New REW measurements, hope knowledgeable people will help me analyse them.

[gorman]

DSP can not correct reverberation.

Ok, perfect. Clear and concise. I will surely test removing those two, then. Thanks once more.

 

[gorman]

DSP can not correct reverberation.

Removed back panels, remeasured. Mic position was carefully done in order to be as close as possible to previous measurements (don't think it matters as much for RT60 but I could be wrong).

Results don't seem to warrant the hassle to me (hassle in not having back panels, from an aesthetic point of view).

Orange is without back panels. Yellow is with back panels. I show from 496Hz, as your calculations suggested.

Furthermore, reading this discussion: https://www.audiosciencereview.com/...nding-domestic-listening-rooms-measure.11672/ brought to my attention that for multichannel lower values are considered better, for better positioning. As my interest in multichannel (5.1 and Atmos) music is quite high and ever growing... I think I'll settle. Especially considering that sound doesn't sound "dead" to me.

I found this interesting as well:

According to the formula offered, my range should be between 205ms and 277ms. So, yeah, I'm a bit low but maybe not as low as your calculations suggest (plus the multichannel consideration above). Not saying your calculations were wrong, mind, just reporting on an additional piece of info I stumbled upon.
Further listening suggests that the formula applies to omnidirectional speakers, and that with front firing ones you get lower measurements. Which would put my numbers in a better light too.

 

[Chris A]

First off, it looks like your subwoofers are phase aligned with the mains but the impulse peak is 10ms late. Worth trying to adjust that with distance in the AVR settings. Should bring down your group delay and improve bass quality around the crossover region.

I get more like ~70ms late. That's way too much delay. I'm not even sure how you got that much delay. Clearly, finding a way to get that delay under 10 ms or less should be a priority.

Your acoustic treatment is too thin, I wouldn't even call those bass traps because they won't be doing anything below 100Hz. For your size of room I would actually reduce the number of panels and make them 2-3x thicker.

Most Owens-Corning 703-sized bass traps are only good down to ~72 H due to their finite length. If you want to extend their effectiveness to a lower frequency, you need to put your bass traps end-to-end to increase their effective length.

That deep shelving unit is causing a strong reflection for your centre channel just 1ms after the initial impulse. Pulling the speaker forward may help dialogue quality. Alternatively, covering the area in front of the speaker with 2-3" of absorption. It looks like you have some thin felt currently.

I, too, noticed the nearfield high frequency reflections at 1.15ms, 2.03ms, 5.91ms, and 6.91ms. These early reflections significantly affect the stereo imaging, IME.

Putting "stuff" (equipment and racks/"entertainment centers") between your stereo loudspeakers has significantly detrimental stereo imaging/soundstaging effects, especially if your loudspeakers do not have controlled directivity to below 90-100 degrees horizontally and about the same vertically (or less--like 60 degrees). I'd try placing absorption all over the acoustically reflective objects between your loudspeakers, then have a good listen to your best stereo imaging recordings. You'll figure out what you'll have to do after that exercise. The step response plot shows significant "echoing" in the first 20ms. The Schroeder Integral plot shows a couple of step-downs at 1.19ms and 8.89ms. If you convert those delays to distance (332 m/s speed of sound), I think you'll find where the big nearfield reflectors are that you'll need to do something about.

I would NOT add to the absorption in the center and rear portions of the room. In fact, I'd experiment (as I see you're now doing) removing some of those absorption panels.

I also recommend flattening the bass amplitude response once you find a way to reduce the bass delay to less than 10ms. It's a bit too much presently, but time aligning the sub(s) will give you a subjective sense if more bass--by as much as 3 dB. (!)

Group delay gets really noisy below ~300 Hz, so any EQ that you're trying to use to tame bass peaks below 300 Hz is not going to work very well if you move your microphone (or ears) move than a few millimeters left or right. This is a consequence of having a small listening room, so there's not much that can be done. Some of those group delay peaks below 300 could be related to your nearfield acoustic reflections between your loudspeakers.

I'll stop there. But there are probably more things that can be done to improve the listening experience in-room.

I encourage you to post new REW measurement files after you've made acoustic changes to your room. Most of the observations I made above were enabled by that .mdat file--not your screenshots.

Chris

 

[gorman]

I get more like ~70ms late. That's way too much delay. I'm not even sure how you got that much delay. Clearly, finding a way to get that delay under 10 ms or less should be a priority.

I honestly wouldn't put too much attention to the LFE channel measurement. It appears Dirac does stuff there which doesn't interact well with Windows when outputting only to LFE. And there wouldn't be a point in me measuring with Dirac deactivated. All the measurements of the speakers have LFE integrated.

