Gold Planar GL600 - 7.1 audio using binaural reproduction technics

[Jean Ibarz]

Hello,

I here propose, for possessors of headphones Gold Planar GL600, a DSP configuration usable in the free software EqualizerAPO, allowing to "mimic", at some degree, the sound that you would perceive when listening a stereo (or 7.1) system in a small/medium living room.
I simply want to share the files, without entering too much into the technical details, so that anyone can try it and, maybe, give positive or negative feedbacks about it.

Required:

• gold planar GL600 headphones,
• windows OS with EqualizerAPO software installed

Instructions:

1. install, if not already done, EqualizerAPO. It is a free software available here: https://sourceforge.net/projects/equalizerapo/. EqualizerAPO allows to apply sound effects on the signal path before being outputted from the soundcard. Among other effects, it is mainly used here to convolve the 8 channels with binaural room impulses responses (BRIRs) that I've made myself.
2. download the folder that contains every settings and files required to be used in EqualizerAPO here: https://drive.google.com/file/d/18QD-__K...sp=sharing
3. unzip the folder here: "C:\Program Files\EqualizerAPO\config"
4. configure the soundcard setup to stereo or 7.1, with all speakers being fully capable to cover the entire spectrum, and configure the sampling rate to 44.1khz or 192khz. Warning: if you load a configuration file with an unappropriate number of speakers and unappropriate sampling rate, the convolution will mess up and the audio output may be excessively high, potentially causing hardware damage or ears damage. I recommand to UNWIRE the headphones and/or headphones amplifier from the soundcard output, or to lower the volume if it can be done in the hardware (volume button from headphones amplifier for example), before loading any configuration files into EqualizerAPO. Configuring the soundcard output to 7.1 allows to make windows send the 8 audio signals, before being convolved and mixed down to stereo with EqualizerAPO and sent to your headphones through any stereo output. If your soundcard doesn't support 7.1 setup, or you don't want to listen 7.1 channels audio content, you can still use stereo.
5. load the appropriate configuration file into EqualizerAPO by adding into the "config.txt" file an effect "include configuration file" and set it to load the appropriate file. For exemple, if your soundcard is configured to stereo and 44.1khz, you should load the file "config_gl600_44100hz_stereo_room_blocked_ear_with_reverb_v2020-12-30.txt" that should be located into "C:\Program Files\EqualizerAPO\config\".

The configuration editor of EqualizerAPO should look like this:

If it's not the case, for example if the frequency response is overclipping by more than 30dB or is more smooth or is looking like a comb filter, than it is likely that the soundcard output is not configured with the appropriate number of speakers and/or the appropriate sampling rate. Verify the soundcard output settings, then try to deactivate/reactivate the output, and eventually exit and open again the configuration editor of EqualizerAPO.

For comparison purposes, I have also included a configuration file "config_clear.txt", that you can switch to to compare 'binaural listening' and 'standard listening, with no correction'.

Configuration for Hifiman HE400i
If any of you possess the headphones Hifiman HE400i, old version -not the 2020 version-, just tell me and I can also upload the settings for these headphones.

Friendly,
Jean.

 

[Sonnie Parker]

Interesting... does this do something similar to the Symth Research Realiser that cost $4K?

 

[Jean Ibarz]

Hi,

Yes it does something similar, but the major difference is that what I propose is binaural without headtracking. Some researchers (most of researchers ?) believe that headtracking is needed to allow a good quality of binaural experience. This is quoted from https://smyth-research.com/ :

Without head-tracking virtual loudspeaker sources will move, causing a perceptual anomaly between the auditory and visual scenes. This frequently leads to either a front-back reversal or collapse of the externalised virtual sound-stage, in which the sound sources appear to be either behind the listener or inside their head.

To be honest, without headtracking, I am not able to achieve a perfect spatialization experience. However, I experienced during more then 2 years binaural listening, first using the database from Shanon Pearce and multiple different earphones/headphones. I bought around 20 different headphones/earphones:

• lowcost earphones: apple earphones, sony earphones, jvc earphones (<25$ each),
• supra-aural closed: sony closed headphones (45$), DIYed to try supra-aural open,
• intra-auriculars closed: etymotic ER4B (250$), KZ AS16 (80$), KZ ZS10 Pro (50$), KZ ZSX (25$),
• intra-auriculars open: Final Audio E2000 (50€),
• circum-aural open: AKG K702 (70$), Hifiman HE400i 2014 (180$ in promotion but worth 500$ in my country), Gold Planar GL600 (150$)
• circum-aural closed: Sennheiser HD8Dj (250$), Quiet Comfort QCII (250$), Hyper X Cloud II (90$)

I also tried to construct my own headphones using differents headphones drivers, but also miniature full range speakers like the Fountek FR59EXE:

even tried with separate tweeters mounted in a hat:

I tried to make my own HRTF filters, but they were so bad: I was not able to reproduce a loudspeaker playing in my room, from far. I believe there were a problem in the way I measure/make the HRTF. I've learned in the process that the reverberation of a living room like mine modifies A LOT (I mean, its really huge) the sound perceived. But I was stuck in a problem: by removing the reverberation (by windowing the HRTF), the sound perceived was unbalanced and spatialization accuracy decreased. But keeping the reverberation was not ok because, and I don't know really why, my brain was not able to "cancel", or "compensate" the reverberation. I've also learned that placebo effect is so huge and so long that it is nearly impossible to correct subjectively, in fine details, some bad HRTF filters. After a lot of time and some research, I think I've understood why HRTF measurements requires blocked ear meatus. Then I tried to make again HRTF using blocked ear meatus measurements. But the result was not satisfying too, unless one day I tried the Hifiman HE400i. With these headphones, the result was nearly good, still far from perfect, but not as far as before. So I tried to correct the perceived sound, and I've finally come to a result that satisfies me. I've also conclude, from my listening experiences, that dynamic headphones are not able to reproduce low frequencies without distorsion, even if no distorsion appears in measurements. One exception to that is for the intra-auriculars KZ ZS10 Pro and KZ AS16, that are able to reproduce clean bass with good enough SPL for me. But planar drivers (and probably electrostatic too) does the job. However, I'm not still sure why it is so.

So, I got good results with the hifiman HE400i, but these headphones are not anymore available at a fair price, they are replaced by a 2020 version, and I don't know what changes between these 2 versions (are the drivers the same or nor ?) and I want people to be able to get good listening experience using headphones at an affordable price, I tried with cheaper and available headphones: the Gold Planar GL600. I had to adjust the filters by +3dB below 80hz and +12dB above 4khz to get satisfying results again.

Today, I'm convinced that I have come to a nearly realistic result with my HRTF, or I should better say BRIRs, filters: I am able to keep the full reverberation of my room, and my brain is able to integrate it completely. When listening with my headphones, I get correct perceived sound, and I'm not able to discern the room reverberation that is simulated by the headphones.

I don't want to accept, yet, the fact that headtracking is a "hard requirement" to trick my brain, and to be able to reproduce correctly a loudspeaker in my room that I would be listening in a static position, not moving my head at all.
I prefear to believe that there are other technical issues in the way HRTF are generated/designed and/or the way headphones are calibrated, and I believe that more improvements can be done in this process.

 

[Jean Ibarz]

I've added configuration files to support stereo and 44.1khz. The initial post have been updated accordingly.

I strongly recommand to remove the configuration files that are inappropriate to avoid loading an unappropriate configuration file by error. For example, if you use stereo and 44.1khz, I recommand to remove configuration files for 192khz or 7.1 setup.