Share your in-room measurements?

[Chromatischism]

That is probably related to our interaural time difference and the ability to detect the source direction of a sound (in other words, if you can't detect the source direction, wouldn't it just all mesh together?) You're right, it does depend on frequency because the "resolution" that we have to work with is determined by the spacing between our ears, and wavelengths vary.

There is more to it but I think that is one factor.

 

[abdo123]

(if you can't detect the source direction, wouldn't it just all mesh together?)

Yes any reflections arriving later than 1 ms after the direct sound increase the perceived level and spaciousness (more precisely the perceived width of the sound source). Unless it's delayed enough to register as echo.

I'm under the impression that if you dynamically gate the response over the frequency range you might actually arrive at something that represents what people hear in a room (and maybe even correct the response more effectively) but it seems that everyone's echo thresholds are different (perhaps that's why my partner keeps complaining about how loud it is lol) and there will never be a one size fits all.

 

[Jon AA]

I've seen you mention this twice now – I do not think raising the high frequencies would be wise. I am very treble-averse. Did someone suggest that? Maybe it was buried in one of Andy's posts...hehe. I would have cut them just like you did.

If I lower the lower midrange from 200-700 Hz, it's the same thing as raising the treble relative to it.

Mid to high? I'm talking about 100-200 Hz. The point where the bass comes down to meet with the upper bass/lower mids in an in-room measurement.

That is where my Bass Boost is. It's just small for this current calibration. I think the misunderstanding here is you are seeing the step from 400-700 Hz and thinking it is part of the bass boost. It is not. Here is the curve with a larger bass boost so you can see it better:

Changing the value of just 1 PEQ in the MultEQ-X file results in the above change. How big or small I make that has nothing to do with the step from 400-700 Hz. That step is about having the correct tonality through the midrange--it's what I estimate a flat line anechoically will look like when measured in a room, as it transitions from very wide to very controlled dispersion through that frequency range.

 

[Jon AA]

Those 44% are not what we perceive (which we don't really know) but the way to approximate the measured PIR.

That's a good point and makes things even easier to see. If you look at the spin of the M2 Erin measured, you can see a large divergence between the listening window and the ER curve starting at 600 Hz. According to one, the HF section should be lowered 1-2 dB. According to the other, if EQing to a generic curve, you'd raise the HF section 1-2 dB. I'm betting on lowering it being the better sounding adjustment. Sound reasonable?

 

[abdo123]

That's a good point and makes things even easier to see. If you look at the spin of the M2 Erin measured, you can see a large divergence between the listening window and the ER curve starting at 600 Hz. According to one, the HF section should be lowered 1-2 dB. According to the other, if EQing to a generic curve, you'd raise the HF section 1-2 dB. I'm betting on lowering it being the better sounding adjustment. Sound reasonable?

It’s very obvious that Harman targeted a flat listen window so i would stick to that in my EQ

 

[Chromatischism]

That's a good point and makes things even easier to see. If you look at the spin of the M2 Erin measured, you can see a large divergence between the listening window and the ER curve starting at 600 Hz. According to one, the HF section should be lowered 1-2 dB. According to the other, if EQing to a generic curve, you'd raise the HF section 1-2 dB. I'm betting on lowering it being the better sounding adjustment. Sound reasonable?

I would agree.

Also I found an example of what I think is near perfection and close to my own measurements. Just so we're on the same page. I bet he's on here so I hope he doesn't mind...

Hope that gets across better what I was referring to.

It also depends on what volume level you listen at:

You can see that between 80-100 "phons" there is indeed a knee at around 200 hz where ear sensitivity increases start to level off, however that isn't the case at lower levels. Below 80 phons you would want a smooth decrease all the way to 1 kHz. Just going by this, your curve would be superior at lower levels.

 

[ernestcarl]

I added a very rough trend line from the HF on down. It shows you don't get an average solidly above the trend until bass frequencies. If this is a reasonable indication of direct sound through the mid-high frequencies for that mic location, it again shows I should make the step down lowering the high frequencies even bigger. Going the other way--elimiating the step down by raising the HF up 2 dB would make a big hole in the lower midrange.

Yes, the suck-out visible between 300 Hz - 1 kHz would still be perfectly audible, though it's arriving later in time (room acoustics) and simply discarded by the set windowing filter. The apparent over-elevation found in the simple steady state FR magnitude graph you originally showed can't be considered as a perfect, direct equivalent translation to what we would actually perceive in person. There is a correlation, for sure, but actual perception of the person sitting in the listening room (doing the EQ calibration) matters more.

...

The "rise time" setting can be altered if you click the options gear/sprocket wheel found on the upper right-hand corner of REW. All the other options can be kept within their original default values:

Below are some measurements I took recently of my Fostex 6301 speaker mounted above my monitor screen display which you may (or possibly not) find interesting:




A better FDW algorithm should employ a more adjustable scale and windowing method like what DRC does. Nevertheless, I think what REW has right now can be enough -- one just needs to cycle through the different settings rather than using only one fixed filter setting.

