Millisecond Video Timing for the Masses
Contents
1. to be acquired With such resolution a problem arises for users of the KIWI OSD if a star dims over several fields when did the actual occultation occur The guidance from David Dunham on this is For most stars the frame to use is the one with the star at 1 4th its unocculted level that is the point corresponding to geometric occultation of a point source according to calculations of the Fresnel diffraction pattern But for a few very large red giant stars such as Antares and Aldebaran the star s angular diameter is significantly larger than the approximately 10m scale of diffraction so for them the geometrical occultation occurs when the star is at 1 2 its unocculted level Note substitute field for frame in the first sentence Determining times of the geometric occultation is one of the many areas where LiMovie excels but that is another story in itself back to the VTI Geoff was adamant about ensuring the accuracy of this VTI and wrote several validation checks into the code that controls the unit as well as a backup procedure if the final validation reports a problem Whenever a timing run is ended by pressing the info button the KIWI OSD verifies that its time information is still within 1 millisecond of the current UTC If not a message is displayed to use the field counter to calculate the exact times for the fields containing an event see the user manual online for how this is done Pressing the button again resta2. Millisecond Video Timing for the Masses By Sam Herchak Updated July 2007 It seems every few years products become affordable that greatly increase the accuracy of occultation timings In the late 1990 s the GPS averaging function became available in handheld units which allowed a user to determine their location with far greater accuracy than maps More recently inexpensive and easily adapted security cameras like the Supercircuits PC 164 have allowed modest telescopes to record asteroid occultations on video down to 11th magnitude without expensive light intensifiers Now we have an extremely accurate and affordable VTI Video Time Inserter which will stamp any video recording with the correct Universal Time UTC down to the millisecond When using a VTI the real limit in getting an exact time of an occultation is the slow field rate of common video cameras and problems with starlight such as scintillation and the Fresnel effect More on that later Many years ago IOTA developed a handful of unique units which could insert time stamps onto a video recording after the fact triggered by the unique tone broadcast every minute over WWV and similar provided the quality of the recorded audio and video were adequate This worked pretty well since only the recordings with hits needed to be processed But it still required shipping the tape to an individual who had the equipment and knowledge to process it Besides the inconvenien
3. ce a much larger factor has motivated others to find another way the loss of time broadcasts by nations trying to cut costs An obvious source for accurate time signals was already in place GPS whose satellites transmit time signals accurate to few hundred nanoseconds But when would an accurate and affordable unit be produced The summer of 2005 is the answer As far as I know a few real time VTI units have been developed previously in Japan Germany the United Kingdom and the United States but each had issues such as cost availability or accuracy Those problems seem to have been overcome with the KIWI OSD manufactured in the US and distributed by PFD Systems LLC The KIWI OSD II KIWI comes from Kiwi Geoff Hitchcox who lives in New Zealand the developer and creator of the all important processor code while OSD stands for On Screen Display The video output from any NTSC PAL SECAM or MESECAM video camera is fed to the KIWI OSD which then time stamps the video fie ds in milliseconds before passing it on to be recorded by a device of your choice A key component of the timing device is the sophisticated Garminl8 GPS unit which has many features such as GPS averaging and WAAS Wide Area Augmentation System developed for aircraft GPS approaches All the specifications and details of the KIWI OSD are available online at the resources listed at the end of the article I will provide an overview however based on several mont
4. hs of research and testing A plastic box contains the guts of the unit and the latest version KIWI 2 has an info button formerly the reset button at one end with 4 connection ports on the other Two are for the video cables another powers the unit and the fourth connects to the GPS The sequence you make the four connections is unimportant although I prefer to connect the power cord last so I can monitor everything the VTI does from the start Additionally there is a plugged hole on the side that allows the Garmin 18 to be connected to a computer via the supplied serial cable for reasons to be described later The complete unit uses a miniscule amount of power from any 12 VDC source 9 15 VDC technically far less than even the PC 164 camera Standard RCA video cables are used to connect your camera and recording device while the Garmin 18 is connected and powered via a DIN plug Since the unit s primary purpose is timing information the Garmin is pre configured to provide information immediately after it acquires a less accurate 2 D position fix this can result in a position error for mobile observers which PII address later As soon as the Garmin has valid GPS time information accurate to millionth of a second it drives a red LED that flashes precisely at the beginning of each second and the KIWI OSD begins its validation process for time stamping You then get several messages visible at the bottom of the video displa
