I don't really understand the physics behind this, to be honest. For me, cancellation is cancellation, even if it's not complete (not exactly 180° of phase difference when the waves meet), boosting would maybe narrow the dip, but would it really fill it?
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Genelec GLM Review (Room EQ & Setup)
- Thread starter amirm
- Start date
I see. The stability of this method seems dependent on the exact listening position.
the first null is only infinitely deep if the boundary is infinitely rigid (20cm of reinforced concrete), the null with drywall bullshit US houses is probably much much lower in Q than what we experience with our manly European brick and concrete walls.
Now if your boundaries are drywall AND it's not the first null then the nulls could be less than 10 dB deep and they can be potentially managed with EQ.
Basically it all depends on how much of a boundary your boundary is, if your neighbors are complaining about your subwoofers then you probably have it much easier than someone living in a skyscraper made out of concrete.
Yes - and thats the main dilemma correcting all kinds of reflections above 80 Hz , including SBIR . The correcting results gonna be worse everywhere in the room, except at the exactly listening position. You also destroy the direct sound from the loudspeaker . In my experience, It dont work well at all. Its much better:
1. try different loudspeaker positioning for best sound before even thinking about dsp corrections.
2. Use damping material on certain places on the sidewalls, damping out the reflections.
3. Use a thick rug on the floor
4. With dsp - correct only the 3 fundamental resonances in the room, below 80 Hz. My experience: Use dsp shelving for the area 80-600 Hz.
+1
The sbir :the frequency change with the position.
ernestcarl
Major Contributor
A time corrected IR was applied using Trinnov by the acoustician...
Depending on the nature and severity of the SBIR, even simple minimum phase PEQs may be all that is needed to produce a spatially consistent -- within the "box" area that is measured -- audible improvement.
For example, I can compensate the cancellation effect caused by my desk and keyboard tray partially using only the ff. two basic min phase PEQs:
410Hz -2.0dB q=6
550Hz +3.8dB q=3
For those with keen eyes, HF slope of MMM is a bit steeper above due to inexact hand movements.
Okay, that is more than one verification method used! I could post 50 swept measurements for each channel in my virtual head box space, for sure -- though, MMM should really be sufficient enough.
1/6 res, 40 dB scale Wavelet spectrograms
(LP = linear phase HPF to cut off the low bass)
Frankly, I don't see any obvious harm caused by the aforementioned PEQs applied to the critical mids section to get all so nervous and worked up about -- in this one particular case example, at least.
The only thing I'm wondering with this method, is what the direct sound's phase looks like after.
One has to remember also the precedence effect and the HAAS effect.
One should be very careful with messing up the direct sound of the loudspeaker doing corrections from the listening place in the frequency area where the ear is most sensitive. We hear the direct sound and the room-sound from the loudspeakers in a different way than a microphone does. The brain selects sounds after 5 ms, ( 1,7 metre ) and the microphone takes up all the sound.
Last edited:
ernestcarl
Major Contributor
You mean the sort of mirrored IR time correction technique explained in that linked by Frgirard? If you mean that... Ummm, that type of DSP is out of my depth, honestly, and I use relatively "mild" peak filters for the most part -- except there was a layer of (equally very mild) EQ & FIR phase correction left in place which I didn't bother to disable in my previous set of measurements. See below for the sake of transparency:
Effect is visible in the flat GD (i.e. peak energy time) curve in the wavelet graph.
If you just mean the two simple minimum phase filters I used, see below:
Not much damage, if any, I think.
Yep, which is why I use multiple measurements and frequency dependent windowing to double check -- listening tests (say, looping snare drum samples and vocal only tracks) also to triple check.
I was inspired to have another crack at "RoomEQ" today given stuff I've read here & generally, I'm still of the opinion that Anechoic Flat EQ speakers in a room are better than using a mic to correct the frequency response (I was correcting it up to 1kHz based on room measurements whilst leaving an Anechoic Speaker EQ active). Anechoic EQ retains more clarity, RoomEQ dulls it. This is for my JBL 308p Mkii speakers with Anechoic EQ based on Amir's measurements. I haven't given up on Room EQ and I'm gonna read @mitchco 's book that he wrote to get some further understanding, but it's not working for me with REW & UMIK and using 9 tripod measurements within a 1.5ft squared listening area. It's really as if the mic is not capturing what you're hearing. I'm also quite convinced under even the best case scenario of optimising your 2 channel speaker system to a listening position that it will therefore sound worse in the rest of the room vs just an Anechoic speaker EQ, but to this point I've yet to even have a better listening experience at listening position with room EQ with my current attempts with REW & UMIK. For the moment I'm a proponent for Anechoic Speaker EQ without Room EQ but I'll give mitchco's book a read before my next approach attempt........my two current concerns/conclusions being that for 2 speaker 2 channel speaker listening that the mic doesn't seem to capture what I hear and any optimisation attempts will make it sound worse away from the listening position thereby making it less optimal for general "room listening" rather than at a specific point.
ernestcarl
Major Contributor
As long as the anechoic curves are reliable (say, from Amir's Klippel data), and maybe verified from your own gated and/or windowed measurements as well (because of unit to unit variation)... Why not?
