[dhjdhj (OP)]

My wife like millions of others wears hearing aids to assist hearing. They amplify sound but eliminate the brains ability to tune out unwanted noise such as background music or other remote conversations with the result that cocktail party environments are impossible. Despite claims to the contrary by aid manufacturers this is true no matter how much you pay for your equipment. Does anyone know why this is and can any 'brain training' be developed to help? A second related question concerns lip-reading. We are trustees of a lip-reading trust and we are struggling to find any research as to how effective lip-reading is. We know it works and significantly improves communication, but we don't know how much formal lip-reading training improves the natural ability we all possess just by observation? Can any one help?   

[evan_au]

The human ear has an amazing capability in signal processing, including the ability to locate conversations of interest, and tune out unwanted conversations.

Some of that directionality comes from the higher frequencies, but unfortunately, with hearing loss, we tend to lose the higher frequencies first.

I recently attended a presentation on a Phd thesis that was looking at cochlear implants. They have developed algorithms that can locate one or more interfering sound sources, and null them out, to some extent. Similar algorithms have been used in military communications to reduce the impact of signal jamming.

Such algorithms have limited ability when there are just 1 or 2 microphones behind each ear, but if additional microphones could be arranged (eg along the frame of a pair of glasses), much better directionality could be obtained.

While the brain has direct feedback to the ears, that feedback does not extend to external hearing aids.

So for now, when organizing a party, make sure that there is a quiet area, for those who prefer to talk and listen.

[dhjdhj (OP)]

Thank you for your information we do have attendees at our lip-reading classes who have cochlear implants and they have exactly the same issues. In fact they have other problems relating to the sounds being different through the implant to those experienced through their normal ear system. Their brains need to relearn how to cope with one type of sound coming into one ear from the implant and the other coming from the ear with an aid as typically only one ear(the worst) is fitted with an implant. The idea of multiple speakers sounds good has any one built such a device and if so how effective is it?

[dhjdhj (OP)]

I would add why we are so interested in any body who has data regarding lip- reading. My wife recently was tested under controlled conditions at the queen Elizabeth hospital in Birmingham and the results were very interesting. On a standard voice test with a male speaker at standard volume her basic hearing allowed her to correctly identify about12% of words, with her hearing aids this increased to 50% when lip-reading was added the figure went up to 90%. Now much of this is identifying words by context because the standard test works with sentences. But these figures are showing that it is possible that lip-reading may be more than an enhancement it may a multiplier in the ability of deafened people to communicate especially in noisy environments. Given that it is a hell of a lot cheaper than ever more expensive hearing aids we are very keen to find out how much lip-reading training improves ability and how much training is required. We know from our own experience the short 10 week courses don't work that well. 

[Colin2B]

There have been some interesting experiments with eye movement detectors which show we look at someone's mouth while listening to them. Even more interesting is that researchers showed video of people speaking syllables but changed the syllable in the audio, many people 'heard' the sylable mouthed rather than that in the audio. We obviously use lip reading more than we realise.
There is a famous sequence in Butch Casidy and the Sundance Kid, where a phrase is mouthed but not heard, most of the audience get it.
My mother-in-law is profoundly deaf and won a number of competitions in lip reading, I'm told they asked her not to compete anymore to give others a chance! In her day the hearing aid was a box the size of an attache case!

[dhjdhj (OP)]

That's great do you know where this research was carried out? We know that some regional accents are much harder than others to lip-read. A soft southern American accent for instance is pretty impossible to lip-read. It also varies very much depending on the angle of observation. We do know that formal lip-reading does improve ability to communicate, but learning is not easy. We just don't know by how much and how long in terms of training hours are needed before the law of diminishing returns kicks in. Given that some 2 million people in the uk wear hearing aids and probably four times that have hearing loss of some kind it seems this that some one some where should have some hard numbers. If not it surely ought to be researched

[Colin2B]

It was shown on a TV documentary, Horizon? I remember seeing the demo, but not who did it. Perhaps someone from TNS staff might remember.
I know soft accents are hard to read because the lips don't move much, my mother in law found me hard to read when I had upper lip hair.
I'll ask my sister in law, she's a speech therapist.
Do you know what training methods are used? I wonder if flash cards alongside each pronunciation of a word would speed the process?

[dhjdhj (OP)]

I don't attend the training sessions, there are two teachers associated with our trust although one has recently gone private.They are fully trained and qualified specialists, who not only teach mouth shapes and sounds but also associated coping methods such as sitting positions, environmental selection and finger spelling. They seem to practise a lot by silent discussion of current affairs and any suggested topic of interest and one in particular finishes each session with a silent joke. The methods do seem to work but a bit like learning to drive its very difficult at first until they suddenly get it almost without realising. We operate in maximum groups of 12 but we are not even sure that is the optimum number to train as many people as we can within our limited budget. We do need data. Universities please note we have loads of volunteers for testing if any one is interested. Whilst I am on my soapbox can any one tell me why hearing aid manufacturers insist on making aid as inconspicuous as possible? Talking to a deafened person requires care and technique. You NEED to know someone has hearing loss to communicate with them. If there are any representatives of Phonac or Siemens reading this post make hearing aids fashion items like spectacles so people want to  wear them.