5. ial cable a computer and Garmin s own Sensor Configuration Software With 3 D fix required selected the KIWI unit will not start a timing run until the elevation has been updated The delay I have experienced in the field by requiring the 3 D fix is under five minutes changing your location a significant distance or poor sky coverage can slow the process down further If you don t move around much this change isn t necessary nor recommended The kit form of the KIWI OSD is no longer sold and in the near future the early versions 1a and 1b will also be gone But the ready to go KIWI 2 is immediately available and is a bargain at 165 USD That s much less than many eyepieces these days The KIWI OSD has been used in the field now for almost two years mine included and all the feedback shows it to be extremely reliable and accurate I highly recommend it for anyone timing events with video Check out the resources below which are full of information on video GPS time insertion and the KIWI OSD Your comments and experiences are welcomed at lt annespam 1 flica net gt www pfdsystems com kiwiosd html www geocities com kiwi_36_nz kiwi_osd kiwi_osd htm http groups yahoo com group kiwiosd www edu observatory org gps height html www ngs noaa gov cgi bin GEOID_STUFF geoid03_promptl prl www geocities com kiwi_ 36 nz www geocaching com mark nearest aspx http www005 upp so net ne jp k_miyash occ02 limovie_en html
6. ming run in the shortest possible time the Garmin 18 is pre configured for 2 D fix This means the unit will display information before the GPS has acquired enough satellites to determine its current elevation Because GPS units store the last known position and elevation when powered off in the 2 D fix mode the GPS has to assume the elevation hasn t changed since last stored This can severely skew the calculated surface position In one of my tests I moved the setup about 60 nautical miles NM and climbed about 1 000 meters in elevation When the KIWI OSD displayed the initial 2 D position it was off by 0 7 NM a gross error by any standard The best solution is to GPS average Take a 5 minute average wait 10 minutes reset the unit and take another 5 minute average Do this 3 times mathematically average the position and elevation readings and your position should be accurate to within 5 meters Be sure to report both the Mean Sea Level elevation and the Geoidal separation An averaged position is so accurate IOTA may need to adjust the elevation figure with a GEOID correction the Earth s surface is not a perfect sphere Also if you use an extension cable on the Garmin to raise it outside your dome for example be sure to correct for the distance between your telescope objective and the GPS receiver Personally I chose to change the Garmin 18 configuration from 2 D fix to 3 D fix required This is done with the included ser
7. rts the unit and begins the whole process from scratch This is also the way to trigger new and easily recorded position information So now is a good time for me to summarize GPS positions which change constantly even when your receiver isn t moving The primary GPS satellites are moving constantly in the sky 12 hour orbits so the number distance and sky position of the satellites used by the GPS receiver are always changing Sometimes the satellite constellation is optimal and allows a precise fix sometimes it is not What is easily demonstrated however is that over a period of time a GPS can provide a position that is accurate to Jess than 5 meters 16 feet This is not possible with a topographic map which for the most part are not accurate to 15 meters to begin with The key is to use the GPS to take several short averages spaced minutes apart Thanks to Scott Degenhardt I began doing this about 8 years ago Even before the signal scrambling that only allowed the military full precision was turned off selective availability or S A Scott and I went to benchmarks certified in accuracy to millimeters and proved simple GPS averaging techniques provide extremely accurate fixes If you have a GPS unit see for yourself the geocaching com link at the end will help you locate a certified benchmark in the US near you For mobile users of the KIWI OSD there is a possible trap with the initial position from the unit To begin a ti
8. y cycling through in a specific sequence giving position information before the video signal is stamped with time information Unless you are in an area with many obstructions resulting in exceptionally bad sky coverage like inside an observatory dome time stamping will begin in just a couple of minutes Except for video experts most people aren t aware that video cameras integrate and output in fields referred to as odd and even you may hear these referred to as half frames Both of these fields are combined to make up each frame For NTSC cameras common in the US there are almost 30 frames per second and just under 60 fields per second Many playback devices are capable of stepping through each field by using what they mistakenly call the frame by frame function This effectively doubles the time resolution of the camera Another option that is becoming more common is extracting times by using the application LiMovie created by Kazuhisa Miyashita Check the e book Chasing the Shadow The IOTA Occultation Observer s Manual for an excellent overview of LiMovie KIWI OSD time stamped video field from a lunar occultation 17 hrs 54 min 02 528 to 02 545 sec 15078 field count The VTI is so precise it marks the millisecond each field begins and ends Each NTSC field has a duration of about 17 milli seconds which is the only reason you can t time any event to millisecond accuracy the camera fields take too long
Download Pdf Manuals
Related Search