This paper presents a proposal for the recording of radio astronomical observations in digital form. The terms used are those familiar to computer users.
The purpose of recording radio astronomical data is to preserve its integrity for future use, by the experimenter or by somebody else. For this purpose, extreme care should be used to avoid destroying the characteristics of the data by too low a sampling rate (Nyquist Theorem), or by improper handling.
This proposal is based on the use of the standards that already exist for the communication between different computers, as well as on the equipment that is readily available or in the hands of most computer users. Consequently, this proposal is based on the use of ASCII for the encoding of the data and a modem for its transmission. The idea is that, instead of connecting the modem to a telephone line, it is connected to an audio tape recorder and the tones stored on tape. In order to generalize the use of this method low speed transmission will be used. Then, the idea is to store data on tape, through a 300 baud modem, in ASCII characters. The only additional equipment needed for producing or reading the recordings is a connector from the tape recorder to the modem, probably only a connector from the miniature plugs of the recorder to the modular plugs used for telephone connections.
It is herein proposed that the recording of radio astronomical data for future use be done as text through a 300-baud modem using ASCII characters, that will include a decimal point and sign if any, and that each value will be separated from the other by a delimiter, space, comma, or better, a carriage return. The data itself will be preceded by a preamble that will indicate all the general information (name, equipment, RA Dec. time ext) required for all SARA recordings. In addition digital data recording will have the following items:
The sampling rate of the data; that is, the original rate at which the data was sampled to convert it to digital form. This sampling rate must include the units in which the rate is given. Accepted ways are, as samples per second, or the period between samples with the proper units given. The sampling rate will be recorded as:
SAMPLING RATE: 12000 SAMPLES PER SECOND
SAMPLING RATE: 5 Ks/s
SAMPLING AT 12.3 MILLISECONDS
The precision of the sampling rate. In many types of studies it is necessary to know how precise was the sampling of the data. If you know a figure, give it; if you don't know a figure, indicate medium, medium low, or low. The precision is better indicated by describing the method used to determine the rate. Proper indications of the precision are:
PRECISION: +/- 1 NANOSECOND
PRECISION: LOW (not too good description)
PRECISION: WWVB RECEPTION DIVIDED WITH TTL LOGIC. NO MEASUREMENT OF PRECISION EVER MADE.
PRECISION: VIDEO COLOR CRYSTAL OSCILLATOR DIVIDED BY TTL LOGIC. OSCILLATOR FREQUENCY MEASURED AT 3.5796 MHZ.
PRECISION: COMPUTER TIMING LOOP. NO DOUBLE BRANCHES. NO MEASUREMENTS EVER MADE.
The preamble and the data should be two standard blocks of text that could be read onto the screen of the computer with minimum of programming, normally the communication package that comes with the modem should be enough. The preamble should also indicate something about the way the data has been recorded, integer numbers, signed integers, floating point values produced by BASIC program, etc. Any additional information that will help judging the value of the data and simplify its consequent use should also be included in the preamble. The preamble should be separated from the data by several carriage returns, followed by a single line indicating the number of data points in the block, and followed by the block of data.
It is recommended that good quality tape be used in the recordings since the difference in cost between a good quality tape and a poor one is not that great. The same is true with the language used for the preamble of the data. Keep in mind that you are not sending a telegram and you will not be paying by the word. An additional word in the preamble might clear an ambiguity making your data more valuable. Something important is judging the precision. Do not under sell your data by presenting it as low precision when it is really quite good. Be honest and when in doubt, explain as completely as possible the details of your procedure.
A procedure for storing radio astronomical data in digital format has been presented. It appears this proposal has the following advantages:
It is clear that no procedure can be perfect. This procedure has the following disadvantages: