Video recording with a strobocopic camera
Author: MSc Marzena Zacirka, Ph.D. Marcin Just, Ph.D. Michał Tyc (DiagNova Technologies)Kymography is an analysis of the vocal folds functioning by assessing phonation movements over a period of time and in any chosen cross-section of the folds.
This type of examination is performed by appropriately recording the image of the vocal folds from the perpendicular direction using an endoscopic camera. A video recording intended for kymographic analysis should mainly contain vocal folds during phonation. Recording folds while inhaling or exhaling is pointless, since such fold movement does not contain any relevant information in kymographic analysis. Due to the fact that the vocal folds move very quickly during phonation, the recording should contain at least 2,000 frames per second, which is not possible with the use of ordinary cameras coupled with endoscopes. A full two-dimensional image with such time resolution can be obtained using expensive specialized high-speed cameras. An image of a single selected section of the vocal folds with an appropriate temporal resolution can also be obtained using a much cheaper high-speed line camera (containing a set of photosensitive elements in the form of one pixel ruler instead of a full two-dimensional matrix), which is a cheaper solution than a camera for high-speed photos, but still requires additional specialized equipment. An alternative to high-speed cameras that record continuous image is to use a regular camera in the strobe mode. With good quality equipment and properly made recording, the image obtained in this way is usually sufficiently reliable material for performing kymographic analysis, and in relation to the use of a high-speed ruler camera, it is possible to select the cross-section of the folds for examination even after recording.
Performing a satisfactory videostroboscopic recording for kymography requires appropriate training and practice from the user. This is due to the complexity of the method of making such a recording. High-class software allows the process of kymographic analysis to simplify and improve the readability of the final results, but the importance of the method of videostroboscopic recording cannot be overlooked.
In order to obtain a recording that will allow for a complete and possibly reliable kymography of the vocal folds, the most important thing is to properly match the moments of collecting individual frames (camera shutter or the moment of flashing the strobe illuminator) to the frequency of the vocal folds movement. Each subsequent frame captured by the camera must be collected at a slightly later stage in the movement of the vocal folds. In other words, the phase difference between the movement of the vocal folds and the strobe must be linearly increasing. This allows registering a video like slow motion. Obtaining such an effect is not always easy, because the frequency of the folds' vibration may change slightly over time, depending on the height of the tone generated by the examined patient, or they may even vibrate very unevenly, even from period to period, due to the existing pathology, and it is not possible appropriate selection of the moments of flashes or shutter. Incorrectly made recording makes it impossible to carry out kymographic analysis or may distort the results. The videostroboscope (its internal program) is responsible for the appropriate adjustment of the moments of collecting the frames to the current voice frequency of the examined patient. For this purpose, it has to determine the fundamental frequency of the patient's voice (the algorithms used for this are not as accurate as those used for computerized voice analysis, but usually provide greater stability). Therefore, it is necessary to provide the videostroboscope with the appropriate quality signal correlated with the movement of the vocal folds. The best, but also the most inconvenient, solutions are a glottograph or a laryngeal microphone (throat microphone), but microphones attached to the endoscope are commonly used because of their simplicity. In the latter case, it is very important to provide good-quality sound to the microphone of the videostroboscope. It should be close to the test person's mouth and should not be exposed to external disturbances. It is also advisable to provide sufficiently long phonation duration at a constant height so that the videostroboscope is able to achieve a state of flash or shutter synchronization with the movement of the vocal folds.
Incorrect synchronization of the collection of frames (too small or too large difference in the phase of the folds movement between consecutive frames) may lead to the recording of the vocal folds appearing stationary or moving very quickly. An example of an incorrectly selected phase difference can be the video below and the kymographic cross-section obtained from this recording (Fig. 1).
Fig. 1. Time section of the vocal folds function (kymographic cross-section) with an incorrectly selected difference in the phase of collecting the frames; too little phase difference between consecutive frames
In this case, the phase difference between consecutive frames is almost zero. This gives the impression of non-moving vocal folds, which may erroneously lead to a diagnosis of complete regurgitation of the vocal folds. The visible gentle vibrations are related to slight differences between the consecutive periods of the movement of the vocal folds and at best can be used for a very rough assessment of the intertemporal uniformity of the work of the folds (i.e. the similarity of the movement of the vocal folds for adjacent periods of their movement).
As mentioned above, irregular differences between the phase of the strobe’s work and the phase of the movement of the vocal folds, resulting from the uneven work of the folds or, less frequently, from stroboscope errors, are manifested in the trembling of the vocal folds in the recording, which results in a temporal profile with visible irregular narrowing of the glottis as shown in the figure below (Fig. 2).
Fig. 2. Time section of the vocal folds operation with discrepancies between the moments of collecting the frames and the frequency of the folds movement; sensation of trembling vocal folds
Fig. 3. a) (on the left) The sensation of trembling of the vocal folds caused by incorrectly performed kymography; b) (on the right) the same kymographic cross-section after correction
The disorders associated with problems in making the correct videostroboscopic recording (shown in cross-section kymography - Fig. 2) should be distinguished from not quite correctly performed kymographic analysis of a correct videostroboskopic recording. The difference is subtle and is essentially limited to the existence of symmetry in the case of a correct kymographic analysis of a not completely correct recording (Fig. 2) and shifts in the case of an incorrect analysis of a good recording (Fig. 3a). In the second case, it is possible to obtain the correct kymographic cross-section after careful correction at the stage of kymographic analysis (Fig. 3b).
