Welcome, everyone, to the webinar, Hearing Aid Compression, Digital Microphones and Noise Reduction. We're so glad that you could be here today to learn more about compression concepts and how they are applied to today's hearing aid. Your moderators for today are me, Ted Annis, Senior Marketing Specialist. And me, Keri Peterson, Member Services Supervisor. Our expert presenter today is Ted Venema, Ph.D. In addition to Ted's many years as a practicing audiologist, he is also a seasoned professor having taught at several colleges and universities throughout Canada. In fact, Ted created Canada's fourth hearing instrument practitioner program while at Conestoga College in Kitchener, Ontario. Ted is a passionate speaker and continues to give presentations on hearing, hearing loss and hearing aids across North America and beyond. Ted is also author of the textbook, Compression for Clinicians, now in its second edition. We're very excited to have Ted as our presenter today. But before we get started, we have just a few housekeeping items. Please note that we are recording today's presentation so that we may offer it on demand through the IHS website in the future. This webinar is available for one continuing education credit through the International Hearing Society and the Association of Hearing Instrument Practitioners of Ontario. We've uploaded the CE quiz to the handouts section of the webinar dashboard and you may download it at any time. You can also find out more about receiving continuing education credit at our website, IHSinfo.org. Click on the webinar banner on the homepage or choose webinars from the navigation menu. You'll find the CE quiz along with information on how to submit your quiz to IHS for credit. If you'd like a copy of today's slideshow from today's presentation, you can download it from the handouts section of the webinar dashboard or you can access it from the webinar page on the IHS website. Feel free to download the slides now. Tomorrow you will receive an email with a link to a survey on this webinar. It is brief and your feedback will help us create valuable content for you moving forward. Today we'll be covering the following topics. Linear gain, compression in analog hearing aids, compression for severe to profound hearing loss, compression for mild to moderate sensory neural hearing loss, and compression in today's digital hearing aids. At the end we'll move on to a Q&A session. You can send us a question for Ted at any time by entering your question on the question box on your webinar dashboard, usually located to the right or the top of your webinar screen. We'll take as many questions as we can in the time we have available. Now I'm going to turn it over to Ted who will guide you through today's presentation. Take it away, Ted. Hello, hello. And it's my pleasure and privilege to once again do a webinar for IHS. This one's on compression. Compression and I know that the title says compression, directional mics, and digital noise reduction. We're going to clarify that as we move along through the presentation. As Ted was saying, the linear gain is the first thing we're going to just take a quick pass at. Compression in analog hearing aids needs to be appreciated in order to understand how it is used in today's digital hearing aids, and I'm a big believer in knowing the historical development of compression because it helps us understand where we are today. There's two populations, severe to profound hearing loss and mild to moderate sensory neural loss. These two clinical populations receive very different kinds of compression and for very good reason. And then the compression used in today's hearing aids is a combination of the various types of compression that were initially used in yesterday's analog hearing aids. So here's our good old title, hearing aid compression. It's a gain issue, not a signal to noise ratio thing. I'll repeat that. Compression is all about gain. It doesn't address signal to noise ratio at all. Hearing aids have to do two things. They have to provide gain for the hearing loss, and then they also need to strive to increase the signal to noise ratio. In other words, separating speech from background noise. And as we said, compression is a gain issue. Signal to noise ratio, however, is a D-mic, directional mic, and digital noise reduction issue. And really, to be honest with you, there's no way we'd be able to squeeze all of this stuff in a one-hour webinar. So here's my tongue-in-cheek offer or, let's see, suggestion to have IHS let me do another webinar in the near future on directional mics and digital noise reduction. But they are twins of a team. Compression is the gain issue, dealing with the hearing loss per se, but we also need to try and separate signal to noise ratio, and that's best done with directional mics and digital noise reduction, which will really not be covered much in today's webinar. But let's look at compression and gain. And to do that, let's compare the eyeball to the cochlea. On the left, I wonder if you can see my cursor here, look at the back of the eye, the retina. The retina is really, in quotes, the hair cells of the eye. The retina, in one sentence, changes light into electricity, and electricity is the language the brain understands. In most vision problems, the eyeball is too short or too long, and so you can see that light with these white lines here may not be properly focused on the back of the eye. And so we get eyeglasses or lenses to refocus the light upon the retina where it should be. You see, 95% of vision loss is conductive in nature. The light isn't conducted properly to the retina where it's got to go. Now let's move to the ear. It's exactly the opposite with hearing. 95% of hearing loss is due to damage to the, in quotes, retina of the ear, the hair cells. And back to the eyeball, how would optometrists feel if someone scratched the back of your eyes and said, okay, now go ahead and fit the lenses. How do you like me now? Well, that's our situation. That's why we have so many courses and seminars in counseling, because the counseling for hearing aid is so radically different from the counseling required for eyeglasses. Most vision loss is conductive. If most hearing loss was conductive, we wouldn't have compression, because your thresholds would increase with conductive hearing loss, and so would your loudness discomfort levels. They would ride up like an elevator. The ceiling and the floor of the box would simply go up together. And you wouldn't have the necessity for compression. But the fact is, with sensorineural loss, the ceiling has stayed the same, and the floor has been elevated. So our dynamic range is now shrunken. Normal inner and outer hair cells. This can be appreciated more by people living in Michigan and in Ontario, Canada. I call the top here, these inner hair cells, those would be the American side of the Niagara Falls. And here, these outer hair cells, we can call them the Canadian side of the