speaker and microphone quality?

Forums › Home Theatre (incl HTPC) › speaker and microphone quality?

nathanbarlow

48 posts

Geek

Topic # 185607 29-Nov-2015 08:45

Hi all,

Wondering if any of the audiophiles on this forum might be able to help me out with a rather odd question.

So I run a blog for people with cochlear implants (artificial hearing implants) (Link) and one of the challenges for us (I have one of these implants) is the telephone. A big chunk of the difficulty is to do with the physics and engineering of the phone. A small part is to do with the hearing loss.

I was wondering if anyone knew the spec's in a bog-standard telephone, for the speaker and for the microphone? I'm trying to explain to my readers why even though they may be great at hearing live voices in person, they find the phone is much harder. The blog is all about evidence-based and science-based explanations, so I'm doing my homework at the moment

So far, I've managed to find out that phones use an 8-ohm speaker (cheap 50c jobbie) but the impedance of a speaker doesnt explain why it's a poor quality speaker. Can anyone help either with an answer, or with type of properties of a speaker which influences sound quality?

Hoping there might be some telecommunications engineers on geekzone :)

Cheers

Nathan

Sideface

4216 posts

Uber Geek
+1 received by user: 2450

Lifetime subscriber

Reply # 1437168 29-Nov-2015 11:21

from wiki Voice frequency

"In telephony, the usable voice frequency band ranges from approximately 300 Hz to 3400 Hz.

"The voiced speech of a typical adult male will have a fundamental frequency from 85 to 180 Hz, and that of a typical adult female from 165 to 255 Hz.

"Thus, the fundamental frequency of most speech falls below the bottom of the "voice frequency" band as defined above.

"However, enough of the harmonic series will be present for the missing fundamental to create the impression of hearing the fundamental tone."

People with cochlear implants may not detect the harmonics referred to above - thus their problem with telephone voices. (That's my guess)

Sideface

Technofreak

2842 posts

Uber Geek
+1 received by user: 480

Reply # 1437179 29-Nov-2015 11:54

As Sideface mentions the bandwidth for the telephone is 300hz to 3.4khz. This will limit some of the voice definition however this bandwidth is plenty wide enough to be intelligible and also be able to recognise the person talking.

I wonder that the problem is the relationship between the position of the phones speaker and the ear and the implant.

Placing the phone handset over the ear now encloses the implant in a closed space which probably affects the audio signal in that space thus changing how the implant reacts to the audio signal. To give an example, in some rooms voices seen to echo about making it hard to hear what's said yet in other rooms you won't have a problem.

Does moving the handset position help. In other words not fully covering the ear, or moving the handset away from the ear slightly.

Sony Xperia X running Sailfish OS. https://sailfishos.org The true independent open source mobile OS
Samsung Galaxy Tab S3
Nokia N1
Dell Inspiron 14z i5

Batman

Mad Scientist
18928 posts

Uber Geek
+1 received by user: 2460

Trusted

Lifetime subscriber

Reply # 1437184 29-Nov-2015 12:37

Found this http://www.michiganear.com/audiology-cochlear-implants.html

Batman

Mad Scientist
18928 posts

Uber Geek
+1 received by user: 2460

Trusted

Lifetime subscriber

Reply # 1437185 29-Nov-2015 12:40

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4490929/

nathanbarlow

48 posts

Geek

Reply # 1437186 29-Nov-2015 12:41

Sideface: from wiki Voice frequency

"In telephony, the usable voice frequency band ranges from approximately 300 Hz to 3400 Hz.

"The voiced speech of a typical adult male will have a fundamental frequency from 85 to 180 Hz, and that of a typical adult female from 165 to 255 Hz.

"Thus, the fundamental frequency of most speech falls below the bottom of the "voice frequency" band as defined above.

"However, enough of the harmonic series will be present for the missing fundamental to create the impression of hearing the fundamental tone."

