As I understand it, the System N parameters were developed in Japan during its initial debate about what TV system to use. The choice was apparently between NTSC (System M) and CCIR/Gerber (System B), and to some extent the debate was characterized as one between 6 MHz and 7 MHz channels. Thus the CCIR/Gerber proponents demonstrated that the 625/50 system could be fitted into a 6 MHz channel if required. If that indeed were the case, then reconciliation of the timing might be an issue, as Argentina started broadcasting with System N in 1951, whereas Japan did not start (with System M) until 1953. So it might have been the other way around, with the early System N work done for Argentina, where the requirement was for a 50 Hz system that fitted the standard American 6 MHz channelling pattern.
The CCIR/Gerber system was already something of a compromise. The Russians had developed the 625-line system for an 8 MHz channel with 6 MHz video bandwidth, the latter allegedly to provide the same horizontal definition as 16 mm film. Otherwise they had generally followed NTSC (I) precepts. It would appear that an 8 MHz channel was more than most Western European countries were prepared to accept, particularly knowing that the USA had chosen a 6 MHz channel. Philips had developed its 567-line system as the best-fit for a 50 Hz system in a 6 MHz channel, but that was going nowhere. Thus came about the Gerber compromise, 625/50 in a 7 MHz channel with 5 MHz video bandwidth. One assumes that 5 MHz video bandwidth was considered to be as low as it was prudent to go with the 625/50 system. As such it provided horizontal definition that was only barely greater than that provided by the British 405 line system with 3 MHz video bandwidth, and about the same as that of the French 441 line system with 3.5 MHz video bandwidth.
In the UK, the BBC started its experimental UHF 625-line transmissions in 1958 May using the System B parameters (as recorded in Wireless World 1958 May). One may surmise that soon thereafter it must have become apparent that Europe was heading towards adopting a standard 8 MHz channel, with standardized vision carrier frequencies, for UHF transmissions. In turn this prompted the BBC to look at the optimum use of the 8 MHz channel. The obvious choice might have been simply to follow the Russian (OIRT) precedent, but what came out of this study were the System I parameters, as embodied in the 1960 TAC report. Evidently the best overall performance increment, as compared with System B, was the combination of a bigger vestigial sideband with a video bandwidth of 5.5 MHz. The latter also allowed NTSC (and derivative) colour subcarrier sidebands to be accommodated with asymmetry.
One may question the wisdom of introducing yet another 625-line transmission system, rather than picking the best of those already available. Perhaps a defence here could have been that firstly, the underlying system was the same as used across Europe, thus was not inimical to easy program exchange, and secondly, that as future British receivers would necessarily be dual-standard 405/625, they were unique anyway, so that differences in 625-line transmission parameters would not be of major consequence.
Whether the modulation question was seriously considered in the run-up to System I is unknown. Perhaps the BBC and the TAC simply followed the established majority choice of negative/FM. Or perhaps there was concern about oscillator drift at UHF, which would have pointed to negative/FM because it allowed intercarrier sound. But given that System I was a new transmission system, there was no fundamental reason why positive/AM could not have been chosen for it, just as the French did with System L.
NTSC (I) in 1941 chose negative vision modulation primarily because it on its face it allowed simpler black-level referenced (i.e. sync tip) agc. Intercarrier sound was not a factor because it was not invented until the late 1940s. In practice simple agc did not work very well at VHF because of impulsive interference, so that line-gated (and preferably also noise-gated) agc was required. This reduced the advantage of negative over positive, although negative still required less agc system gain (exacerbated by the fact that whereas sharp cutoff valves were often used in negative/FM receivers, there was a greater tendency to use remote cutoff types in positive/AM receivers).
Intercarrier sound as typically executed was in and of itself a compromise, and where high quality sound reception was required, particularly multi-channel, it was necessary to revert to split sound, or use the quasi-split technique, or “true intercarrier” as used in some BBC receivers. Once economic PLL synchronous vision demodulation techniques became available, intercarrier sound could be “clean” enough that it could also be used for positive/AM systems (e.g. look up the Motorola MC44302A IC).
Allegedly the French chose positive/AM for System L to simplify dual-standard receiver design. As part of this there was also a common sound IF (39.2 MHz) for Systems E and L. This required System L UHF receivers to be oscillator-low. (System E receivers were a mix of oscillator-low and oscillator-high.) When System L was extended to VHF, whilst oscillator-low was possible for Band III, it was not for Band I, so that the Band I channels were inverted (as System L’), with vision carrier high, so that oscillator-high could be used with them. As anyone who has been involved with corporate complexity reduction exercises will likely know, complexities can be moved around and hidden from immediate sight, but they usually turn up somewhere else in another form. It has been said that the French chose positive/AM as a non-tariff import barrier, but the counter there is that for the Outré-Mer territories, where there had been no use of System E and therefore no need for dual-standard receivers, the French chose System K’, which was negative/FM. (Although Algeria had started out with System E, including bilingual sound, but somewhere along the line that was abandoned.)
For Systems K’ and L, the French chose a video bandwidth of 6 MHz (as in the Russian System D/K) coupled with a 1.25 MHz vestigial sideband. By the usual reckoning, this combination would have required a 8.5 MHz channel, but was used with 8 MHz channels. There was a System E precedent for this kind of “squeezing”, though. Originally System E was designed for a 14 MHz channel, as evidenced by the original channel that became F8A (174 to 188 MHz). But the definitive tête-bêche VHF channelling system actually spaced the vision carriers at 13.15 MHz.
