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

[101geo]

@mitchco 's Hang Loose DSP Suite will let you do that.

A quick question would you run room correction before or after crossovers? My instinct says before but I have been wrong before (Sooo many times..)

 

[john61ct]

A quick question would you run room correction before or after crossovers? My instinct says before but I have been wrong before (Sooo many times..)

Been answered above.

Get all the other DSP multichannel etc in place and dialed in, run Gsonic last as "stereo-only"

 

[OCA]

I thought this thread was about the free version?

Maybe paid stuff could get its own thread, would be less confusing for those who just want to test...

It was wrongly posted here instead of direct message, sorry!

 

[DenonMate]

Sorry, my mistake. I misunderstood the thread to be about Gsonic in general, since a user had started a discussion about what the next iteration of Gsonic would be. You're right; the thread should focus on the stereo version.

 

[Keith_W]

A quick question would you run room correction before or after crossovers? My instinct says before but I have been wrong before (Sooo many times..)

AFTER crossovers.

 

[101geo]

AFTER crossovers.

Sorry I got a little confused, I understood the measurement and filter generation order, I went down the road of which filter to run first for playback before I realised you don't do that...

 

[flexy123]

Are you using Equalizer APO? It might be lowering the volume in pre-amp settings. Also check Windows volume mixer for hidden mic/output volume settings. A1X uses AVR internal sweeps which are at 75dB and independent of AVR volume, is equal to -30 otherwise)

I had Equalizer APO running and of course I did something stupid a few days ago, I measured for Gsonic and forgot EQ APO still had old FIR from an earlier run active So I disabled it entirely. But level on PC HDMI output in general is still a bit low (not a big problem), but I also had this many months ago when I compared when playing Spotify via Spotify Connect, versus outputting from PC via HDMI. Still a bit low on HDMI.

My question is whether increasing src input level on the AVR, +5db or so would have any negative effects, rather than just cranking up Windows/Spotify etc volume?

Also..in my first tests with Gsonic (with proper measurement FL/FR, XO etc. settings from Acoustix, but then Audyssey off so bass management is kept), it sounded REALLY good.

Edit: The observed behaviour however..is strange. Because when I sweep with REW from the PC, which uses same HDMI and AVR input, I dont have the impression that level is low. I also get loud tones when I do the Windows speaker tests where you click on speakers in Windows sound card settings. Just for the sweeps in Gsonic Reference, I have to crank entire system 100% everywhere, Gsonic at 0db (rather than -12db) AND AVR to like -20db just to get moderate output level.

 

[ShadowBoxer]

Just for the sweeps in Gsonic Reference, I have to crank entire system 100% [volume].

Fwiw, I get the same behavior with Gsonic (always too low).

Side note: when I switch to CamillaFIR I don't observe this same low volume behavior for reasons TBD.

 

[AudioJester]

Well, after 2 weeks with GSonic filters - I continue to be amazed!

 

[LionIT]

Gsonic creates the filters. Another DSP software is required to utilize the filter for playback.

Yes, this is clear. I was talking about the image posted by OCA showing the MIMO convolution engine. It should have an input to receive the audio to be processed, this is why his previous answer is not clear to me.
Anyway, maybe it's not so relevant in this thread...

 

[Kihmski]

Any recommendations for a vst plugin that can load the generated filters? I want to use this in cubase

 

[tifune]

Any recommendations for a vst plugin that can load the generated filters? I want to use this in cubase

HangLoose from accuratesound.ca

 

[_giacomo_]

Dear @OCA, I ended up with a dedicated Windows installation to work around the bit perfect issue affecting the UMIK-1 under Linux.

after loading all the target curves for inspection, I see that none of them is flat. I'd like to start with a flat one, just to avoid taming the high frequency response. I perfectly understand the psychoacoustic motivation behind a non flat target. Yet, how to set a flat target on GSonic? If there is no predefined setting, I can obviously write one myself
With regards

 

[OCA]

Dear @OCA, I ended up with a dedicated Windows installation to work around the bit perfect issue affecting the UMIK-1 under Linux.

after loading all the target curves for inspection, I see that none of them is flat. I'd like to start with a flat one, just to avoid taming the high frequency response. I perfectly understand the psychoacoustic motivation behind a non flat target. Yet, how to set a flat target on GSonic? If there is no predefined setting, I can obviously write one myself
With regards

Save this file under target curves folder and it will show up in target curves list in the app.

 

[longts]

I’m seeking a multichannel version compatible with a 2.1 or 2.2 system. Integrating sub and main speakers is a challenge for me.

Apart from Roon, are there any DSP solutions that can use Tidal or Qobuz and be controlled via a mobile application? I previously used Minidsp Flex with Dirac Live, but I found the sound lifeless.

Currently, I’m using PEQ in my Wii Pro Plus (measured by RTA REW, with only a cut peak from 20-300Hz, following the Harman curve). I find it much better than Minidsp, even though I don’t have a high-pass filter for my SVS SP1000 and Kef R3 speakers (crossover at 85hz, Kef R3 ported).
I’m genuinely interested in DSP solutions like this, but a streamer like Wii is too convenient for me. If there’s a solution that can work with Tidal and a mobile app, I’ll definitely buy a Motu 4 and be ready for it!

Thanks everyone!

 

[OCA]

I’m seeking a multichannel version compatible with a 2.1 or 2.2 system. Integrating sub and main speakers is a challenge for me.

