GSonic Reference – Free Stereo Room Correction Tool (Measurement + FIR Export)

[kharan]

How does Gsonic correct the massive dip at 50 Hz on the right channel? On the other hand the minor dip around 180 Hz in the right channel seems to be mostly uncorrected.

"4 R Raw" is the uncorrected measurement and "2 R AcoustiX" is the Gsonic corrected measurement, channels are all offset.

Is it an instance where the left channel that doesn't present the same dip at 50 Hz is used to correct the right channel? Does Gsonic have any mechanism to avoid excessive excursion when it increases some bass frequencies?

Side note, did I just massively increase the headroom of my speakers considering Gsonic EQ'ed away a massive amount of bass that was due to room gain?

 

[Plompudu]

View attachment 524076

How does Gsonic correct the massive dip at 50 Hz on the right channel? On the other hand the minor dip around 180 Hz in the right channel seems to be mostly uncorrected.

"4 R Raw" is the uncorrected measurement and "2 R AcoustiX" is the Gsonic corrected measurement, channels are all offset.

Is it an instance where the left channel that doesn't present the same dip at 50 Hz is used to correct the right channel? Does Gsonic have any mechanism to avoid excessive excursion when it increases some bass frequencies?

Side note, did I just massively increase the headroom of my speakers considering Gsonic EQ'ed away a massive amount of bass that was due to room gain?

You can drag and drop the .wav files to see what they do

Mine look like this for example (High frequency difference cause I'm using a adjusted Target for the left channel to fix driver variance):

I agree, that a max boost option would be nice, cause the boost at 20 / 30Hz in my case is increasing the distortion a bit more then I'd like.

 

[kharan]

You can drag and drop the .wav files to see what they do

Mine look like this for example (High frequency difference cause I'm using a adjusted Target for the left channel to fix driver variance):
View attachment 524091
View attachment 524094
View attachment 524093
I agree, that a max boost option would be nice, cause the boost at 20 / 30Hz in my case is increasing the distortion a bit more then I'd like.

I see, it is just a massive bump in the convolution file where the dip is. I'm surprised that it manage to correct a dip with boost at the frequency response though, it is really well corrected at 50 Hz, nothing of the dip is left.

Which also brings me back to my other question, does the software make any judgement call for when to add boost and when not to do so, especially in the bass region. @OCA ?

 

[OCA]

cause the boost at 20 / 30Hz in my case is increasing the distortion a bit more then I'd like.

There's no amount of distortion one would like. The problem simply means your target curve is forcing too much bass for your speaker's design limits and you need a less bass heavy curve.
GSonic uses a hyperbolic soft clip (no brickwall tearing) at 9dB in bass/mid frequencies, about 3dB in treble.

Edit: For the creative speaker placements out there, the cutoff limit is -30dB

 

[OCA]

does the software make any judgement call for when to add boost and when not to do so, especially in the bass region

Yes of course, it will only boost minimum phase dips but that was discussed earlier in the thread.

 

[kharan]

Yes of course, it will only boost minimum phase dips but that was discussed earlier in the thread.

I missed that I will read the thread again, so if it's a minimum phase dips, it will try to match to the curve as long as the correction stays within +9 dB?

 

[Plompudu]

There's no amount of distortion one would like. The problem simply means your target curve is forcing too much bass for your speaker's design limits and you need a less bass heavy curve.
GSonic uses a hyperbolic soft clip (no brickwall tearing) at 9dB in bass/mid frequencies, about 3dB in treble.

9dB is right around the max I'd personally use for my system (above 30Hz; below I usually limit it to <6dB)
But I'd appreciate if you allowed us to set the variable ourself.

Or if you clarified it somewhere in the software.
Maybe something like "Filter 1 boosted up to X.XX dB at XX Hz, the FIR filter therefore lowers the level by X.XX dB" and "Filter 1 tried to boost above the limit of XXdB at XX Hz - using hyperbolic soft clip to limit the boost"

Edit: For the creative speaker placements out there, the cutoff limit is -30dB

What do you mean by this?

 

[OCA]

I missed that I will read the thread again, so if it's a minimum phase dips, it will try to match to the curve as long as the correction stays within +9 dB?

It tracks excess phase at every frequency bin (and its neighbors) before it attempts to fill a dip. That's the only right way to use boost.

 

[OCA]

9dB is right around the max I'd personally use for my system (above 30Hz; below I usually limit it to <6dB)
But I'd appreciate if you allowed us to set the variable ourself.

Or if you clarified it somewhere in the software.
Maybe something like "Filter 1 boosted up to X.XX dB at XX Hz, the FIR filter therefore lowers the level by X.XX dB" and "Filter 1 tried to boost above the limit of XXdB at XX Hz - using hyperbolic soft clip to limit the boost"

What do you mean by this?

Filters already lower their volumes based on maximum boost used for clipping protection. The distortion you're hearing is simply your speaker's woofer giving up at a certain volume level. That's why I suggested using a less bass heavy curve if you want to reach a certain volume level. Bass will feel heavier in higher volumes anyway.

What do you mean by this?

I am only kidding. A 30dB peak above target doesn't happen that easy in any room but I have seen a number of such set ups with A1 Evo where Audyssey's -20dB cutoff limit wouldn't suffice.

 

[kharan]

GSonic calculates the natural roll off point of the measured speakers very precisely and boosts or cuts LF response accordingly.

Found that again, I now know why the log says "speaker roll-off: 20 Hz". Having said that, measured anechoically, it's more like 3- dB at 30 Hz and -10 dB at 20 Hz, the rest is room gain.

