05-02-2019, 03:14 AM
Here is the summary version of the outcome of my intermittent delvings into the origins and history of the 625/50 TV system over quite a few years. This quest started quite some years back with the question – why were there so many 625/50 variants? None of the books and magazine articles that I had read really addressed it. About the only aspect that was covered in the literature that I could access was why the system I parameters were considered to be the optimum use of the 8 MHz channel. And of course, however the 8 MHz channel was disposed, the underlying implication was that it was better than the 7 MHz channel. By no means have I found all of the answers, but then my research has leaned more to the “armchair” type than to scholarly and rigorous.
The 625/50 system was a direct derivative of the American 525/50 system, so it is worthwhile to first consider the origins of the latter.
The first NTSC in 1941 developed an analogue TV system by considering all of the then-available options for nearly all of the parameters, and choosing what it saw as the best for each. The only exception was the channel width, which was fixed at the previously decided RMA number of 6 MHz, although distribution of the signals within the channel was not predetermined. A field frequency of 60 Hz, to match the predominant US power supply frequency, was probably a given, but the case was worked through anyway. Negative picture modulation and FM sound were in fact independent decisions, although of course since then indelibly associated.
As to line count, the initial thinking was that 441 lines was appropriate given the 60 Hz field frequency and the vision bandwidth of 4 MHz, the latter chosen as part of the best use of the 6 MHz channel. (The later extension to 4.2 MHz came with the advent of the colour system developed by the second NTSC.)
The change to 525 lines that came late in the deliberation process, and was driven largely by Donald G. Fink. He outpointed that there was quite a bit of flexibility in setting the ratio of vertical to horizontal definition. Then he stressed the benefit of flatness of field, achieved by minimizing or completely eliminating the unilluminated space between the lines. He showed that this could be achieved with 525 lines, but not with 441 lines. Within reason, this effect was largely independent of screen size, as line width tended to increase with CRT size. This was a key point that I don’t think has always been stressed in the histories. The conventional wisdom was that line count was something of a continuum, the more the better, up to the point where cost, complexity and bandwidth requirements provided a practical limit. But Fink saw that there was also a threshold number within that continuum. Thus the NTSC (1st) system survived essentially unchanged until the end of the analogue era. The 6 MHz channel was its weakest part. In a 1976 retrospective paper (*), Fink made the comment: “The RMA allocations committee had the burden of setting the width of the television channel. The figure they chose, 6 MHz, was ambitious at the time but it has since tuned out to be one of confinement. If they had known what the future had in store, they should, in this author’s opinion have opted for an 8-MHz-channel. But that is hindsight!”
The 625-line system was initially developed by the Russians circa 1944 as a 50 Hz field frequency derivative of the NTSC system, retaining many features of that system. The line count of 625 was chosen on the basis of reasonably matching the NTSC field frequency. Absent a pre-existing channel width constraint, the Russians chose an 8 MHz channel with 6.5 MHz intercarrier and a vision bandwidth of 6 MHz to provide a suitable level of horizontal definition, said to be that which matched what was obtainable from 16 mm film. For reasons unknown, the vestigial sideband was kept at the NTSC number of 0.75 MHz, rather than increased proportionally.
The Russians were said to have presented their 625/50 system at the ITU Atlantic City 1947 meeting, where it was proposed as a standard, although as yet I haven’t yet found any mention of it in the meeting documents that I have looked through.
Post-WWII the Germans, including W. Bruch, picked up on the Russian work. It would appear – but needs to be verified – that they either accepted or endorsed the Russian choice of vision bandwidth and channel width.
However, others thought that an 8 MHz channel was too extravagant, and looked to the American 6 MHz channel as being preferable, at least from the viewpoint of getting an adequate number of channels into the assigned VHF bands. Philips was in that camp, as is evident from this submission that it made at the CCIR Stockholm 1948 meeting:
Out of that thinking came the Philips 567/50 system, apparently derived as the best 50 Hz system fit to the American 6 MHz channel, using the same general approach as the NTSC. If one does the numbers, then the (Iine-count squared x field frequency) products are closely similar. I should imagine though that whilst the Western European countries might have been reluctant to embrace the Russian 625/50 system as such, perhaps for political as well as channel width reasons, then neither were they likely to settle for a standard system with a lower line-count. Still, that Philips developed the 567/50 system does lead to the inference that there was at least some unease at squeezing 625/50 into a 6 MHz channel.
