 |
|
 |
|
|
|
           
|
COMMUNIKAY Vol. 3, No. 1
Motor Speech Profile IntroducedKay introduced a new software option for the powerful CSL platform at ASHA '95 in Orlando called the Motor Speech Profile (MSP), Model 4341. The new program, which is the result of collaborative efforts between Kay and clinical speech scientists, provides acoustic analysis of motor disordered speech based on a variety of specific tasks elicited from the patient. These results are then quantified and/or graphed to help objectify patient performance. MSP is similar in concept to another CSL option, the Multi-Dimensional Voice Program (MDVP), which objectifies many voicing parameters. MSP is focused specifically on motor speech assessment. Both programs were developed by Dimitar Deliyski, Ph.D., of Kay's engineering department. The evaluation of motor speech disorders is challenging for the speech-language pathologist and neurologist. While perceptual analysis is valuable, there is also a need to objectively evaluate a variety of parameters of the subject's speech to fully assess speech production. These can include, among others, breath support, voicing behavior, nasality, rate of speech, and vowel and consonant production. Each of these can be evaluated objectively to more fully describe the subject's speech. The clinician can supplement the more subjective impressions of overall intelligibility with data that can be measured and monitored over time. Ideally, even subjective evaluation of intelligibility should be rated by a new listener unfamiliar with a patient's speech patterns. MSP provides built-in protocols to analyze separate parameters. Says Deliyski, "The program uses the acoustic parameters most closely associated with a perceptual attribute to objectify speech behavior. For example, distorted vowels are often characterized by neutralized second formant transitions. MSP sets up a task to measure and assess this behavior." In another example, diadochokinetic (DDK) rate and its periodicity have been shown to be closely associated with articulator mobility. A DDK task is also included in the protocol which provides a model followed by the subject's attempt, and then automatic analysis of the production. According to Deliyski, a new feature (not displayed at ASHA) that will ultimately be included in MSP is the analysis and graphic display of long-term frequency and amplitude modulation (e.g., tremor, vibrato, etc.) in the patient's voice. These modulation parameters will also be derived from a simple, protocol-driven elicitation of the task to facilitate sample collection and analysis for the speech clinician. The results from other instrumental procedures (e.g., aerodynamic, nasometric, etc.) can also be entered into the analysis report summary. Kay received thoughtful input from a large number of clinicians and speech scientists during the design of MSP. Though the following list is not comprehensive, we wish to especially acknowledge the valuable contributions from: Klaas Bakker, Ph.D., SW Missouri State University; Frank Boutsen, Ph.D., University of South Alabama; Eugene Buder, Ph.D., University of Memphis; Carla Gress, Ph.D., University of California at San Francisco; and Shimon Sapir, Ph.D., Northwestern University. Please note that suggestions/contributions from these individuals do not imply their endorsement of this or any other Kay product. Researchers Adapting MDVP for Sung VoicePoliticians. Actors. Lawyers. World-class opera singers. Cantors. Music educators. Even 7-year-old singers on cruise ships. Professional voice users all, and all among the scores of patients who routinely visit the Texas Voice Center in Houston. But making bold new strides in the research arena is as much a part of the routine at the Center as is restoring and maintaining the vocal health of their patients. Under the direction of Dr. C. Richard Stasney, clinical associate professor of otolaryngology at Baylor College of Medicine, the Center uses a team approach that successfully interfaces the three professions of vocal pedagogy, laryngology, and voice science/speech pathology. Among the most exciting and ground-breaking work under way is that of Sharon Radionoff, the Center's singing voice specialist and voice technologist. For her dissertation, Radionoff is completing one of the largest studies ever performed in the singing world. She has collected hundreds of voice samples of singers that she has analyzed using Kay's Multi-Dimensional Voice Program.
MDVP extracts reliable and accurate analysis of voicing parameters. The analysis results are graphed against normative thresholds. These thresholds are appropriate for flat-tone sustained phonation of the /a/ vowel. "But here," Garrett says, "when a voice patient comes in, we're doing a spoken /a/, which we then follow by doing the sung /a/. The problem is that we have no normative data on singers." Radionoff's work will redefine these normative standards for singers. "Right now, the MDVP displays analysis results outside of normative thresholds even in the most perfect singing voice," she says. "But if we and other labs are going to be using this instrumentation for the singing production," she says, "it's critical that we have new normative thresholds that we can compare the population of abnormal singers to. We need to know what is normal for a singer." Radionoff tracked 28 voice students at Rice University's Shepherd School of Music during a year of voice study. She collected data on three occasions (i.e., in October, January, and April). On each occasion, she collected phonatory and respiratory data, and a series of five trials for spoken and for sung phonations. All of these samples, multiplied by the 33 parameters extracted by the MDVP, will be used to develop what Dr. Garrett calls, "the ultimate normative database." Radionoff is lavish in her praise of the reliability of the MDVP/CSL instrumentation. "Without reliable instrumentation, all of the numbers that have been collected would mean nothing." Garrett and Radionoff look forward to having optional normative thresholds on the MDVP that would allow the user to "Set for Singer's Values." This would display all of the different parameters, but with each one scaled according to Radionoff's data. Then, they explain, what now appears to be deviation from normal might just be a pattern that naturally occurs when using vibrato, for example. "It will also give singers a feedback mechanism which," Garret claims, "will never replace their voice teachers or the human ear, but which will make their lives easier." "We're at the threshold of a brand new era. It's an exciting time to be in the field," says Garrett.
