A Simple Method for Timing
by Steve Preston
This document describes a simple method for achieving good accuracy (less than 0.1 sec) in timing occultation videos. I have adopted this methodology as a replacement for the "stopwatch method". To use this methodology you must have recorded WWV1 on the HiFi audio track of the video tape while recording the video of the occultation. For more information on recording WWV and videotaping occultations in general visit IOTA's web site.
2006 May 29 Update
Adobe has purchase Syntrillium and they no longer offer a free version of Cool Edit so it is harder to find the free version of Cool Edit. As an alternative freeware program I recommend Audacity: http://audacity.sourceforge.net/ . Unfortunately, my text descriptions below still refer to CoolEdit.
Also, I recommend that you avoid HDTV or other digital TVs when performing this test. Use an standard, inexpensive "analog" TV monitor instead. I have measured delays of up to 100ms when displaying standard video signals on an HDTV monitor.
Basically, you connect the VCR's video output to a TV (so you can watch the video) and the VCR's audio output to the PC via the A/V selector switch. You route the VCR's audio output through one input of the A/V selector switch and leave the other A/V input unused. In this manner, you can cutoff the audio from the VCR by selecting the unused A/V input. The PC is setup to record the output from the VCR (via the A/V selector). Here is a description of how I do this setup on my PC.
Connect the VCR's video output to the TV. There are two possibilities for most TVs: via a video input or via the antenna/cable input. You probably already know how to do this part of the setup, but I'll write it down anyway.
Video IN: If your TV is a recent model, it probably has a video input. In this case, connect the VCR's Video OUT to the TV's Video In. To view the VCR output, you set the TV to Video IN instead of using the channel Tuner.
Antenna/Cable IN: If your TV does not have a video IN connector, you will need connect the VCR's Antenna/Cable OUT (coax) to the TV's Antenna/Cable IN. In this case, you view the VCR output by setting the TV to channel 3 or 4 (depending on the setup of your VCR).
Connect the VCR's Left and Right Audio Out to input A of your A/V selector switch using a standard stereo audio patch cable. Then connect the A/V switch output to the LINE IN of the PC's audio card (the microphone input will probably work but the line input is better) using the 1/8" stereo PIN to stereo RCA phono cable. Don't forget to set the A/V switch to position "A" at this point.
<tbd .gif/.jpg diagram of system hookup>
<tbd: pictures of (1) my A/V switch, (2) my cables>
3.2 Setup PC for recording
Next you will setup the PC for recording the occultation audio track. For most PC audio recording applications, this step is fairly simple as long as your remember to setup the Windows Mixer properly. The Windows mixer is sometimes listed as "Volume Control" in the Entertainment or Multimedia section of Programs->Accessories. Before trying to record, you should setup the recording section of the Windows mixer. I'll give a brief description of the Windows mixer setup here.
Windows Mixer Settings
Start the Windows Mixer/Volume Control via Programs->Accessories->Entertainment->Volume Control. The mixer actually has separate controls for recording and playback, but it will show only the playback controls on startup. You should check the settings for both playback and recording. Since the starts up with the Playback controls, check to make sure the sections labeled Volume Control Balance (overall volume control), Wave Balance, and Line-In Balance are active (the Mute checkbox should be empty) and the volume levels are somewhere near the middle range (mid-way up). Next, you should setup the recording controls. To get to the recording volume controls, you must use the following menu item to change to the Recording controls view: Options->Properties. In the Properties dialog, you will see a section labeled "Ajust volume for" that has three options: Playback, Recording, and Other. Select Recording, then OK to see the Recording control. The first column of the recording control, Recording Balance, sets the overall recording volume and balance. Start by setting the overall recording volume to the middle position (mid-way up the slider). Next, notice that each of the other columns has a little checkbox in the lower left corner labeled "Select". These "Select" boxes determine which inputs are fed into the audio card for recording. For the occultation recording, you should select only the Line-In Balance section. In all the other sections the checkbox should be empty (deselected). Set the Line-In Balance volume level to the middle level (mid-way up) for now. Now you are ready to setup your PC audio recording software (e.g. Cool Edit 2000).
In case you find my Windows mixer description confusing, Cakewalk has an excellent description here.
PC Recording Software settings
Setting up the recording software varies from program to program, so I'll just mention that you should try for 16bit, 44.1Khz, stereo PCM recording. In Cool Edit 2000, this dialog pops up when you select File->New. Find the "RECORD" button and read the next section to learn how to proceed from here.
To time an occultation, you record the VCR's audio track (WWV) with the PC while watching the videotape. When the occultation event happens, you "blank" the audio by selecting the other input (e.g input "B") on the A/V switch. To determine the time of an event, you will review the audio recording on the PC and measure the time from a WWV minute mark to the position on the audio recording where the audio signal (WWV) stops (at the "button" press). Since you pressed the "B" button when you saw the occultation event, this process gives you the time of the occultation.
