CED VideoDisc and Player Discussion

[Rick314]

(Mostly to Tom Howe, but anyone feel free to respond.)

I am a computer science and electrical engineer with 30 years experience in telecommunications. I recently got interested in the Voyager Golden Record, and think there are some errors on the cover. I found http://www.cedmagic.com/featured/voyager/voyager-record.html and I am hoping you are interested enough to confirm or deny my opinions.

To prove what I am thinking I need the 2 1-hour analog waveforms that created the Golden Records. I am trying to get these from JPL but am not getting very far.

1) A least-significant bit error: I am guessing that the image scan sawtooth waveform is ideally 120 Hz. (This is key, and can't be proven without the source waveforms.) If it is, its period is exactly 8 1/3 ms. This is 722.46 * 2^14 hydrogen-line periods. 722.46 should round to 722, not 723. So the binary indicating 1 frame should be 1011010010(14 0 bits), not 1011010011(14 0 bits) as is on the cover. If the ideal period being indicated is 8 1/3 ms, do you agree?

2) Graphic incorrect: Again, assuming the scan sawtooth is 120 Hz, the 8 1/3 ms period applies to 1 sawtooth period, not the time indicated in the graphic of one sawtooth period plus one retrace. I think the left graphic duration boundary should have been 1 retrace-time to the right. (This just looks like an error in any context -- the period of a periodic waveform is what is of interest, not the period plus 1 ramp of the sawtooth.) Do you agree?

For both the above, they are unimportant to decoding the images, since whatever timing adjustments are needed will be done to make the first image be a circle. But once this is done and the exact timing is known by the decoding alien, they will realize both the above errors. How embarrassing for our race.

3) On the scan sawtooth, there are 6 little circles, and the middle 2 are lower on the scan line than the left 2 or right 2. Do you know what these are indicating?

4) On your page you say "There is also a binary number of 10000 at the left end of the picture that converts to decimal 16. The interpretation of this seems somewhat cryptic..." I think you have this wrong, and those bits are meant to be read with the same orientation as the "512" on the right. So you have "1, 2, 3... 512" read from left to right. What do you think? (It is sort of asking for trouble to make the 1s and 0s the same but rotated 90 degrees, and then write some values horizontally and some vertically.)

I have a copy of the Murmurs of Earth paperback coming that I got used from Amazon.com.

Thank you,

Rick Nungester
Liberty Lake, WA, USA
WA6NDR (amateur radio)

[Rick314]

Regarding "1) A least-significant bit error", if the image scan sawtooth waveform frequency is ideally 2 times the standard NTSC (TV) scan rate of 60/1.001 Hz, then the value on the cover is correct.

(1/(2*60/1.001)) / (2^14 / 1420.405752e6) =~ 723.17, which should round to 723, and the value on the cover image is correct. (1420.405752e6 Hz is the frequency of the hydrogen line.)

But regarding "2) Graphic incorrect", I still think the graphic is incorrect in indicating what this time period corresponds to.

[cedmagic]

Rick:

1) I checked the computation on the cover vs. the possible error and came up with these figures:

101101001100000000000000 = 8.3396 ms

101101001000000000000000 = 8.3281 ms

So if the period is 8.3333 ms, then the second binary value would be closer as you surmised.

2) I agree the graphic does take some artistic license in its depiction.

3) I've never seen an explanation of those circles on the sawtooth, and it is not explained in Murmurs of Earth.

4) I agree reading that as three vertical binary numbers instead of one horizontal number is a good interpretation. When writing the page I remember reading online about the binary 10000 = 16 grays interpretation, but after eight years the source is long gone. It's actually cool that it can be read in two directions to mean two different things. I may update the web page with both interpretations.

You might also want to contact Jon Lomberg about his design of the cover. Although it's been 30 years, he may have some notes on it.

--Tom Howe

[Rick314]

The "16 shades of gray" logic doesn't make sense to me, since the source signal is analog. I would assume its lowest level was no light (black) and highest level was lots of light (white, whatever this means to the ET) with intermediate values mapped continuously between those two.