What is the advantage of the new three-chip CCD camera offered by Kay? For stroboscopy, general endoscopy, and FEES, Kay's standard packages include a single-chip CCD camera. Kay now also offers a three-chip camera, though it is not standard due to its higher price. The major advantage of the three-chip camera technology is that instead of integrating red, green, and blue on a single chip, the three primary colors each have their own separate CCD chip and circuitry. The result is truer color, higher signal-to-noise ratio (56 dB), and higher resolution (700 lines) than are possible with single-chip cameras (46 dB signal-to-noise, and 430 lines of resolution). Some customers will undoubtedly view the improved clarity worth the extra expense. And although the cost difference is still substantial (approximately 70% higher for the three-chip models), it has declined steadily over the last two years, a trend that will likely continue. On the RLS (Rhino-Laryngeal Stroboscope), the audio volume knob [located in upper right section of the front panel] controls the audio microphone signal level. Do adjustments of this knob affect extracted amplitude measurements? No. The audio microphone (attached to the top of the CCD camera and plugged into the strobe unit) has two functions. One is to provide voice amplitude measurements in real time as the patient phonates during the stroboscopic examination. This measurement is completely independent of the audio volume knob's position. The second function is to provide the voice signal to the VCR so you can hear the patient's voice on playback of video recordings. The volume knob controls the audio microphone's signal level being fed to the VCR. Observe the VU meter on the VCR during phonation, and make adjustments so that the meter is at mid-range during comfortable phonation. You may need to make adjustments in extreme cases: for patients with very soft voices (e.g., unilateral vocal fold paresis), turn the knob clockwise to increase output level to the VCR; for very loud voices (e.g., professional singer phonating fortissimo), adjust the knob counterclockwise. How can I import Palatometry data into the CSL to compare physiologic and acoustic data? The Palatometer acquires data from 96 electrodes embedded in a custom-fitted pseudopalate worn in the mouth as well as the acoustic signal (using microphone input). Within the Palatometer software, the speech waveform is time-linked to the linguapalatal contact patterns made during a speech utterance. These two sets of data are stored to a single file (.nsp format) when saved to disk. Users may wish to perform in-depth acoustic analysis (e.g., spectrographic, pitch, amplitude, etc.) concurrently with analysis of the physiologic data provided by the Palatometer. The CSL allows importation of data for this purpose from both Kay's Palatometer as well as the Reading (Hardcastle) Electropalatograph. Within CSL, load the file of interest; the acoustic waveform will load to view A. Now create a new view (preferably square-shaped) to accommodate palatometric data. With the new view active, click Analyze on the Main Menu, then "Show Palatometer Display". The file's palatometric data will appear in this view. You may now create additional views for spectrographic, pitch, and other data; link all views. Cursor positions in all views of data are precisely linked in time to the palatometric data of the spoken utterance. Loading the latest CSL software on 4300A hardware CSL software has been updated on an ongoing basis to add new features and functionality; the more recent releases (5.0 and higher) take advantage of the newer 4300B hardware architecture introduced in August 1993 (see summary in CommuniKay Vol. 1, No.1). However, these later software versions also work with the 4300A hardware. Although a few software features specific to 4300B hardware (e.g., DAT interface, up to four channels of input, etc.) do not function with 4300A hardware, users can take advantage of many other advantages in these later versions. When updating to CSL version 5.0, observe the following points to ensure the program will install and run properly. First, the recommended entry level computer for the newer software is a PC 486 DX. Second, your computer's available conventional (DOS) memory must be at least 565K (less was required in prior versions). Type mem from the root directory to verify. If not, type memmaker (MS-DOS 5.0 or higher required) from the root directory to optimize your conventional memory. If this is still not adequate, manually edit the autoexec.bat file to free up memory. Third, the new HARDWARE.CFG file, which contains the I/O and IRQ default values set at the factory, must be altered if your previous CSL software was set to non-factory settings. Configuration modifications that you may have made originally are not automatically carried over when you install the CSL 5.0 software. Note that any modifications made in your prior CSL software (versions 4.x and earlier) can be viewed in your old CSL.CFG file. Finally, any options you have purchased will also be updated when you upgrade to the CSL 5.0 version. These should be installed along with the core CSL software. Call the factory (ext.160) or your local representative for additional information. Tips for Accurate Signal Acquisition When you capture or import acquired signals into Kay's CSL or Visi-Pitch II, an understanding of how unwanted noise