21st Century Audio DSP for the hearing impaired

Go To Last Post
91 posts / 0 new

Pages

Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Its been a decade or so since last I looked at audio dsp and I'm willing to bet that there is now a cheap easy solution that will give me multi-channel equalization and various special filters to handle things like echos in a room with bad acoustics.

I want to build my self a hearing aid where I can control as many parameters as possible to see what difference what factors make. I could pay somebody $5,000 and assume that they know what they are talking about and not just lying bastards trying to rip off old folks entering the Say-What? club, but I'd rather play with this myself.

Any suggestions?

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The Texas Instruments DSP units are quite nice. Also take a look at Alteras website as they are doing some amazing things

But I would go with the TI DSP. The evaluation/development kits are reasonably priced.

Just my opinion. :)

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Wad'ya say?

+1 TI DSP

"I may make you feel but I can't make you think" - Jethro Tull - Thick As A Brick

"void transmigratus(void) {transmigratus();} // recursio infinitus" - larryvc

"It's much more practical to rely on the processing powers of the real debugger, i.e. the one between the keyboard and chair." - JW wek3

"When you arise in the morning think of what a privilege it is to be alive: to breathe, to think, to enjoy, to love." -  Marcus Aurelius

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A display and interface will be necessary to adjust the hearing aid while it's in use.
A tablet or an open equivalent in the OLinuXino (Allwinner series, LCD, enclosure).
OLinuXino A20 (Olimex)
The DSP could be done in software maybe with some Xiph (freedom-and-open) software to build upon; for a demonstration:
Xiph, Digital Show and Tell, Episode 2 (video link, transcript)

An alternate way using OTC hardware is to use Android and Android's interface to DSP hardware.
KitKat Feature Spotlight: Audio Tunneling To DSP Dramatically Reduces Battery Consumption When Playing Audio by David Ruddock (Android Police)
Or, Android Open Accessory from an OTC tablet to your DSP.

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I went for a hearing test and they wanted about $10K for a hearing aid :shock: I found something cheaper but I'm debating if I should get a mono version for my right ear or go for broke and get a stereo version.

Attachment(s): 

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hey Joe,I just did a google search of "Texas Instruments DSP and Equalizer" and wow the hits it returned are impressive. Plenty of already done complete code to get you going etc.

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jgmdesign wrote:
Hey Joe,I just did a google search of "Texas Instruments DSP and Equalizer" and wow the hits it returned are impressive. Plenty of already done complete code to get you going etc.
Already done sounds good to me - I'll check that out tomorrow.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Mr Samperi,

Does the stereo version come with a graphic equaliser option?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Joe,

I appreciate what you are trying to do and respect your knowledge, but the latest in hearing assist tech is pretty high level - it is literally some of the most impressive technology I have ever seen. I am a sign-language interpreter with a 2 year electronics degree who has done a great deal of research on both hearing aids and cochlear implants. I do not have your knowledge level in the field of electronics but I would say if you truly need hearing assistance you should embrace what is available...

There are a great number of options available that are low cost - here is one from Bass Pro!! This is less than two hundred dollars.

Still, I doubt there is much margin (profit) in even the $5000 technology you describe above - the market drives all these prices (as I am sure you are aware).

Here is a higher end version from Bass Pro (yes - BASS PRO) and when the big box stores are charging this much then the cost is there somewhere...notice the detailed specs!! Imagine what the $5000 version you are referencing can do! (also consider the software developed to adjust the device, the specialists who spent years in school learning how to assess your loss and adjust the tech, the one-off construction and custom molding of the device itself, etc etc etc)

(Disclaimer: I do not recommend Bass Pro for hearing aids!!)

Just some guy

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

smileymicros wrote:
Its been a decade or so since last I looked at audio dsp and I'm willing to bet that there is now a cheap easy solution that will give me multi-channel equalization and various special filters to handle things like echos in a room with bad acoustics.
ASF, 3rd Party; search for equalizer, DSP.
I thought there's also an equalizer for AVR32 UC3 but I couldn't locate it.
SAM4, SAM G, and AVR32 UC3 have DSP code from or through Atmel.
Some of these have an I2S interface to/from a hardware codec; some software codecs will or may be functional on these MCUs.
A possible power source for these is one or two coin cells or a small lithium-ion cell; digital microphones, analog earphones.
ADC and DAC (or ABDAC) may be good enough without an external codec.

