A Real, Audiophile-Grade, Solution for Hearing Loss on All Devices

[jjaskuna]

I have high frequency loss.

My experimentation indicates it is more like a brick wall filter.

EQ doesn't change perception.

The range that I do hear works well enough, though.

I blame genetics, the loss discovered as a child, and similar to Mother's.

Thank you for sharing and being the first person to do so. I was not sure if anyone would respond.

I can't blame my genetics. Only myself and car stereos.

 

[jjaskuna]

I think this is an extremely interesting project, and hope it will be a useful and popular product. I’ve had just two similar products in the past. One was built into the native music player app in an HTC phone (now defunct), and the other is part of the iOS app for JBL earbuds. In both cases, the user takes a hearing test of five or so bands, and the software calculates compensation filters. Both worked pretty well. By the sounds of it, ARIA will have much broader usability.

There is a hearing test involved in this. Good catch. It's used to get individual hearing test data into the algorithms and return the data to the individual, such that it can be ported to other products/uses as desired.

 

[jjaskuna]

Thank you for this - I will set aside time to explore your invention. I have hearing loss and use hearing aids for work related meetings etc., but will not use them for for music listening. Despite hearing loss I am still acutely conscious of the quality (or lack thereof) of music sources and playback equipment.

Good point to highlight. I agree and have a similar experience assessing gear.

And, thanks for considering listening.

 

[jjaskuna]

Like Ray, my hearing loss is more like a brick wall filter over 1650kHz compounded by tinnitus which I've had since youth. So your app won't do much for me personally but I applaud your efforts and I'm sure there will be many who could benefit.

I had a buddy with more broad spectrum hearing issues who went to an audiologist and got expensive in-ear devices. I took the data from the audiologist and created an EQ profile from it as a preset on his gear for music and movies which he said worked great albeit for solo listening/viewing. Unfortunately he only got to enjoy it for a couple months before passing.

You read right into the limitations of this and the targeted application.

 

[jjaskuna]

More like a zero....

Or poor explanation on my part. The high sample rate is getting at the absence of a decimation stage. Turn off oversampling to achieve this.

The goal is that input sr == output sr, such that you are not passing through a required resampling. I wanted something that can process high SR or bit depth audio natively, as I do listen to these formats occasionally and would prefer to get the same quality on the output as the native file format.

 

[Hayabusa]

.

Our Solution: We've developed a calibrated, HD hearing test that finely maps your hearing profile from 125Hz to 20kHz, enabling our hearing refinement technology, ARIA, to deliver lossless & high-resolution sound personalization (hearing correction). Lossless and high-resolution are used technically here. ARIA works with any digital audio source and all listening devices (no hardware dongles) without the fidelity compromises found in BT sound personalization embodiments.

Nice topic, I have been working on this also.
As doing this correctly relies on knowing absolute sound levels at the ear I wonder how you can claim this works for all listening devices?

 

[jjaskuna]

I'd have to think boosting highs to restore hearing loss will do two things. Increase that hearing loss, and be uncomfortable to listen to if you have that loss. One of the issues with damage to the cochlea is those variable filter bands (critical bands) in your hearing (equivalent to something like 32 bands) become less selective with age and hearing loss. So even fixing the frequency response doesn't fix this overlapping band problem. This is why older people with hearing loss have trouble picking out details in noisy environments or get uncomfortable from the general noise in say large sporting events. What once was loud, becomes a loud cacophony. In a sense masking ability of hearing is compromised so everything sounds all run together. I'll admit not reading the OP's detailed info. Did I miss something addressing this?

My experience is that I listen with 4-5dB lower volume on my preamp/dac when correcting for my hearing loss vs the native audio. If you are reaching to increase the volume of the corrected signal, the system is not working as designed in the implementation. You should be reaching to increase the native audio's volume in AB testing. We can investigate if you are interested.

Masking is a critical consideration as you have pointed out, but it's progressive just like age-related hearing loss and the person you are describing is in the more advanced stage of age-related hearing loss. So the question remains, what solution is effective between the time of HL onset and the stage of HL you are describing. This is more likely to help the younger individual who is not even diagnosed with hearing loss (>=20dBHL impairment), but age-related changes, recognized or not, are affecting the fidelity of their listening experience.

