4. Goals of Amplification
1. To make important sounds audible and
Channels comfortable by providing:
ChannelFreeTM – frequency specific amplification
Benefits of no Channels – level specific amplification
Benefits of Speed
Summary
2. Maximize speech understanding
3. Provide high sound quality
5. Single Channel
• First instruments were single channel with
some frequency shaping
Channels • Good sound quality, but reduced flexibility
to fit varying dynamic ranges and non-
ChannelFreeTM
linear targets
Benefits of no Channels
Benefits of Speed
Summary
6. Multiple Channels
Channels • Multiple channels were developed that
ChannelFreeTM
were finally able to achieve frequency and
Benefits of no Channels
level-dependent amplification
Benefits of Speed • BUT these systems didn’t always sound
Summary
so good
7. Multiple Channels
• Why didn’t it sound so good?
• Gain in each channel was varied based on the
frequency and intensity of the signal
Channels
• With fast time constants, the contrast between the
ChannelFreeTM peaks and valleys of the signal are reduced (i.e. it
Benefits of no Channels
spectrally smears)
Benefits of Speed
90 Key
Summary
80 Original
vowel /ee/
Output (dB SPL)
70
F1
F3
60
Amplified
F2 with many
50 channel
system
40
500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000
Frequency (Hz)
8. Multiple Channels
• So the engineers thought again and
found...
Channels
ChannelFreeTM
• It was possible to achieve less signal
degradation via two mechanisms:
Benefits of no Channels
Benefits of Speed
Summary Slower time constants in each
channel
Overlapping channels
9. Slow time constants &
multiple channels
Channels
Speed of compression
ChannelFreeTM
fast Equal sound
Benefits of no Channels
quality
Benefits of Speed
Summary
slow
low high
Number of channels
10. Slow time constants - Drawbacks
• Loss of audibility of soft consonants
Channels
• Errors of undershoot and overshoot
ChannelFreeTM
Benefits of no Channels
Benefits of Speed
Summary
11. Overlapping Channels
• Channels no longer act independently
Channels
• They take into account more of what is
ChannelFreeTM happening in their adjacent channels
Benefits of no Channels
Benefits of Speed
Summary
Intensity (dB SPL)
Frequency (Hz)
12. Overlapping Channels -
Drawbacks
Channels • What you give with one hand you take
ChannelFreeTM
with the other:
Benefits of no Channels
– In other words, what you gain in sound
quality you lose with adjustment flexibility
Benefits of Speed
Summary
13. Multi-channels: The Report Card
• Today’s multi-channel amplification has overcome
problems but has introduced compromises:
Channels
ChannelFreeTM – Single channel sounded good but poor fitting
Benefits of no Channels
flexibility
Benefits of Speed
– Multi-channel fast had good fitting flexibility
but sounded bad
Summary
– Multi-channel slow sounded better but issues
of undershoot and overshoot introduced
– Multi-channel overlapping sounded better but
issues of fitting flexibility introduced
14. Multi-channels: The Report Card
Channels
ChannelFreeTM
Benefits of no Channels
Benefits of Speed
Summary
MIX ED
15.
16.
17. What is ChannelFreeTM?
• Bernafon engineers looked at the mixed
report card of multichannel systems
• ... and they thought out of the box!
Channels
ChannelFreeTM
Benefits of no Channels
Benefits of Speed
Summary
Divided and recombined
sound path
• Instead of dividing the signal into many
channels...