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

johnrk wrote:
... the specialists who spent years in school learning how to assess your loss and adjust the tech, ...
I know a masters degreed audiologist; they have a passion to help the hard-of-hearing.

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Weird - I submitted a post earlier today and saw it and it is gone now?

Anyway I said something about knowing that those guys also have to eat, but for $5000... well maybe it is worth it, but I'm more 'say what' than 'can't hear sh*t' so I'd like to build a device that will let me experiment with equalization and various audio DSP filters to see what makes a difference to my actual ears. There are times when I'm in certain types of rooms listening to a lecture and just can't make out whats being said mostly due to boomy acoustics, but I have no real trouble talking to the person sitting next to me. Also, I can hear the TV okay but often can't discriminate what they are saying especially if it isn't a familiar accent or they are mumbling.

Thanks for all the information.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Equalizer -
Rockbox digital music player (free music player firmware) contains an equalizer and headphone crossfeed.
Rockbox, SanDisk Sansa c200 v2, Manual, Sound Settings
Some digital audio players are ARMv5 (dual ARM7 or an ARM9), SDRAM, flash, maybe a microSD socket.
Some have microphone inputs.
Some may be hackable.
Note the Mini2440 port (ARM9): http://www.rockbox.org/wiki/Mini2440Port
ARM9 appears to be useful therefore SAM9 or SAMA5D3; some low price commercial and industrial boards/SOMs/COMs are available.
SAMA5D3 Software Package contains some audio-specific drivers;
Release Notes - SAMA5D3x Software Package, 1.3.2
AVR32 UC3 - couldn't find a frequency domain application but there's a time domain (scope) application:
Atmel AVR32831: AVR UC3 C Scope Meter and Filtering Demonstration

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

That sounds like a fun project, but you do know that you can 'test drive' hearing aids. You have something like a month to decide if your choice will be acceptable. At least that's what they told me. I ended up going with them, so I never tested the theory.

However, another old coot that I know did just that. He tried several pairs before he made up his mind. .

What I probably will buy, if I ever get over my terminal cheapness, is a controller. Mine have the option of a wireless controller to adjust SOME of the parameters. Currently, when I do want to change something I have to fumble at the nano sized buttons. At least the pair is rf linked so I only have to adjust one.

hj

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My dad resisted the hearing aids for years pissing all of us off. Made going to the theatre miserable, and the tv in the house was at maximum volume.

I finally had enough and purchased him a Sennheiser IR headset similar to this:
http://en-us.sennheiser.com/wire...

The transmittier sits on top of the TV and the receiver headset runs for days on a rechargable "super cell"(It's actually two super capacitors soldered together in parallel). POOF! The tv volume went down to normal levels as the transmitter gets it's feed from the cable boxes audio out jacks thereby letting mom adjust the regular tv volume.

THe nice thing about these is that all the theatres in NYC use the same IR transmission - it's some sort of industry standard so Dad can take them with him to use there. No more irritating everybody

So, now dad has a couple pairs of the headphones and about two years ago he got a real nice set of in-ear hearing aids. You really cannot see them unless you look too. So with the aids and the headsets he is able to hear just about everything except high frequencies, like some female voices which he says can be a blessing.

Expensive, yes, worth every penny? Absolutely.

Edit:poor spelling

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think I'll get started looking at Rockbox and other similar programs to see what I can do on a PC.

I've tried a few apps on Android and the only one that seems to work is very limited in what it will do and it creates a noticeable lag between lip movements and the sound produced. That lag is unacceptable since I find myself somewhat lip reading while folks talk and getting that out of sync just confuses things more.

Next I'll investigate what is available for a Raspberry PI.

And at the moment it looks like I might ultimately to go with the TI PurePath Studio software and the TLV320DAC3120EVM-U Evaluation Module. That EVM is mono and so am I so it seems a good match.