I encourage you to post new REW measurement files after you've made acoustic changes to your room. Most of the observations I made above were enabled by that .mdat file--not your screenshots.

I attach a new file. The names of measurements indicate what's been measured. L, R and C Dirac are measurements with everything untouched. Back absorption in place, Center speaker angled normally. Then there's L and R without back panels and C with the speaker tilted higher (11.5 degrees vs 6.5 degrees) in order to see if it made a difference for early reflections.

I'm not knowledgeable enough to tell but, eyeballing it, it doesn't seem that there are significant differences.

EDIT: after all these days of measuring, remeasuring, listening, relistening... I wonder if I shouldn't just go with the first comments that stated measurements looked great and start using my system. Perfect is the enemy of good, after all. And, using Dirac, there's a limit to what I can do in the digital domain. Plus, overhauling placements it's really not an option. Nor is the removal of further absorption, apart from the panels behind the front speakers, as I've already mentioned.

 

[gorman]

I honestly wouldn't put too much attention to the LFE channel measurement. It appears Dirac does stuff there which doesn't interact well with Windows when outputting only to LFE. And there wouldn't be a point in me measuring with Dirac deactivated. All the measurements of the speakers have LFE integrated.

I remain available to remeasure LFE, just in case. But I would need to have a proper best practice set of rules to follow for my specific case. If it would be better to use a file for the sweep, with included timing reference, let me know (although I'm not sure that would work for subs). It's just that when I measure LFE, all three subs are used, which is why I suspect there's some shenanigans operating in the background, leading to strange results.
I'd be surprised if Dirac missed the mark for subs by 65 milliseconds... and I'm almost sure that 65ms would be quite detectable in normal listening, wouldn't it?

I will try following this list of suggestion for measuring https://www.avnirvana.com/threads/rew-alignment-tool-guides-or-manual.6814/post-50646 I still have three subs to contend with. So... I guess I'll try to measure it separately and together (with main fronts disconnected, as suggested in that list). But not today.

 

[gorman]

I will try following this list of suggestion for measuring https://www.avnirvana.com/threads/rew-alignment-tool-guides-or-manual.6814/post-50646 I still have three subs to contend with. So... I guess I'll try to measure it separately and together (with main fronts disconnected, as suggested in that list). But not today.

And it was today... this place is a disease, I'm telling you.

Attached measurements done following the instructions on AVnirvana.

• Mic at LP
• Lexicon in stereo mode
• Current Dirac settings unchanged (XO, EQ, distances...)
• Current SW settings unchanged
• REW Acoustic timing mode selected
• REW Use L main for the reference channel
• REW sweep range set to 20-20k Hz

Measurements:

1. FL (alone; SW off)
2. FR (alone; SW off)
3. SW (alone; Measure the redirected bass by using FR channel with the FR speaker disconnected)

I don't have Lexicon but I used Stereo mode, no processing basically.

File too big to attach, download it from here (valid 30 days): https://easyupload.io/4bmtp7

It seems to me that Dirac chooses to align on the "start" of the impulse and not on the peak. If that's correct or not, I have no idea. I'll probably end up writing on AVnirvana as well...

 

[Keith_W]

It seems to me that Dirac chooses to align on the "start" of the impulse and not on the peak. If that's correct or not, I have no idea. I'll probably end up writing on AVnirvana as well...

I am sorry I do not recall if you are using linear phase FIR or minimum phase IIR. If it's linear phase then the impulse peaks should be aligned. If it's min phase then align at start of impulse.

 

[gorman]

I am sorry I do not recall if you are using linear phase FIR or minimum phase IIR. If it's linear phase then the impulse peaks should be aligned. If it's min phase then align at start of impulse.

I freely admit the terminology and concepts fly well over my head. What I know, and I can repeat just like a parrot, is that Dirac uses a mix of linear and minimum phase filters.

I found this: https://www.dirac.com/wp-content/uploads/2021/09/On-equalization-filters.pdf where I think they explain the rationale behind the choice.

And this: https://gearspace.com/board/studio-...05-room-eq-linear-phase-vs-minimum-phase.html confirming the mixed approach.

 

[ZolaIII]

I freely admit the terminology and concepts fly well over my head. What I know, and I can repeat just like a parrot, is that Dirac uses a mix of linear and minimum phase filters.

I found this: https://www.dirac.com/wp-content/uploads/2021/09/On-equalization-filters.pdf where I think they explain the rationale behind the choice.

