My journey to a better sounding classical guitar

I have been reading a lot on this forum the last 6 months but not contributed much since I did not have any experience with Smart Fit or hearing aids in general. Now I have spent quite a lot of time with Smart Fit to tailor a music program for my classical guitar, which sounded terrible with the default music program. So now I want to share my journey on how I modified that to a program that actually sounds like a guitar.

The problem with the default program was that EVERY note was buzzing or ”warbling”, so it was useless. Unfortunately my audi was not able to do anything to improve that, so I decided to try to learn to do it myself. The first thing I did was to make sure that all speech enhancing parameters were turned off or minimized. As far as I could see this was already done in the default program, so the problem was not there. Next I decided to reduce compression, since I had read that this could cause my problems. I actually reduced it to 1.0 (no compression) by copying the 50 dB gain curve to the 65 and 80 dB curves. The buzzing was still there but now not on every note. Next I tried to experiment with the shape of the 3 identical (no compression) curves, but no matter how I tried there was buzzing. The only thing that changed was the notes that buzzed. I was stuck.

Then came a revelation. I am a member of audiologyonline.com and have watched a couple of their video courses, and there I stumbled onto Course: #39829 by Tracy Rudd at GN Resound. She showed how she helped musicians with their music program. One of the things she stressed was to reduce compression, which I had already done, but what caught my attention was the shape of her gain curves. they did not at all look like the default target curves in the music program. Instead they looked like a gently curved rainbow without any abrupt slope changes. They did not look like any target curves I had ever seen. Since she did not say what hearing loss they corresponded to, I assumed they only illustrated the the general concept. However, after trying to smooth the default curves in the music program without success, I finally tried curves similar to her curves (still without compression). Suddenly almost all buzzing was gone! But the guitar did still not sound like a true classical guitar.

While trying to figure out why the guitar did not sound natural, I noticed that people had been trying to use chatGPT to make music programs. Since I have used chatGPT for other projects, I decided to try it. Well, it was not capable to create a suitable program based on my audiogram. However, during my discussions with chatGPT it gave me a lot of hints on what to try and also many tips on how to use Smart Fit. One thing I noticed when experimenting with my gain curves was that even though the gain curves were smooth, the output power curves were not necessarily smooth. So I thought that maybe I should smooth them a little. This made the guitar sound more natural. I discussed this with chatGPT and it confirmed it actually is the output power curves that should be smoothed. So now I have smooth output power curves but the gain curves are not as smooth as before since the smoothing of the output curves changed them.

This is where I am now. The next step will be to try to make the guitar sound even more as I remember it by experimenting with the compression (it is still 1.0). I have noticed that if I just parallel shift the 65 and 80 dB curves, then an evenly distributed compression exceeding 1.3 makes the buzzing return. I am currently discussing this with chatGPT to see if I can use higher compression than 1.3 at frequencies bands that may not be so sensitive to compression. Maybe this will make the guitar sound more natural. Now (with compression 1.0) some higher frequency notes seem to dominate too much.

These were my 2 cents. Maybe some of you have further suggestions on what to try. My conclusion on chatGPT is that it is not there yet, but it is a good fact checker and has helped me a lot with using Smart Fit and give me suggestions on what to try when I am stuck.

At least I can now start building and playing guitars again.

I thank the members of this forum for openly sharing their knowlage, and a spciial thanks to @tenkan who helped me getting started with DIY fitting.

2 Likes

Thanks for posting this very helpful info. I am not as happy with how music sounds on my hifi as I would like and thinking self-fitting might get me closer. Are you wearing a Resound aid? I have graduated from the Eargo 5 to the Jabra Enhance Select 200, then, a couple of months ago, to my current Jabra Enhance Pro 20 aids. I asked to have any processing disabled for the music setting but music still sounds mostly thin and artificial regardless of dome type. I hope it can sound better but perhaps this is as good as the EP20s and/or I can process music now.

The advantage of self-fitting, as you’ve illustrated, is the iterative process of trial-and-error and hearing those changes immediately. I’m guessing very few audis would be willing, or have the time, to work with you (or me) on this. Please keep us udated!

Larry

I have the Resound Nexia 9, which as far as I know is identical to the Jabra Enhance Pro 20 and is programmed by the same software (Smart Fit) as the Nexia. I also have the M&RIE receivers and currently the Shurefit 3 closed domes (they are not actually completely cosed but have smaller holes than the open domes).

The single thing that made the most difference to the sound for me was the video I mentioned. For me it was an eye opener that essentially eliminated buzzing or warbling ofthe quitar notes.

I noticed that your audiogram is similar to mine. Since my (and your) ears are receiving a mix of natural and digitally processed and amplified sounds it is not so easy to find the right balance between the two at each frequeyency to make the combined sound as natural as possible. This is where I am struggling now. One thing that the video showed me was that lower frequencies were amplified more than for speech in order for the music to sound less thin.