As previously noted, Japan stayed with a “low” IF, 26.75 MHz vision, until circa 1970, after which it changed to a “high” IF, 58.75 MHz vision, that was noticeably higher than any others.
On the face of it, that might be explained by inertia; that is, in 1953, when Japan started TV broadcasting, it simply adopted a variation of the earlier American “low” IF, at a time when the “high” IF was establishing itself in American practice, and for the most part, was yet to arrive in Europe.
But there might have been more to it. Japan adopted a channelling plan that was different to that used in the Americas. Channels J1 through J3 were in Band II, 90 through 108 MHz. Channels J4 through J12 were in Band III, 170 through 222 MHz, with a slight overlap between channels J7 (188 - 194 MHz) and J8 (192 - 198 MHz).
That could have made the use of the American 45.75 MHz IF problematical. Its 2nd harmonic was 91.5 MHz, very close to the channel J1 vision carrier at 91.25 MHz, and its 4th harmonic was at 183 MHz, similarly close to the channel J6 vision carrier at 183.25 MHz. Then the channel J4 oscillator frequency was 217 MHz, close to the J12 vision carrier at 217.25 MHz, although whether J12 was part of the original allocation or whether it was a later addition is unknown.
On the other hand, it does look as if 26.75 MHz was chosen to fit with the channelling plan. Its 4th harmonic was at 107 MHz, within channel J3, but in a position where it probably would not do too much harm. Its 5th harmonic, the highest usually considered, was well below Band III. It also put the channel J4 oscillator at 198 MHz, right on the boundary between channels J8 and J9, so out of harm’s way. Similarly the J5, J6 and J7 oscillators fell on channel boundaries, although the J8 oscillator was at 220 MHz, enough inside channel J12 to be a potential problem, I imagine. If though J12 was a later addition, part of the Band III “creep” beyond its original 216 MHz upper limit, then it would not have been part of the original deliberations.
As well as 26.75 MHz, both 32.75 and 38.75 MHz would have resulted in channel boundary-positioned oscillator frequencies for the lower Band III channels, and both would have avoided the J12 intrusion. But 32.75 MHz had a 3rd harmonic at 98.25 MHz, rather close to the J2 vision carrier at 97.25 MHz. And 38.75 MHz had a 5th harmonic at 193.75 MHz, very close to the J8 carrier at 193.25 MHz. So neither would have been good choices.
So by that back-of-the-envelope analysis, it looks as if the original Japanese IF choice was made as the “best fit” at the time, and it just happened to be a “low” IF. Going above 45.75 MHz was probably not a wise move at the time, as it would have complicated IF strip design with the domestic type valves that were likely to be available.
Quite why the change to 58.75 MHz was made is not clear, but avoidance of the channel J12 conflict and better allocation of the UHF channels were likely to have been factors. By the time 58.75 MHz arrived, the solid-state era was well established, and the requisite IF gains were easily obtained, particularly with integrated circuits.
With 58.75 MHz, the 2nd harmonic was above Band II. The 3rd harmonic was at 176.25 MHz, just above the boundary between channels J4 and J5, and unlikely to be problematical. Band III oscillator frequencies were all above the band. And the UHF “taboo” channels were well-removed from wanted channels, and so well down the RF selectivity curve.