Falls. Now, the inner hair cells send all sound information to the brain. Without them, we're deaf as a post. But they have a fundamental flaw. They cannot pick up sound below 50 or 60 dB. And so, they need the outer hair cells to help them pick up sounds below 50 or 60 dB. The outer hair cells have a very different role that way. If you're looking at damaged hair cells, they mostly are confined to the outer hair cells. The outer hair cells mechanically help the inner hair cells pick up sounds below 50 or 60 dB. And so, if you haven't got outer hair cells, you've got a 50 to 60 dB sensorineural loss. And that's what most people have. It's called presbycusis, Presbyterian, Church of the Elders, as opposed to the deacons. Presbyopia, your arms aren't long enough to see the page. That hits you when you're 40. Presbycusis, hearing loss in the elderly. The trouble with treble. You can't hear soft sounds below 50 to 60 dB. And that's because of damage to what? The outer hair cells. And so you can see in comparison to this picture here, the damage in the following PowerPoint slide here is mostly confined to the outer hair cells. Now what happens with the outer hair cells is a secondary thing as well. We're going to highlight that right here. This is the floor, this orange belt laying across here is the floor upon which all the hair cells stand. And if you could see my cursor here, the inner hair cells would be all lined up along one edge of this floor. And then the outer hair cells would be all in the rows over on this side here. Now look at here where it says apex. This is actually the floor upon which the hair cells stand. You can see how wide it is. And it's the widest at the narrow tippy top of the cochlea, which is exactly backwards to what you would think. And at the wide base of the cochlea where all sounds enter the cochlea, the floor upon which the hair cells stand is actually narrower. And well, this floor here has more mass than this floor on the right where at the base of the cochlea, which is why this floor at the apex of the cochlea resonates more with low frequencies. At any rate, what you can see in this particular picture is that the peak of this wave, and in this case, is found nearer to the apex of the cochlea. Sounds that would stimulate the cochlea are creating this traveling wave, this ripple, and it's got a peak near the apex. So this person has been stimulated with a low frequency sound. But look at how dull and rounded this wave or this peak here is. So this person may be stimulated with a pure tone, but lots of low frequencies are being stimulated together. This is the second rule now we're talking about with the outer hair cells. Note in this slide it says a wave without outer hair cells. Now let's go to the next slide, and you will see that the outer hair cells amplify and sharpen the peak of the traveling wave. So you can see that the outer hair cells do two things. They amplify in italics, and secondly, they sharpen. And you can see that action in the black right here at the peak of the wave. The wave is amplified, and it's sharpened. So the amplification that takes place, especially for incoming sounds that are below 50 to 60 dB, these are the sounds that the inner hair cells cannot pick up. And so the outer hair cells are lifting that traveling wave, making it bigger. And that's the action the outer hair cells are doing for soft inputs. But the sharpening is something that compression cannot bring back. That sharpening enables this person to distinguish between frequencies that are close together. It's like the coal now has 100 teeth instead of just four or five teeth. The frequency resolution, the ability to distinguish between frequencies right next to each other is enhanced with the outer hair cells, as compared to this prior slide here. And this sharpening is something that allows a person to separate speech from background noise. And that is the issue behind the signal-to-noise ratio. The sharpening is the issue behind the signal-to-noise ratio, covered by directional mics and digital noise reduction. And no D mics and digital noise reduction cannot sharpen a traveling wave. But they can help separate the speech from background noise, so that this person here can separate the speech from noise. By increasing the signal-to-noise ratio, this person is enabled to separate speech from background noise in a way that he could not originally. So let's stick to Snickers here. Let's look at the amplification issue dealing with compression. The sharpening done by outer hair cells cannot ever be remedied. This is why hearing aids for ears aren't like glasses for the eyes. If this traveling wave is a twin set here, you can see two particular peaks here, two frequencies close together. This person can easily distinguish between those two frequencies. When you've lost outer hair cells, you're at the middle panel, not the top one anymore. Now the peaks are dull and rounded. They're smaller in amplitude, and they have less sharpening. If you look at the bottom panel, you'll see what hearing aids might do. They will increase the amplification of the traveling waves, but they will not enhance that sharpening. And again, this just underlines the point that hearing aids cannot, can make the waves bigger again, but cannot sharpen it. That sharpening has to be addressed by directional mics and digital noise reduction. Let's stick to compression here. Oh, I still have another slide here. The sharpened traveling waves increases frequency resolution, and our ability to distinguish between frequencies closer together is enhanced. Our ability is thus enhanced to separate speech from background noise, and to compensate for a loss in ability to do that, we can increase the signal-to-noise ratio. So the sharpened traveling wave increases frequency resolution, and without that, we have a loss of ability to distinguish between frequencies close together, a hard time separating speech from background noise. So to compensate, we can increase signal-to-noise ratio by means of directional mics, etc., a topic for another webinar. We can, however, amplify traveling waves by the use of compression and amplification, and especially with WDRC, wide dynamic range compression, because that is meant specifically to imitate the outer hair cell action. So read the title again here. We can, however, amplify traveling waves in the same manner as is done by the outer hair cells. We can focus the amplification especially below 50 dBHL, focus the amplification on soft inputs, and gradually decrease the gain as the inputs increase, because remember, with sensorineural loss, the ceiling of loudness tolerance hasn't changed much. It's all about the floor having been raised. So this is the focus of wide dynamic range compression, also known as WDRC. Look at reduced dynamic range. On the left, you'll see normal hearing, 0 dBHL. You'll see UCL on the bottom, around 100 dB, and MCL sandwiched in the middle. And I always call this the