Performing a correct recording that can be used to perform a kymographic analysis also depends largely on the correct framing already during image recording and capturing the appropriate fragment of the vocal folds movement during recording with a strobe camera. In general, the image of the working vocal folds should be approximately 50% of the height of the entire frame. If the vocal folds seen in the video are substantially smaller, this may make kymographic analysis difficult, degraded or impossible. The appropriate size of the vocal folds on the screen can be achieved by positioning the endoscope tip at a shorter distance from the folds or by using endoscopes with a longer focal length (with a narrower imaging angle). Optics with the possibility of focal length adjustment, i.e. with a variable imaging angle (ZOOM), can be a good solution. Here, the angle of the endoscope head to the endoscope axis should be distinguished from the imaging angle. The imaging angle (depending on the focal length of the endoscope lens, usually from 50 to 110 degrees) translates directly into the size of the field of view (i.e. inversely to the apparent size of the folds in the video image). The head inclination angle (from 0 to 90 degrees, usually 70 to 90 degrees in ENT endoscopes) forces the endoscope tip to be properly positioned in relation to the vocal folds, which influences the distance between the tip and the folds used by the doctor, which in turn indirectly affects the apparent size of the vocal folds in the image. Smaller inclination angles force the position closer to the folds, but the selection of smaller head inclination angles causes greater problems with obtaining the image of the folds in some patients. For kymography, endoscopes with a head inclination angle of 70 degrees can be suggested, but the choice of an endoscope is an individual matter and should be preceded by a practical test.
Incorrect cropping or positioning of the endoscope tip during image recording may result in poor or complete lack of visibility of the vocal folds; they may be completely or partially obscured by another structure. The analysis of the function of the vocal folds requires full visibility of the vocal folds, because covering them prevents the correct kymography. An example of incorrectly imaged vocal folds is presented below (the endoscope tip is not fully positioned at certain times), which gives a false picture of the cross-section of the vocal folds function (Fig. 4).
Fig 4. The cross-section of the function of the vocal folds that were obscured in the middle of a fragment of the recording by a different structure
The figure shows the time course of the work of the vocal folds, which, roughly in the middle of the recording, were completely obscured by another structure located near the folds. Such a defect of the recording resulted in the loss of some data, which in this case made it almost impossible to analyze the work of the vocal folds.
If the strobe camera has already been properly set up for phase difference when collecting frames and the captured image has been properly framed by optimally positioning the endoscope tip and the folds are fully visible, focusing is another important component of effective strobe video recording.
It happens that the larynx around the folds is as clear as possible, while the folds themselves are blurred and very indistinct, so it is equally important to position the camera so that the vocal folds are the sharpest or one of the sharpest elements in the image. Incorrect focusing of the image will not threaten the correct diagnosis, but it can significantly reduce the legibility of the time cross-section of the folds operation (Fig. 5). It should be mentioned that with most endoscopes, it is possible (and advisable) to focus after the endoscope tip has been inserted into the throat.
Fig. 5. Time section of the function of the vocal folds from a recording where the vocal folds are out of focus
The last factor influencing the quality of the analysis performed is the capture of at least one whole phase of the movement of the vocal folds, i.e. at least one full occlusion and one full opening of the vocal folds. Additionally, in the case when the collected frames contain more than three full short circuits and opening of the folds, the video fragment should not be removed in order to reduce the valuable recording to only three phases of movement, because most good programs that allow obtaining kymographic sections from videostroboscopic recording already allow it already when performing the kymographic analysis, select the most interesting part of the video recording and perform the analysis without losing the entire video sequence.
Fig. 6. The cross-section of the function of the vocal folds obtained from the recording, which contained only one full short circuit of the folds
A recording containing only one incomplete short-circuit or opening of the folds will make it practically impossible to perform kymography, as it does not contain enough information about the work of the folds, as in the example above (Fig. 6). The figure shows the time course of the folds operation, showing only one full short circuit of the folds and two incomplete openings. Based on the kymography of such a video, it is not possible to make a correct diagnosis.
If the video recorded with a strobe camera meets all of the above-mentioned conditions, the analysis of the kymographic cross-section obtained from it will enable a very precise assessment of the work of the patient's vocal folds and the correct diagnosis.
Since it is in fact impossible to record a perfect video image in which the image would be perfectly stable, the program used for kymographic analysis should be equipped with a manual or semi-automatic cropping method (with rotation), which allows user to eliminate any unwanted movements of the vocal folds on the screen caused by undesirable movements of the endoscope tip. We invite you to read the description of the next step in the analysis, i.e. cropping in the DiagnoScope Specialist program.
A more modern alternative to the described stroboscopy is the high-speed video technique mentioned at the beginning. It eliminates the two most important problems of strobe recording: incorrect flash timing and endoscope tip vibration. Sample recordings with the use of high-speed camera and comparisons with strobe recordings in the gallery of video from the high-speed camera.