Certainly dual- and multi-standard receivers that covered both negative/FM and positive/AM systems involved some complexities, but they were hardly new when they became the norm in the UK. They had been used in Belgium since the beginning of TV there in 1953, and also in parts of France (around Strasbourg). Not only that, but some of those receivers covered four standards and managed to have line-gated black-level agc on both negative and positive systems, a precedent that the UK setmakers seemed unwilling to follow. The BREMA standard IFs might have encouraged some compromise on 625, as well. The simplest approach was probably a basic IF curve (before 405 traps) that was of double-Nyquist form, with -6 dB points at 34.65 and 39.5 MHz. This would have limited the 625 video bandwidth to 4.85 MHz maximum at -6dB, maybe around 4.2 MHz at -3 dB, well below the 5.5 MHz transmitted.
Now turning to Australia, System B, as used in most of Western Europe, was the logical choice in the mid-1950s, at which time System I was still in the future. There was no good reason for it to use a rather compromised Latin American version of the 625-line system. By so doing it may have gained one extra VHF channel, but it solved the channel count issue by using Band II, and later on some of the space just below Band III. By the time it needed more channels, UHF was a reality, and uniquely it adopted 7 MHz channelling in bands IV and V.
New Zealand adopted System B on the cusp of availability of System I, but it would have been a brave move for a small country, and one of the world’s most remote, to have been the first to use that system. The available Band I and Band III channels were initially sufficient for a single program, and in fact the seven first program main transmitters were all accommodated in Band I. Not only that but some of the second program main transmitters were also fitted into Band I. Eventually a third program was accommodated in Band III to which two channels were added at the top end, but for this exercise co-siting throughout had to be abandoned. Bear in mind too that in the NZ case, reasonable population coverage does not require anything like full geographical coverage, unlike the UK case where the two parameters are closer together. Band I would still have been limited to three channels whether the latter were 6 or 7 MHz. NZ did adopt European 8 MHz channelling for UHF; that may have been a plus when digital came along.
Now back to the UK case, adoption of a positive/AM version of System N, and taking advantage of its lower bandwidth, would have meant using 6 MHz channels at UHF rather than the European 8 MHz standard. Whilst such unilateral action was not outruled, it would have been somewhat counter-current at the time. System N, with its 4.2 MHz video bandwidth, would have meant that at best, 625-line pictures would have had 90% of the horizontal definition of 405-line pictures, and in practice probably not as good as this. One may see the case (just) for System N where the use of 6 MHz channelling was seen as an imperative, but not where 625-lines was to be introduced as a step-improvement over the existing 405-line system. Back in those days broadcasters usually adopted future-looking standards that might initially be a stretch for deluxe consumer receivers, but they did not pare back to match the assumed “modal” receiver or even further to match the lowest common denominator. At least System I was a potential improvement over System A, both in terms of line-count and horizontal definition.
Agreed that Band III in the UK was probably not fully utilized, whereas Band I was likely oversubscribed. But whether, even with 5 MHz 405-line channels, there was enough room in Band III to provide national coverage for two programs is debatable. Some use of UHF likely would have been required. And mixing 5 and 6 MHz channels in Band III might have brought its own complications and set of “taboos”, perhaps even more so once colour subcarriers were in the picture, so as to speak. Moving to a national chain of UHF transmitters at the time was logical and allowed for future expansion, and consistent with what was happening elsewhere. Starting a third programme by shoehorning it into Band III with 6 MHz channels would have delayed the need to use UHF by only a handful of years, but would have saddled the country with a compromised transmission standard.
One thinks of the Belgian System F case, where the French 819-line system was shoehorned into a 7 MHz CCIR channel. Perhaps justified in 1953, before standards conversion was an everyday event, where Wallonia needed to be able to access French origin programs, and where transmitters needed to be able to relay both 625-line (System C) or 819-line (System F) programs according to whatever was incoming. But as far as I know the System F transmitters were all on System C by the end of the 1960s, and Belgium had adopted System H for its UHF transmissions.
So one might say that whatever the short-term advantages that might be ascribed to System N, its basic parameters were just too much of a compromise.
Negative/FM or positive/AM was really a separate question, and the UK would appear to have had a free choice in this aspect.
PAL-N colour did not arrive until 1978 as far as I know, although no doubt it would have been developed earlier had it been required. An interesting question, had a positive/AM version of System N been adopted, is whether the advent of colour would have required an offset for the sound carrier as it did in the 405-line NTSC case. That would have been another complication, perhaps the more so if afc had become widespread in domestic receivers when colour arrived, with some receivers using vision channel-derived afc and others sound-channel derived. And positive/AM at UHF probably did required good afc.
Finally Donald Fink, who was an eminent member of NTSC (I) and the person who chose the 525-line number, is on record as saying that it would have been better to have adopted an 8 MHz channel than the 6 MHz actually used. (Although 6 MHz was a "given" that NTSC (I) was not free to change.)
In summation, I think that the foregoing amounts to a “no” vote on the modified System N proposition. It definitely would have been the wrong 625-line system.