (which really matters in a large room), setting SPL to match the speakers, etc. It comes with the box. Ranes  can do time align--they can go one step further with moving a jumper or two and time align giant woofers for use with short horns or silk dome tweeters. I prefer discrete circuits--they sound better.

 

I've read up on digital xovers. I put them in the same box as five amp integrated amplifiers. They can't do real summing for subwoofers. You ain't gettin what you think you are.  Is that how you're bi-amping the JBLs? With an all in one surround sound kit? One of my friends has a Rotel like that. Not impressed--but he is always impressed with my gear. Real, fully discrete amps which suck twice as much juice as they put out. Back to what I said before, the D class amp in a subwoofer will not put out audophile grade bass. They're boom boxes for theater effects. Try boosting them off the floor and see if the bass gets tighter.

 

Since all of my gear is analogue, I only require a DAC to go from the Mac to the pre amp.

 

https://www.rane.com/ac23s.html

[mitchco]

Robert, Audiolense does everything the Rane hardware device does, and more, in a software program running on a computer. This site is called Computer Audiophile. There is no box like a Rotel or some AVR/Receiver hardware with a firmware solution like Audessey in this article. There is no "one surround sound kit." In my pic is a computer, like your Mac, but a Wintel box running Windows 10. It just looks like a receiver...

 

I use JRiver Media Center software program to play movies and music. Audiolense measures, analyses, and designs the XO's, time alignment, excess phase correction, frequencies eq etc., all in a commercial software program. The generated output is a 64 bit linear phase FIR filter which is hosted in JRiver's convolution engine in which  the music is convolved with in real time. Therefore, the music arriving at ones ears is time aligned, phase coherent. and frequency response shaped to a known industry frequency response target (e.g. Olive and Toole). This is all in the article...

 

JRiver and the computer are connected via ASIO/USB to a 6 channel DAC and the analog outputs are direct input to 6 separate amplifiers, of different types. The volume is controlled digitally via JRiver's software program. There is no hardware preamp.

 

I think you are stuck in a loop on Class D amps driving subs. If you spend a few moments and continue reading the article, you will see there is an acoustic measurement showing a flat from 12 Hz on up frequency response, matching the preferred target response in both the frequency and time domain. This model Rythmik sub is designed for music, with a very flat measured frequency response. These are not HT subs and there is no boom.

 

How about showing  us an acoustic measurement of your system? How about a measurement of the impulse, displaying a step response like I have above? This will verify that you are listening to a time aligned system. How about a measured frequency response at the listening position? Let's see what you really have so we can compare apples to apples. Use REW, as it is free, but mostly importantly, it is a highly regarded acoustic measurement software program that will run on your Mac and makes it easy to exchange measurements and produce overlays for comparison. Show me the money!

[H_Space]

On 6/4/2018 at 11:09 AM, dallasjustice said:

Yes. That’s what I do. But you need to make sure the propagation delay through your R/L DAC is exactly the same every time. This is a big problem with many DACs. You should check with the manufacturer. Many DACs have goofy jitter attenuation that creates a variable propagation delay. These DACs will never work with this setup. My Benchmark DAC works perfectly. You can check the group delay in REW loopback. You need to take multiple measurements to make sure the R/L DAC isn’t drifting. That’s the only way to make sure. 

By chance, anyone know if there's a variable PD from an RME Babyface (Gen 1) to a Crane Song Solaris? I know the RME has this Steadyclock feature, and the Solaris I would imagine is quite solid.

[H_Space]

On 5/10/2018 at 1:22 PM, dallasjustice said:

What do you mean by that?  I agree that Dirac does a poor job of integrating subwoofers with stereo playback. However, Acourate and Audiolense can accommodate just about any configuration you want. 

 

MSO is very useful for getting smooth bass frequency response across a wide listening area. Some people believe that is all one needs for optimal results. IMO, time domain matters too. I agree with Mitch that linear phase crossovers and time aligned drivers are needed for best results. Either Acourate or Audiolense can accomplish almost anything you need. There are pros and cons between Acourate and Audiolense. I’ve extensively used both for multiple sub integration. 