Be that as it may, the American-origin experimental 625/50 transmitter installed at Torino, Italy in 1949 worked in a 6 MHz channel, namely American channel 6, 82-88 MHz. (That was the origin of what eventually became system N, and also of Italian channel C, 81-88 MHz).
Meanwhile the French had gone their own way in 1948 with their high definition 819/50 system that required a 14 MHz channel and had a 10.4 MHz vision bandwidth and 2.0 MHz vestigial sideband. It retained some features of the earlier French 441/50 system, namely positive vision modulation, AM sound, no equalizing pulses and initially anyway, and inverted channel layout with the vision carrier at the high end. By 1952 it had been compressed into a 13.15 MHz channel (essentially by abandoning the outer guard bands) as part of a tête-bêche channelling system.
Exactly when and by whom the 7 MHz channel was proposed I have not seen recorded in a definitive fashion, but as the resultant system was referred to as the Gerber system and sometimes as the Gerber compromise (and it was a compromise), it may well have been his idea. The available evidence suggests that in the final round, as it were, of the CCIR deliberations, the 8 MHz channel had been excluded and the choice was between the 6 MHz and 7 MHz channels.
The original Russian 625/50 system (later system D) was the first with this line-count to see regular service, in 1948. Next appears to have been the 6 MHz channel version (later system N), with regular service started in Argentina in 1951, I think before any regular public use of what became system B in Europe. It would appear that the Latin American countries with 50 Hz electricity supplies had opted for a system that could use the established American 6 MHz channels. From the viewpoint of not creating mutual interference problems with their 60 Hz neighbours who were or were planning to use the NTSC system, that may have been a wise move.
Belgium had its own problem, in that it needed a transmitter network that could relay incoming programmes both from France on 819 lines and from the Netherlands on 625 lines, with any given transmitter being able to switch between the two systems. This required a common channelling system, so the 819/50 system was squeezed into the 7 MHz 625/50 channel. Why Belgium opted for positive vision modulation and AM sound is unclear. Clearly, both 625/50 and 819/50 had to be the same in this regard. Perhaps it was a case of even-handedness, in that as the Belgian 819/50 system was different to the French prototype, so the Belgian 625/50 variant had to be different to the European prototype. Or perhaps there were technical reasons. (As an aside, Belgium was the only European country, other than the UK, to seriously consider using AM for its VHF sound broadcasting service.)
By 1953, there were four variants of the 625/50 system in service, namely what later became systems B, C, D and N.
The next significant step was the planning of the European UHF TV services and channel allocations, from the mid-1950s. Here the past error of the 7 MHz channel was recognized, and early on there was universal agreement on the use of an 8 MHz channel, with the vision carrier positioned at 1.25 MHz above the channel lower edge for all systems. Those countries already using the 7 MHz channel at VHF were more-or-less stuck with what they had, although they were not averse to putting the 7 MHz system into an 8 MHz channel, as that provided best utilization of the UHF spectrum. Some I think entertained the idea of doing differently at UHF, but perhaps not too seriously. Belgium though did consider what became system I as a possibility.
France and the UK were both unconstrained by existing 625/50 systems, and both initially planned on fully utilizing the 8 MHz channel with 6.5 MHz intercarrier and 6 MHz vision bandwidth. In the UK case, there was a late change, recommended by the TAC in 1960, to the use of a 6.0 MHz intercarrier, with 5.5 MHz vision bandwidth and 1.25 MHz vestigial sideband. The reasoning was that in terms of picture quality on typical domestic receivers, the gain from the extended vestigial sideband (allowing a gentler Nyquist slope) was greater than that from extending the vision bandwidth from 5.5 to 6.0 MHz. That was probably true in the days of distributed L-C selectivity IF strips and diode demodulators, but much less so later on when SAWFs and synchronous demodulation became the norm.
France elected to use both the 6.0 MHz vision bandwidth and the 1.25 MHz vestigial sideband, thus using up the outer guard band space in the 8 MHz channel, as it had done with fitting 819/50 into a 13.15 MHz channel. One argument advanced for the use of a 6.0 MHz vision bandwidth was that when a colour subcarrier was added at the anticipated 4.43 MHz or thereabouts, it would allow full double sideband working, i.e. there would be no early attenuation of the upper sideband. This was probably advantageous with the NTSC system, but may have been more important for SECAM. As originally conceived, SECAM used simple AM subcarriers, in which any sideband asymmetry resulted in quadrature distortion with diode-type demodulators. I am not sure when the change was made to FM subcarriers, but even there, sideband asymmetry would result in distortion. As well, France opted for positive vision modulation and AM sound, as had been used for its 819/50 system E, to simplify the design of dual-standard receivers.