18. What is ChannelFreeTM?
• They developed ChannelFreeTM: a unique
Channels
digital signal processing strategy
ChannelFreeTM
Benefits of no Channels
Benefits of Speed
Unbroken ChannelFree
Summary sound path
• Looks simple… but it offers the benefits of
multi-channel amplification and more.........
without compromises
19. ChannelFreeTM
No channels and extremely fast processing
Channels
ChannelFreeTM
Benefits of no Channels
Benefits of Speed
Summary
20. Benefits of no channels
Maximum Speech Understanding
Channels
ChannelFreeTM • With ChannelFreeTM, speech is amplified
Benefits of no Channels
in its entirety therefore the natural
Benefits of Speed
structure is preserved
Summary • Research shows that spectral
characteristics are most important for
speech understanding
– Van Schijndel et al (2001)
– Boothroyd (1996)
21. Benefits of no channels
Maximum Speech Understanding
Channels
ChannelFreeTM
90 Key
Benefits of no Channels
80 Original
vowel /ee/
Output (dB SPL)
Benefits of Speed 70
F1 ChannelFree
F3
Summary 60
Amplified
F2 with many
50 channel
system
40
500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000
Frequency (Hz)
ChannelFreeTM = no spectral smearing
22. Benefits of no channels
Clear, natural sound
• All sounds processed without breaking
Channels
sound into channels
• Natural characteristics of sound
ChannelFreeTM
Benefits of no Channels
maintained as with speech
Benefits of Speed
- e.g. timbre characteristics of music are
Summary maintained
23. ChannelFreeTM
Extremely fast processing
Channels
ChannelFreeTM
Benefits of no Channels
Benefits of Speed
Summary
24. Benefits of fast processing
Maximum Speech Understanding
• Average speech carries ~10 phonemes per
Channels second
ChannelFreeTM
• ChannelFreeTM analyses and makes gain
Benefits of no Channels adjustments 16,000 times per second to
Benefits of Speed capture individual phonemes
Summary
25. Benefits of fast processing
Clear, natural sound
Channels
Using time domain processing, ChannelFreeTM:
ChannelFreeTM
Benefits of no Channels
3. Analyzes and make adjustments on every
Benefits of Speed sample
Summary – Sampling rate is 16,000/second therefore
16,000 analyses and gain
adjustments/second
– Instant response to the environment
4. Has a fast throughput time
26. Benefits of fast processing
Clear, natural sound
Channels
• Research shows that long throughput time
can interfere with sound quality, especially
ChannelFreeTM
for own voice
Benefits of no Channels
• Additional research shows that frequency-
dependent throughput delay reduces
Benefits of Speed
Summary
sound quality
27. Proof
Premium class sound quality:
NAL study
Channels
•Sound quality comparisons of 5 premium class
ChannelFree TM
instruments for different input signals
Benefits of no Channels
Symbio, Diva, Claro, Canta 7 and Triano S
Benefits of Speed •Objective measurements were also made to see
Summary what correlates with subjective preference (low
time delay, low noise)
•Symbio got highest average rating by hearing
impaired subjects for male and female voice, and
music*
•Symbio was significantly preferred for listening in
quiet room (normal hearing)
* not statistically significant
28. ChannelFreeTM – The Report Card
• ChannelFreeTM is a radical departure from
traditional multi-channel amplification. It has
Channels no compromises and therefore delivers:
ChannelFreeTM
Benefits of no Channels
Benefits of Speed
Maximum speech understanding
Summary
Clear, natural sound
Because:
• Symbio XT doesn't split sound into channels
• Symbio XT is extremely fast
29. ChannelFreeTM – The Report Card
Channels
ChannelFreeTM
TI ON
Benefits of no Channels
Benefits of Speed
OLU
Summary
R EV
I n order to have dynamic range compression, it was also necessary to find a solution that would allow both frequency- and level- dependent amplification, to fit input signals into the dynamic range of the hearing impaired person which also tends to vary with frequency.
Multiple channel systems were developed that achieved frequency and level specific amplification by splitting the signal into distinct frequency regions and applying different compression characteristics in each channel.
When the gain in each channel is adjusted independently, and the compression acts quickly, the contrast between the peaks and valleys of the signal are reduced. This can clearly be seen in this diagram which is a representation of the vowel /ee/. The first, second and third formants (F1, F2 and F3 respectively) can clearly be seen in the original signal as peaks in the spectrum (black line). When amplified through a many channel system with fast compression, the resulting signal (pink line) lacks the spectral contrast (peaks and valleys) of the original signal. This effect is known as spectral smearing.