And this: https://gearspace.com/board/studio-...05-room-eq-linear-phase-vs-minimum-phase.html confirming the mixed approach.

Just keep it simple, listen to his guidelines and still do former (as I doubt it's linear phase and you can make minimum phase FIR it's just it won't have precision on the beginning of the range compared to infinite one).

 

[gorman]

Just keep it simple, listen to his guidelines and still do former (as I doubt it's linear phase and you can make minimum phase FIR it's just it won't have precision on the beginning of the range compared to infinite one).

I think there's some serious misunderstanding on your part... I am using Dirac as room correction. This has been made clear from the OP and in several other posts in this discussion. I am not "making" minimum phase or linear phase anything. Dirac does, and apparently does this with a mixed approach.

Understanding the approach, apparently, it's important to properly evaluate timing alignments, considering what Keith_W mentioned. Interpreting graphs as minimum phase makes the time alignment correct for the subwoofers. Interpreting them as linear phase makes the time alignment seriously wrong. All this if I understand correctly, which is not a given.

So, while I have everything to learn, I beg... keep in mind I use Dirac. I don't have access to many specific adjustments, I can change just the target curve and the crossover point for separate group of speakers (Front, Center, Surrounds, Top Front, Top Back).
Theoretically Dirac does everything else (to find out to what degree of success is why I measure in REW).

 

[ZolaIII]

@gorman it's AVR so most probably FIR and not very high resolution one either.
Understand your filters, I do it by hand (no boundaries). If something (or someone) can do it it doesn't mean it should as that's plane stupid or at very list put into boundaries (taking care that impulse response doesn't go much off and definitely not pre ringing). That's all we told you to check in the first place. Your setup is fine! When the force awakens push the speaker from the side wall and do proper calibration and correction. Take a brake, have a sneakers and remember all it needs is love (and some; time and effort... and hard work).

 

[gorman]

it's AVR so most probably FIR

That does not appear to be the case. https://www.dirac.com/wp-content/uploads/2021/09/On-equalization-filters.pdf

 

[Keith_W]

If it is Dirac on a PC with output to Dirac's own convolver then I believe it is linear phase FIR. If the DSP device is your AVR or MiniDSP, then it will be mixed phase. So it should be aligned to impulse start.

 

[gorman]

If the DSP device is your AVR or MiniDSP, then it will be mixed phase. So it should be aligned to impulse start.

Yes, it's on my AVR. All is good in that regard, then (for a given measure of good but it's not the disaster it would have been with 65ms of difference).
Thank you.

 

[TheZebraKilledDarwin]

Your measurements look fine but only YOU hear, where you want an improvement:

I don't know why it is avoied like the plague, but be aware, when the measurements look decent, or even good, always the ears must be used to tune any setup.
Subjective descriptions are necessary as guidance to look at certain measurement parameters, which might have been ignored.

I give you an example: you have a SPL dip in the bass region. But for your setup, it might sound WAY BETTER with ALL kinds of mixes, to leave the dip instead of trying to achieve a better looking curve.
The flatter response might just cause fatigue with certain tracks, that have lots of energy in a region, the room naturally would cancel out.

Or another typical situation, because in nature nothing comes for free:
Improving one parameter at the cost of creating an adverse effect somewhere else. So it is YOU who must decide, which compromise sounds better.

Again an example of a dip in the bass region at the corssover: by just looking at that dip, and trying to make it vanish with an optimized crossover, might ignore a much more important and problematic ringing at a lower frequency. In fact I had such a situation in my HT.
Ignoring the dip optimization but eliminate the ringing by the 2nd sub, was the correct move.

No automatic room correction system can make these decisions for you. Let your ears guide you, where the biggest problem is.

A simple but good practice to use the ears, is to make sine tone test files for the bass region, in 5 Hz steps. Just listen at lower volumes. you don't need a trained ear, you will immediately feel, if something sounds more or less annoying.
No frequency should make you want to vomit. But if you notice such frquency ranges, you have maybe found a huge problem worth investigating further, looking beyong what the SPL over freqeuncy says.

Another simple way to recognize problems by ear, is to focus on mixes that do not sound good on your setup (its totally wrong to use "reference mixes" - you need to use decent, but not so great mixes and make them sound well; or use mixes that are borderline - i.e. EDM mixes with very loud kickbass for the club).

Then, find out, how these mixes differ from better sounding mixes in your room, by comparing the source material on a hires analyzer. The analyzer might give precious hints, where the problematic frequency regions of your setup are.
Then you have an indicator, can go into REW and check also the TIME DOMAIN (decay analysis might show delayed resonant buildups, while the SPL graph looks fine). Just as example, you get it.