The 1.25 MHz vestigial sideband was also adopted at UHF by some Western European countries who otherwise retained the system B parameters. In this guise it was known as system H. And the Eastern European countries did consider adopting the 1.25 MHz vestigial sideband, although that did not happen in practice.
By the early 1960s, there were seven 625/50 variants, namely:
Western European B/G
Belgian C
Eastern European D/K
Western European H
UK, Ireland I
French L
South American N
The “final round”, as it were, occurred in connection with the setting of TV standards for the African Broadcasting area, at the ITU 1963 Geneva meeting. By then some African countries had already adopted and so were committed to system B. For the French Outré Mer territories, a negative/FM version of system L was developed, which was subsequently allocated letter designation K1. Without the need to consider dual-standard receivers for the African territories, the French could revert to the by-then standard negative/FM combination. The same arguments as for system L were used in support of the 6 MHz vision bandwidth and 8 MHz channel, and unsurprisingly the OIRT also advocated for this combination. South Africa, though, opted for system I. Countries not already committed to system B had the option of choosing an optimized rather than a compromised 625/50 system, and I guess that different groups of engineers had different views as to what constituted an optimized system within an 8 MHz channel.
Now there were eight 625/50 variants in use:
Western European B/G
Belgian C
Eastern European D/K
Western European H
UK, Ireland I
French Outré Mer K1
French L
South American N
Some sub-variants (systems B1, D1 and I1) were added right at the end of the analogue era, but these did not materially alter the overall picture.
In hindsight one could say that the Russians got it nearly right, missing only the proportional adjustment of the vestigial sideband, something that could have been remedied later on without too much pain. Had the Russian system been adopted in Western Europe, then I speculate that we might not have seen the subsequent multiplicity of 625/50 variants, and by the time that UHF transmissions arrived, it may have had too much inertia for anyone to seriously consider alternatives, which in any event would have offered relatively minor improvements. Even the French may have baulked at positive/AM had the rest of Western Europe long been using an 8 MHz channel, 6 MHz vision bandwidth negative/FM combination. However, the Western European compromise left wide open the door for doing better at UHF, and there were different views as to what “better” looked like, with no weight of established practice – in Western Europe anyway - to favour any one. Perhaps all that can be said is that for those countries new to 625/50, not to have done better than system B/G would have been a worse choice than doing better by any of the means proposed.
Be that as it may, system N may have been used anyway if the imperative to use a 6 MHz channel in Argentina and elsewhere in South America superseded all other considerations.
(*) The Donald Fink paper referred to is: “Perspectives on Television: The Role Played by the Two NTSC’s in Preparing Television Service for the American Public”. Apparently it was first published in the IEEE proceedings in 1976 September, then reprinted in 1999 as a Classic Paper. It is recommended reading from one of the outstanding TV systems and standards pioneers who was also an excellent technical writer.
Cheers,
Steve
The 625/50 system was a direct derivative of the American 525/50 system, so it is worthwhile to first consider the origins of the latter.
The first NTSC in 1941 developed an analogue TV system by considering all of the then-available options for nearly all of the parameters, and choosing what it saw as the best for each. The only exception was the channel width, which was fixed at the previously decided RMA number of 6 MHz, although distribution of the signals within the channel was not predetermined. A field frequency of 60 Hz, to match the predominant US power supply frequency, was probably a given, but the case was worked through anyway. Negative picture modulation and FM sound were in fact independent decisions, although of course since then indelibly associated.
As to line count, the initial thinking was that 441 lines was appropriate given the 60 Hz field frequency and the vision bandwidth of 4 MHz, the latter chosen as part of the best use of the 6 MHz channel. (The later extension to 4.2 MHz came with the advent of the colour system developed by the second NTSC.)