I'll do a bit more digging before making my final decision and report back then. This will probably take a long time to really get rolling.

Thanks for all the input.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jgmdesign wrote:
I finally had enough and purchased him a Sennheiser IR headset similar to this:
I already use headphones and an amplifier for the TV, but I think my real problem is not volume so much as my messed up frequency response. Thats one of the reasons I'd like some sort of audio analysis system so that I can experiment with all the commonly available parameters.

And yeah, I'll probably eventually get some real hearing aids. And since I am deaf in one ear, I guess I'll only need one so the cost should be less.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

After years of folks always bitchin that my PA mixes were 'too loud' I can still hear 15KHz thru a sig gen and a tweeter. I think normal hearing loss is a notch that starts getting deeper and wider at 3.5KHz, so the obvious compensation is just boost the 3.5KHz sliders on a graphic equalizer on a stereo playing a record that you liked 20 years ago, and see if some of the twinkle comes back on the guitar leads. Piano and violin have lots of highs that define the tone. I think Audiologists are just interested in the 300-3K speech range. Maybe they just throw in the towel when trying to get hi fi flat freq resp back for music listening.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Bob may have the simpler, yet not so cool idea on measuring your frequency response. Function generator, amplifier and speaker. Just roll through the frequency ranges and determine where the sound becomes noticeably lower.

The TIDSP is pretty cool though.

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In case you didn't want to turn it into an engineering project (and really why wouldn't you!), there is a company called "Blamey Saunders Hears"blameysaunders.com.au that have some pretty impressive hearing aids that are user adjustable for a comparatively low price.
There was a review in Silicon Chip magazine July 2011 and again in March 2013 for some newer models.

Steve.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you want a quick idea of what equalization might be able to do for you, try Audacity. Record some sounds of the type you're having difficulty with. Audacity has an equalizer that lets you adjust the spectral content to just about any precision you want. In addition you can control expansion and compression of the sound which seems to be a big feature of expensive hearing aids.

Audacity's free and quite capable.

Greg

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

Weird - I submitted a post earlier today and saw it and it is gone now?

There's something "odd" going on at Freaks - changes/posts/edits keep disappearing. I have a feeling that it's something like a process that detects database corruption so winds back to a slightly earlier copy and misses the most recent changes or something.

Wonder if this is part of the "Change is coming" process or it's just the old beast in its final death throes?

(sorry for off topic!)

On the subject of DSP it's kind of the game I'm in too these days only we use them for heavy processing of video rather than audio. The problem you face with the most capable devices is that they are physically huge. They tend to come in 300-600 pin BGAs and packages are as much as 1" square which won't lend itself for being poked into your lug-hole though I guess one could implement the old style of hearing aid with a small box in a top shirt pocket? So I guess it's not just a powerful DSP capable of audio processing you require but something that's available in an extremely small package? That might actually be the major challenge here. With a large budget it might be worth getting one or more examples of commercial units like John suggests then break them open to see if there's a possibility of re-use of their electronics or at least to find the kind of devices they are using.

Quote:

I already use headphones and an amplifier for the TV, but I think my real problem is not volume so much as my messed up frequency response. Thats one of the reasons I'd like some sort of audio analysis system so that I can experiment with all the commonly available parameters.

Could you not then run some experiments with commercial audio processing equipment and headphones to profile the range/changes to audio spectrum you need then approach an existing hearing aid producer with those specs. and see if they already have something to match or maybe they could produce an adapted firmware for your requirements.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Elektor have a project this month using an Analog Devices ADAU1701 in a LQFP48 package. AD appear to supply a drag-and-drop programming environment for it.

#1 This forum helps those that help themselves

#2 All grounds are not created equal

#3 How have you proved that your chip is running at xxMHz?

#4 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand." - Heater's ex-boss

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'll check out Audacity and the Elektor project.

And as Cliff points out, the really capable systems won't fit in the ear. But I'm not sure I need an in the ear hearing aid at the moment since I do okay in face to face conversations. It is just when there is some distance in a room. I was in a meeting yesterday with 35 folks in a large room and used my TV amplifier headphones and did okay, but I am very aware of the boominess that gets amplified. I've heard DSP filters that add boominess to simulate auditorium spaces and it seems to me that the reverse could be done and I'd bet that there are DSP functions in software freely available that do this.