4.1 Position videotape for recording
Now review the videotape on the VCR and find the first audible WWV minute mark PRIOR to the occultation event you are timing. Make a note of the UT minute for this minute mark as announced by the voice prior to the start of the minute. Note: if you hear a female voice announcing the time, you are likely to be listening to WWVH out of Hawaii instead of WWV out of Fort Collins, CO (this will be important later). You might also find it useful at this stage to play the tape to the occulation event and note the approximate time of the occultation event. This will make it easier to "be ready" for the event when you do the actual "timing run".
Now position the videotape approximately 20 seconds before the UT minute mark prior to the occultation event.
Start recording audio on the PC (hit the "record" button in Cool Edit). Start the video (play) and watch the video on the TV for the occultation event.
4.3 "Mark" the event
You "mark" the occultation event on the audio track by switching to the "empty" A/V input when the event happens. If your VCR audio input is connected to input "A" and input "B" is not connected to anything, you will hit the "B" switch after you see the occultation event. Don't try to "guess" and hit the switch at the same time as the event, just try to react quickly and hit the switch as soon as possible after you see the event - this should provide more consistent results. The "B" switch is stop button on your "PC stopwatch" .
After you hit the "B" switch to mark the event, you can stop recording with the PC (stop button on Cool Edit) and you can stop the VCR.
4.4 Calculating the time for the occultation event
Review the recording with Cool Edit 2000 to find the minute mark for WWV and note the time within the recording for the start of the tone that marks the beginning of a minute. Go to the end of the recording and you will see where the data ended as you switched to input "B". Note this time - the time of the "button press". Subtract the minute mark time from the button press time and you have the time from the WWV minute mark to the event. This gives you the time of at which the button press stopped the audio to the PC. <tbd show an example using Cool Edit 2000 screen shots>
Since the audio actually cutoff after you saw the occultation event on the video, you must now subtract some time to account for this time lag. This time lag between the event on the video screen and the audio cutoff is your "personal equation". It varies from person to person and the variations can be several tenths of a second, so you really should "calibrate" yourself to determine how much time to subtract for your "personal equation". Obtaining a good esitmate of your personal equation is probably the most challenging part of this method. To do a good job of estimating your personal equation you really should do a couple of test runs with a videotape of an occultation event for which you already have an accurate time. You try the "PC stopwatch" method of timing several times and compare this result against the real time to determine your personal equation. My personal equation is about 0.2 seconds and my times varied by about 0.02 seconds over ten trials2. Since many people do not have a videotape of an event for which they know the accurate time, feel free to contact me if you would like me to make a copy of my "calibrated" occultation event for you. Alternatively, you can test your personal equation against the WWV minute mark - listen to the tape a stop the recording when you hear a WWV minute mark.
Your personal equation will vary from day to day and depending on various aspects of an event (disappearance, reappearance, contrast of star, etc.). So, even if you have been careful in determining your personal equation, I recommend that you account for these variations in two ways. First, report your timings as less accurate than your estimates show. I measured my time lag at 0.2 seconds, with a variation of plus or minus 0.02 seconds. However, for hand timings, I will only report my times as accurate to 0.1 seconds. Secondly, always do a couple of timings (trials) for each occultation event and watch for large variations. Of course, if you want to go one step further, you can check your personal equation against the "calibration" video before every session with the "PC Stopwatch". Personally, I plan to trust the 0.2 seconds and increase my +/- 0.02 error estimate to 0.1 seconds to account for the daily variations. I'll probably only re-check my personal equation once a year.
OK, once you have determined your personal equation, you subtract the personal equation (e.g. 0.2 seconds) from the "button press" time. And, you are almost done. You have now determined what I call the "videotape" time for the event. However, to get the "real" time, you must make two more adjustments. You must account for a delay within the CCD video camera and account for the radio propagation delay of the WWV radio signal.
To factor in the radio propagation delay of WWV, you must add time. Specifically, you must add 3.335 microseconds per kilometer distance from the WWV transmitter in Fort Collins, Colorado. Note: if you were actually recording WWVH out of Hawaii you must compute the distance from the transmitter location in Hawaii rather than Fort Collins. To factor in the internal delay of the CCD video camer, you must subtract the "personal equation" of the CCD camera - the time difference between the exposure of the CCD and the signal on the video output of the camera. I have measured this at approximately 17 milliseconds for the PC-23C and the Astrovid 2000. I you don't have one of these, 16 ms is probably a safe bet - it won't be any larger. After including these two factors, you should have the real event time- finally!.
I've put together a sample Excel spreadsheet which shows my data and calculations for an occultation I timed using this manual method (PC stopwatch).
So far, this has been much easier than using a stopwatch to time events off the videotape in play mode. Currently, I only use this method for the dim occultations which I can't follow in field by field review on the videotape. For brighter events, I have used an LTC timecode inserter to indentify the video fields and apply a variation of the audio recording methodology of this document to establish the times for video fields. Using LTC timecode and field by field review provides more accuracy and eliminates the observer's personal equation. Recently, I purchased an GPS based device to overlay time directly on the video. This is a more expensive solution (approx $300) but much easier for reducing the data for bright stars. I'll put up a web page on this device once I have more experience with it.
<Back to my Astronomy Page>
Last updated: 2004 Jun 07