Sonar for the visually impaired - Audio ideas

Projects / Project Swiftlet  "Pump up the volume" Original post date: 06/19/2015

I'll share all my notes and my thoughts for the audio aspect of my design here.

This is the frequency response curve of the Knowles 2403-260-00001. It is 11 x 15 x 3.5 mm rectangular speaker. From the frequency response curve, it is useful for part of the voice frequency (300 and 3000 Hz). I would expect similar sized speakers would have similar curves. To generate any volume below 800Hz would require a much larger speaker and a much higher power consumption due to the displacement.

According to wiki: Psychoacoustic model

• Frequency resolution of the ear is 3.6 Hz within the octave of 1000 – 2000 Hz.
• Typically, the ear shows a peak of sensitivity (i.e., its lowest ATH) between 1 - 5 kHz, though the threshold changes with age, with older ears showing decreased sensitivity above 2 kHz.

- Speaker THD is 1% or less at 2kHz

1kHz to 2kHz is the sweet spot for raw sonar playback for sensitivity and frequency resolution. The ultrasonic helmet slows down it ultrasonic samples by 20 times. So if I were to do the same for my device, the 40kHz signals would be playback in the 2kHz spot.

The echo frequency from a moving object shifts due to the Doppler effects and with training, someone might be able to extract that information from the tone change.

The human frequency resolution is +/- 3.6Hz at 2kHz (sample is slow down by a factor of 20), so it should be able to detect +/- 72Hz at 40kHz (original speed).

Calculator from here: http://hyperphysics.phy-astr.gsu.edu/hbase/sound/dopp.html

FYI: walking speed is about 1.23 m/s (4.4 km/h; 2.8 mph)

There is probably around 5kHz bandwidth depending on the ultrasonic transducer, bandpass filter design etc. That would limit the top speed to 20m/s (45mph).

One of the questions I have on my mind is the shortest burst of sound recognizable as a tone and not a click. The tone detection is needed for the brains to extract the frequency shift for doppler effects.

Haven't find anything solid, but seems like I would need something around 100ms or more. That length/20 is the burst of the ultrasonic.

References: 

Neural networks simulating the frequency discrimination of hearing for non-stationary short tone stimuli.

This paper addresses the question of frequency discrimination of hearing

for non-stationary (short) tone stimuli (duration < or = 125 ms). 

Shortening of the stimulus duration leads to widening of the frequency 

spectrum of the tone.

For pulse duration less than 200ms, sensitivity of the signal is dependent on the pulse duration. 

Graph reference from sstackexchange

http://music.stackexchange.com/questions/31981/whats-the-shortest-that-a-note-can-be-and-still-have-a-recognisable-pitch

How long must a tone be heard in order to have an identifiable pitch? Early experiments by Savart (1830) indicated that a sense of pitch develops after only two cycles. Very brief tones are described as "clicks," but as the tones lengthen, the clicks take on a sense of pitch which increases upon further lengthening.
It has been suggested that the dependence of pitch salience on duration follows; sort of "acoustic uncertainty principle" ,
Δf Δt = K,
where Δf is the uncertainty in frequency and Δt is the duration of a tone burst. K which can be as short as 0.1 (Majernik and Kaluzny, 1979), appears to depend upon intensity and amplitude envelope (Ronken, 1971). The actual pitch appears to have little or no dependence upon duration (Doughty and Garner, 1948; Rossing and Houtsma 1986). In this demonstration, we present tones of 300, 1000, and 3000 Hz in bursts of 1, 2 4, 8, 16, 32, 64, and 128 periods. How many periods are necessary to establish a sens of pitch?

http://www.feilding.net/sfuad/musi3012-01/demos/audio/413_pitch_salience_tone_duration.htm

-------------------------------------------
Bought 3 x 2 of these speakers from here at $0.98 a pair with "Free shipping" via registered mail.

Model Number: AX032 Earpiece Speaker for iPhone 4/4G.

• Item size: 23*9*4mm
• Net weight: 1g

Chances are that these requires some mechanical mounting. A possible way is to have the membrane side protrude through a slot on the PCB and sandwiching the back side contacts with a PCB daughter card.

As I scrolled down the Feed (on hackaday), I saw the Project AMIKETO with the red circles around that say "Speakers". I guess I should have checked on Digikey for that! Having a datasheet to work with take some guesswork out of the mechanical design.

The speaker has spring loaded contacts and will probably be one of the parts that needs to be replaced during the life time of the device. For the prototype design, I'll need to come up with a design that sandwich the speaker between 2 PCB with a grill. Looks like I'll need something beyond a headphone amplifier (class D) for driving the speakers and need some volume control.

Someone must have read my complain about how slow Canada Post was... The speakers from China magically show up today as regular parcel. The tracking shows that it arrived at at Vancouver 20 odds days ago and then it went dark. Normally the packages became Canada Post's problem around the time it leaves on a flight from China and doesn't take 20 days or so from the west coast.

(1/4W Resistor for size)

These speakers have a gasket already and they are even smaller 10mm x 7mm x 3.1mm (0.392" x 0.278" x0.125") vs 11mm x 15mm x 3.5mm of the one I looked at on digikey. Smaller speakers means that the resonant frequency will be higher.

Even though the resonant frequency is a few times the voice frequency, I can actually watch and understand what they say on TV because of the higher harmonics.

FYI voice frequency from wiki: Adult male: from 85 to 180 Hz, Adult female: 165 to 255 Hz

The speaker is very inefficient. You can feel it vibrates, but because of the small amount of air displacement, they are too quiet. They are really designed for mobile phone where the speaker needed to be directly next to the ear along with a resonant cavity. I'll need to play with bigger sized speakers.

I really don't to require the user to wear headphones. :(

3W Class D stereo amplifiers for a different project that I ordered a month ago finally arrived. Probably going to use it to test out the iPhone speakers on the bench as they don't exactly comes with datasheet. Also, the volume is not going to be as loud on a bare board than inside a properly design enclosure.

The part is from Diodes Inc. PAM8403 3W Class D Stereo Amplifier I replaced the cap on the lower right corner between L/R with 2uF to eliminate power on pop noise. Mouser seems to be the only retail place that sells this chip for $0.90.

Found a 2.5cm (~1") dia. speaker from a toy. Even though it is a cheap one with ceramic magnet, it works a lot better. Now at least I know a size that would work without a proper acoustic enclosure.

I am looking at oval/rectangular shaped speakers as they have resonant frequencies peaks corresponding to their two axis. The interactions between the two peaks allows for a wider frequency range with the longer axis for the lower frequencies. I sorted the speakers by sound pressure level.

Here is my preliminary speaker mounting location. The speaker is at 22.5 degrees and it sort of pointed at the outer ear like a periscope. This is what I have to make do until I have the basic functionality and ready to tackle a case/acoustic design.