Mathaudio Room EQ users (tips & tricks)

[Sokel]

Thanks. I expected the worst. It works!))

Had a look at your bass array thread so here's your favorite wavelets as a bonus:


without


with

 

[Flaesh]

By the way, "without" doesn't look so bad either))
I wonder if the subject can fix something like this:

 

[Sokel]

By the way, "without" doesn't look so bad either))
I wonder if the subject can fix something like this:
View attachment 399397

It's big-ish ported 3-way (about 160 lit.) but can only go down to 30's,1 meter is not a good distance to measure it but you can see that Math fixes things.
What's the above?Seems like something pushed too hard.

 

[Snarfie]

MathAudio Room EQ 3.0.5 – my impressions.

I’ve been beta testing the latest update of Room EQ from MathAudio and wanted to share some brief, subjective impressions and clarifications from the developer.

What’s new in 3.0.5​

According to the developer, version 3.0.5 introduces measurable improvements in room-correction accuracy above ~5–7 kHz, primarily when using 44.1 kHz and 48 kHz sample rates. The goal is more accurate correction in the high-frequency range.

My listening impressions​

I compared v2.8.5 (32-bit) and v3.0.5 (64-bit) on several tracks.

Subjectively, I noticed:

• Slightly improved midrange clarity, especially on vocals
• A more natural and spacious decay on high-frequency content (hi-hats, vocal reverbs)
• Marginally tighter attack on kick drums and toms

However, the differences are subtle. In a strict blind test, I’m not sure I could identify the new version reliably. Still, the changes seem directionally positive.

High sample rates​

I asked how this update affects higher sample rates (88.2/96 kHz and above). The developer clarified:

• The measurable accuracy improvement exists at all sample rates
• At higher sample rates, the improvement is so small that it’s not audibly noticeable, even to him

Takeaway​

• The update appears technically meaningful, especially for 44.1/48 kHz users
• Audible differences are present but small
• No downside observed in upgrading, so I’ll continue using 3.0.5 as it is a improvement.

If others have measurements or blind-test results, it would be great to see how they align with these impressions.

---

 

[HDavidson]

I've been using MathAudio Room EQ for a long time, along with Convolver correction. My only suggestion is to use VAR-type smoothing in EQ Wizard (REW).
Precision (Q) is necessary for low-frequency correction. For midrange and high frequencies, 1/3 octave smoothing is sufficient.
The current version only uses two types of smoothing across the entire frequency range, and in my opinion, this remains the main problem with this software.

 

[Snarfie]

I've been using MathAudio Room EQ for a long time, along with Convolver correction. My only suggestion is to use VAR-type smoothing in EQ Wizard (REW).
Precision (Q) is necessary for low-frequency correction. For midrange and high frequencies, 1/3 octave smoothing is sufficient.
The current version only uses two types of smoothing across the entire frequency range, and in my opinion, this remains the main problem with this software.

Are you using the three combined Mathaudio Room eq, REW an Convolver?

 

[HDavidson]

It depends on the situation (and mood). For one pair of speakers, I need positive correction to eliminate the speaker's own dips. MathAudio Room EQ doesn't offer this option. In this case, I use a convolver with the appropriate correction.

 

[Snarfie]

It depends on the situation (and mood). For one pair of speakers, I need positive correction to eliminate the speaker's own dips. MathAudio Room EQ doesn't offer this option. In this case, I use a convolver with the appropriate correction.

Interesting I’ve run into similar issues with notches/dips in the past.

At one point I tried to correct a sharp dip around 120 Hz using a parametric EQ (EasyQ) together with MathAudio. On moderate listening levels, the response looked and sounded more balanced. However, when playing louder I started to hear distortion and some audible strain from the speakers.

This behavior matches MathAudio’s design philosophy. Their Room EQ intentionally avoids boosting deep notches and mainly reduces peaks. As they explain, boosting deep dips can require large driver excursions, increase distortion, and lead to overcompensation of room effects.

In the first measurement (see my first screenshot), you can see the sharp dip around 120 Hz. That was the dip I initially tried to compensate with EasyQ. It “worked” at low levels, but at higher SPL the distortion became noticeable.

In the second measurement (after changes to the room), the dip is much less severe. Instead of forcing correction electronically, I addressed the acoustics by adding a new blind and a curtain. This reduced the null naturally and eliminated the distortion issue at higher volumes.

So from my experience:

• Boosting deep, narrow dips can work at low levels but may cause distortion at higher SPL
• Many deep nulls could be room-related cancellations and not truly fixable with EQ
• Acoustic treatment or repositioning is often the safer and more effective solution

That said, if a dip is clearly due to the speaker itself rather than the room, it looks like a different approach (like convolution) make more sense.

 

[HDavidson]

Many deep nulls could be room-related cancellations and not truly fixable with EQ

This is obvious. There's no point in touching them.
But any gaps associated with speaker design issues can and should be corrected.
Too much correction in Mathaudio Room kills the dynamics, so it's pointless.
It's enough to cut only the main resonances, but moderately. And the mids and highs should be leveled out at the 1/3 octave level. That's why it's missing VAR-type smoothing.

 

[Gorgonzola]

I've been using Room EQ in Foobar for several years with great success as I deem it.

I don't seem to have any problem substantially correcting fairly deep notches. Isn't it matter of not over-driving one's amplifier trying to fight against room nulls? Maybe I get away with it because I listen at very moderate volumes with a decently powerful amp.

 

[HDavidson]

In my opinion, even this level would be enough, or even less.

 

[Snarfie]

I've been using Room EQ in Foobar for several years with great success as I deem it.

I don't seem to have any problem substantially correcting fairly deep notches. Isn't it matter of not over-driving one's amplifier trying to fight against room nulls? Maybe I get away with it because I listen at very moderate volumes with a decently powerful amp.

When starting using DSP like Mathaudio i changed my amp for a NAD C370 for enough juice which gave me more stable headroom using DSP. Curious are you listening close farfield or both. Picture before curtains.

 

[Gorgonzola]

When starting using DSP like Mathaudio i changed my amp for a NAD C370 for enough juice which gave me more stable headroom using DSP. Curious are you listening close farfield or both. Picture before curtains.
View attachment 510791

Both, in different location ... main stereo listening room: far; computer desk: close.

 

[Snarfie]

In my opinion, even this level would be enough, or even less.

View attachment 510787

Are these measurements from your farfield speakers?. Guess you did multiple measurments.

 

[HDavidson]

Are these measurements from your farfield speakers?. Guess you did multiple measurments.

These are not my measurements, but the example given above.

Gorgonzola​

 

[Snarfie]

These are not my measurements, but the example given above.

Gorgonzola​

Ah got it

 

[ShadowBoxer]

I’ve been beta testing the latest [3.0.x] update of Room EQ from MathAudio

Is it known what the v2.x -> v3.x upgrade charge will be for paid (VST) versions? It doesn't seem to be posted on the web site.

 

[Snarfie]

Is it known what the v2.x -> v3.x upgrade charge will be for paid (VST) versions? It doesn't seem to be posted on the web site.

They could clearify it better but under download it says Foobar plugin is for free not mentioning that for the VST versions.

 

[ShadowBoxer]

Yes, but I license the paid version. There is no mention of a version upgrade fee like with most software (and the necessary license key does not migrate). Ofc, for those that use the free Foobar version it's a non-issue.

 

[HDavidson]

Mathaudio Headphone EQ version 2.0 updated
It's strange, but they removed the windows for entering digital filter values that were in the last version. Strange decision.