There's nothing particularly scary about my own speaker EQ (above 1kHz):
*anechoic curves I extracted from Neumann's Klippel graphs using vituixcad.
Yeah, it certainly helps if the Anechoic measurements are accurate! Certainly Amir's measurements of uncomplicated 2-way bookshelf speakers are accurate, then we're just left with unit to unit variation. In my journey with my JBL 308p speakers I started out with "Room EQ" based on REW & UMIK making changes all the way from 20Hz-20kHz, then Amir measured the 308p at which point I did a smooth Listening Window Anechoic EQ whilst keeping On-Axis Anechoic flat and this proved to be better than any UMIK REW derived EQ's......a number of times I've attempted to combine my Anechoic EQ with Room EQ but each time it ends up worse....the only step up was the Anechoic EQ. I must admit that UMIK/REW "Room EQ" was better than no EQ, but Anechoic EQ trumped all scenarios. My current position of understanding is that room EQ is for people without Anechoic EQ or people with horrible rooms that mean their bass is significantly bloated/distorted in areas. To be honest, my UMIK measurements show some serious variation in bass, but to be fair the room corrections really don't improve that element much without dulling the rest of the frequency range. I'm kinda thinking a room is a room, and you have musicians or people speaking in it, and they don't get EQ'd, your brain accommodates to it - and sure not every room is the best room for a group of musicians and not every seat in the house is the best from an acoustic point of view, but I think there's enough latitude & uncertainty contained within those equations that you can't really nail it down better beyond Anechoic Speaker EQ, certainly if taking the whole room acoustics into account (ie wanting it to sound natural throughout all places in the room). Praps with multiple subs you can cover a significantly large area without deviation after Room EQ, but certainly my attempts at 2 channel room EQ do not exceed the subjective experience of Anechoic Speaker EQ alone. Currently I'm of the fairly loose opinion that Room EQ is a little unnatural, and especially when taking all possible positions within the room to account.
Try to correct only for the rooms fundamental resonances , below 80 Hz and you shall se that the correction is better for the ear .
Anechoic flat is good , you can then use shelving from 600 Hz downwards to get the proper timbre balance in a normal room.
The direct sound from the speaker is more important for the brain, than the room response from listeningposition.
The brain always selects the first incoming sound which is usually the direct sound from the loudspeaker . The brain then lowers other sounds up to - 10 dB coming from the room.
Its called ” The law of the first wavefront” ( The HAAS effect )
The mic doesnt select anything.
Read more here :
https://en.m.wikipedia.org/wiki/Precedence_effect
Last edited:
That's a pretty easy thing to try, so you're saying only correct up to 600Hz based on in room measurements in terms of cutting peaks, and praps combine that with a subjective Low Shelf 600Hz filter to experiment with tonality. I can't say I've tried that exact approach, certainly not combined with the Low Shelf 600Hz subjective filter. I can certainly see how the latter filter would change overall tonality though. It would be counterintuitive though to my measurements, because after Anechoic EQ my measurements don't suggest I need a Low Shelf 600Hz filter, but I might just try it next time.Try to correct only for the rooms fundamental resonances , beyond 80 Hz and you shall se that the correction is better for the ear .
Anechoic flat is good , you can then use shelving from 600 Hz downwards to get the proper timbre balance.
The direct sound from the speaker is more important for the brain, than the room response.
The brain selects the first incoming sound which is the direct sound from the loudspeaker .
Its called ” The law of the first wavefront” ( The HAAS effect )
The mic doesnt select anything.
ernestcarl
Major Contributor
Here's the thing, I am not going to listen for any extended period of time in the worst listening spots of a room. Equalization is applied primarily to the listening areas/spots of concern. Why on earth should I EQ for the space behind my speakers, behind the couch, or the empty corners of the room?
If the listening room is without a real designated listening spot, okay, limit the EQ to the most offending modal issues and/or speaker only (adjusting for HF tonal balance) EQ.
There's also the possibility that the speakers may not be amenable to the kind of EQ you are attempting to use. I cannot speak for your own LSR308p, but I do not apply any boosting EQ to my LSR305s (non-matching, BTW, so individual speaker EQ is required). There are multiple audible resonances in those speakers... and a ringing around 315Hz which I cannot fix with any reasonable amount of EQ or physical mod.
Maybe in the past, but nowdays, I do not EQ to just to get pretty graphs that fit nicely to a target curve. I listen afterwards and do some AB or ABX tests; and if it really sounds bad/unnatural, fine -- just disable the offending corrections.
*How many times did I use the word "EQ" above? LOL! Too, many
Last edited:
sorry, lost in translation. I ment below 80 Hz
Ah, I was just about to EDIT my post to ask you that same angle. OK, so you're saying just use measured mic EQ below 80Hz, then rely on Anechoic EQ above 80Hz? That's the lowest I've ever heard that suggestion, but it's easy to try. I have a few peaks under that figure, and here's a pic showing it, I currently had a +5dB Low Shelf boost at 200Hz included within that measurement that you see, so I might need to remove that boost in order to increase clarity, but the filters you see there are not the total of my attempt at Room EQ, I did a few other filters up to 1kHz. (You can see the peaks are cut down to the blue target curve).
