[Okay! Rolling up my sleeves to learn what every adolescent already knows about websites and blogs, etc.]
This is the first entry of the new Refined Audiometrics Laboratory Blog. I’m implementing this because what I have to share with the world is just too important to allow it to fade away. The past 16 years have produced a hard won victory to the understanding of how we perceive loudness, and how we can largely overcome damaged hearing with respect to accurate musical perception.
Anyone with damaged hearing knows that hearing aids help with speech perception. Heck, just about anything helps, including shouting. But there isn’t a hearing aid capable of full spectrum, accurate restoration of musical sounds with their rich harmonics, and timbre preservation. Even the most sophisticated digital hearing aids will turn live concert oboes into muted jazz trumpets. When that happened to me during a Mozart concert, once I located the “jazz trumpet” in the orchestra and discovered the oboe soloist, I tore out my hearing aids and swore I’d never again use them when listening to music.
The past 16 years has seen huge progress, not only in my ability to restore accurate musical perception, but also in the technology available for running the massive digital signal processing chore. When I began, it took an expensive, dedicated, special purpose DSP computer to run the Crescendo code (my own codename for the project). But over the past decade, modern computers have evolved to the point where they are almost trivially capable of running a full-up Crescendo system.
The first five years of this endeavor saw me starting from zero knowledge about hearing and spending day and night in the lab making careful measurements. By the end of 2006, I had found the equations that govern the whole of our hearing experience – not just the cochlea, but the cochlea, afferent 8th nerve, brain processing, and efferent 8th nerve interactions.
Using those equations as my guide, I then set out to create a physical system that will dynamically modify the spectrum of sound so that it sounds to damaged hearing the same way it sounds to people with undamaged hearing. And I succeeded beyond my wildest dreams!
When I speak, I can’t even hear my own sibilance, let alone from anyone else. But when I turn on my Crescendo system and listen to my wife speak, I hear her almost magical sibilance at 10 kHz (!!) as clear as can be. I find it mesmerizing to listen to her speak.
Over the years I gradually improved the performance and detailed behavior of Crescendo. I have helped a number of musicians across the globe who have impaired hearing. I have even been featured in CD liner notes by one artist, as having helped him accomplish his recording goals.
Now I want to make Crescendo generally available to everyone who needs it, and at a reasonable and highly affordable cost. If you are interested, after reading through this blog, just contact me at dbm at refined hyphen audiometrics dot com.
I have made the entire C++ source code for the DSP processing into free and open source code on GitHub. So if you have the chops in programming and math and physics and electronics and music, then have at it on your own. But if you need support, or just want an already available solution, then you’ll need to contact me and pay a minor toll on the bridge.
I haven’t decided yet on the price. I invested 16 years of my life in this project, and spent quite literally hundreds of thousands of dollars on equipment and lab facilities. I’m thinking that you can go out and buy a personal amplifier, for speech perception, for $150. So a fair, and still affordable solution for the finest musical listening experience in your life, should be worth at least an extra $100.
I’m now retired and so this isn’t the way I make my living. And I really want everyone who needs this to have one for themselves. Asking $250 for my personal attention and lifetime updates and assistance should be acceptable to everyone. You’ll still need a personal computer to run the Crescendo system, but most people already have that.
Once you listen through Crescendo, you will never leave. I assure you. I have used my own Crescendo system for the past 10 years, every single day – listening to recordings, mixing & editing audio, performing music. I can’t be without mine.
In fact, I need it so desperately that I have devoted huge amounts of time looking for alternate ways to achieve the same thing, or at least some approximation of it. And I will share some of those ways with you in this blog, giving details on how you can set up your own zero-order Crescendo approximation. They work much better than nothing, but given a choice, I’d still leap at the chance for a full-blown Crescendo system.
Okay, so what is Crescendo? Damaged hearing behaves in a manner referred to as “recruitment hearing”. You don’t just lose volume, but you lose it at a rapidly increasing rate as the sound loudness fades away. It happens nonlinearly, meaning not in a straight line. At some level of sound you lose it entirely, and that varies with the pitch or frequency of the sound. Near that threshold of perception, the apparent loudness increases quite dramatically, and then as the sound grows louder you gradually achieve normal hearing.