audiogram, O-D-D-I-O-G-R-A-M, because it's the upside-downogram, where the floor is illustrated at the top and the ceiling is illustrated at the bottom. I do wish it would be flipped around, but that is something that, thank goodness, is done in today's real ear. A different topic. On the right, look at the reduced dynamic range for this person's left ear. Now the floor is at 50 dB, and the UCL hasn't changed from the audiogram on the left, and the MCL is squashed in the middle again. The dynamic range, dynamic range always deals with decibels. It doesn't... We're not talking frequency here. Dynamic range always deals with intensity. And you can think of dynamic range as the decibel distance between someone's floor and someone's loudness tolerance. So this person's decibel distance is reduced. This is the focus, by the way, of all fitting methods, half-gain rule, and the reason for compression, the reason for the season of compression in the first place. Look at the reduced dynamic range. This means we cannot mirror an audiogram with average speech inputs. Look at this. You've got on the left here soft inputs. You can see the green dashed line at the top of the audiogram. Look at the person's X's here. The left ear thresholds are at 50. So this person, the soft green line at around 5 or 3 decibels or whatever, we can, C-A-N, amplify those inputs by the full degree of the loss. Follow the green arrow. Yup, you betcha, we can. But what happens if the input is at the yellow dashed line? What happens if the sound coming into the hearing aid microphone is at around 55, as I'm showing with my cursor here? How now, brown cow? We cannot amplify that input by the full degree of the loss. Witness, we can only amplify it by the distance of the yellow arrow. That was the whole fundamental foundation for the half-gain rule as established by Leibarder decades ago. We cannot mirror an audiogram with average speech inputs because average speech inputs are going to be around 55, 60 dBHL, and we can only amplify those by about half the degree of the loss. Look at the red dashed line right under the word MCL. That can be amplified by only even less, by the red arrow. So the amplification goes down from green to yellow to red in accordance with the reduced dynamic range. The whole idea behind that old half-gain rule, look at what they had to deal with. They didn't have real ear. Look at this person's thresholds from near normal on the left dropping to a moderate degree on the right. And they would take the person after he wore headphones and raised his hand under headphones. Then they would take a hearing aid, and yes, it was a monaural hearing aid because in those days ear, nose, and throat doctors didn't like binaural fittings. And so the person was seated in front of a speaker in the same room where he was tested under headphones, and he was seated a meter or a yard in the U.S. from a speaker. And warble tones would be emitted from the speaker, and the person would raise his hand every time he heard the warble tone. And you would ask the person to set the hearing aid to a comfortable loudness volume, and in that quiet room with the hearing aid at a comfortable volume, the listener was asked to once again raise his hand when he heard the tones, only this time from a speaker. And you felt like you were playing an organ because 250 hertz would come out like, and 500 hertz, and 1,000 hertz, people are thinking I'm having a spasm here. But at any rate, the aided thresholds, the letter A's, would be drawn across the audiogram. And you would hope that the letter A's were not restored all the way up to zero. You would hope that the A's were halfway lifted. And you can see that by the numbers here. They are half the degree of the person's hearing loss. And the bottom line was, if you amplify by that amount, then the goal would be achieved. Secondly, aided speech outputs, note the words below the thresholds here, aided speech outputs would be nicely sandwiched between the thresholds and the loudness discomfort levels for speech. If you amplified those warble tones by half the degree of the loss, then with average input speech of 50 to 60 dBHL, that would be aided so that it would be situated nicely in the middle of the guy's dynamic range. And that's what was done with linear hearing aids. Because linear hearing aids gave the same gain for all input levels. If you had the volume set at halfway or two-thirds, and you had the linear gain hearing aid giving X amount of gain at that volume, well then, by gum, that linear hearing aid was giving that gain to a whisper, to average conversational speech, and to a yell. Look at the picture here. In this example, we've got input on the horizontal, output on the vertical. And the frequency is always 2,000 Hz, unless otherwise stated. So frequency is not listed on input-output functions. Input-output functions are the language of compression. They are the way that compression is usually described. And look at the input here from 0 to 100, and the corresponding outputs from 60 up to 120. In this case, the linear hearing aid is giving 60 dB of gain. You can see in the lower left corner, 0 in, 60 out, 20 in, 80 out, 40 in, 100 out, and so on. The line or function on the graph is at a 45-degree angle. It's tit-for-tat. For every 20 dB of increased input, my output correspondingly increased also by 20. It's 20 for 20. The first number on an input-output ratio is always input, and the second number is output. So for every 20 decibels of input increase, I got a corresponding 20 decibels of output increase. It was 20 over 20, and that boils down to 1 over 1. This is a linear gain, a one-to-one ratio. And then, if the inputs keep increasing from 60 to 80 to 100, you don't want the outputs to increase anymore, or you'll cause more hearing loss. You'll ruin more hair cells. So they limited that maximum output at, say, for example, 120 dBSPL, and they limited it by means of peak clipping. Now, what is peak clipping in English? You think of a sound wave, like a letter S laying on its side, and you think of the sound wave having peaks and valleys. Well, they literally cut the peaks right off, just cut them off and made them square. Little Jack Horner sat in the corner and got a square rear end, if you know what I mean. So literally, peak clipping occurs in the receiver of a hearing aid. The receiver is the speaker. I know receiver's a weird name. They should just call it a speaker at any rate. The speaker has a diaphragm in a metal box, and the speaker changes electricity into sound. Microphones are backward speakers. Mics change sound into electricity. Speakers are backward mics. They turn electricity back into sound. Now, the diaphragm wiggling back and forth like an S-shaped movement, if it starts to slap the metal sides of the box, well, then you can see that the S-shaped motion of that diaphragm speaker is going to have its peaks clipped. And so, you're going to get square waves instead of round waves. And