In this implementation critical hearing bandwidth is fixed. You are correct. It is easily tunable, however. You will be able to select different models of critical hearing band calculations, as there are many and even within a single model computations have changed over time. These need not be fixed and are fully tunable, albeit down the road a month. We've taken the approach of allowing for wide band adjustement down to the level of individual hearing band adjustment on a signle ear, and many things in between. With the ability to adjust between the traditional model, bark, ERB, and user defined, a flexible use of critical hearing bands is an idea we are on board with.

 

[jjaskuna]

No, good hearing will go up to 20,000Hz, so 1,650Hz, is very low, but 1,650,000Hz or 1,650kHz is insanely high, even for your average bat.

This is my mistake. That crazy high sample rate is the audio's sample rate, not the maximum tested frequency.

I've seen some interesting research on thresholds above 20kHz, but nothing this high.

 

[Golf]

Be prepared that after playing your stereo set with hearing corrected output from the loudspeakers, your wife and your dog may finally move to another district

 

[jjaskuna]

I have not found specific information about how it actually works. Could you share more detail, especially regarding how it differs from something like Peace Equalizer? That one lets you find your threshold of hearing for different frequencies (preset ones by default, but can be edited iirc), then corrects those thresholds based on ISO equal loudness curves for different volumes and creates an EQ based on that.

Yes, absolutely. I was hoping someone would ask.

Here is the framework: Start with calibrated hearing threshold results spanning at least from 125Hz to 20kHz for a listener. Test more frequencies in the spectrum, not less. 4 frequencies does not provide enough information. Currently, the software uses template profiles and will move to user test results soon. Next, compare the test results against a reference, which is full spectrum from 0-20kHz (truly nyquist), to derive the users correction curves.

Here is a plot of my test results against the reference curve.

Currently, this is completed in a middle/sides optimization stage, not on the left and right channels. I've included two models of this optimization step, which apply different M/S models behind the "character" toggle. This will be fully tunable by the user and allow for additional modes, such as a decoupled stereo mode and a linked stereo mode. There are additional optimization steps governing the application of shared regions of hearing loss and those not shared. This is important for stereo or multichannel listening, as we are now listening in the environment and HRTFs become important considerations.

Next the filters are optimized or deisgned. Generally, we are tying to minimize the error between predicted and target response. Currently, there are both IIR and FIR embodiements with this one being an IIR topology. FIR is next and also presents optionality for things like phase specification. With IIR filters, operating rate can have dramatic effects on sound, so this is another tunable parameter based on what you feed in or if you use the "oversample" mode. This is a conveince feature. If you really want to experiment with different sample rates, turn it off and try different audio files or resample with something better--HQPlayer or Roon are options.

Apply the correction with the designed filters and then cascade (or remove) the dynamics and tuning sections. These aim to 1) control true peaks to prevent clipping in a musical way (user defines musical way, but we have some suggestions); 2) allow adjustment of the imaging post correction, as applying correction will modify the imaging perceived. These processing stages are optional, should one want to use just the filters. Within the dynamic section you can obviously control how the dynamics are applied, if at all, with the threshold, attack, release, ratio and knee parameters.

Re: Peace Equalizer
I'm not clear on what kind of filters they are using or how they handle freqs above 12.5kHz, as ELC is only defined to this point. It looks quite flexible, however, and seems to have some filter topology optionality, which I'm a fan of. I'm on mac most of the day, but I am going to give this a try. Here's my thought: audition using a common set of test results. If you want to try, let me know and I will share my tabulated results to compare in Peace and in my software--hearing profile #0.

One aim of mine is portability of test results and this is a strong use case.

 

[jjaskuna]

No surprise as it is Audiophile-Grade.

If you are willing, take a pass of my response to @V17 and let me know where you would want to see more detail broken out.

 

[Golf]

Just a thought: When using a software based room correction in connection with a calibrated microphone like the MiniDSP UMIK-1 for the measurements, could it make sense to modify the mic’s calibration file according to one’s individual hearing curve? In order to make the mic »hear« with the same impaired hearing ability as the user?

 

[jjaskuna]

Nice topic, I have been working on this also.
As doing this correctly relies on knowing absolute sound levels at the ear I wonder how you can claim this works for all listening devices?

tbh, writing about this in a non-technical way is hard for me. What I am trying to express is that it works with any listening device you can connect to a computer. With a DAC, ADC, and BT, you could connect to just about anything in theory, though I am sure there is something I have not thought of to disprove this This was important to me in a software solution, as I don't really agree with locking the solution behind a hardware-dongle where it can only be used if one had purchased that piece of hardware.