Luckily, I don't have any rea dips below 40 Hz, but if Gsonic attempted boosts below 30 Hz, I suspect my speakers really wouldn't like them, especially since they are active and there may be some boost baked in the crossover.

Another question, since my speakers are already linear phase with FIR filtering for crossovers, does doubling up on the FIR filtering result in unwanted consequences beyond additional delay?

Also, I forgot earlier, but many thanks for your contribution.

 

[OCA]

the rest is room gain.

Correct, also the bass tilt of the target curve has an effect.

unwanted consequences beyond additional delay?

In fact quite the opposite. Any driver crossover related delays are being corrected. Your total latency increases of course but if you have no video to watch simultaneously, that's not an issue.

 

[_giacomo_]

First, a hint that may help other Linux users with older machines. If GSonic does not start and the log ends with:
"Creating window...FAIL: glfwCreateWindow returned null GLFW error 65543: GLX: Failed to create context: GLXBadFBConfig"
Then execute GSonic with the variable:

> LIBGL_ALWAYS_SOFTWARE=1 ./GSonic

@OCA

I may have missed it, just wondering if there is a way to limit the "boost" the filters apply when trying to fill dips in the bass. I just want a very limited boost, if necessary, and make sure it's under 2-3 dB. Is that somehow related to the "Correction Strength" setting? Is the slider value somehow quantifiable in terms of dB. Of course there is no corresponding need to limit the reduction of a peak.

Suggested HF correction and phase limit? I guess a few hundred of Hz... but I am sure you discussed it in the video...

 

[kharan]

Correct, also the bass tilt of the target curve has an effect.

I think you understood correclt, but in case I was unclear, what I wanted to say is that my speakers start rolling off a bit over 30 Hz anechoically but because of the room gain, Gsonic detects the roll-off point much lower.

In fact quite the opposite. Any driver crossover related delays are being corrected. Your total latency increases of course but if you have no video to watch simultaneously, that's not an issue.

There shouldn't be any driver crossover delay baked into the speaker itself. With that said, my speakers have a low delay mode which uses IIR filtering instead of FIR filtering.

One idea would be to switch to the low delay mode on the speakers and let GSonic handle all the phase stuff to lower delay.

 

[kharan]

Never mind, the latency with FIR filtering on the speaker themselves is about 20 ms. Negligible compared to the latency induced by Gsonic.

 

[OCA]

Is that somehow related to the "Correction Strength" setting?

Yes, but only above room transient frequency...Below that it's "hyperbolic tangent" capped at 9dB with selective boost, correctly clip protected for any boost required. I don't intend to revise the very carefully calibrated algo unless you had an audibly bad experience with the results that requires revisiting. The +9dB cap of GSonic is not the same as boost limit of typical target matching least square methods.

Suggested HF correction and phase limit? I guess a few hundred of Hz... but I am sure you discussed it in the video...

No need to use any advanced settings (defaults to full range) unless you have electrostatic speakers.

 

[_giacomo_]

Yes, but only above room transient frequency...Below that it's "hyperbolic tangent" capped at 9dB with selective boost, correctly clip protected for any boost required. I don't intend to revise the very carefully calibrated algo unless you had an audibly bad experience with the results that requires revisiting. The +9dB cap of GSonic is not the same as boost limit of typical target matching least square methods.

I just do not want to put any extra strain on my woofers.. 9dB sounds immense

No need to use any advanced settings (defaults to full range) unless you have electrostatic speakers.

ok I remember from your tutorials that we do not want to correct above a couple hundred Hz..

 

[OCA]

I just do not want to put any extra strain on my woofers.. 9dB sounds immense

ok I remember from your tutorials that we do not want to correct above a couple hundred Hz..

 

[Plompudu]

Yes, but only above room transient frequency...Below that it's "hyperbolic tangent" capped at 9dB with selective boost, correctly clip protected for any boost required. I don't intend to revise the very carefully calibrated algo unless you had an audibly bad experience with the results that requires revisiting. The +9dB cap of GSonic is not the same as boost limit of typical target matching least square methods.

I would appreciate if you could explain how you got to the 9dB value.

There are a few scenarios, where a different value would be useful.
For example at low level listening I'd be fine with a boost higher then 9dB, if it resulted in less audible Nulls / smoother response.

But the +7dB at 20Hz, like in my case, eats a lot of Amp power / reduces my headroom more then I like - I'm fine trading of linearity at the Subbass, if it meant I could keep the overall tonality of the Bass Shelf (Dolby-Spec of +10dB for the "LFE" with a 80Hz Crossover to the Fronts) and the ability to get reference level above ~30Hz without the risk of overdriving the woofers

 

[john61ct]

View attachment 524146

Please correct the words I am putting into your mouth if this isn't what you meant:

With most past auto-DRC implementations, things are too likely to get messed up if you let it apply to much above Schrödinger.

However Gsonic's algorithms to prevent unintended artifacts is so sophisticated, in most cases you can trust it to process the full spectrum.

 

[OCA]

Dolby-Spec of +10dB for the "LFE" with a 80Hz Crossover to the Fronts

That's not the spec for crossing over a sub with speakers. The +10 dB boost applies specifically to the mono LFE track to enhance rare low-frequency effects that are unique to movie content. Subs work at equal volume level with speakers crossed over with them for all other content. If you are using something like that as your target curve, you cannot expect it to be distortion free at very high volumes. Such curves are only ever used for low volume listening. You ignored my suggestion to try a less bass heavy curve. I also don't know what gain setting you are using with which sub and why. Regardless, the tool aims to optimize for correct sound production not for maximum headroom. One can get always the best headroom with no filters. And FWIW Dirac ART boost default is +18dB adjustable to +24dB (including satellite speakers) and with very little selectivity at what it boosts.