But unlike a person shouting at you, or just cranking up the volume on your playback, Crescendo knows how your hearing recruitment behaves at each frequency, and it bends the spectrum of the sound accordingly, so that you hear what people with unimpaired hearing can hear. Crescendo squeezes the entire dynamic range of music, from whisper soft levels to blaring loud, all into your reduced dynamic range brought about by recruitment hearing. And it does so while preserving the balance among the harmonics of musical sounds. So oboes continue to sound like oboes.
Crescendo operates in stereo, and provides 100 Bark bands of nonlinear compression, with each band spaced at quarter-Bark intervals. It corrects the sounds more than 300 times per second as they change, in real time, with very tiny delay (< 6ms). And it provides full spectrum processing so that you hear everything from deepest bass to highest treble and the “air” above. Finally it automatically adapts to your particular hearing by using only one adjustment knob. Once set, you never touch the control again. You just sit back and listen like never before.
Those last sentences deserve deeper contemplation… If you have ever been to an audiologist, you know they lock you inside an isolation booth and play very soft beeping sounds in each of your ears, to map out how your hearing varies with frequency. Those measurements are called “Threshold Elevations”. The numbers the audiologist records are the amount of gain needed at each frequency for you to just barely hear a sound.
And when you get fitted for a hearing aid, she uses those gains to adjust your hearing aid to your hearing. Lots of dials and knobs on her computer screen. And you go back after a week and have her readjust them for you, and maybe even a few more times.
In all, conventional audiology and hearing aid fittings are a time consuming, elaborate, and fraught experience. But we don’t live in a world of threshold level sounds. And they still turn music into intolerable crap.
I learned from the experiences of my users that most people with hearing damage exhibit nearly the same slope to their threshold elevations as a function of frequency. But only when looked at in the right reference frame. I found that nearly everyone exhibits threshold elevations that grow by about 3.5 dB / Bark. The difference between individuals is in how much damage there is. But if I know the degree of impairment at only one frequency, say 4 kHz, then I can pretty well tell you what your damage is at all other frequencies.
That means that we don’t need elaborate adjustment knobs for tuning up a Crescendo. All I need to know is one number and all the rest can be figured out by Crescendo. And it works incredibly well. And you can even do it for yourself. You don’t need me…
Now I also found the following: as a scientist, I know there is a “correct” setting for every individual. But I also witnessed the futility of helping people set things up “correctly”. Listening isn’t science, it is emotion and art. What I saw was that people quite naturally find their own best “correct” setting – no fancy laboratory equipment needed during calibration. Just let the user tune the knob for him or herself. And everyone who has ever listened through Crescendo has liked it – even when they didn’t really have hearing damage.
[ Time for a footnote…
Among the hearing impaired population, about 80% of us have “Sensioneural” hearing impairment. That arises from listening to overly loud noises, rock concerts, some illnesses. But there are still 20% of us who have other kinds of hearing impairment.
Crescendo was designed specifically to accommodate sensioneural hearing loss. That kind of loss exhibits increasing amounts of impairment as you go higher in frequency. Can it help otherwise? Perhaps…
I have one client who is a retired Army artillery officer. In addition to high frequency loss, as you could imagine after a career standing near firing canons, he also had low frequency loss. He had a U-shaped audiogram, and could only hear in the narrow frequency range between 500 Hz (C above middle-C) to 1 kHz (2 octaves above middle-C). He had only one octave of hearing, unassisted.
I was quite unsure about offering him a Crescendo. But you know what? It was a heck of a lot better than he heard with nothing. So he was tickled pink to have a Crescendo.
So, I can only make brash claims about the 80% of us with sensioneural hearing loss. I cannot guarantee improvements for the other 20%. But anecdotally, every one of them has been happy with Crescendo anyway.
Finally, even if you have sensioneural impairment, but your degree of impairment is “Profound”, meaning >90 dB of threshold elevation, then I may not be able to help very much. But who knows?
End of disclosure…]
I’ll sign off this entry for now. If you want to see some more details and background on Crescendo, you can read an Audiophile Magazine article I wrote a few years ago, called “The (Un)Broken Audiophile” .