can be mixed with the signal of interest will improve your ability to interpret the analysis of a signal compromised by noise. Noise is generally taken to mean random fluctuations, added to or modulated with, a wanted signal. Here, however, noise means any unwanted signal, periodic or not, existing with a desired signal. For example, if you import a voice signal acquired using a multimedia card, you may unwittingly report voice parameters (e.g., jitter, harmonic/noise ratios) which have been significantly altered by noise added during signal acquisition. In another example, if you connect other components to your CSL during capture (e.g., capturing through a DAT player simultaneously), you may be adding noise without realizing it. Any method of acquiring and storing signals can affect the signal quality. Noise can be introduced during signal acquisition in a number of ways. System components (e.g., microphone, cabling, pre-amplifier, amplifier, anti-aliasing filters and A/D) could be of poor quality. The CSL and Visi-Pitch II use professional-level system components and careful design which minimize system-generated noise. However, if you import signals acquired with other products, you should evaluate their performance. Additionally, noise sources (e.g., fan noise, electromagnetic signals from monitors or fluorescent light fixtures, power supply hum) may inadvertently be acquired along with the signal. Poor room acoustics can also add noise. A digital storage system (e.g., computer acquisition, DAT recorder) converts the incoming continuous analog signal to a discrete digital signal. Converting an analog signal to a digital representation is performed by an analog-to-digital converter (A/D converter). Most systems today use a 16-bit A/D converter which produces a 16-bit binary number to represent the range of incoming values. A binary number with 16 places has a range of 216 or 65,536 possible values. This range of values equals 96 dB of possible signal level variation (20 log10 65536). This is the maximum achievable range, also called the dynamic range, with a 16-bit linear converter. The full 96 dB range, however, is seldom achieved because the associated electronics can rarely take advantage of the full dynamic range available. For example, most sound cards which plug into a computer can lose half of the dynamic range to system noise and large DC offsets. One simple way to perform a partial check of a recording system, without the need for test equipment, is to acquire a signal exactly as you would during your work except with the microphone turned off. Then, with the Visi-Pitch II Waveform Editor or CSL, analyze the acquired signal for noise by scrolling the cursor along the waveform to note the noise level. If you repeat the above test with the microphone on (but quiet), you can separately measure the noise added by microphone pickup of acoustic and electromagnetic signals. For example, if the values range -200 or +200, under either condition, the eight or nine least significant bits (about 50 dB) are corrupted by noise. The true dynamic range of the 16-bit A/D has been reduced from 96 dB to about 46 dB! A useful dynamic range of above 85 dB is recommended by the National Center for Speech and Voice for voice measurements. This means that the noise should fall within the range of -4 to +4 using the tests described above. Please look for Kay products on display at the following conferences,
workshops, and congresses.
Workshops
International Conferences
Add Video-Otoscopy to Your Strobe Are you getting the most out of your Kay stroboscopy system? Components used for stroboscopy (e.g., light source, CCD camera, lens/adaptor, VCR, monitor, printer) can also be used for other endoscopic procedures. An increasingly popular one is video-otoscopy. Otolaryngologists routinely perform otoscopic exams to evaluate the ear canal and tympanic membrane. For physicians who would like a convenient means of sharing their observations with their patients, or for educating residents at teaching hospitals, video-otoscopy provides a large, clear image of the structures of interest on the video monitor for all to see. The physician can point to critical findings in the otoscopic exam and be certain that all are observing the same features. The exam can be video recorded for later viewing (e.g., after treatment) or used for pedagogical purposes at a later time. Hard copy color prints can be placed in the patient's chart or sent to a referring physician. Kay's stroboscopy systems (both the Basic and Computerized) contain all of the components, except the rod otoscope, to perform video-otoscopy. However, Kay now offers a rod otoscope, and others are available from a number of manufacturers. These otoscopes contain a rod lens for transmitting the image and a circumferential fiberoptic bundle which transmits light for structure illumination. Most commercially available devices are easily coupled to the standard lens/adaptor eyepiece used for other rigid and flexible endoscopes. Kay's Computerized System has the added advantage of allowing image digitization for storage in the computer. Left ear versus right ear can be placed side by side on the screen and annotated to clarify specific findings. Remember that components of your stroboscopy system can be shared across a variety of endoscopic applications.
|
| Copyright © 1996-2008 KayPENTAX. All rights reserved. Site Map | Contact Us |
