Reference no: EM132349644
Sound Processing Assignment (Matlab) -
Introduction - Fig. 1 (adapted from Feng et al. [1]) illustrates the performance differences of the various front-end algorithms (sound processing strategies) utilised in cochlear implant speech processors - that is, the translation from a captured sound to electrical impulses delivered via metal electrodes placed within the cochlea.
In the laboratory in Week 2, formant-based sound processing strategies will be explored. In Fig. 1, these strategies are represented by the F0F2 and F0F1F2 strategies utilised in the 1980s. In those strategies, the electrodes closest to formants 1 and 2 were stimulated according to the amplitude of the relevant formant. The stimulation rate was at F0, however note that for this assignment we will keep the stimulation rate fixed.
In subsequent strategies, the features of the sound were reassessed with a focus on finding the most important components of the sound as opposed to its dominating frequencies. This latter approach made the difference between cochlear implants being essentially an aid to lip-reading and finding their place in human achievement as a means of truly restoring hearing. Practically speaking, the transition from the formant-based processing strategies to the spectral peak strategies allowed cochlear implant recipients to (among other important things) regularly converse on the telephone.
For those of us who can hear well, it may be useful in relating this to our own experience by considering the challenges of learning a new language. In the early stages of learning, one might expect to detect the occasional word in a new language - akin to the 3M/House single electrode implants of the 1970s. With a bit of practice, some - perhaps several - very basic words can be understood or recognised, but the full scope of sentences using these words are only occasionally understood because the missing parts of the sentence have not yet been learned. This is akin to the state of the art in the 1980s with the formant-based strategies, where only a few words were able to be successfully represented in the cochlea using electrical pulses.
When the spectral peak approaches came as we entered the 90s, it was as if overnight the learning phase of the new language was almost over, and suddenly more than 75% of everything became understandable.
Notice that for the past two decades though, the advances have become small and, in some cases, less successful than the preceding strategies. For example, in the device manufactured by Cochlear Ltd, ACE is still the strategy used today as it gives many people almost 100% speech intelligibility in quiet conditions. There is still significant amount of work to be done in enabling cochlear implant recipients to hear well in noisy conditions and achieve better pitch perception. Perhaps your role in the future will be to deliver that remaining 25% of sound that will make the difference between being able to understand, to being fluent in the language of hearing.
Each of you through the associated lectures, the laboratory and this assignment, will study and implement into practice, three sound processing strategies:
F0F1F2, SPEAK, and CIS
Note that the latter two are similar "N of M" strategies that vary primarily in their presentation rather than processing, but they have a separate history, and are both essential components of modern cochlear implant therapy. Similarities and differences should be noted carefully and described in the various reporting steps described below.
Expectations -
The Major Assignment is in three parts. Parts 1 and 2 relate to assessment of your individual performance, and Part 3 relates to assessment of the performance of the group to which you will be assigned.
Part 1: Unique and individual MATLAB code written by you to implement the sound processing strategies listed above.
Part 2: An original, individually written report describing the operation and performance of your code, as well as describing the methodology, results and conclusions from your group study in your own words. The main aim will be to compare the performance of the three sound processing strategies. Note, while the results of the study in your report will be the same as each of your group members, it is imperative that the discussion and conclusions reflect your own thoughts and in your own words.
Part 3: A group presentation that describes your group study and results.
Note - See attached files for more details for each of the parts.
Attachment:- Sound Processing Assignment Files.rar