My intuition states that if I try your method I would remove the other filters that aren't pictured here and see how it sounds as well as potentially trying what you're saying with my 200Hz +5dB Low Shelf Filter removed as well. My Anechoic EQ doesn't include that 5dB Low Shelf Filter - I added it to fill in some bass holes whilst cutting the peaks. (My Anechoic EQ is showing it's benefit from 1kHz+ in this screenshot). (My Anechoic EQ would be the orginal measurement shown in this pic minus 5dB Low Shelf at 200Hz).
EDIT: this is the result of my long time Anechoic EQ result measured at my listening position, just pay attention to the highlighted green line, ignore everything else including the pictured filter, it's hard to explain why....but long & short this is my Anechoic EQ that I prefer that's been measured as an average at my listening position (ignore the pictured filter which is a fudge factor because I've not remeasured, too long story to explain), just pay attention to the neon highligted green line which is the result of what I supposedly hear:
Last edited:
Yes , you can use the measured mic EQ below 80Hz, then rely on Anechoic EQ above 80Hz .
Depending how near the frontwall your speaker stands, you have to use shelving from 600 Hz and below , that must be set by ear listening to music.
The shelving filtering should be somewhere between - 2 dB to - 6 dB from 600 Hz and below.
Remember that a good roomresponse curve from the listening place, will have a rising bass from 150 Hz and below, about + 4 to + 7 dB . A flat curve will have to little bass, because your eye/brain sees the room and expects some gain in the bass.
Try it
Edit : Remember that the direct sound from the loudspeaker is very important. Every correction you do in dsp from listening place ( room response ) changes the frequency response in the direct sound somewhat from the loudspeaker , making it sound worse than the mic measurements shows, because of the HAAS effect .
Last edited:
Well I might try messing around with a 600Hz Low Shelf (subjectively) after I've cut the bass peaks under 80Hz, I'll see how it goes, I'll report back after I've given it a shot. (The Low Shelf certainly doesn't want to be in the minus figures though, as my room is not bass heavy overall - the Anechoic EQ is about bass amount perfect in a rough estimate before peak cutting, so if anything that 600Hz Low Shelf would be positive for me after cutting sub 80Hz bass peaks.)Yes , you can use the measured mic EQ below 80Hz, then rely on Anechoic EQ above 80Hz .
Depending how near the frontwall your speaker stands, you have to use shelving from 600 Hz and below , that must be set by ear listening to music.
The shelving filtering should be somewhere between - 2 dB to - 6 dB from 600 Hz and below.
Remember that a good roomresponse curve from the listening place, will have a rising bass from 150 Hz and below, about + 4 to + 7 dB . A flat curve will have to little bass, because your eye/brain sees the room and expects some gain in the bass.
Try it
Edit : Remember that the direct sound from the loudspeaker is very important. Every correction you do in dsp from listening place ( room response ) changes the frequency response in the direct sound somewhat from the loudspeaker , making it sound worse than the mic measurements shows, because of the HAAS effect .
Last edited:
The purpose of room compensation in professional monitoring is to trim the frequency response of perceived-direct sound to become flat, at the listening position. It depends on the room, the loudspeaker, the placement and the listener what is perceived as direct sound.
"The use of single point equalization is a safe choice for measurement of studio monitoring rooms having relatively low reverberation times and well controlled room modes… Spatial averaging across a wide area may come with the risk of compromising the result at the main listening position.”
Quotes from our AES and Tonmeistertagung studies (2016-18), recognising also how spatial averaging is more relevant in other applications, e.g. recreational.
Another topic discussed in this thread, bass management using distributed sub woofers, is generally not ideal in pro monitoring. You want to know exactly what is in the source, and a flat frequency response is not the only goal. Room + monitors (“setup”) create a structure capable of conveying a qualitative experience of envelopment, so a fine setup should also faithfully reproduce envelopment-contrasts in the source, rather than wishy-washy pleasantness.
In a good room with good monitors, nothing beats acoustical summation with listener movement, also at VLF. To me, the experience of envelopment is one of the most precious sensations/feelings of listening to music in a great concert hall, or from a decent reproduction setup.
"The use of single point equalization is a safe choice for measurement of studio monitoring rooms having relatively low reverberation times and well controlled room modes… Spatial averaging across a wide area may come with the risk of compromising the result at the main listening position.”
Quotes from our AES and Tonmeistertagung studies (2016-18), recognising also how spatial averaging is more relevant in other applications, e.g. recreational.
Another topic discussed in this thread, bass management using distributed sub woofers, is generally not ideal in pro monitoring. You want to know exactly what is in the source, and a flat frequency response is not the only goal. Room + monitors (“setup”) create a structure capable of conveying a qualitative experience of envelopment, so a fine setup should also faithfully reproduce envelopment-contrasts in the source, rather than wishy-washy pleasantness.
In a good room with good monitors, nothing beats acoustical summation with listener movement, also at VLF. To me, the experience of envelopment is one of the most precious sensations/feelings of listening to music in a great concert hall, or from a decent reproduction setup.
Similar threads