People with cochlear implants may not detect the harmonics referred to above - thus their problem with telephone voices. (That's my guess)

that actually really helps! People with CI's can hear the fundamental frequency, but the algorithm in the implant processor puts emphasis on the harmonics above the fundamental frequency. So basically, an implant can hear the fundamental frequency if that frequency is on its own (eg pure tone sounds) but as soon as it gets to complex harmonics (like music or speech), the processor dampens down the fundamental frequency. It's actually why alot of Ci users like country music - not as many harmonics when picking a guitar I guess

So the phone not transmitting fundamentals, might mean, the first harmonic is being treated as a fundamental, and being dampened. Hmmm might have to look into that (its my day job as a grad student reseacher with CI hearing abilities) more seriosuly, but not for the blog.

nathanbarlow

48 posts

Geek

Reply # 1437198 29-Nov-2015 12:44

Technofreak: As Sideface mentions the bandwidth for the telephone is 300hz to 3.4khz. This will limit some of the voice definition however this bandwidth is plenty wide enough to be intelligible and also be able to recognise the person talking.

I wonder that the problem is the relationship between the position of the phones speaker and the ear and the implant.

Placing the phone handset over the ear now encloses the implant in a closed space which probably affects the audio signal in that space thus changing how the implant reacts to the audio signal. To give an example, in some rooms voices seen to echo about making it hard to hear what's said yet in other rooms you won't have a problem.

Does moving the handset position help. In other words not fully covering the ear, or moving the handset away from the ear slightly.

we have to place the speaker of the phone to our microphone of the implant, which is above the outside of the ear (pinna). There is a brand of CI called Advanced Bionics, who designed a mic to go in the ear canal, and then got this exact problem you describe, with phone usage. SO theres a phone mode for their mic, I believe.

Fred99

7223 posts

Uber Geek
+1 received by user: 3764

Reply # 1439765 3-Dec-2015 10:50

As well as limited frequency range and generally crappy speakers, there's some audio compression going on in telephony transmission.
I suspect a few things are going on here. The limits of phone speakers, the audio compression - and how that might relate to the way that the CI sound-processor then organises the sound before sending it to the implant itself, the background noise, plus I guess due to cost constraints, most have a single CI - not two, speech recognition is impaired, particularly when there's background noise, with unilateral hearing. (however, people with bilateral hearing don't usually have problems hearing speech through a telephone pressed to one ear).

Anyway, I assume that what you're doing has a purpose to improve the ability of people with CI to use (normal) phones. I presume that the CI sound processor has line-in inputs (my son's Cochlear brand BAHA does) to bypass the mics and direct-input a device, iPod, whatever. He never uses this. The sound processor is however located on an abutment in about the same position as where a CI sound-processor sits, a pair of good quality open-back studio headphones works pretty well, apparently, with the cup sitting over the sound processor. He'd never use a phone against/near the sound processor, but of course CI and BAHA are different in how the sound gets to the brain, and he's got a "good" ear on the other side, so there's no need.
If the phone isn't transmitting and reproducing the speech in a linear manner, but clipping, equalising, frequency shifting, splitting the signal or whatever for compression, even if this is to "optimise voice frequencies/recognition etc" then there's probably plenty of scope for it to conflict and not to work very well. As I understand it, CI splits the audio spectrum into a number of frequency "slices", each slice then sent to the corresponding electrode on the cochlea. IMO there seems to be potential conflict between an already processed/compressed signal, then what the CI sound processor does with the signal to get the CI working. The CI sound-processor could be trying to take slices of sound which have been removed or tampered with by the digital sound processing / compression in telephony.

Perhaps identifying the "weak link" with an audiologist, with tests of recorded phone voice DI'd to the sound processor, perhaps compare to DI'ing a recording made using a mic near the phone speaker itself. Perhaps phone speech recognition can be improved by re-programming and saving a user-selectable "phone mode" on the sound processor. Perhaps audio spectrum analysis of both the recorded direct and recorded via the phone speaker and a mic can provide clues as to why the CI / phone combination isn't working very well.
Probably, as well as sound quality differences between phones, there's going to be a difference between cellular and landline phones, skype etc, so at best a "compromise" might be possible. Perhaps just a simple equaliser app on a smartphone can be of some benefit, and relatively easy to set up.