Room Equalization: From Brute-Force Flattening to Perceptual Optimization

[Axo1989]

That's true. On the other side, the more complex algorithms are not on the cheap side: Bacc, Trinnov, proper multisubs etc all needs specialised very expensive processors and or licenses. If you have a lot of channels in your home cinema, same thing. It is much easier and much faster to develop it in software. It is not as convenient as a HW device with a well tested software inside that's for sure.

Yes, this may be resolved I think by using a good computer as source. An M-series Apple Silicon SoC for example is considerably more powerful than those typically used in hi-fi devices (not to mention potentially using GPU and/or neural network resources as well as CPU). Of course your OS, audio player and DSP software need to behave as well.

 

[YouDontHaveEnoughBass]

I generally understand why only EQing the bass is the way to go with room measurements, but I've always wondered whether I should be EQing something like a bump around 500hz from the speaker being very close to the wall, or whether I should just leave this be - what would you suggest for this? (i know the obvious answer is 'move the speakers from the wall' but its a living room and I'm following Genelecs recommendation to have them as close to the wall as possible if big distances aren't an option)

 

[Digital_Thor]

Yes "prior to DRC efforts" that is clear.

But besides that, I'm not sure which you mean, inside a multi-driver enclosure passively?

Or via active DSP using "as anechoic as possible" measurement feedback, independent of crossovers?

I mean literally any type of DSP, DRC included. Yes, you can beam-control by having a multitude of amplifiers and drivers like in Beolab 90, Kii 3, for example. But any "normal" speaker with whatever amount of drivers - be it active or not - you can't change the dispersion of a speaker with any type of line-level adjustments.
Again - 2 dimension vs 3 dimensions - they don't transition fully between each other.
Anyone using any set of stereo speaker setup, can only change the dispersion(3D) by changing the design of the speaker. Just take a set of B&W speaker and a set of KEF's(whatever version, almost). No matter what you throw at the B&W - with any DSP in the entire world - then it will still never ever sound like a KEF. A DSP can't do that.
What some might do, is to mix up acoustics and any given speaker, and then force some kind of mix of reflections and creative DSP trickery, trying to mimmick a different type of speaker with a different type of dispersion - but that will still never be the same, even though you might "trick" the measurement microphone into showing you a different steady state listening position response - it still won't fly, sorry
If any of these DSP promises were absolute truth, then we could just throw any driver in any box, and then just smack the living crap out of it with enough power and DSP, and expect excellence - which we of course all know is not true.
We can definitely take less perfect speakers and smooth them out a bit, especially if we are within several limits of psycho-acoustical theories, because we can definitely not hear everything that we might be able to measure - just look at all the test's done with DAC's and amplifiers. So, sometimes - good enough, is just good enough.
I have a fully active 10 channel setup with loads of power and so on.... but I only use 2 PEQ's for the tweeter, 3 for the midrange and one in the bass - for tweeter and midrange, this is not room correction but the filter to make it anechoicly smooth and flat(spinorama style) The bass is just to slightly tame one reflection from the room. With the subwoofers, I'm still working on it, but it's pretty smooth and flat by just having 4 subwoofers scattered around the room and one big EQ for the room mode.
According to MSO, I can get a very smooth at pretty flat response with no use of delay, FIR, all-pass and so on - but just by using gain, different over-lapped frequencies(Geddes approach) and a bit of EQ.
I sometimes believe that people think, that just because we now have lots of DSP power, it HAS to be used, rather than building good speakers, and only correct what is needed - no more - no less.

 

[john61ct]

you can't change the dispersion of a speaker with any type of line-level adjustments.

Of course, I did not know by "speaker problems" you were talking about just dispersion issues.

Nor can I imagine trying to make a speaker of a certain technology sound like another completely different? bizarre idea...

> According to MSO, I can get a very smooth at pretty flat response with no use of delay, FIR, all-pass and so on - but just by using gain, different over-lapped frequencies(Geddes approach) and a bit of EQ.

Do you agree?


> only correct what is needed - no more - no less.

I agree with that, but realise some issues simply cannot be fixed, without creating new problems, swapping speakers, changing the room

"Good enough for now" needs to be accepted sometimes

 

[pierre]

Next iteration of RoomEQ is done and available soon!
What's new:

- Merged RoomEQ with XTC (Cross Talk Cancellation): if you have 2 mics or ideally a binaural headset you can use your HRTF with your room and optimise for it. Code is done, testing is starting. I expect it will need a few iterations to work well. This has a lot of potential. I have 2 mics but the binaural mics/headset are on their way from Thomann.
- Replaced optimisation algorithm by a new one which is as good but much faster for the problem at end.

Future sounds pretty cool, now that the code infrastructure is in place, I can iterate fast.

Questions for the community:
- First item requires proper phase measurement and so far i get different answers from different microphones, SPL is consistent but Phase is not (which means there is an issue somewhere with the mics). Better ones are expensive. If someone can advise on what to try/buy, i am interested. I currently have one measurement Senn, a few umik 1 and 2.
- Can I do a poor man head with 2 mics and some foam in-between? what would be the ideal density?

P.

 

[john61ct]

I am a strong believer is both the test signal output, and the measurement mic going through the same interface + preamp,

high-precision, electronic loopback referenced measurement, specially for precise time-of-flight delay and phase measurements

Rather than relying on acoustic timing references which are affected by room reflections and varying temperatures, eliminate the transfer function timing

which means XLR mic

Read on from here for mic reco's.
https://www.audiosciencereview.com/forum/index.php?threads/asr-acourate-users.47057/post-2541445

I think the more expensive investments are for businesses, and maybe pay off more for FR than timing

 

[Keith_W]

- Can I do a poor man head with 2 mics and some foam in-between? what would be the ideal density?

You could use a MiniDSP EARS?