For the version using your own test results, I do strongly request that the hearing test be taken on calibrated earbuds/headphones & computing devices/smart phones. At one time I had a hard block preventing completion of the test with uncalibrated devices, but was convinced to use several strong notices instead. This is an accessibility limitation that needs resolution, as most performant devices are >$100. Though, many, many devices costing a lot still fail to meet performance requirements, so price is not a good correlation metric.

 

[dualazmak]

Just for your reference,,,

To compensate slight/moderate (and/or natural?) age-dependent hearing decline in high Fq (say above 7 kHz?), I implemented flexible and safe on-the-fly relative gain control for my super-tweeters; in my setup, the compensation should be "subjectively" determined for audience to audience, though;

- Excellent Recording Quality Music Albums/Tracks for Subjective (and Possibly Objective) Test/Check/Tuning of Multichannel Multi-Driver Multi-Way Multi-Amplifier Time-Aligned Active Stereo Audio System and Room Acoustics; at least a Portion and/or One Track being Analyzed by Color Spectrum of Adobe Audition in Common Parameters: [Part-11] Violin Music: #643

My DSP-based multichannel multi-SP-driver multi-amplifier active system (ref. here for the latest setup) also have flexible and safe analog level on-the-fly relative gain controls (in addition to on-the-fly DSP gain controls) for L&R subwoofers, woofers, midrange-squawkers, tweeters, and super-tweeters, all independently and remotely. The analog level 5-Fq-zone relative gain controls have no effect/modification on upstream DSP-domain gain/XO/EQ/phase/time-alignment!

I believe these flexible and safe analog level on-the-fly relative gain controls (tone controls in five Fq zones) for audience(s) would be the best way especially when I would invite audio and music enthusiastic friend(s), younger or senior people, in my listening room. I can very easily reset the relative gains fit for myself after the guest(s) would left my home.

Spoiler

 

[Hayabusa]

tbh, writing about this in a non-technical way is hard for me. What I am trying to express is that it works with any listening device you can connect to a computer. With a DAC, ADC, and BT, you could connect to just about anything in theory, though I am sure there is something I have not thought of to disprove this This was important to me in a software solution, as I don't really agree with locking the solution behind a hardware-dongle where it can only be used if one had purchased that piece of hardware.

For the version using your own test results, I do strongly request that the hearing test be taken on calibrated earbuds/headphones & computing devices/smart phones. At one time I had a hard block preventing completion of the test with uncalibrated devices, but was convinced to use several strong notices instead. This is an accessibility limitation that needs resolution, as most performant devices are >$100. Though, many, many devices costing a lot still fail to meet performance requirements, so price is not a good correlation metric.

So lets say you connect a PC running your algorithm to a high-end headphone amp with its own volume control. How would you feedback information about the volume level back to your algorithm to do the correct corrections?

 

[Ivanovich]

Being basically deaf above 9K with tinnitus seems unfixable. However, I find the brain fills in what‘a missing very well and most music doesn’t have much above 10K.

I once tried EQ’ing up the frequencies where I had less sensitivity and the result was that those frequencies were over emphasized. The brain adds gain already, and adding more energy there doesn’t make music sound better for me.

What’s needed is a way to regrow the stereocilia in the inner ear! Apparently, mammals can’t do this but fish, birds and reptiles can. So jealous!

 

[Hayabusa]

I once tried EQ’ing up the frequencies where I had less sensitivity and the result was that those frequencies were over emphasized. The brain adds gain already, and adding more energy there doesn’t make music sound better for me.

Correcting for hearing loss does not involve eq-ing but compressing the low levels that sink below the higher hearing threshold.
So the higher levels will remain basically the same.

 

[middlemarch]

Downloaded the app to my Macbook Pro. It failed during the install. Tried twice. Sent a message to the developers. Tried to send my system log file which details the install failure, but the message app on their website failed when an attachment was present. Sent the message without the log file. Have not had any response yet. Has anyone had success running the app, on any machine?

 

[Keith_W]

Just FYI.

Download blocked by Windows and automatically deleted. Please investigate and resolve this.

I also had a look at the user manual. I do not see how this is a "solution for hearing loss", it looks as if there are a number of presets (specifically, 12+1) that you choose subjectively to "match your hearing profile" according to the manual. I did not see any description of a method of measuring your hearing loss. I thought it may include a David Griesinger style DGSonicFocus which plays alternating tones which you adjust until the tones are subjectively equal in volume. This is why I wanted to download the software to have a look.

 

[middlemarch]

Here's hoping the actual software is a bit more robust than the website. Not off to a great start