[SomeJoe7777]

Rick,

I know this thread is old, but I was recently looking for information on the Voyager golden record myself, and came across the page on the CED web site and this thread on the forum.

1) Really, the time bases given in the decoding instructions are more of a guide than anything else. The images on the record are still frames, not moving frames, thus they have no temporal relationship between them. As such, the exact record playback speed and/or the exact waveform period are somewhat arbitrary.

In other words, I could play the record back at twice the speed indicated on the decoding plaque, and the scan line waveform period would then be halved. But as long as I understand that each image is a group of 512 scan lines, all that does is make the picture decode twice as fast. In fact, I can pick any playback speed/waveform period, and as long as I decode in groups of 512 scan lines per frame, I still retrieve the images.

2) Yes, the graphic is not semantically correct. The border lines delineating a scan line have been extended one retrace interval too far (in either the left or right direction). However, if I were an alien race attempting to decipher this, it would be clear to me that what is being communicated is the waveform period, despite the inaccurate delineation points. And again, the period is actually arbitrary. Assuming the aliens have some sort of device that can perform the function of the oscilloscope, once they are able to see the waveform it becomes clear that it's grouped into 512 regions, each of which represents a scan line that has to be assembled together.

Also keep in mind that they have an additional clue - two Vidicon cameras are on board Voyager. Analyzing the circuitry of the camera will reveal the concepts of image scanning.

Realistically, the aliens are not going to build a CRT with these scan frequencies to see these images anyway. They will either use the stylus to recover the voltage waveform and record it onto their own media/computer, or possibly just photograph/scan the surface of the record with a tool to recover the analog waveform, and then assemble the images digitally. (This is probably the way we would do it today if we found this record.)

4) Definitely the binary numbers at the left of the scan raster image are 1, 2, and 3, written vertically just like the 512 at the end. They indicate how the scan lines are supposed to be assembled.

The images should not be limited to 16 grays. The waveform is analog, and has theoretically infinite representations of the luminance level. This, however, is a key point that is not communicated. Figuring out that the voltage level of the waveform corresponds to luminance intensity will require a leap of intuition on the aliens' part.

Figuring out that three consecutive images have to be put together with color filters to form a color image would be very difficult. We humans use red, green, and blue to represent full color images because our eyes have 3 different color receptors, and they can be fooled into thinking they're seeing a full color image using only these 3 primaries. If the aliens' eyes are anything except trichromant, they won't get it. The 3 successive images will seem to be more than is necessary (if they're dichromant or monochromant), or will seem to not convey all colors (if they're tetrachromant or higher).

3) I have absolutely no idea what the circles on the sawtooth waveform are. They don't seem necessary. At first, I thought they might have something to do with interlacing, i.e. If the sawtooth waveform was to represent the frame scan waveform (what we would call the vertical retrace waveform on NTSC TV) vice the scan line waveform (what we would call the horizontal retrace waveform on NTSC TV), then the circles could convey that the first pass over the frame works with odd scan lines and the second pass over the frame works with even scan lines. However, I reject this interpretation on the following basis:

a) By all accounts, the images from this record are progressively scanned, not interlaced. The Vidicon cameras on the spacecraft also progressively scan.
b) Interlacing would add a high amount of complexity to the decoding process, which is counter to the goal of making the decoding process as simple as possible for the aliens to figure out.
c) Interlacing is not necessary for this application - Interlacing was done in NTSC to allow sufficient temporal resolution (i.e. frame rate) while working within the bounds of the bandwidth limitations. Since this application has no temporal component, there is no reason to interlace to save bandwidth.
d) Proper interlacing requires an odd number of scan lines per frame to precisely align the even and odd fields. Since these images are composed of 512 lines, this does not meet this requirement.

The only other explanation I can come up with is the following:

If the aliens decode this record and then begin searching the cosmos for us, they may receive some of our NTSC transmissions. These are similar, but not identical, to the decoding method used on the record. To decode NTSC, they would need to understand interlacing.

This writeup is only my interpretation, and is based solely on my engineering background and problem-solving skills. As such, it is only an opinion.