then, you've introduced distortion. You've introduced harmonics that aren't even in the input signal. And so, the output is going to be bigger than the input because the hearing aid created gain, but it's also going to sound quite distorted, like I'm calling you from the bottom of the swimming pool. And that's why peak clipping sounds terrible, okay? So, you don't want the hearing aid to clip peaks. You want the linear gain itself sounds clean, no caffeine. It actually sounds quite nice as long as the output does not exceed the level where you introduced peak clipping. So, peak clipping is used to limit the maximum power output in linear hearing aids, but it gave a lot of distortion at the same time. Then came output-limiting compression. In the 80s, okay, the early 80s, output-limiting compression entered the scene. It had an input-output function that would be similar to linear gain. Again, in this example, I'm just showing 60 dB of gain, okay? They could give 30, it could give 40 or whatever. I'm just showing an example here. So, notice that this input-output function looks a lot like the linear one did on the previous slide. It's just that now look at the peak clipping portion here, and now we'll move to the next slide. Notice how peak clipping is no longer used. Now you've got compression, and you can think of compression like a rapids in a river, a backward current that sort of slows the river down a bit. And so, instead of using peak clipping to limit the maximum power output, they used compression. And note, on the horizontal axis, as the input increases from 60 to 80, the output, look at the red line now, only increased by about 5 dB. So, not much output increase. My input may have increased by 20. My output, look at my cursor, may have only increased by 5. So, you had a 20 to 5 ratio, which boils down to a what? A 4 to 1 ratio. So, now you're starting to get compression, 4 to 1. You can have 8 to 1, 10 to 1. Here's something I want people to really realize, okay? You might even want to jot this down in your notes. Going from a 1 to 1 to a 2 to 1 is a big change, or from a 1 to 1 to a 4 to 1 is a big change, okay, in gain. But you start to get a law of diminishing returns. Going from a 4 to 1 to a 5 to 1 is less of a change, to a 6 to 1, to an 8 to 1, to a 10 to 1, to a 20 to 1. Now you've got barely any differences. There's hardly any difference between a 10 to 1 and a 20 to 1 compression ratio. But there is quite a difference between a 1 to 1 and a 2 to 1. And that diminishes as the compression ratio increases. At any rate, look back at the words on the left side of the screen. Output equals input plus gain. Input is the sound coming into the hearing aid. Gain is the muscle of the hearing aid added to it. And the sum total is the output. And this little formula is important to bear in mind. Output limiting compression has a high knee point. Let's say it begins, for example, at 60 decibels. That's where the function takes its bend, like a knee in a leg. Linear gain is to the left of the knee point. You can see that in blue. 45 degree angle. 1 to 1 compression. Okay? In other words, no compression. As my inputs increase by 20, my outputs increase correspondingly by 20. Output increase equals input increase. But then once to the right of the knee point, you've got a high compression ratio. For inputs greater than 60 dB SPL, compression dramatically limits the MPO. So you can think of the previous slide, think of someone jumping up and down on a bed and he's hitting his head against the cement ceiling in the red peak clipping area. Think of output limiting compression as someone tacking a sponge to the cement ceiling. And now when Teddy jumps up and down on his bed and hits his head on the ceiling, it doesn't hurt quite as bad because there's a little bit of a give. The output is giving just a little bit as the inputs increase. So you're limiting the MPO, but not by hard peak clipping here, but by a gentler means with output limiting compression. And output limiting compression is associated with two things. One, a high knee point, and two, a high compression ratio of at least 4 to 1. Loudness growth at the end of the 80s, early 90s became all the rage in clinical fittings. And this is because the role of the outer hair cells began to become more and better known by clinicians. Outer hair cells became the knowledge base of outer hair cells increased. The knowledge that they namely, that they increase, that the cochlea is actually like an amplifier. That inner hair cells cannot pick up sounds below 50 to 60. They need the help of the outer hair cells to do this. So outer hair cells amplify soft sounds so that the inner hair cells can pick them up. So this picture shows loudness growth. DBHL is shown along the horizontal and your perception of those decibels is along the vertical axis. Now normal loudness growth occurs with the red line. The person, for the normal hearing person, 10 to 20 dB is perceived as very soft. And for the normal hearing person, 50 to 60 is perceived as comfortable. And for the normal hearing person, 90 to 100 is perceived as too loud. So this person's loudness grows like a tree. Its growth rate is a gentle increase of slope. Now look what happens with mild to moderate sensorineural hearing loss. Presbycusis, the most common hearing loss in the world. We look at that by the blue line. Now this person can't hear until 50 or 60. And so 50 or 60 sounds very soft. And yet his ceiling, 90 to 100, still sounds too loud. And so where the red and blue lines meet is called recruitment. And look at the orange arrows. The job of a hearing aid then to imitate the outer hair cells should be to amplify soft sounds by a lot, amplify average sounds by less, and amplify loud sounds by little or nothing at all. Now where have we heard that before? Okay. Here came the compression to address this specific observation. Along with loudness growth came wide dynamic range compression. And now think of the words. Shrinking a wide dynamic range into a smaller one. And you can think about that by holding your right hand horizontally by your eyeballs and hold your left hand by your chest and keep the right hand where it is and raise your left hand up to your nose. You're raising your floor and your ceiling is staying the same. That's what happens with a reduced dynamic range. So WDRC has to address this. Focus on imitating, we'll look at the title, imitating outer hair cell amplification with soft inputs. The cochlea is actually a WDRC amplifier. And look at this input output function compared to the prior ones shown here. This is output limiting compression. Notice the high knee point, high ratio as shown by the red line, which is very shallow. And now look at WDRC. The knee point is lower and the red line has a shallower slope. So the knee point is at around 40 dB SPL. And as with output limiting compression, linear gain is shown to the left of the knee point. The compression ratio, however, is also