Technoarchaeology: Finding The Right Image in a Room Full of Tapes

Image: a portion of our set of Lunar Orbiter data tapes at McMoon’s – an abandoned McDonalds onsite at NASA Ames Research Park, home of the LOIRP – Lunar Orbiter Image recovery Project.
Here at the LOIRP (Lunar Orbiter Image Recovery Process) project there are two different phases of the image retrieval process that are distinct from each other. The second phase, the production of the vast majority of all the of the Lunar Orbiter images, will simply involve putting tapes on the tape drive machines, acquiring the data, and processing them into images.
However, we’re still in the first phase of the project where we need to search through tapes in a painstaking fashion just to find the images we are interested in downloading. Once we find what we are looking for, downloading is a snap and can be done in a matter of hours.
Finding the images using a jumbled nomenclature and labeling system last used more than 40 years ago is part of what we call “Technoarchaeology”.

There is a critical piece of information that we have never been able to find in all of our searches of the tapes: something that shows the linkage between the numbers on the tape cans. Each can has a label and number plus a letter designating the Earth ground station where they were recorded as the were sent back to Earth from a Lunar Orbiter spacecraft: W for Woomera, M for Madrid, and G for Goldstone. There is also an audio track on each tape where the tape drive operator often adds in operational notes such as date, time, etc. Indeed, you can often tell by the accent of the operator where the tape was recorded. Listen to this audio track and you can easily tell where it was recorded.
The images themselves have been identified in all of the existing databases (e.g. LPI, USGS) using a nomenclature that actually makes sense. In the case of the lunar south pole image we released today LOIV-94-H1 for example means Lunar Orbiter 4, Image 94, High resolution, subframe 1 of 3). So, while we know which image is which, correlating that to which tape(s) that image is on is another matter all together.
Therefore, what we have to do in our process of locating an image is to slowly look for the right images on the right tapes. This is quite labor intensive.
In the case of LOIV-94-H1, we first went through several tapes we thought that it would be on, including the subframes (including the subframes equals a total of 3,480 images) and recorded the first and the last framelet number (each framelet has an identifying number). Since each framelet has a number, this allows us to capture a couple of framelets and then look for the right image in the LPI database.
The problem with this approach is that the numbering system for the numbers on the framelets goes from 001-999. Since each high resolution image has about 28 framelet numbers (some framelets have two numbers due to the way that the calibration strips were written to the film), and each high resolution image has about 88 framelet numbers.

Image: This is another 2 inch tape reel on which the recently released Lunar Orbiter V image of the Earth was stored. Note on the label [much larger image] that “L.O.E.” refers to “Lunar Orbiter E” aka “Lunar Orbiter V”. The tape was recorded on day 238 of 1967 local time i.e. 26 August 1967. Starting at 06:34:38 GMT and ending at 07:26:10 GMT
On the spacecraft, the high resolution and medium resolution images were taken at the same time. This means that theoretically, we have 88+28 or 116 plus a few blank framelets, per image. About every 8 images the numbers repeat!! This is why we had to copy every beginning and ending framelet number from each image into an Excel spreadsheet in order to do our matching of the images on the tapes with the LPI database numbering system.
On top of this, we had three ground stations recording images (Woomera, Madrid, and Goldstone). We quickly discovered that we have more data tapes than what we should have for the number of images that NASA’s documentation says that we should have. Another factor is that with three ground stations, the way that the images were obtained should have been systematic. But the process was not systematic as far as we can tell – and the tape labeling suggests.
Think of it this way. The Moon is over any one location on the Earth for 13 hours. The original process of loading a tape on the tape drive and doing the downlink of an image took approximately an hour per image. As such, each ground station should be able to get 13 images per pass, right? No. One-third of the time, the Lunar Orbiter spacecraft is behind the Moon and therefore cannot send a signal. As such, two-thirds of 13 hours is about 8.5 hours of viewing time per pass per ground station. This is interrupted periodically by the spacecraft going behind the Moon.
In theory this means that they should be able to get at least 8 high and medium resolution image pairs per pass over a given ground station on Earth. Each high and medium resolution image pair takes 43 minutes to receive (plus overhead time to load the tape, set things up, and then remove the tape), which means that they should be able to get one complete high resolution and medium resolution image per lunar near side pass (smart fellows they were to figure this out). However, this is not always the case.

Austin Epps sitting in the LOIRP lab at “McMoons” at NASA Ames Research Center downloading imagery from an original Lunar Orbiter data tape using a restored FR-900 tape drive on 18 August 2009.
Some of the tapes that we have found only have a couple of minutes of video on them. One tape that we found had the same framelet repeated for the entire tape (the advance mechanism stuck on that image). Other tapes have partial images – and this is what we found with this most recent image LOIV-94-H1.
The first part of LOIV-94-H1, which was not the subframe that we needed, was on a tape from Woomera (W-30). The second part of the image was on a Goldstone tape (G-38). Hmm… what is wrong with this picture? Well, if the image was just cut off because of the end of a lunar pass over the site, the next station that received the image should have been Madrid!! (the Moon sets in the west and from Australia that means Madrid was the next station. As such, we lost almost an entire day looking for a the rest of image LOIV-94-H1 .
We are rather certain that as we go through the corpus of tapes that we will probably find some images – or fragments thereof – that are not in the established Lunar Orbiter datasets. These may well be images that no one has ever seen. How many there may be or what they are of will only be clear as we begin phase two of our project.
We are really looking forward to the time where we can just put tapes on the tape drive and let it go and worry about what image number it is after the download has been completed. Right now, however, this hunt and seek process we use is what our daily life is like here at McMoon’s. However, when we have both tape drives running and some new computers and software in place (soon), our output will dramatically increase.
So, this was what our day was like today, August 19, 2009, as we were looking for images from the Lunar Orbiter program – the first of which were received on Earth exactly 43 years ago – almost to the day.