 

There are are many different subwoofer setup techniques. I think there are 2 categories:

1.  Mono/summed arrays. 

2.  Stereo sub arrays. 

 

However, there are variations within each category. For example, some mono sub arrays are simply time aligned to seated position. Either Acourate or Audiolense can handle these arrays. All stereo sub arrays should be time aligned to listening position. When I say time aligned, I mean flat group delay throughout the crossover region. As you can see from Mitch’s plots, it’s just about impossible to get flat group delay down to 20hz. It really doesn’t matter that much as long as both subs are consistent and the group delay is consistent throughout the crossover. Stereo subs can get a little more complex though. I personally use a 4 stereo sub array (cascaded subs). I believe Mitch linked to another thread which shows how and why I do that. 

 

Finally, there are the non time aligned subwoofer arrays. These are mono/summed sub arrays. Some folks advocate the use of non-time aligned mono subs. I personally don’t see the advantage of using those types of setups. These include Welti. That’s a different topic.

 

However, there is a very effective mono sub array which is not time aligned. It is called “source/sink.”  It’s mostly done using only two mono subs. The frontwall/midwall sub is the “plane wave.”  It is time aligned with R/L speaker. The rearwall/midwall sub is set to opposite electrical polarity from front sub. It is also delayed so that the plane wave and the rearsub wave meet each other behind seated position. The phase rotation of the rear sub needs to be adjusted using RTA function in REW while both subs are playing a LF pink noise. The phase rotation is carefully dialed in until all the room length modes are eliminated. Source/sink has two huge advantages for those with rectangular rooms who have nasty length modes. 1. When properly setup, it can mostly eliminate all length modes, without any DSP using only two subs. 2.  It will eliminate any rearwall boundary interference at listening position. Most people in rectangular rooms sit behind the room length midpoint. In these cases, the rearwall will likely destructively interfere at a specific frequency based on its distance from listening position in relation to the front wave source distance to listening position. This is called the “Allison effect”. Others call it SBIR. Still others call it a “null.”  They are all the same thing. It is NOT a room mode. Because it is non-minimum phase, DSP can’t fix it. Only speaker placement can overcome this issue. Of course a rearwall sub setup in a “source/sink” array will eliminate this boundary interference. 

 

Back to your question about Acourate vs. Audiolense. The only sub array I know about that Acourate can do which Audiolense cannot do is this “source/sink” array. The reason is that Acourate Convolver can be setup to simultaneously measure two mono subs with delay added to rear sub. OTOH, Audiolense can not measure in this way. Audiolense can only measure one channel at a time. 

 

I’ve tried just about every subwoofer array I’ve described, except MSO.   In my room I’d rank the 4 stereo cascaded sub array first. Second place goes to “source/sink.”  Other rooms are different. There is no ideal or perfect setup.  You have to tryout different arrays in your room, measure them and see what measures (frequency and decay) best. 

 

Subwoofery done right can be a very iterative process. This is true for most any array. Because there may be a lot of move-and-measure, it’s important to have an easy/fast method to loopback measure each array. This is where Audiolense beats Acourate. From the time one setups up a speaker array with crossovers to the time the .cfg files and FIR impulses are in a folder for Jriver/Roon, it may take 5-10 minutes when you get the hang of it. Acourate won’t go that fast. You’ll need to create your own .cfg files, crossovers and the speaker setups in Acourate Convolver will take some serious practice to get really fast. I know Uli can do it very fast. But my brain works much slower. 

 

I think both Acourate and Audiolense are outstanding. I’d say buy both. That’s what I did. I still use both of them; best audio money ever spent. 

 

Michael. 

45

Where can I learn more about this 4 stereo cascaded sub array and “source/sink” array?

[nefilim]

hi @mitchco - thanks again for all your contributions. i'm trying to monkey-see monkey-do but not faring too well

 

i'm quite uncertain about my phase results - i made a few a short video to show where i'm at below.

 

donny is out of his element here but that looks quite awful to me and quite unlike the results you got, is that a huge amount of pre ringing or am i misunderstanding? 