 

[thewas]

Not going to get into semantics, but it's both.

Obviously this would be frequency specific and I was thinking that this would probably be the best approach for room EQ above ~100Hz.

As Toole writes the target is flat direct sound and smooth directivity, a smooth PIR is just the result of those but on its own not enough (for example non flat direct sound compensated by the according to directivity to create a smooth PIR).

It’s very obvious that Harman targeted a flat listen window so i would stick to that in my EQ

Yes, that's the flat direct sound target above the modal region.

 

[thewas]

That's a good point and makes things even easier to see. If you look at the spin of the M2 Erin measured, you can see a large divergence between the listening window and the ER curve starting at 600 Hz. According to one, the HF section should be lowered 1-2 dB. According to the other, if EQing to a generic curve, you'd raise the HF section 1-2 dB. I'm betting on lowering it being the better sounding adjustment. Sound reasonable?

When you have to compromise like this it depends usually also on the listening distance and reflectivity of the room (thus the direct to reflected sound ratio) how much you concentrate just on the direct sound, but it is always a compromise.

 

[YSC]

Did some move around to tinker with my PC and wifi.. done remeasurement after EQ below 400hz:

*Edit, I messed up the L+R and R plot,
L:

R:

Both channel

 

[Thomas_A]

Did some move around to tinker with my PC and wifi.. done remeasurement after EQ below 400hz:
L:

View attachment 224144
R:
View attachment 224145

Both channel

View attachment 224146

How is your gated direct response of the speaker around 0.75-100 cm from the speaker listening axis? Treble too hot?

 

[YSC]

How is your gated direct response of the speaker around 0.75-100 cm from the speaker listening axis? Treble too hot?

did't do that as it is very near field listening, and the room is wayy too messy to put gated response gear...
2-4k looks like some comb filtering with all the keyboard and gaming gear on desk, in direct listeninig I don't detect treble peak though, tried to EQ the treble shelve down but my ears told me something was wrong so I only EQ below 500hz

 

[Thomas_A]

did't do that as it is very near field listening, and the room is wayy too messy to put gated response gear...
2-4k looks like some comb filtering with all the keyboard and gaming gear on desk, in direct listeninig I don't detect treble peak though, tried to EQ the treble shelve down but my ears told me something was wrong so I only EQ below 500hz

Ok nearfield explains it. I wonder if near and farfield room curves should differ or not.

 

[Zaireeka]

This is the best I could get in my very slightly treated room (2 bass traps, 2 acoustic pannels), speakers (Genelec 8350A) toed-in at 45° and 180cm from listening position. It sounded amazing, now I need to do all the process again for my freshly installed 8351As...

 

[YSC]

Ok nearfield explains it. I wonder if near and farfield room curves should differ or not.

no idea, but I am in a gaming desk with monitor slighttly in front of the L+R speakers, and the sub moved a bit to place some stuffs below the desk, General trend sens the comb filtering seems right to me and pretty flat, only thing is if I eq it down the center of image seems off, maybe the measurement up there have too much random reflections from stuffs right next to the speakers.

 

[Faxe]

I am using the Lyngdorf TDAI 3400 amplifier together with Paradigm Persona 3F's. This is what the measurements look like using Roomperfect in focus and global modus and the bypass mode.

Right channel

Left channel

 

[dasdoing]

I wonder if near and farfield room curves should differ or not

the less room you hear (as you get closer), the more a neutral target would aprouch flat

 

[Chromatischism]

Those 44% are not what we perceive (which we don't really know) but the way to approximate the measured PIR. As Toole says our hearing apparatus and processing is not the same as a typical omnidirectional mic and a computed linear transfer function.

I wanted to add some more to this. To those who say the direct sound always dominates, the answer is that will depend on the frequency in question.

From the epic thread at AVS: https://www.avsforum.com/threads/ho...-what-the-science-shows.3038828/post-58634362

Some “room EQ” algorithms, and some DIY persons, time align the first arriving, direct sound, from each of multiple subwoofers. The goal is to ensure delivery of a synchronized direct sound. But:
1. In small rooms the direct sound is swamped by reflected sound – it is not the dominant factor in what we hear.

Because the room has a large effect on frequencies up to at least 300 hz and to a lesser degree 500 Hz (insert whatever number you feel like within this range depending on room size), I would not count on the region I called out to be dominated by direct sound. In fact, it is probably more than 50% sound power.

 

[thewas]

Yes, direct sound dominates our perception above the transition frequency, that's why bass correction should be done not based on anechoic but on LP measurements.

 

[TangoDJ]

This Google Drive Link has:

1. The pictures of the room, 3D rendering with measurements and equipment info.
2. The processor (AVM 70 ARC Genesis) calibration details. I ran the REW measurements WITH ARC calibration ON.
3. The "House Curve" settings. I have two rows of seats. The main listening position is in the Front.
4. Raw REW measurements for both Front and Back seats.
5. Few spectrogram and decay pictures

Dealing with a room mode at 52Hz and relatively lower decay times in the Bass area.