In order to maintain good sound quality, a relationship must be observed between the number of channels and the speed of compression applied in each channel. The line may not be a straight line as depicted here, but there is a general rule of thumb that as the number of channels increases, the speed of the compression should decrease. It is when this rule is violated that spectral smearing can occur and degrade sound quality. As the number of channels increases, the compression speed should be slowed down.
The problems with slowing down the compression speed are: Loss of audibility for soft consonants, if they’re preceded by a louder sound e.g. a vowel (not an uncommon situation in running speech). Undershoot and overshoot. Undershoot is when the gain is reduced for a louder sound which is then followed by a softer sound before the gain has time to recover: it’s more or less the same as loss of audibility for soft consonants. Overshoot is when a soft sound is followed by a sudden loud sound - the system can’t adapt quickly enough to reduce the gain and so too much gain is applied to the louder signal.
Another approach is to increase the amount of overlap between adjacent channels. This means that instead of the gain in each channel being controlled only by the intensity of the frequencies falling within the channel, adjacent frequency regions also influence the gain. Spectral smearing is further reduced by this “coupling” between the channels.
T he drawback of channel coupling is that some of the flexibility that the channels introduced is removed: what you give with one hand by offering so many channels, you take away with the other by coupling them so they act less independently. Slowing down the compression and increasing channel overlap can be used at the same time, and do reduce signal degradation and improve sound quality, but they do not prevent signal degradation.
H ere is an overview of the performance of multichannel systems as ‘premium class’ amplifiers While they perform well in some respects, it’s clear that they demand compromises.
ChannelFree ™ allows frequency- and level-dependent amplification, but does not split the incoming signal into channels or bands to achieve this. Therefore the natural spectral structure of the speech/music/other sounds is preserved. This is particularly important for speech because research has shown that spectral characteristics are more important than other characteristics (intensity and timing) for speech understanding (refer to „More Channels are Better Right?“ article by O‘Brien (www.bernafon.com).
This is the example we saw earlier of the /ee/ vowel, amplified now with two different signal processing strategies. The black line is the original signal, the pink line is a many channel system which has flattened out some of the valleys of the sound, and the blue line is the same sound as processed with ChannelFree ™ . As you can see, it differs from the black line by the frequency-specific amount defined for the hearing loss and input level. However, the spectral contrast is well maintained (and corrected for hearing loss) with ChannelFree ™ , but is largely lost with the multichannel system. For this reason, we state that ChannelFree ™ offers exact reproduction of speech.
How does ChannelFree ™ achieve clear, natural sound quality? Sound is processed without being broken up and divided into channels. This maintains the natural characteristics of all sound, so that the tonal quality /timbre is preserved.
Average speech carries about 10 phonemes/second (which is 3-6 syllables/second). Slower systems can't adjust the gain quickly enough to apply the appropriate gain to individual phonemes. However because ChannelFree not only analyses, but also makes gain adjustments, 16,000 times/second, it’s possible for Symbio XT instruments to capture and amplify individual phonemes.
The processing speed is faster than FFT-based processing (lower throughput delay). All digital systems delay the sound but the lower the delay, the better the sound when an amplified signal mixes with an unamplified signal as is the case whenever an instrument is fit with a vent.
Outline of the Dillon et al (2003) study... Comparison with different inputs: male voice (in quiet and in impulse noise) female voice piano music own voice quiet room Objective measurements were also made to see what correlates with subjective preference and it was found that low throughput delay and low noise were correlated with subjective preference for sound quality. Refer to www.bernafon.com/NALstudy
To summarize, ChannelFree ™ offers two main benefits: Maximum speech understanding Clear, natural sound and there are two main reasons for that: ChannelFree ™ processing doesn't split sound into channels ChannelFree ™ processing is extremely fast
And in contrast to our report card on multichannel amplification, the report card on ChannelFree is: It's a revolution!