lower. As inputs increase, it says on the left here, the gain is slowly decreased. Witness, as we go from 40 to 60 dB of input, my output, look at my cursor, only went from 100 to 110. So let's look at 40. The output is now 100. And now look at 60. The output is now 110. So my input increased by 20 here, but my output only increased by 10. That's a 20 to 10 ratio, which boils down to a 2 to 1 ratio. So you can think about, again now, I'm going to give you an analogy and this isn't mine. This is actually from Francis Cook, KUK, who described this years ago. Output limiting compression, you can think of a kid in his dad's car and he's speeding down the road and he sees the stop sign. And as he gets near the stop sign, he slams on the brakes. WDRC, on the other hand, the same road, the same car, but now it's a little old lady driving the car. And halfway down the road, she already begins to slow, slow, slow down for the stop sign, angering everyone behind her. So these are very different types of compression. Now we're going to compare the two side by side. Look at output limiting on the left and WDRC on the right. Output limiting, high knee point, high ratio, in this example, 10 to 1. Okay, what the hay. And WDRC, low knee point, low compression, 2 to 1. Low knee point, low compression. Back on the left, for whom would this bell toll? For whom would this hearing aid be most appropriate? One would think severe to profound hearing loss. How come? Because you want to provide a lot of gain and you want to provide a lot of gain for soft inputs, look where my cursor is, for medium inputs. So you're providing linear gain, which is more gain than compression. Always know that. Linear gain is more gain. It's tit for tat. It's 1 to 1. And provide lots of gain for soft to average inputs and then suddenly decrease the gain and decrease the output for loud inputs. So with output limiting compression, you've got a strong, high degree of compression over a narrow range of inputs. Which inputs? Loud ones only. Now go to the right. You've got linear gain, but only for soft inputs. Just the soft ones. You're trying to imitate the outer hair cells, which amplify sounds below 50. And then you've got a weak degree of compression for medium inputs and for loud inputs. So WDRC provides maximum or linear gain for very soft inputs only and then it provides a weak degree of compression over a wide range of inputs. Wide dynamic range compression. The purpose is very different for WDRC specifically addresses which hearing loss? Mild to moderate. Because that's the hearing loss caused by outer hair cell pathology. And outer hair cells amplify soft sounds by a lot. So you want to focus your linear gain, read most gain, for soft inputs only. And then the gain gradually decreases as we follow this red line here. Now when you go back to the, some white might say, well what kind of hair cell damage does severe to profound hearing loss have? Well they've got not only outer hair cell damage, but also now some inner hair cell damage. Outer hair cell damage tends to precede inner. You lose outer hair cells due to noise and you lose outer hair cells due to aging. And if a person has severe to profound loss, rest assured that person's got hair cell damage to both populations of hair cells. So that's why I divide the clinical hearing loss population really into two camps. And it follows with hearing aids too. People with mild to moderate sensory neural loss, they need a lot more bells and whistles on their hearing aids. They're actually harder to fit because they can still hear without their hearing aids. They can hear one on one. They just have trouble in noise. Whereas severe to profound hearing loss, the glass is always half full. They simply cannot hear without the hearing aids. So they've got to have the amplification. And usually the hearing aids for that population are simpler. They're more powerful, but they usually have fewer bells and whistles because the person hasn't got enough hair cells in order to utilize all the different other accommodations that come with the complexities of today's hearing aids. At any rate, let's move on from input-output functions and move on to how output-limiting compression on the left would be compared to WDRC on the right in terms of frequency response, not input-output. Let's get up to a more friendly graph, the kind we see on ANSI or the kind we see on our screens usually when we're doing manufacturer software, when we're fitting. You've got frequency along the horizontal. And once again, now you have gain along the vertical. So frequency is now on the graph and so is gain, whereas input-output didn't show you that at all. So these are just two languages. You can think of this as German and this as Polish. They're telling exactly the same story, just in two different ways. Output-limiting would provide lots of gain for soft inputs of, say, for example, 40, lots of gain to average inputs of like 60, and then suddenly it would provide less gain for loud inputs like 80. The breaks are suddenly scrunched. WDRC on the right is different. Same maximal gain maybe for soft inputs, but then less gain for average inputs and less gain for loud inputs. The little old lady slowly applying the breaks. So WDRC hearing aids, by the way, if I was truly honest here, I would have lowered these whole three lines a bit because WDRC hearing aids are usually less powerful as well. But this is just to illustrate the differences in compression used. Compression is also adjusted in different ways. On the left again, output-limiting. Now look carefully at this slide. This is a very quizzable slide. Ha, ha, ha, hint. Look at you've got linear gain as shown with the blue diagonal 45-degree angle line, and now you've got adjustments to the output-limiting. You can lower and lower the maximum power output. Look, I'm lowering the knee point from the top. I'm lowering it over here. I'm lowering it over here. And in so doing, I've reduced, reduced the maximum power output. I didn't change the ratio of compression. Notice the angle of the red lines remains the same. I've just lowered the maximum power output. On WDRC, the adjustment is completely different, or as Monty Python would have said, it's completely different. Because here, witness as I lower my knee point, something totally different is happening. My gain is changing. What gain? My linear gain. And take this point home carefully as well. Put a star by this one. As I lower my knee point with WDRC, I'm increasing my gain. Increasing. So you might think, oh, this is maximal gain. Uh-uh, it's not. And we'll show you that in a second here. Let's see if I can show you that. Let's show you right here. Look at here. If I look at the rightmost vertical, look at this gray vertical line. This amount of input is required, follow my cursor, to give me this amount of output. Now, if you could put your finger on that part of the screen, hold it there. If I lower my knee point, I need way less