Decruitment and Hyper-recruitment are fundamentally different phenomena. If you have significant hearing impairment, you may find frequency regions that exhibit both of these conditions.

Decruitment is the situation where no matter how loud the sound grows, you never perceive it as loud as it is to someone with normal hearing. In contrast, with normal recruitment, while your hearing may be impaired for weak sounds, as the sounds grow louder, your hearing becomes more normal.

Decruitment behaves very much like conduction loss, and it may be a sort of conduction loss as a result of physically impaired hair cells covering that frequency range.

You can liken decruitment to standing in a closed closet while someone just outside the door is shouting to you. You can hear them alright, but you will never hear them as loudly as if you were standing just outside the closet.

So, what does an audiologist find when they test your hearing at threshold levels? They cannot tease apart the separate issues of recruitment and decruitment. All they see is that your threshold at those frequencies is elevated. And they mark your audiogram as impaired to some degree of dBHL.

It wouldn’t matter if your hearing were perfect, except for decruitment loss. In that case you would have no recruitment in your hearing at that frequency. You just need to amplify the signal above your loss. But if you have any recruitment then you also need some nonlinear Crescendo compression.

Measuring your hearing at just one loudness level can never tease apart these separate mechanisms. What the audiologist obtained was the sum of your recruitment and your decruitment loss, but can’t tell how much of each.

That said, we must now realize that the audiology measure is an upper bound on your recruitment. If you have no decruitment then your hearing is really that badly recruited. But if you have some decruitment then your recruited hearing is not as bad as indicated. Your recruitment is less by the amount of decruitment.

But audiology has no way to measure decruitment. It is a guessing game. Can we do better? Perhaps we could measure hearing at two levels and from those measurements, tease apart the separate contributions of recruitment and decruitment?

On our own, we don’t have access to isolation booths, to measure threshold hearing levels. And besides, we don’t live in a threshold level world.

Perhaps we could estimate our hearing by quickly toggling between two frequencies, where at one frequency we have a better estimate of corrections needed. You need to be quick because our short-term memory for sounds is remarkably leaky. Then adjust the second frequency elevation until they sound about the same loudness.

Do this same exercise for two signal levels, perhaps 20 dB apart, and you will probably find a different correction setting at the two levels. That is a hint that some of your problem may be decruitment in addition to recruitment.

Crescendo is formulated on the basis that sensioneural hearing loss is purely recruitment hearing. I have extended the models and processing to additionally treat decruitment and hyper-recruitment.

The way to treat decruitment is to back off the threshold level by some amount and increase post gain in that frequency channel by the same amount. That’s what the +D settings do in each frequency channel.

Dialing in some positive amount of +D simultaneously decreases your threshold elevation and adds some post processing gain by the same amount. If you indicated that you had a threshold elevation of 75 dB, then dialing in 6 dB of Decruitment makes Crescendo process that channel using an elevation of 75 – 6 = 69 dB, and then adds 6 dB of post processing gain.

(But this all actually occurs in Phons space, not dBSPL space. I.e., while you specify an audiologist’s measurements in dBHL (same scaling as dBSPL but different zero point), Crescendo converts that to dB elevation above the absolute threshold of hearing (ATH), then converts to Phons (different scaling from dBSPL), and then subtracts off 6 Phon for decruitment. Later, at post-processing time, the post-gain of 6 Phons is applied before converting all back to dBSPL space where the gains can be physically applied.)

The way to tell whether you have too little recruitment correction is to notice if fading sound levels decline too rapidly up to loss of signal. Too much recruitment correction makes the sound level out at faint levels – they stop fading before cutoff. The right amount of recruitment correction makes the sound decline in step with the decreasing signal level till cutoff – no leveling out, and no premature disappearance.

So this suggests another test: take two adjacent channels. You should have a good idea of the lower frequency channel already. Play two pure tones centered in the two channels, and fade them together. If they both sound like they are fading in lockstep, then you have the correct recruitment settings in the higher frequency channel. Then, if they don’t sound about as loud as each other, dial in some more decruitment and increase the threshold setting to compensate.

I’m just tossing around ideas here. I have no idea what a diagnostic measurement for decruitment will be, and neither does anyone else. In fact, I’m probably the only one even looking…

  • DM

Author: dbmcclain

Astrophysicist, spook, musician, Lisp aficionado, deaf guy