 

i'm using the default TTD filter procedure as shown below but if i were to play with a custom procedure, i'm unclear how to adjust the windows based on the measured phase response, i'm wondering how you made the decision in your first article:

 

"First, let’s try a shorter time correction window in the bass, so now I have 2 cycle window both in the low and high frequencies as entered in the True Time Domain Subwindow:"

 

is there something in the measured data that can guide one on how to adjust the time windows? especially since i have huge peaks from the L sub in the phase response as far out as 45ms and 67ms! i do see in the simulated response that those peaks have disappeared... so that appears to be a good sign at least?

 

 

just to add a little bit of background:

 

• my room is similar to yours, i'm against the long wall left of the center, on the right rear it opens up to the kitchen
• 2 x Kef (semi) full range speakers driven by a power amp (channels 1 & 2)
• 2 + 1 x JL Audio subwoofers
  • right channel is two cascaded subs, both should be receiving the same signal (channel 3),
  • left channel is a single sub (left of the center of the room) (channel 4)
    • anywhere here is just terrible, both in freq response and phase response - i really want to find a reasonable location for the left sub somewhere here but it just haven't been able to yet
• RME ADI2-Pro in multichannel USB mode, channels 1/2 goes to power amp L/R and channels 3/4 goes to 3 subs
• i've manually adjusted the phase knob on the left sub which is closer to the seating area so that the first peaks of the sub impulses lines up (they go out of phase later...) - i'm sure AudioLense can correct for that but figured might as well since these subs (and speakers) are used in a home theater system too (Audyssey isn't the best:) 

 

here's my speaker setup, i've gone quite a different way than you have (since the mains have integrated crossovers - unfortunately). coincidentally, going this way i have noticed i cannot enter a low pass value on the third tab for the subs, i believe AudioLense just uses the high pass value for the mains as low pass for the subs (judging by the XO graphs), i would have liked to play with them overlapping a bit more. 

 

thank you!!

peter

 

[Ralf11]

Any thoughts on how well this would work if Magnepans were used as the main speakers?

[nefilim]

oh also, wondered what does the measured impulse response say about the polarity of the mains vs subs and if that's a problem?

[mitchco]

Hi Peter,

 

Nice gear! And thanks for the information. That's not preringing you are seeing, but the result of using a linear phase target, which does looks correct. Try this, load your linear phase target, don't generate a correction, turn off all chart view details, but leave the target button enabled. Switch to the impulse view and then select from the Analysis menu, Simulation, Step Response Simulation plus target and you will see what I mean...

 

I would use the same target, then click new target and switch to a minimum phase target, save that and use that for your correction. You can check what that looks like with the procedure above before you generate a correction. Then you should see a nice right triangle step response in the simulation. Give that a listen too. I ended up liking a mixed phase target the best. 80% minphase on bass and 100% on top. Look for mixphase menu item in the target designer.

 

Up to you, but I would (initially) turn off any phase adjusts on the subs and let Audiolense do it's thing. The simulated step response looks good over time and you will see with a minphase target, the right triangle step will be there. Also, I would possibly drop the overall subwoofer levels down a bit to better match the level in the mains. You kinda want to shape the response with levels before you add a target. That way, there isn't so much filter insertion loss as the filter needs to attenuate the overall sub level. to be in line with the mains...

 

Under the measurement menu, you should see Automatic Polarity Correction enabled, so no worries on polarity and won't cause any issues, regardless of polarity of subs.

 

Under the Correction menu, is the Correction Procedure Designer. Click on that and select TTD measurement and click on new procedure and give it a name.  I would uncheck Prevent treble and bass bass boost. I would up the Max correct boost to 12 dB. I would turn on TTD correction per driver in addition to TTD correction. From there I would enter some values in the TTD subwindow only. This where you want to play around a bit. Try some small values like 3/2 then 6/3 and note the frequency and step response differences and then give the filters a listen. The longer correction in the bass, will tighten up the bass considerably and may get rid of that dip you have, which is not likely audible anyway.

 

Have a look at the manual for XO help with respect to width of the overlap. I went with the default values, but it is another area for a bit of experimentation.