input to give me almost exactly the same output. Do you see that? And that's what I mean by saying as you, as you, with WDRC, as you lower the knee point, and as these 45-degree angle lines then go to the left, your gain is increasing. Which is, again, backwards to what people would think. But the only thing that takes this point home is knowledge that input plus gain is output. And with the highest knee point on the right here, I need way more input to give me this output. Hold your finger there now. And then if I lower the knee point way to the left, I need a lot less input to give me almost the same output. So this is just something to keep in mind. To summarize here, a clinical, a clinical spectrum of compression. On the left, you've got linear gain. The middle pairs has, I should say, this is linear gain on the left. Output limiting is the center pair. And WDRC would be the right pair of lines. Look at the horizontal lines. These are the inputs. Soft inputs, the bottom yellow, is lifted a lot by linear gain. Average input is the light blue, lifted a lot by linear gain. Loud inputs would be lifted a lot, but suddenly you've exceeded this dashed line, which is the listener's loudness discomfort level, and so you've used peak clipping, which distorts the outputs. The middle, output limiting. Once again, soft inputs amplified by a lot. Once again, average inputs amplified by a lot. But now, loud inputs are suddenly compressed by a lot. Output limiting compression. You've limited the MPO without peak clipping. Now on the right, soft inputs lifted by a lot. Average inputs lifted by less. Loud inputs lifted by even less. And so this would be the analogy. The arrows can be seen as the gain. Input plus the gain is the output. And the WDRC, slowly, you can see the blue here is already given less gain than it was under output limiting compression. So WDRC provides a low degree of compression over a wide degree of input intensities. And you can see that the lines on the right here are more evenly spread out as well. It's called wide dynamic range compression. So to summarize, WDRC is appropriate for mild to moderate sensory loss. Output limiting would be appropriate for severe to profound sensory neural loss. Today, because analog was either output limiting or WDRC or linear, today's hearing aids are digital. We can combine everything together. Linear for soft inputs, WDRC for average inputs, and output limiting for loud inputs. We've got two threshold points here. And then you've got something called expansion. Today's digital hearing aids do something really interesting. They tend to provide linear gain and expansion. I'll explain this in just a bit. And then they provide linear gain again for slightly louder inputs. And this would be to address a listener's complaint. I hear people at tables further away better than I do the person sitting across the table from me. That's because he's wearing WDRC. And remember, WDRC amplifies soft inputs by more than average inputs. So it's gonna focus its amplification on softer sounds coming from further away. And it's gonna decrease the gain for the listener right across the table from the speaker right across the table. So that complaint I hear better in restaurants. I hear people further away than I do, I hear them better than people close up. Is linear gain all that bad? So this is why manufacturers have started using linear gain again at a second time for certain inputs. As long as linear doesn't distort, it can sound delightfully clear. That's why hearing aid manufacturers use it for some inputs. For example, average to slightly louder speech inputs to address the complaint. I can hear others at tables further away better than the guy across the table. An interesting thing that's come up is ADRO, Adaptive Dynamic Range Optimization. Interesting departure, it focuses on linear gain. Look at this picture here, WDRC. Read the slide with me. Note, linear gain only for very soft inputs up to 40. WDRC occurs for medium and louder inputs. As inputs increase from 40 to 100, the outputs increase only by half as much. This is a compression ratio of two to one. This is how WDRC amplifies soft sounds by a lot and louder sounds by progressively less and less. Well, 40 dB input plus 50 dB gain is a 90 dB output. 40 in, 90 out. 60 in is 100 out. So the gain is now decreased to 40. 80 dB input is now only 110 out. So now on the left, it's a 30 dB gain and so on. ADRO, Adaptive Dynamic Range Optimization says nothing doing. They stay using linear gain. Look at what they do. They increase and decrease gain accordingly. Greater and less, read in the blue at the top. Greater and lesser amounts of linear gain are provided depending on the listener's sound environment. In loud environments, less linear gain. In soft environments, more linear gain. The focus is the listener's comfort. Note, look at the bottom here, that a shift to the right actually shows a decrease in linear gain. Remember that? For example, a 40 dB SPL input results, look at the rightmost line where the arrow is shown, okay? The rightmost diagonal blue line, 40 in is 70 out. As the diagonal line moves to the left, 40 in is 90 out. You see the gain now increased to 50. Thus highlighting the point once again that on an input-output function, as the linear gain line moves to the left, your gain is increased. Anyway, ADRO decreases and increases its linear gain. It never uses compression. It just uses greater or lesser amounts of linear gain, and here's why. Unaided speech is shown at the top. Here's a waveform. It might be the sentence, my father can beat your father at checkers. Time, it's along the horizontal, and amplitude is the vertical, and this is the sound wave of a sentence. The top sound wave represents an example of a sentence spoken at an average conversational loudness level. The peaks are the louder parts, namely the vowels. The valleys might be the softer parts, namely the unvoiced consonants, like S, S, CH, everything else. Aided speech with WDRC, remember WDRC amplifies soft sounds by the most, so WDRC is gonna amplify the valleys, the soft parts, by a lot, and it's not gonna amplify the peaks by very much. The bottom sound wave, read with me, represents the same sentence amplified with WDRC, and note how the overall wave is amplified, but the peak to valley contrast is decreased. WDRC amplifies soft sounds by a lot, and louder sounds by less. ADRO differs. It sticks to linear gain. Here's the same unaided speech sentence on the top panel, but because ADRO uses linear amplification, it's gonna preserve the peak to valley contrast. Note how the overall sound wave is amplified, and the natural peak to valley contrast is preserved. ADRO just simply uses greater or lesser amounts of linear gain, but they never use compression, and this is being used a little bit by other manufacturers now as well. It's starting to enter the scene because of this particular situation. When the peak to valley contrast of the speech