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

[jeffaegrim]

Looks really interesting !
Thanks

Is there a way to limit corrected frequency range ?
It's very important to be able to set at least a high frequency limit, in my experience. Low limit may also be usefull.

341ms latency clearly exclude any studio use.
Do you have any plan to include a "low latency" option, using a combination of PIR and FIR ?

There is no notable latency with GSonic filters applied.

Honestly, I may be out of line, but most of the questions being posted now have been answered earlier in the thread. This is a FREE resource provided by one person who is releasing updates faster than I can download them. I think we should extend more good will towards OCA - for example, by reading his thread first and not making him constantly justify his rationale.

 

[OCA]

my RME ADI-2 DAC shows something that doesn’t look at all like white noise.

Looks like DAC has an octave band spectrum analyzer. White noise has flat energy per Hz and each higher octave contains 2xmore Hz bins causing that slope and attenuation of lower bands. I will replace it with pink noise in the next update which has equal energy per octave.

how do I measure their effect in REW?

First generate a measurement sweep as .wav file with REW's Generator:

Then use that file to measure "From file" in REW and play it in Roon with the filter engaged:

 

[ShadowBoxer]

Is there a way to limit corrected frequency range ?

 

[OCA]

GSonic Reference v1.0.22 released!

• Test tone updated to pink noise
• Added variable "Intelligent Boost" under Advanced Settings:

Max (default): Matches behavior of previous versions - applies up to a hyperbolic +9 dB boost below the room’s transient region, gradually tapers based on spatial coherence between measurement positions, includes a linear crossfade from 10 kHz–24 kHz, reaching up to +3 dB at Nyquist.
Medium: +6 dB / +2 dB limits
Low: +3 dB / +1 dB limits

Important: This tool is designed around physically realistic room and speaker behavior, including natural low-frequency roll-off. Systems that heavily alter this (e.g., aggressive pre-EQ, extreme bass extension, or non-standard DSP chains) may produce misleading results. Requests to adapt the algorithm to compensate for such setups are outside the scope of GSonic’s design goals.

 

[Jero-n]

Here, Equalizer APO is used to do the loudspeaker crossover . Stereo 4-way



On the APO website:

- Applications must not bypass the system effect infrastructure

(APIs like ASIO or WASAPI exclusive mode cannot be used)



Can this sound system be measured?

If possible, will it result in a usable measurement?

 

[OCA]

Here, Equalizer APO is used to do the loudspeaker crossover . Stereo 4-way



On the APO website:

- Applications must not bypass the system effect infrastructure

(APIs like ASIO or WASAPI exclusive mode cannot be used)



Can this sound system be measured?

If possible, will it result in a usable measurement?

Just use Wasapi and APO should work. GSonic doesn't use "exclusive" mode.

 

[Jero-n]

Thanks, I'll give it a try

 

[Sir Sanders Zingmore]

Looks like DAC has an octave band spectrum analyzer. White noise has flat energy per Hz and each higher octave contains 2xmore Hz bins causing that slope and attenuation of lower bands. I will replace it with pink noise in the next update which has equal energy per octave.

First generate a measurement sweep as .wav file with REW's Generator:
View attachment 524349
Then use that file to measure "From file" in REW and play it in Roon with the filter engaged:
View attachment 524350

I discovered that Roon won't play the recorded sweep with convolution on if the recorded sweep is mono.
Not sure why, it will play it with convolution turned off in Roon but just silence when it's on

 

[OCA]

I discovered that Roon won't play the recorded sweep with convolution on if the recorded sweep is mono.
Not sure why, it will play it with convolution turned off in Roon but just silence when it's on

I guess Roon expects a mono filter (GSonic generates both versions) instead of the single stereo wave file when using the mono sweep for measurement.

 

[_giacomo_]

GSonic Reference v1.0.22 released!

• Test tone updated to pink noise
• Added variable "Intelligent Boost" under Advanced Settings:

Max (default): Matches behavior of previous versions - applies up to a hyperbolic +9 dB boost below the room’s transient region, gradually tapers based on spatial coherence between measurement positions, includes a linear crossfade from 10 kHz–24 kHz, reaching up to +3 dB at Nyquist.
Medium: +6 dB / +2 dB limits
Low: +3 dB / +1 dB limits

Important: This tool is designed around physically realistic room and speaker behavior, including natural low-frequency roll-off. Systems that heavily alter this (e.g., aggressive pre-EQ, extreme bass extension, or non-standard DSP chains) may produce misleading results. Requests to adapt the algorithm to compensate for such setups are outside the scope of GSonic’s design goals.

lovely! Much appreciated!