I think what I need to do now is find an affordable box that attaches to a laptop and has software that lets me test standard DSP audio functions and frequency equalizer settings - basically I'd like to test everything that is available (and affordable) primarily because I'm interested in what these changes will sound like to me. I can optimize the settings on the box and them use them in situations like the classroom. Later if my hearing gets worse, I can test hearing aids and have knowledge of what can actually be done so I'll know if I'm getting a song and dance or being told the truth.

So anybody know of one of these magic boxes? At the moment I'm thinking about getting one of the TI boards and playing with it, though I wouldn't mind paying more for something easier to use.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

and it seems to me that the reverse could be done and I'd bet that there are DSP functions in software freely available that do this.

Isn't that just a high-pass filter?
Quote:

So anybody know of one of these magic boxes?

Isn't this just a "graphic equalizer"?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think the algorithms are much more complex than 'high-pass filter'. They have to produce or kill echos, differentiate between what is probably background noise and what is being attended to, that sort of thing.

In addition to frequency these things deal with separating out signals by loudness across a frequency range - like being able to figure out the difference in someone talking directly to you and folks chatting at a distance in the background. And repeated signals like fans or echos that sort of thing.

I think they would more accurately be referred to as psycho-acoustic algorithms - but that is just a guess. Maybe in a few months I'll have enough knowledge to talk about it.

Though various adjustable band-pass filters would probably go a long way to help with my problem.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Echo cancellation has to do with long transmission lines. I've never heard that in a hearing aid context. Nor do I think there is any way to have a hearing aid sort out one conversation from another in a 'cocktail party'. Winamp has a graphic eq. Put on a song and the headphones and turn each fader up n down. 'Boilermaker's Disease' is a notch at 3.5 KHz. Boost this. Sound better?

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DSP audio echo cancellation is old news:
http://en.wikipedia.org/wiki/Ech...

Cocktail party - just point one mic at the close speaker and another 90 degree into the room and subtract that signal from the first. Their are DSP algorithms that do similar things with as single mic input.

You might skim:
http://en.wikipedia.org/wiki/Aud...
and
http://en.wikipedia.org/wiki/Psy...

Last Edited: Sat. Jan 18, 2014 - 02:47 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Not meaning to belittle the seriousness of the topic (I have just done the online test suggested by Steve), but thought that this offering might be amusing...

Morris an 82 year-old man, went to the doctor to get a physical.

A few days later, the doctor saw Morris walking down the street with a gorgeous young woman on his arm.
A couple of days later, the doctor spoke to Morris and said, 'You're really doing great, aren't you?'
Morris replied, 'Just doing what you said, Doc: 'Get a hot mamma and be cheerful.''

The doctor said, 'I didn't say that ... I said, 'You've got a heart murmur; be careful.'

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well Joe, I'm not always wrong about everything. But evidently telling you how to straighten out your auditory frequency response because I know something about audio was a mistake.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hey Bob:

There's enough sour grapes in your last post to make you ill. Think before posting.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks so much Rick. Ask cliff to show you the nice reply that Joe took back. I can Take It and I can Dish It Out. Bring it on (G Bush). The part I'm interested in is Loudness Perception when doing a selfie freq response. Third octave pink noise is probably a much better test signal than a sinewave from a generator, but lots of old silverback engineers have a sig gen. Only obstreperous old sound men like me have pink noise generators and 31 band equalizers. You cant measure your own hearing acuity. You start out with a nice loud signal at 1KHz and start turning it up. Yep! I can still hear it at 10K! Still hear it at 12K! Still hear it at 14.5K! But at that point its 30 dB down. If you bought speakers that claimed to go to 14.5KHz but didn't tell you they were 30 dB down at that freq, you'd punch him. The audiologists put you in a sound proof room and play a tone so soft it is only about 10dB above detectability, so when it drops 10dB, its gone. Zilch. Nada.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Since this Thread is about hearing aids and audio signal processing, anyone have any links to Bluetooth earwigs?