 

Lastly, you may want to reach out to Bernt on the Audiolense support forum with respect how to best handle your 3 subs from an Audiolense configuration point of view. What is the recommended approach and tradeoffs... 

 

Your results are looking good! Just a bit more fine tuning with the info above should get you most of the way there.

 

Hope that helps.

 

Mitch

[mitchco]

19 hours ago, Ralf11 said:

Any thoughts on how well this would work if Magnepans were used as the main speakers?

 

Yes, works just fine. There are folks on the Audiolense support forum that have used this for both Maggies and electrostatic panels. Also, one of the reviewers of my book at Amazon uses it on an Open Baffle design...

[Ralf11]

Thanks Mitch!

 

BTW, my search on their forum did not turn up any hits on magnep or maggie

[nefilim]

Thanks so much @mitchco - played around a fair bit and found the default TTD filtering works best for me, as far as I can tell at least, I'll share another video soon :) 

[RobertinMn]

If you haven't installed bass traps in the corners of the room, you should. Makes a huge difference in the quality of the both the bass, mids and highs. Corners include the wall to ceiling points as well.

[blue2]

Thanks for another superb article. As a beginner thinking of going down this route in a 'conservative' manner I'm thinking of 2 or 3 steps:

1. Buy a calibrated microphone and measure the main speakers and sub separately. Adjust (single) sub crossover frequency and volume to integrate with main speakers guided by frequency response curves and remeasure combined sub and mains. Try out various test tracks with and without the sub - schematic would be as in 'Full Range' picture.

 

2. Buy some software (HQPlayer/Accourate/Audiolens/Audirvana/JRiver) or use freeware foobar2000/Sox/REW and create convolution filters treating the sub and mains as one full range system. So calculate convolution based on the system set up in step (1). Evaluate system with and without DRC. Probably try LF range alone first since mains and sub overlap roughly 50-100Hz.

 

3. Progress to splitting sub as separate output as in 'Split Range' picture. If I'm spending money on software I'd like to include upsampling in the feature requirements. Since convolution needs to be handled in the computer, so must the master volume control so the sub and mains maintain integration. So I need a music player such as JRiver which handles convolution and sub output. This could be connected by line out or by a computer sound card/onboard DAC. My DAC is 2 channel and I wouldn't want to spend more on a multi channel. Since volume control is software controlled the preamp is unnecessary, and the DAC could be set at a safe max volume (for speaker protection).

 

So the questions - my concerns are volume control and which software package to buy. Is my reasoning correct and is this a workable plan? Does Audirvana have the required features to replace JRiver as used by @mitchco? I believe HQPlayer could do the job but not sure about mains + sub output via separate devices as seen by the computer.

 

TIA

[mitchco]

Hi @blue2 Thanks for the kind words. I see no-one has got back to you on this. Purchasing a calibrated measurement mic and REW is excellent acoustic measurement software to get you going, is a great idea. I don't know how you are going to get around the 2 sounds devices as seen by the computer though... I don't have any experience with Audirvana or HQPlayer,... However, getting the analog split off your preamp to the subs should work. With REW you will be able to assess your setup and then make a plan from there... Of course, I recommend Audiolense as the fastest way to integrate the subs with some room correction...

[blue2]

I've now bought a miniDSP UMIK-1 and using REW measured my Tannoy Kensington SE's and Wharfedale SW150 HT subwoofer. It's apparent I have a room node at 40Hz which main speakers and sub are exciting. The sub is mainly for HT and I was hoping would supplement the Tannoy's but they don't go low enough and just reinforce the 40Hz node.

 

So my next consideration is a better sub. I've noted @mitchco is using Rythmik Audio F12 (about $1100 inc. shipping) but couldn't find  anything for @dallasjustice or @3ll3d00d who also seem to have invested in hifi audio sub set ups. Another option is maybe 2 x BK's XXLS400-FF Subwoofer 400W RMS 12" but this is non-servo controlled.

 

Any comments/recommendations from folks that have tried/demoed subs in this budget range i.e. 1 or 2 subs totalling <£1000?