ALL parameters, frequency and time domain must be taken into account, but only ears in your room with your setup, can tell you, what parameter is causing a problem.
Acusticians always use their ears to tune setups, nobody is just looking at measurements. They can be extremely misleading.

 

[gorman]

I don't know why it is avoied like the plague, but be aware, when the measurements look decent, or even good, always the ears must be used to tune any setup.
Subjective descriptions are necessary as guidance to look at certain measurement parameters, which might have been ignored.

Oh, believe me... that's all I've been obsessing about for the past two weeks. I've got a by now very long playlist of tracks (normal stuff I like, not reference stuff I don't care about) and I'm going back and forth with different curves (sometimes slightly different curves). With Dirac I don't have many parameters on which to act, just crossover point and target curve. Nonetheless I've spent a crazy amount of time fiddling with those. Measuring once in a while.

Currently experimenting on a slight dip in the 1.4k-2.5k region, to try and medicate a bit of harshness (it might be coming from the reflections we talked about before).

Moving speakers and subs in the other position I would have available will be a last resort item, still don't know if I'll manage to find the will to try it.
Speakers closer to the center, moving the left one from the wall a bit. Subs more to the side. Basically inverting speakers and subs from what you see in the first post picture.

Another thing I might want to try is changing the facing of the subwoofer sitting beside the couch. This had been discussed previously but now I'm thinking of changing the facing because of the way sub bass "resonates" inside the couch. When I'm sitting and listening to bass heavy content is almost like having a tactile transducer. Which... I don't know, don't think it's good. It's kinda disconcerting, to be honest. Maybe it won't change a thing. Probably I could try the effect without recalibrating Dirac. EDIT: tried right now, it doesn't change a thing. It probably has more to do with how the couch is structured internally. Ok, then, free tactile transducer is gonna be... (I'm exaggerating here a bit).

 

[TheZebraKilledDarwin]

Oh, believe me... that's all I've been obsessing about for the past two weeks. I've got a by now very long playlist of tracks (normal stuff I like, not reference stuff I don't care about) and I'm going back and forth with different curves (sometimes slightly different curves). With Dirac I don't have many parameters on which to act, just crossover point and target curve. Nonetheless I've spent a crazy amount of time fiddling with those. Measuring once in a while.

Currently experimenting on a slight dip in the 1.4k-2.5k region, to try and medicate a bit of harshness (it might be coming from the reflections we talked about before).

Moving speakers and subs in the other position I would have available will be a last resort item, still don't know if I'll manage to find the will to try it.
Speakers closer to the center, moving the left one from the wall a bit. Subs more to the side. Basically inverting speakers and subs from what you see in the first post picture.

Another thing I might want to try is changing the facing of the subwoofer sitting beside the couch. This had been discussed previously but now I'm thinking of changing the facing because of the way sub bass "resonates" inside the couch. When I'm sitting and listening to bass heavy content is almost like having a tactile transducer. Which... I don't know, don't think it's good. It's kinda disconcerting, to be honest. Maybe it won't change a thing. Probably I could try the effect without recalibrating Dirac. EDIT: tried right now, it doesn't change a thing. It probably has more to do with how the couch is structured internally. Ok, then, free tactile transducer is gonna be... (I'm exaggerating here a bit).

It doesn't matter where a sub is placed, because we can't determine the direction of the wave.
Or: place a sub in corners is another beauty.

Contrary to widely held believes, some even spread by companies, even at 80 Hz it makes a huge difference where the sub is placed.
One sub on one side of the couch might be sounding great, because its close, but the asymmetry will never cease to be a problem.

IME a sub HAS to be placed symmetrical, the closer to the center channel the better.
Because the time domain carries the dynamics, distance is very important, as you can easily hear, by placing the ear closer to the sub. But the sound perception is also different, if the pressure wave reaches both ears at the same time, or one ear first. Especially at the extremes of a sub on one side only. It could even create a nauseating, asymmetrical feeling in the ears.

No matter how much an EQ was flattening the frequency response for the single point microphone membrane, we are listening with two ears and our brain recognizes this difference.

 

[gorman]

You are welcome!

Let us know what Dirac will answer

I don't think a bass trap would make any difference but I will let others chime in too

Got an answer from Dirac.

They state that "The Bass group delay problem is something the dev team is already investigating". From this, I gather that they think they could be doing better in this regard. Hopefully they will.