waveform is preserved, you've got better speech intelligibility. We now are, this is, we are going over our time level here, just about, but I'm gonna just persevere. I'm almost done. Let's talk about expansion. Remember we looked at that one slide, and I said I'd explain expansion later. Expansion is used along with WDRC, only with WDRC. Expansion is known as an internal noise squelch. Remember what we said about WDRC. It amplifies soft sounds by a lot. Well, in dead quiet, WDRC is gonna be amplifying its butt off, and in so doing, it's gonna amplify internal circuit noise of the hearing aid, and the listener who has good low-frequency hearing is gonna hear that amplified circuit noise by the mic and the amplifier. That's gonna provide, so expansion is known as an internal noise squelch. Here's how to understand it. Here's W on the left, input-output function. Once again, input on the horizontal, output on the vertical, and WDRC is the black dashed 45-degree angle line, one-to-one with its knee point, and here's WDRC going to the right, two-to-one compression. Expansion is the dark blue line going down at a really steep rate, all the way down to here. It's a greater, it's the opposite of compression. It gives a greater-than-one-to-one ratio. You could call expansion, instead of a two-to-one ratio, it's a one-to-two ratio. For every one dB of input increase, I've got a two decibel increase for outputs, and the way to understand that is to take a chill pill, adopt some zen, and let the picture tell the story. Look at how zero in is zero out. 20 in, follow my cursor, is 40 out. 30 in, 60 out. 40 in, 80 out. Look at the words in the black words at the bottom center. Zero input is zero output. The gain is nothing. 20 input, 40 output. Here, 20 input, 40 output is a gain of 20. 40 input is an 80 output. That's a gain of 40. And then 60 input, now your WDRC is kicking in, is a 90 output. That's a gain of 30. 80 input is an output of only 100, gain is 20. Look at how your number's here. Gain of zero, 20, 40, 30, 20. Your gain is maximal at, and only at, the knee point. And for softer and softer and softer inputs, my gain is less and less and less. Hopefully, your knee point is set at a soft level like 40 or 30, so that soft speech is given the main gain. On the right, you're looking at the very same story, only here, the horizontal is input, and your vertical, instead of being output, is gain. Look at how WDRC on the left, a one-to-one ratio, the gain is 40 here, the gain is 40 for a 20 input, the gain is 40 for a 40 input. That is shown here on the right. WDRC is giving a steady 40 decibels of gain for all inputs from zero to 20 to 40. And then the gain goes down as the inputs increase. Well, expansion is the blue. It gives no gain for zero inputs, 20 gain for 20 inputs, 40 gain for 40 inputs, and then less gain as the inputs continue to increase. That is how expansion suppresses internal noise. And it's always used along with WDRC for whom people who have good low frequency hearing. I finish with this. Yes, there is an end to this nattering diatribe with dynamic compression characteristics. Dynamic always refers to your watch, time, always. So sudden changes to input intensity. The top panel represents environmental sounds. They're soft here, then suddenly they get louder, and then suddenly soft. Here's the hearing aid. On the bottom is the hearing aid's reaction. It's giving gain to the soft inputs, and then all of a sudden, because the input sound increased in the environment, the hearing aid went into compression, and it took some time to go into compression. They call that the attack time, the reaction time of the hearing aid. How long in seconds did it take to go into compression? And then when the environmental sounds were soft once again, how long did the hearing aid take to release from being in compression? So dynamic characteristics regard or address the time it takes for a hearing aid to react, to go in and out of compression. There's various types of it. The top here says automatic volume control. That is a deliberately slow and long attack release time. Meant to imitate the length of time it takes for a client to raise his hand up to his ear and adjust the volume to compensate for changes in the environment. So automatic volume control types, those have long attack release times. The opposite is syllabic compression, quick attack release times, really snappy, very fast, almost like watching an antique roadshow when you can hear the flutter in the loudness of the background people while the announcer is talking to the lady about this vase is worth $20 million, and you found it in the garage, and you can hear the puffing, the squeezing and separating. You can hear the kind of the effect of fluttering of syllabic compression. Average detection is varying attack release times, slow attack release times for sounds that take a long time to become loud, and quick attack release times for a door slam. So the attack release times vary depending on the time it takes for the sounds to become louder and softer. Transient sounds receive quick attack release, slower inputs receive longer attack release. The bottom line is today's digital hearing aids tend to default to syllabic detection for the low frequencies, and they tend to use average detection for the highs. They may differ, but here's a word to the wise. Don't mess with Texas. Don't mess with dynamic compression on your manufacturer software. Unless someone guides you through it, I wouldn't touch it. I would leave that puppy well alone. I'm done now. I hope you've enjoyed it. It's been a slice. Maybe half the people have now left the room going, this guy never stops yapping, but I can now entertain questions if we like. Meanwhile, I'd like to say thank you on behalf of the group and I hope we pass the audition, as the Beatles would have said. Ciao. It's been a slice. Thanks, Ted. Ted, we're so excited that we've had over 250 of your fellow colleagues that have joined us today on this webinar. As Ted said, we do have some time for questions, albeit not a lot of time, but we do have a few minutes for questions. If you do have a question for Ted, please enter it in the question box on your webinar dashboard. Ted, our first question is from Steve. And Steve asks, what is really meant by peak clipping? I hear this term all the time, but don't really know what it involves. Okay, that's a good question. We went over that, but I'm gonna explain it again. Peak clipping is literally what you would think of if you close your eyes and listen to the term. You're clipping the peaks of the sound wave. Think of a sound wave as a letter S laying on its side. And what you're doing with peak clipping is literally taking a knife and slicing off the rounded peaks. Now you no longer have a letter S anymore. Now your letter S looks like a square. It's got squares in it, okay? And when you've done that, if your input was a pure tone of let's say 1,000 hertz, if 1,000 hertz is