Sonic Enchant Speech Variable Processing (SVP) Technology:
The Sonic SVP technology is rooted in the Cochlear Amplifier model. The outer hairs of a healthy cochlear helps amplify soft sounds and control the levels of loud sounds. But damage to the outer hair cells affects the cochlear amplifier’s ability to properly amplify sounds, resulting in sensorineural hearing loss (SNHL). SNHL shows significant loss of low intensity sounds, some loss of moderate intensity sounds, and near normal perception for high intensity sounds -> a non-linear loss at varying intensity levels.
Wide dynamic range compression (WDRC) is the hearing aid industry’s solution to SNHL. It amplifies low intensity sounds more than higher intensity sounds. But it has limitations because it doesn’t address some of the other ways the cochlear functions, where 1) the cochlear acts as a very fast (and therefore accurate) signal analyzer and amplifier for incoming sounds, and 2) it processes frequency-specific contents to impart clarity.
All modern digital hearing aids have some sort of WDRC. Sonic is no exception, and SVP is their WDRC system. Conventional WDRC systems can only process audio events in slower time frames due to the lack of computational power needed to analyze incoming signals and adjust gain fast enough, especially for speech signals. With speech, individual phonemes occur on the average of ten times per second. A phoneme is any of the perceptually distinct units of sound that distinguish one word from another, like p, b, d, t in the English words Pad, pat, bad and bat.
The SoundDNA platform developed for the Enchant line claims that it has enough computational power to process speech signal fast enough, analyzing incoming speech signals and adjusting gain thousands of times per second. Other systems are not fast enough to accurately measure rapid differences in the sound intensity of speech. They also commonly have fast attack but slow release times, which typically apply too much or too little gain because they do not react quickly enough to track dramatic changes in speech sound intensity. The end result is homogenous application of gain and under-amplified soft speech sounds. SVP, on the other hand, claims that with its very fast attack and fast release times (thanks to the high computational power), can accurately measure and compensate for sudden changes in sound pressure levels, resulting in very accurately amplified reproduction of the speech signal. This addresses the limitation (1) as mentioned above that plagues many conventional WDRC systems.
If speed is key to speech accuracy, then frequency contrast is key to speech clarity. This is the limitation (2) mentioned above for conventional WDRC systems. Frequency contrast is the difference in intensity across frequencies of an incoming sound. In speech, the frequency contrast between individual phonemes provides clues for listeners to tell different phonemes apart. In a SNHL patient, the ability to amplify correctly at different frequencies is diminished, resulting in a loss of frequency contrast, hence loss of speech clarity.
Most traditional multichannel compression systems (a) split the acoustic signal into separate frequency bands, (b) measure the signal at separate levels for each band, and c) compress each band based on its own prescribed gain setting. But in doing so, they flatten the overall spectrum of the sounds, losing the frequency contrast in the process. SVP takes a different approach by measuring and applying gain to the entire wideband acoustic signal. The signal is never split up, processed and recombined. This helps avoid the deleterious (flattening) summing effect of applying gain in multiple bands that traditional multichannel compression systems employ. The result is that the frequency contrast is preserved by SVP, and with it, speech clarity (clarity of individual phonemes is preserved). Of special importance as well is the vowel identification performance for hearing impaired people that is very dependent on the peak-to-valley ratio preserved by this frequency contrast.
SVP as a WDRC amplification strategy was started by Sonic in 2011 and maintained until 2014. Its strategy of using fast time constants (i.e. fast compression) to achieve phoneme-level clarity is originally called SVP, then later dubbed “Phoneme Focus”. Fast compression allows a system to rapidly apply gain to low-level phonemes (consonants) and then immediately reduce gain for high-level phonemes (vowels) that follow.
Although successful with a majority of patients, believe it or not, the Phoneme focus is not for everyone. Sonic has found that if the patient is over 75 years old, has a severe to profound hearing loss, or is starting to display cognitive issues with memory or attention, then they may be losing their ability to use TFS.
TFS is Temporal Fine Structure, carrier waves that contain useful information for sound identification because its rapidly varying oscillations carry details about the signal’s fundamental frequency and short-term spectrum. Without going into too much details of TFS, it’s basically what the Phenome Focus strategy (as explained above) was designed to go after (fast signal analysis and amplification for speech accuracy, and frequency contrast preservation for speech clarity).
For the patient profile described above who aren’t suited for the Phoneme Focus strategy, Sonic has developed an alternative strategy they call Envelope Focus (since 2015 to date). It basically uses slow time constants (slow compression) which adjusts amplification based on the long term average changes of the auditory environment. Concerning speech, gain does not change from one phoneme to the next. Instead, the natural peak-to-peak differences between louder vowels and softer consonants are preserved over time, and retain a greater loudness contrast to each other, compared to a faster system.
To date, the Sonic SVP as we know it then, now provides two options for selection in its ExpressFit Pro software. The hearing professional will determine which profile criteria their patient fits into, and select either the Phoneme Focus or the Envelope Focus for that patient accordingly.
In summary, Speech Variable Processing (SVP) is Sonic’s Wide Dynamic Range Compression strategy that’s tailored specifically for speech, originally with a Phoneme Focus strategy, then later with an Envelope Focus strategy.
SVP is the compression strategy that Sonic uses when there’s only speech with no noise (Speech in Quiet). When noise is added to speech, a slew of issues present themselves, like the undesired amplification of noise for one, because SVP does not know better if the signal is speech or noise. To handle compression for speech in noise, Sonic came up with a new compression strategy/technology called Smart Compress, which we’ll explore in a future post here.
Compression is only one side of the equation, the amplification side. Noise reduction is the other side that Sonic addresses with their SPIN (Speech In Noise) noise reduction strategy. Traditional systems treat directionality and noise reduction as two separate ways to improve speech in noise issues. Sonic’s SPIN stragegy combines the operation of directionality and noise reduction together to come up with what they think is a much more effective integrated Speech In Noise solution. That will be the third post in this series of posts covering Sonic Enchant technologies.