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I havent bothered to look, but i'd lay money that there is an iphone/android app that can take the audio, shove it through an equaliser and shove it out the headphones.

If i forget my glasses, i take a photo and enlarge it on the screen of the iphone so i can read. I felt old the other day when i wanted to purchase some non glycol based coolant and i needed to read the chemical list on a bottle of motul coolant. The young staff at the shop were probably shaking their heads at the old coot doing this. They also had little idea of what i was wanting.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello again JP. You have an iphone? I have a version 1.6 android phone with no data plan.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
I havent bothered to look, but i'd lay money that there is an iphone/android app that can take the audio, shove it through an equaliser and shove it out the headphones.
I'm using a Galaxy Note 3 (Android) and I've tried a couple of apps, but they all have a serious delay that messes up lip sync to the adjusted sound - and lip motion is important to me in understanding what is being said. Also they don't provide enough volume increase to make much of a difference anyway.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

smileymicros wrote:
... and various special filters to handle things like echos in a room with bad acoustics.
WM5102 (Wolfson Micro) codec states:
"The programmable DSP is optimised for Wolfson's wideband voice processing which includes multi-mic noise reduction and echo cancellation."
Linux device driver in kernel 3.6 and sub.
137 ball 0.4mm pitch W-CSP (not easy).
Some of Atmel's 32-bit MCUs have an I2S interface.
I2S may be on some Atmel Xplained boards.
There's a touch screen LCD for Xplained.

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

gchapman wrote:
smileymicros wrote:
... and various special filters to handle things like echos in a room with bad acoustics.
WM5102 (Wolfson Micro) codec states:
"The programmable DSP is optimised for Wolfson's wideband voice processing which includes multi-mic noise reduction and echo cancellation."
Linux device driver in kernel 3.6 and sub.
137 ball 0.4mm pitch W-CSP (not easy).
Some of Atmel's 32-bit MCUs have an I2S interface.
I2S may be on some Atmel Xplained boards.
There's a touch screen LCD for Xplained.
Great find! On their website they state:
Quote:
Ez2 hear™ features Ambient Noise Cancellation (ANC) technology which dramatically removes both stationary and non-stationary background noise whilst still maintaining voice quality. It also features high performance Acoustic Echo Cancellation (AEC) and an Echo Suppression engine, as well as Wolfson’s unique receive path (Rx) noise reduction, Rx Automatic Gain Control (AGC) and multi-band equalisation, ensuring the best possible two-way experience.
This is exactly what I was talking about. Now to find something like an evaluation board for such a system. I'm still leaning toward the TI products if I can find something that does what the Wolfson does via open source software on their chips, but if Wolfson has something to tinker with that is affordable I may go with them. At least I'm seeing some attractive options for playing around with this stuff.

Thanks,
Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

This video provides an example of background noise cancellation - the sort of thing that would help in a Cocktail party or as in the example, a large noisy hall:
http://www.voicegurus.com/high-q...

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

smileymicros wrote:
I'm still leaning toward the TI products if I can find something that does what the Wolfson does via open source software on their chips, ...
TLV320AIC3256EVM-U Evaluation Module
An introduction video on that page.
Some of TI's miniDSP codecs come in QFN packages.
Don't know how open TI's software is; I didn't open the two ZIP files.
That may not be a concern; there may be open software for the driver that uses a generated closed binary blob.
ti-codecs (GITORIOUS, ASoC Codec drivers for TI devices.)
"The miniDSP cores are fully software controlled. Target miniDSP algorithms, such as active noise cancellation, acoustic echo cancellation or advanced DSP filtering are loaded into the device after power-up."
http://www.ti.com/product/tlv320aic3256

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dalpilot wrote:
Audacity's free and quite capable.
A lot of audio applications run on Linux.
Another one is PulseAudio (freedesktop.org); it has a number of loadable modules (ref.1) and uses ALSA (Advanced Linux Sound Architecture).
ASoC (ALSA System on Chip) can be used to interface with an application processor's built-in audio (codec, codec interface, audio DAC, etc.).
Note that Atmel SAM9 has some ASoC support.
A lot of ARM9 in ASoC.
There is some open hardware for ARM9 and Cortex-A.
Ref.
1. module-echo-cancel (PulseAudio)

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yeah, that looks like a really good place to get started. TI at least want's my businiess...

