I think there's some confusion between what is considered original and what is considered backup. The Memorex telemetry tapes that were overwritten with new data were those from Apollo 11 and a bunch of other missions. These were the tapes running through the telemetry recorders during the Apollo missions and, as such, constitute the first recording of the data. That is the sense in which they could be considered originals. For Apollo 11, this also constituted the embedded television signal. Intuitively we would think of these as "originals" and therefore worthy of special protection. But the facts are slightly more complicated.
None of the data were directly usable from the tapes. To access the data on them, you needed telemetry tape machines, which were comparatively expensive and uncommon and most in demand for recording new data. You needed a technician to configure the heads and channel processors. This was especially true for the embedded video signal, which was a complete one-off used only on Apollo 11. You needed custom decoding hardware and technicians to operate it.
Therefore the first thing that was done with the telemetry was to read out the different data channels onto standard 9-track, 6250 bytes-per-inch computer tape or on paper strip charts. The charts can be read directly and photographically copied, and the standard computer tape could be run through any of the computers at NASA. For Apollo 11, the video signal was decoded and converted to a standard NTSC signal and recorded on standard reel-to-reel video tape. This tape could be played on any video tape machine.
These secondary recordings were considered the best sources available because they could be easily copied and accessed. Strip charts and computer tapes are reasonably faithful reproductions of the "original" data. If some spacecraft sensor reads 42 and this is preserved on the Memorex tape as a PCM-encoded number 42, then it will read out on the 9-track tape as 42 and on the strip chart somewhere accurately between 40 and 50 on the pre-printed grid. There's no loss of precision. The spacecraft told you 42 and you have 42 in your data stream. This was a well-developed, well-tested method because it was used on previous missions and would be used for subsequent missions. It was the general case.
But the process of reading out the video signal was entirely different. There was no magic way of electronically converting the embedded slow-scan signal to a standard video signal. You had a slow-scan monitor that could display the video, and then they just pointed a standard video camera at the screen. This is a fairly ingenious and reliable method, but it obviously accepts some degradation in the picture. It was deemed acceptable because of the one-off nature of Apollo 11's television solution. Previous missions had the benefit of the CSM's high-gain antenna and wider bandwidth. Subsequent lunar surface missions would rely on the erectable S-band antenna. (Apollo 11 carried one, but would have set it up only in an emergency.)
The usable form was considered the primary video record, even though it was technical a secondary format. If you wanted to re-read the video portion of the telemetry signal, you would have needed to reassemble all that one-off equipment: the slow-scan decoder and monitor and a camera. At no time did NASA contemplate a less lossy method of obtaining the picture from Apollo 11 television. And this was not considered irresponsible back then. Kinescopes and telecines were standard methods of converting between picture formats—film and videotape. These were all lossy methods and considered acceptable and common at the time. NASA considered the primary visual records of the mission to be the film photography, not the live television.
NASA did retain the original tapes for as long as possible. The decision to reuse them for a later mission, however, was not irresponsible. It was the proper decision in the lurch they had been put in by the supplier of new (faulty) tapes. Viewing this decision in light of late and improbable data recovery techniques is unfair. The situation is unfortunate, but utterly unrelated to any sort of credible claim of a hoax.
But to summarize all this in an answer to your question, the Memorex tapes were both original and backup. They are the earliest and best recordings of the video signal, but they were never meant to be used as a primary source. The primary source was the slow-scan-to-NTSC conversion, and the (relatively unusable) telemetry tape played the role of backup.