The change to 525 lines that came late in the deliberation process, and was driven largely by Donald G. Fink. He outpointed that there was quite a bit of flexibility in setting the ratio of vertical to horizontal definition. Then he stressed the benefit of flatness of field, achieved by minimizing or completely eliminating the unilluminated space between the lines. He showed that this could be achieved with 525 lines, but not with 441 lines. Within reason, this effect was largely independent of screen size, as line width tended to increase with CRT size. This was a key point that I don’t think has always been stressed in the histories. The conventional wisdom was that line count was something of a continuum, the more the better, up to the point where cost, complexity and bandwidth requirements provided a practical limit. But Fink saw that there was also a threshold number within that continuum. Thus the NTSC (1st) system survived essentially unchanged until the end of the analogue era. The 6 MHz channel was its weakest part. In a 1976 retrospective paper (*), Fink made the comment: “The RMA allocations committee had the burden of setting the width of the television channel. The figure they chose, 6 MHz, was ambitious at the time but it has since tuned out to be one of confinement. If they had known what the future had in store, they should, in this author’s opinion have opted for an 8-MHz-channel. But that is hindsight!”
The 625-line system was initially developed by the Russians circa 1944 as a 50 Hz field frequency derivative of the NTSC system, retaining many features of that system. The line count of 625 was chosen on the basis of reasonably matching the NTSC field frequency. Absent a pre-existing channel width constraint, the Russians chose an 8 MHz channel with 6.5 MHz intercarrier and a vision bandwidth of 6 MHz to provide a suitable level of horizontal definition, said to be that which matched what was obtainable from 16 mm film. For reasons unknown, the vestigial sideband was kept at the NTSC number of 0.75 MHz, rather than increased proportionally.
The Russians were said to have presented their 625/50 system at the ITU Atlantic City 1947 meeting, where it was proposed as a standard, although as yet I haven’t yet found any mention of it in the meeting documents that I have looked through.
Post-WWII the Germans, including W. Bruch, picked up on the Russian work. It would appear – but needs to be verified – that they either accepted or endorsed the Russian choice of vision bandwidth and channel width.
However, others thought that an 8 MHz channel was too extravagant, and looked to the American 6 MHz channel as being preferable, at least from the viewpoint of getting an adequate number of channels into the assigned VHF bands. Philips was in that camp, as is evident from this submission that it made at the CCIR Stockholm 1948 meeting:
Out of that thinking came the Philips 567/50 system, apparently derived as the best 50 Hz system fit to the American 6 MHz channel, using the same general approach as the NTSC. If one does the numbers, then the (Iine-count squared x field frequency) products are closely similar. I should imagine though that whilst the Western European countries might have been reluctant to embrace the Russian 625/50 system as such, perhaps for political as well as channel width reasons, then neither were they likely to settle for a standard system with a lower line-count. Still, that Philips developed the 567/50 system does lead to the inference that there was at least some unease at squeezing 625/50 into a 6 MHz channel.
Be that as it may, the American-origin experimental 625/50 transmitter installed at Torino, Italy in 1949 worked in a 6 MHz channel, namely American channel 6, 82-88 MHz. (That was the origin of what eventually became system N, and also of Italian channel C, 81-88 MHz).
Meanwhile the French had gone their own way in 1948 with their high definition 819/50 system that required a 14 MHz channel and had a 10.4 MHz vision bandwidth and 2.0 MHz vestigial sideband. It retained some features of the earlier French 441/50 system, namely positive vision modulation, AM sound, no equalizing pulses and initially anyway, and inverted channel layout with the vision carrier at the high end. By 1952 it had been compressed into a 13.15 MHz channel (essentially by abandoning the outer guard bands) as part of a tête-bêche channelling system.
Exactly when and by whom the 7 MHz channel was proposed I have not seen recorded in a definitive fashion, but as the resultant system was referred to as the Gerber system and sometimes as the Gerber compromise (and it was a compromise), it may well have been his idea. The available evidence suggests that in the final round, as it were, of the CCIR deliberations, the 8 MHz channel had been excluded and the choice was between the 6 MHz and 7 MHz channels.
The original Russian 625/50 system (later system D) was the first with this line-count to see regular service, in 1948. Next appears to have been the 6 MHz channel version (later system N), with regular service started in Argentina in 1951, I think before any regular public use of what became system B in Europe. It would appear that the Latin American countries with 50 Hz electricity supplies had opted for a system that could use the established American 6 MHz channels. From the viewpoint of not creating mutual interference problems with their 60 Hz neighbours who were or were planning to use the NTSC system, that may have been a wise move.