amplified so that peak clipping occurs, now you're gonna clip the peaks off of that amplified 1,000 hertz tone and suddenly you have, by doing so, you have introduced harmonics of 1,000 hertz. You've introduced 2,000 hertz a little bit. To a lesser degree, 3,000 hertz. And to a lesser degree, 4,000 hertz because those are harmonics of the fundamental frequency, which was 1,000 hertz. It's like asking Mary to come to the party and Mary brings her two kids along, okay? You didn't know, all right? Uninvited guests came to the party. That's what happens with peak clipping. It introduces distortion. And what kind? Harmonic distortion. The addition of added harmonics to the input sound frequencies. And you've done that because you've clipped the peaks. And they got clipped because the diaphragm of the speaker was moving so much side to side that it slapped the metal walls of the speaker box. And you can construe that as clipping the peaks. Thanks for your question, Steve. Thanks, Ted. Ted, our next question looks like it is from Alan. And Alan asks, do you find some manufacturers are superior to others in the use of compression and digital mic technology? Or are they all about the same? They are all about the same. I'm gonna just give that bottom line for two reasons. One, I don't work for any particular manufacturer and it would be unfair for me to say that. But two, I've been a clinician as well. And knowing from experience, the fitting of the adjustment and hearing aid manufacturer software, many of the same trends are used by many of the manufacturers. They all provide expansion. They all provide WDRC. They all provide output limiting compression. So really, you know, that is pretty well, compression is fairly well addressed by most manufacturers. I don't find a huge difference that way, except in the use by which some might address linear gain for those mid to slightly louder than average inputs to address that question I was mentioning about the listener in a restaurant, hearing people further away, that use of linear gain for other inputs other than really soft inputs, that is being used by several manufacturers because they know that they don't want the speech waveform to be distorted. Thanks, Ted. Ted, our next question is from Eva. And Eva asks, when should we change attack and or release time? You shouldn't. Leave it. You really, you know, unless you had some strange kind of distortion and you were calling up the manufacturer to get help on the phone, I would let that manufacturer guide you through it. Because if you go in willy nilly and start messing with attack release times, again, experience has taught me to leave well enough alone. And if I am going to adjust it, I'm going to do it with my hand being held by the manufacturer's audiologist or person on the other end of the line in their help desk. So make a long answer short, I would try to leave that alone as much as I possibly could. They've got it fairly well set up in their software for their particular product. Good question though. Thanks, Ted. Ted, our next question is from Bradley. Bradley asks, when a client complains that he can hear farther away than a person talking to him at the same table, I find that increasing the gain for loud sounds helps. What does that do to the compression? Sometimes the best thing is to provide linear gain for that person. Sometimes the idea is that you can crank up the volume and decrease the compression. The idea is generally when you've got, what brings that issue to the fore? And it is a very, it's a common complaint because WDRC left by itself is gonna act like an idiot. It's gonna just go according to the rules, what it's supposed to do. I've got to amplify soft sounds by a lot because I've got to imitate the outer hair cells, don't you know? And then in so doing, you're getting softer voices becoming amplified. Some people say that the best way to address that is to reduce the compression, reduce the compression ratio from two to one to a more like a one to one for that input intensity, for average speech intensities of 50 to 60 dB HL or 60 to 70 dB SPL. Reduce your compression and use linear gain for that area of intensity. And then you will decrease that problem. And the nice thing is digital hearing aids let you do that today. We couldn't mess around with as easily address that with analog, but with today's digital hearing aid, you can. Thanks, Ted. Ted, our last question is from Susan. And Susan asks, you mentioned that to read an input output graph, gain increases are seen as the 45 degree angle line moves to the left. Shouldn't that be the other way around? No, it really isn't. It's like, and I can I take you back a couple of slides here? Let me show you something here. Follow this slide right here. Do you see that? Take a good look at this now. Look at these two diagonal lines here, 45 degrees. They're both giving linear gain. The right most line, if you follow from 40 input up here with my cursor and go to the left, my output is 70. Input plus gain is 70. 40 plus 30 is 70. Now let's keep your input at 40 and follow the vertical line all the way up to the right most diagonal blue line. And now you'll see that my cursor, that when I move it to the left, it's at 90. Input of 40 plus a gain of 50 equals 90. So with my left most line, I've got 50 dB of gain. And with my right most diagonal line, I've got 30 dB of gain. Again, let the picture tell the story. And it's just, those are the rules of reading input-output functions. It's a great question though, but I think it's good that you asked it because it underlines this seminal concept of input-output functions. I've got a little book on compression. Check that out, Compression for Clinicians. But this has been a great webinar. I've very much enjoyed giving it and the questions today. Thank you so much, IHS, for letting me do this. Absolutely, Ted, always a pleasure. Ted, I'd like to thank you for an excellent presentation. And I'd like to thank everyone for joining us today on the IHS webinar, Hearing Aid Compression, Digital Microphones and Noise Reduction. If you'd like to get in contact with Ted, you may email him at tvenemaatshaw.ca. For more information about receiving a continuing education credit for this webinar through IHS or AHIP, please visit the IHS website at IHSinfo.org. Click on the webinar banner or find more information on the webinar tab on the navigation menu. IHS members receive a substantial discount on CE credits. So if you're not already an IHS member, you will find more information on IHSinfo.org. Please keep an eye out for the feedback survey you will receive tomorrow via email. We ask that you take just a moment to answer a few brief questions about the quality of today's presentation. Thank you again for being with us today and we will see you at the next IHS webinar. ♪

webinar

compression

digital microphones

noise reduction

hearing aids

audibility

WDRC

expansion

outer hair cells

linear gain

speech intelligibility