I’m attaching some text in blue below explaining some more details of the thinking behind Speech Rescue in case anyone is interested in reading it. It’s from the white paper by Oticon on Speech Rescue published in 2015.
Also click on this attachment (hope it’s big eough, if not, type Ctrl + several times to magnify it)
to see how Speech Rescue can be configured in 10 different ways.
The explanation on how Oticon comes up with the 10 configurations is in the blue text below. Basically they pick it out based on the natural frequency selectivity of the cochlea measured in ERB with a logarithmic increase toward the high frequencies. Anyway, as you can see, the lowest destination region (configuration 1) is at around the 2KHz range. If you look at my hearing loss at 2KHz, my left ear is at 75dB and my right ear is at 85db. When I tried out Speech Rescue for the first time, I had the 85dB receivers on, and I could hear the lowered sounds clearly on the left ear and not the right ear. At that time I didn’t know why. So I went home and looked up this white paper and it shows that even for the lowest destination band, that’s already at 2KHz range. I assumed that Speech Rescue must have selection configuration 1 for my hearing loss for both ears (since this is the lowest destination band already and my loss is very significant there already). So I deduced that there was not enough amplification by the 85dB receiver for my right hearing loss which is also at 85dB. So I thought that I’d need a more powerful receiver so I can hear the lowered sound on the right ear, hence the upgrade to 105dB receivers.
BUT, after I upgraded to the 105dB receivers, although I was able to hear the “s” and “sh” on my right ear, it was NOT the same lowered sound on my right ear. So in my next visit I made the audi show me what configuration Speech Rescue selected for my hearing loss. For some unknown reason, it selected configuration 7 for my right ear and configuration 1 for my left ear. That was the reason why I didn’t hear the same lowered sounds in my right ear, because the lowered sounds for the right was lowered to 4.1KHz, while the lowered sound for the left was lowered to 2KHz. I assumed that Speech Rescue would have lowered sounds to the SAME destination reason (same configuration) for both ears, but I found out later that it was not the case. After I had the audi do a manual override and lowered my right ear to the same 2KHz destination region (configuration 1), I began to start hearing the SAME lowered sound on the right ear just as well as how I hear it in the left ear.
Long story short, had I known this up front, I might not have had to upgrade to the 105dB receiver. I could probably have stayed at the 85dB receiver, manually lowered the right ear setting to configuration 1 (from the default configuration 7), which is the default for my left ear, then I would have most likely been able to hear the same lowered sound in my right ear with the 85dB receiver as well.
Hope that explains it. But now that I already have the 105dB receivers, I’m just going to stick with them for future proofing in case my hearing loss gets worse because the 85dB receivers are already barely adequate for my current hearing loss in the first place.
Excerpt from the Speech Rescue white paper:
The specific Speech Rescue settings complement thenatural frequency selectivity of the cochlea. The frequencyselectivity is defined by the width of cochlear bandpassfilters (auditory filters). Their widths, measured in ERB(equivalent rectangular bandwidth), increase approximatelylogarithmically towards higher frequencies (thebase of the cochlear). Following this natural perceptualarrangement, Speech Rescue captures several high-frequencybands and re-codes them in a lower-frequencyband according to a logarithmic scale. Consequently, on aperceptual frequency axis the width of the destinationregions (~ 3 ERB) is only a little smaller than the width ofthe source regions (4-5 ERB) (see log scale in Fig. 3b). Thisarrangement thus introduces minimal distortion of thelowered signal. For configuration 1 to 5, the total sourceregion is split into three sub-source regions to fit into thevery narrow destination region (on a linear scale). Forconfiguration 6 to 10, division into two sub-source regionssuffices since the destination region is positioned closerto the base of the cochlea and is therefore naturally alittle wider (on a linear scale).
Speech Rescue strives to protect the information naturallycarried by the low frequencies in three tangible ways.First, as the destination band is kept narrow (only 800 to1600 Hz) (Fig 3a). Second, the Genie fitting softwareassists the clinician in how to place the destination bandat the edge of the patient’s hearing range. Third, the sourceband is set so that it never goes lower than 4 kHz (Fig. 3b,c), which allows it to capture the primary energy from thefricative consonants such as ‘f’, ‘s’ and ‘th’ without alteringthe relationship between formant frequencies.
In addition, Speech Rescue takes advantage of the opportunitiesoffered by the natural dynamics of speech. Voicelessconsonants naturally contain almost no low frequencyenergy. Thus, for a single talker, high- and low-frequencyspeech tend to be mutually exclusive in many situations(Fig. 4). This means that even though Speech Rescuesuperimposes the source region with the destinationregion, information from the two will rarely overlap in time.For example, Speech Rescue will position the high-frequencyenergy from the voiceless fricative, f, in a naturallyavailable space in the lower frequency position. Conversely,when voiced speech cues occur in the low and mid frequencies,there will often be no high-frequency speechenergy present in the input signal to be lowered by SpeechRescue.
Speech Rescue takes advantage ofthe opportunities offered by thenatural dynamics of speech.
Finally, since the algorithm copies and keeps the originalhigh-frequency sound, the hearing aid retains the fulloutput bandwidth to provide the patient with maximumbenefit of high-frequency sounds in case they can extractany sound information (e.g. musical or environmentalsounds) from the high frequencies. In this way, the Oticonfrequency composition also acts to reinforce the originalhigh-frequency speech. For frequency compression andfor some variations of transposition there is no output forhigh frequencies when frequency lowering is enabled inthe device