Apart from Roon, are there any DSP solutions that can use Tidal or Qobuz and be controlled via a mobile application? I previously used Minidsp Flex with Dirac Live, but I found the sound lifeless.

Currently, I’m using PEQ in my Wii Pro Plus (measured by RTA REW, with only a cut peak from 20-300Hz, following the Harman curve). I find it much better than Minidsp, even though I don’t have a high-pass filter for my SVS SP1000 and Kef R3 speakers (crossover at 85hz, Kef R3 ported).
I’m genuinely interested in DSP solutions like this, but a streamer like Wii is too convenient for me. If there’s a solution that can work with Tidal and a mobile app, I’ll definitely buy a Motu 4 and be ready for it!

Thanks everyone!

I think you can use JRiver in WMD mode and play Tidal with convolution filters and it has a remote as well but I am not sure. @Keith_W is the expert on this.

 

[Keith_W]

Since I was mentioned, I will give a quick reply and hopefully not drag the thread off-topic.

If you want (1) streaming (2) a FIR capable convolution engine (3) a remote control, you have a few options:

1. Use Roon. It works with Tidal and Qobuz. Problem with Roon: the license is expensive and so is the subscription.
2. Use Audirvana. Problem: the version that enables streaming is also subscription only, with no option to purchase a license.
3. Use JRiver with its DLNA feature enabled, and a UPNP remote. For Android, use BubbleUPNP or M-Connect. For iOS, M-Connect only. Getting this feature to work requires several settings, PM me if you want me to walk you through it. This is the cheapest option, but it's not elegant, suffers from severe latency, and is unreliable. Every now and then it stops working for no god damned good reason and you have to reboot both PC AND tablet.
4. Choose a streaming service where the iOS/Android app is able to control the PC app. There are only two: Spotify and Qobuz. I think Apple Music as well but it only works on Apple products (so typical of them). NOT Tidal, Amazon Music, YT Music, Idagio, or anyone else!

If you remove the "remote control" requirement it is much simpler. You need (1) playback software, e.g. Tidal app, Qobuz app, etc. and (2) convolver and (3) some way to connect the two apps, preferably bit perfect. "Bit perfect" means you avoid WASAPI Shared like the plague and use either WASAPI Exclusive or ASIO.

Since you have a Wiim, you could continue using it if you purchase a Motu. This is how:

1. Connect your Wiim to a digital input on the Motu, and the Motu to your PC.
2. Set up your convolver to take input from the correct channel on your Motu. The only convolvers capable of doing this NATIVELY are Hang Loose Convolver and Acourate Convolver. Acourate Convolver is out because it uses proprietary .CPV files and not .WAV, so your only option is HLC. Technically JRiver can also do it, but I would not wish it on my worst enemy since setting JRiver up to take input from a sound card is needlessly painful.
3. Load the GSonic filters into your convolver*
4. Set the convolver to output to the correct channels on your Motu.

* NOTE: Hang Loose Convolver requires a .CFG file to be written. I don't know if GSonic does this for you, but if it doesn't, you will need to read HLC's manual very carefully and write your own .CFG file. I can help you with this if you PM me.

 

[_giacomo_]

Here are my results: L and R correction is excellent (yellow curves in the lower part of the plot). Original measurement (yellow, top). Correction (cyan).
I haven't spent too much time yet, but I think a phase issue is causing a cancellation when L and R are combined. In the second attachment (graph with white background) the result obtained manually with REW+REPHASE. The valley from 35 to60Hz is wider but less deep. Observations?
One more thing: loading the GSonic filters in REW: 128dB offset. I added -128dB to the filters in order to produce comparable curves. Do we need to add the offset to the GSonic filters before loading them into the convolver? NOTE: I use a low level convolver: brutefir under linux. I don't think it applies any offset and so I think the convolution will clip immediately.

@OCA any observation about the different results obtained with gsonic and REW/Rephase?
I used a flat target curve and +3dB max correction in GSonic. Full scale correction, 1.0 Strength.

 

[OCA]

Here are my results: L and R correction is excellent (yellow curves in the lower part of the plot). Original measurement (yellow, top). Correction (cyan).
I haven't spent too much time yet, but I think a phase issue is causing a cancellation when L and R are combined. In the second attachment (graph with white background) the result obtained manually with REW+REPHASE. The valley from 35 to60Hz is wider but less deep. Observations?
One more thing: loading the GSonic filters in REW: 128dB offset. I added -128dB to the filters in order to produce comparable curves. Do we need to add the offset to the GSonic filters before loading them into the convolver? NOTE: I use a low level convolver: brutefir under linux. I don't think it applies any offset and so I think the convolution will clip immediately.

@OCA any observation about the different results obtained with gsonic and REW/Rephase?
I used a flat target curve and +3dB max correction in GSonic. Full scale correction, 1.0 Strength.

That dip looks like a 50Hz AC ground loop random phase shift issue. Are you in EU? Check if it persists or keeps shifting between measurements. If so, there's not much you can about it and it's not nearly as audible as it seems.

Filters are 0dBFS. REW needs a reference to convert them to dB SPL and usually uses 120dB - 3dB (division safety) = 117dB but that will change based on your PC's sound card and audio driver type, etc.

No need to add any offset to the filters in any convolution engine. They are normalized and clip protected.

 

[3ll3d00d]

Technically JRiver can also do it, but I would not wish it on my worst enemy since setting JRiver up to take input from a sound card is needlessly painful.

You just save it as a library item and play it no?