[evan_au]

Now much of this is identifying words by context because the standard test works with sentences.

One of the tests for cochlear implant patients used sentences with a word replaced by an unexpected "random" word. This is played in the presence of background noise.

It is a better indicator of normal speech to understand normal sentences with "expected" words than sentences with random replacements!

Many of the algorithms applicable to cochlear implants are also applicable to more conventional hearing aids.

lip-reading may be ... a multiplier in the ability of deafened people to communicate

My elderly father denied that he had a hearing problem. As he had gradually become deaf, over many years, he had unconsciously relied more and more on lip reading.

It became obvious when he could understand sentences while looking at you, but he could not tell you were talking to him when he was looking elsewhere.

One simple test from a professional was to ask him to repeat back a sentence. They did it again, but placed a sheet of paper over their mouth while speaking. This demonstrated to him that he had significant hearing loss, and he needed hearing aids.

Hearing aids helped, but as you say, they are far from perfect!

The more advanced (expensive) ones "talk to each other behind your back", using wireless signals to locate and accentuate sounds from straight a head. But they can't differentiate very well between sounds straight ahead and sounds straight behind.

[alancalverd]

Many mechanisms at work here.

We know that animals locate the source of a sound from the phase difference between the pressure waves reaching the ears. This is going to be larger for medium frequencies (500 Hz - 5 kHz) than for bass notes but above 5 kHz the wavelength is much less than the width of the head so you can get "aliasing", with multiple solutions to the phase-position equation. However it happens that the key elements of speech are in the 500 - 5000 Hz range, so we can train ourselves to discriminate the speech of the person we are looking at, from ambient noise. The intelligent binaural hearing aids Evan mentioned, presumably use a similar processing algorithm to improve directionality.
 
The ear itself also has a directional "dispersal" characteristic that modifies the perceived spectrum  so that even a single ear provides some directional clues. A single hearing aid with a fairly nondirectional microphone can't inject all the information needed to locate sound sources as accurately as two functional ears, and a single cochlear implant probably won't generate useful phase information at all.

There is also a strong predictive and trigger element in speech recognition. I often fly with a very experienced pilot who depends on lipreading on the ground,  but even in heavy radio traffic he picks out every call for  our plane because he is acutely sensitive to the callsign that precedes each message, and, being aware of our situation in time, space and proximity to other aircraft, also has a high degree of anticipation. This can help in a formal situation, in an office or conference room, but clearly isn't much use on the classic "cocktail party" environment, unless you have latched on to one or two key words.

The inconspicuity question is a good one, and I think fashions may change. Kids used to listen to bulky Walkman tape players with tiny insert earpieces, but current fashion seems to be to match a postage-stamp sized ipod with aviator-style headphones that announce "I am cool". Binaural hearing aids could surely be built into spectacles so even those with perfect eyesight could wear plain lenses in a frame that comfprtably carries a bigger battery (hence more signal processing power) and a better microphone.   

[dhjdhj (OP)]

Thanks for the replies and interest. We know that lip-reading is a skill acquired by many people by a kind of osmosis as they slowly lose hearing as your examples demonstrate but these people are the ones who have the confidence to continue to communicate despite hearing or paradoxically those who deny they have a problem. We train either those who have had rapid hearing loss like my wife or those who lack of confidence cuts them off from any situation where they may feel inferior or insecure. These people do not develop lip-reading skills. One of our biggest challenges is too persuade some people to come to the training groups at all as they have become too isolated. Interestingly my wife has a microphone that blue toothed straight into her hearing aids cutting out background noise. This improved her word recognition to 68% which is sufficient for meaningful conversation, but even this has big drawbacks. The mike has to be passed around if we are in a group.(this creates a topic in itself) or if we attend talks or lectures she gets the speaker to wear it. Questions then become interesting as she only hears the answers and has her on version of jeopardy. That is why we are so interested in any research. 

[evan_au]

Questions then become interesting as she only hears the answers

You could ask the speaker to always summarize the question, for those who could not hear it.

[dhjdhj (OP)]

Not very practical The answer is as I asked when starting this discussion is either very much better hearing aids or training the brain of users to tune out unwanted noises. The posts give some indication that this might be possible as deafened people appear to be able to tune in to specific sounds like the airline pilot mentioned above and also my wife ability to understand familiar voices at quite low amplitude. We will soldier on thanks to any one who replied we appreciate it