 

[WeekendWarrior]

I discovered that Roon won't play the recorded sweep with convolution on if the recorded sweep is mono.
Not sure why, it will play it with convolution turned off in Roon but just silence when it's on

Umm...I downloaded left, right and stereo sweeps with timing reference from REW into Roon and played them with Gsonic stereo filters active. It all worked and I recorded the outputs back into REW.

 

[Sir Sanders Zingmore]

Umm...I downloaded left, right and stereo sweeps with timing reference from REW into Roon and played them with Gsonic stereo filters active. It all worked and I recorded the outputs back into REW.

Umm yeah that’s what I ended up doing too.
My comment was simply expressing surprise that Roon wouldn’t play a mono file with convolution.

 

[WeekendWarrior]

I must have misunderstood, but all's well that ends well!

How did you create your file, it's not something I've done before, although I am a relative novice with DSP/REW.

 

[kharan]

@OCA I mentioned it a few days ago, if you could make it so that the user can add a custom delay before the sweeps start playing, it would be very useful to a lot of people.

 

[TheKoC]

Would it also be possible to allow users to set the starting frequency of the sweep? I don't like how much the woofers of my C8Cs move below 12 Hz, although it is probably just a visual problem, not a risk of damaging them

 

[WeekendWarrior]

I've spent a bit of time with GSonic (using HK target curve) and overall I'm pleased with the results. For my right speaker it has worked wonders below 150Hz (red trace) with more subtle changes above that frequency.

Left speaker correction a bit less successful (purple), although my left speaker has a cabinet nearby which is probably adding reflections that can't be removed. There's an odd effect between 3kHz and 4kHz, although this could be because GSonic uses three measurements and I have only shown one here.

I've created some other filters correcting only at lower frequencies, but I've not listened to them all yet to find my favourite.

Overall, what a great tool, and easy to use compared to the alternatives.

Wondering @OCA what the premium version will include?

EDIT: Just to complete the picture, here is the left correction trace (purple) compared to the average of the three left side measurements.

 

[OCA]

I've spent a bit of time with GSonic (using HK target curve) and overall I'm pleased with the results. For my right speaker it has worked wonders below 150Hz (red trace) with more subtle changes above that frequency.

View attachment 525102

Left speaker correction a bit less successful (purple), although my left speaker has a cabinet nearby which is probably adding reflections that can't be removed. There's an odd effect between 3kHz and 4kHz, although this could be because GSonic uses three measurements and I have only shown one here.

View attachment 525103

I've created some other filters correcting only at lower frequencies, but I've not listened to them all yet to find my favourite.

Overall, what a great tool, and easy to use compared to the alternatives.

Wondering @OCA what the premium version will include?

EDIT: Just to complete the picture, here is the left correction trace (purple) compared to the average of the three left side measurements.

View attachment 525112

If you can hear new sounds in your old tracks and locate instruments better with a central sound stage that's about ear height and sticks then it's good DSP. Bass strength depends on the volume level you are listening and unfortunately also on the track itself.

FWIW, nice FR graphs and target tracking is not nearly as meaningful as many believe and takes just one line of code to achieve the perfect target response but will unfortunately sound horrendous. Green filter below will measure the blue response when convolved with the red original speaker response and it takes a single operation in REW to generate:

Here's the phase response:

But despite the epic graph performance, this filter will cause audible pre-echo up to 400ms before the actual sound arrives and is not just bad but literally un-listenable.

 

[Calypso1]

Has anyone encountered these mic error? Thanks. I have attached the Umik calibration that works with REW and the error.

 

[kharan]

Has anyone encountered these mic error? Thanks. I have attached the Umik calibration that works with REW and the error.

Your calibration file is corrupted.

 

[Calypso1]

Your calibration file is corrupted.

Thanks Kharan. Downloaded it again from the miniDSP website and I am back in business.