I got all excited about the Wolfson WM8280 based on some videos I saw of devices it is being used in but I couldn't find the data sheet or an evaluation board on the Wolfson site so I sent them an email. And I'm going to show the response here even though it means I'll never be able to visit Scotland for fear of winding up in some dank gaol because I violated this part of the email:

Quote:
-------------------------------------------------
This message and any attachments contain privileged and confidential information that is intended solely for the person(s) to whom it is addressed. If you are not an intended recipient you must not: read; copy; distribute; discuss; take any action in or make any reliance upon the contents of this message; nor open or read any attachment. If you have received this message in error, please notify us as soon as possible on the following telephone number and destroy this message including any attachments. Thank you.
-------------------------------------------------
Oh, gee I've done it now...

What I sent them was this highly provocative query:

Quote:
I couldn't find the data sheet for the WM8280 product nor could I find
and evaluation board. Can you tell me where I can locate these and if
they are not available could you provide me with a link to a similar
product that does have a data sheet and evaluation board?

Thanks,
Joe

And I received this response:
Quote:
Hi Joe
Wolfson is not currently able to support our DSP enabled audio SoC portfolio in the hobbyist market. These products are currently targeted only toward our OEM customers
Thank you for your interest.

Please note how clear it was from my request that I am a hobbyist.

And it would be freaking wonderful if I come up with an application that sells a million so I can tell the folks at Wolfson that they really messed up. I'm envisioning a Kickstarter page where I request 10k and get 10M in funding then right after my Wired interview I give the folks at Wolfson a call and tell them to kiss my Southern Fried You Know What.

Yup, TI it is. At least being Texans they'll know what my 'Southern Fried You Know What' is.

And speaking of you know what? This has been one bitch of a day with two major personal cockups and this little cherry on the top of the whole pile just made my day. I think I'll go sit on my pity pot. Please somebody express some sympathy for me as I mope and whine. Maybe I'll go out and have tea with the Chickens - they at least don't mind that I'm a not obviously an OEM - that is if I have mealworms to toss them while I drink my tea.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

smileymicros wrote:
Oh, gee I've done it now...
Oh, that's OK.
Any room on the bench or do I have to sit on the floor of this dank gaol?
smileymicros wrote:
Please somebody express some sympathy for me as I mope and whine. Maybe I'll go out and have tea with the Chickens - ...
Nope. None given.
Though your self prescription is excellent.
An advice is to go to ground; Mother Earth and her beings will aid.
Just don't Get Low (2009, IMDb).

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think you should drive down to the local Guitar Store like Sam Ash or Guitar Center. There will be a bunch of cool helpful employees with sleeves and piercings, but they are probably all great musicians. Get a pair of headphones and a used graphic equalizer. Play Around. Learn what 1KHz sounds like. Learn what 4KHz sounds like. It will kill an afternoon, then you can go to the local Blues Club and trade notes with the PA dude. He's probably a guy like me that's been running sound for thirty or so years and he'll dig talkin about dBs and MegaMortems ans SPLs. Its all good.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

gchapman wrote:
An advice is to go to ground; Mother Earth and her beings will aid.

The chickens were very sympathetic. They totally agreed that the DSP based echo and noise cancellation along with algorithms designed for specific situations like in that background noise cancellation video I showed them is exactly what I needed. And they think that your suggestion for the TI EVM along with PurePath is a great tool for my further audio studies. Though they then followed up with, 'and give me some mealworms - now damnit!'

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

smileymicros wrote:
And they think that your suggestion for the TI EVM along with PurePath is a great tool for my further audio studies.
Hope that works out for proof-of-concept.
If the goal for your alpha test is as small as possible on a coin cell or two then that (miniDSP codec + MCU) is one way.
A concern is ease of tuning; advantage to an ARM9 or Cortex-A.
Using Linux, or real-time extensions to Linux, and ALSA is handy.
The following relays audio (source to sink) with optional filtering and display:
GTK-Bounce (xiph.org, wiki)
An excellent video is linked on that page.
Another concern is recording and playback; I2S bit rate is significant.
Another way to ease tuning is to use the DSP that's in some application processors.
Though the following is dated (recent additions to Android) it's an introduction to another way to DSP:
Understanding Android's strengths and weaknesses by Juan Gonzales, Darren Etheridge, and Niclas Anderberg (Texas Instruments) (embedded.com; September 29, 2011; search for "Digital signal processing")
"Visualizer showing loudness enhancer audio effect
Visualization of how the LoudnessEnhancer effect can make speech content more audible."
Android KitKat (search for DSP then scroll down a bit)
"Audio preprocessing

Developers can apply preprocessing effects to audio being recorded, such as to apply noise suppression for improving speech recording quality, echo cancellation for acoustic echo, and auto gain control for audio with inconsistent volume levels. Apps that require high quality and clean audio recording will benefit from these preprocessors."
Android Jelly Bean (select Android 4.1)
"Low-latency audio"
Android Jelly Bean (select Android 4.2)
A20-OLinuXino-MICRO wiki, Android (Olimex, currently Android 4.2.2)

smileymicros wrote:
Though they then followed up with, 'and give me some mealworms - now damnit!'
:D
Scary chickens!

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Again - thank you, you've given me a lot to look at.

I'm sort of convinced at the moment that even an ARM won't do the real time processing needed to run the various filters I'd like to test, so unless I learn other wise I'll probably stick with the TI product and do the testing with PurePath. If I find that my personal experience with Android apps isn't necessarily the only thing they can do and that they can in fact filter out background noise while maintaining real time sync with a persons lips then that will be the way to go, but I'm not hopeful at the moment. Maybe I'll learn better from one of the links you've given me.

I'm thinking that if I can put together a system for $100 that let's one experiment with all these audio filters and algorithms that there would be a lot of techie old-farts who would also be interested so I might do an article on it - eventually...

And you have really been a help and I hope you know how much I appreciate it.
Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
filter out background noise while maintaining real time sync with a persons lips

From a standpoint of near total ignorance, I woud suggest that real time sync is always going to be a problem, not necessarily because of the processor speed, but because most(all?) digital filtering techniques require a certain buffer size for historical samples. A bit like the delays encountered in digital radio or TV reception, which are an unavoidable result of the error correction techniques.

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The threshold of perceiving an 'echo' is 10s of ms. If you can understand someone talking 30 feet from you, you can tolerate the 30ms delay involved.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John_A_Brown wrote:
Quote:
filter out background noise while maintaining real time sync with a persons lips

From a standpoint of near total ignorance, I woud suggest that real time sync is always going to be a problem, not necessarily because of the processor speed, but because most(all?) digital filtering techniques require a certain buffer size for historical samples. A bit like the delays encountered in digital radio or TV reception, which are an unavoidable result of the error correction techniques.
Well there is real time and then real time. I think that the algorithms running on an Android or Linux are probably fine for recording sound with various filters applied, but since these OS will randomly stop to garbage collect or send your passwords to Nigeria and you can't control that the delay is noticeable. Where as the dedicated DSPs have special features that do nothing but manage fast memory, spin the algorithms, and get the signal through with highest priority. IIRC these dedicated DSPs have fast RAM large enough to hold the samples right adjacent to the logic that processes the data. I'm guessing that on these devices that it is fast enough since they are being used for hearing applications and I've yet to hear a lip sync problem. But I'm just starting to learn so I don't know yet.

Smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

and I've yet to hear a lip sync problem.

In most AV systems the audio and video streams of data (that arrive independently on different PIDs in an MPEG2 data stream) have presentation time stamps so after any processing of the audio or the video or both they can still be delivered to the user in sync simply by synchronising PTS at the final moment of play-out. (except when the DTV firmware throws a wobbly and this fails and 10,000 people phone up to say that Audrey's voice in Coronation Street did not match her mouth movements ;-))

The same is true in RTP supporting V(ideo)oIP over TCP/IP.

Pages