Belgium had its own problem, in that it needed a transmitter network that could relay incoming programmes both from France on 819 lines and from the Netherlands on 625 lines, with any given transmitter being able to switch between the two systems. This required a common channelling system, so the 819/50 system was squeezed into the 7 MHz 625/50 channel. Why Belgium opted for positive vision modulation and AM sound is unclear. Clearly, both 625/50 and 819/50 had to be the same in this regard. Perhaps it was a case of even-handedness, in that as the Belgian 819/50 system was different to the French prototype, so the Belgian 625/50 variant had to be different to the European prototype. Or perhaps there were technical reasons. (As an aside, Belgium was the only European country, other than the UK, to seriously consider using AM for its VHF sound broadcasting service.)
By 1953, there were four variants of the 625/50 system in service, namely what later became systems B, C, D and N.
The next significant step was the planning of the European UHF TV services and channel allocations, from the mid-1950s. Here the past error of the 7 MHz channel was recognized, and early on there was universal agreement on the use of an 8 MHz channel, with the vision carrier positioned at 1.25 MHz above the channel lower edge for all systems. Those countries already using the 7 MHz channel at VHF were more-or-less stuck with what they had, although they were not averse to putting the 7 MHz system into an 8 MHz channel, as that provided best utilization of the UHF spectrum. Some I think entertained the idea of doing differently at UHF, but perhaps not too seriously. Belgium though did consider what became system I as a possibility.
France and the UK were both unconstrained by existing 625/50 systems, and both initially planned on fully utilizing the 8 MHz channel with 6.5 MHz intercarrier and 6 MHz vision bandwidth. In the UK case, there was a late change, recommended by the TAC in 1960, to the use of a 6.0 MHz intercarrier, with 5.5 MHz vision bandwidth and 1.25 MHz vestigial sideband. The reasoning was that in terms of picture quality on typical domestic receivers, the gain from the extended vestigial sideband (allowing a gentler Nyquist slope) was greater than that from extending the vision bandwidth from 5.5 to 6.0 MHz. That was probably true in the days of distributed L-C selectivity IF strips and diode demodulators, but much less so later on when SAWFs and synchronous demodulation became the norm.
France elected to use both the 6.0 MHz vision bandwidth and the 1.25 MHz vestigial sideband, thus using up the outer guard band space in the 8 MHz channel, as it had done with fitting 819/50 into a 13.15 MHz channel. One argument advanced for the use of a 6.0 MHz vision bandwidth was that when a colour subcarrier was added at the anticipated 4.43 MHz or thereabouts, it would allow full double sideband working, i.e. there would be no early attenuation of the upper sideband. This was probably advantageous with the NTSC system, but may have been more important for SECAM. As originally conceived, SECAM used simple AM subcarriers, in which any sideband asymmetry resulted in quadrature distortion with diode-type demodulators. I am not sure when the change was made to FM subcarriers, but even there, sideband asymmetry would result in distortion. As well, France opted for positive vision modulation and AM sound, as had been used for its 819/50 system E, to simplify the design of dual-standard receivers.
The 1.25 MHz vestigial sideband was also adopted at UHF by some Western European countries who otherwise retained the system B parameters. In this guise it was known as system H. And the Eastern European countries did consider adopting the 1.25 MHz vestigial sideband, although that did not happen in practice.
By the early 1960s, there were seven 625/50 variants, namely:
Western European B/G
Belgian C
Eastern European D/K
Western European H
UK, Ireland I
French L
South American N
The “final round”, as it were, occurred in connection with the setting of TV standards for the African Broadcasting area, at the ITU 1963 Geneva meeting. By then some African countries had already adopted and so were committed to system B. For the French Outré Mer territories, a negative/FM version of system L was developed, which was subsequently allocated letter designation K1. Without the need to consider dual-standard receivers for the African territories, the French could revert to the by-then standard negative/FM combination. The same arguments as for system L were used in support of the 6 MHz vision bandwidth and 8 MHz channel, and unsurprisingly the OIRT also advocated for this combination. South Africa, though, opted for system I. Countries not already committed to system B had the option of choosing an optimized rather than a compromised 625/50 system, and I guess that different groups of engineers had different views as to what constituted an optimized system within an 8 MHz channel.
Now there were eight 625/50 variants in use:
Western European B/G
Belgian C
Eastern European D/K
Western European H
UK, Ireland I
French Outré Mer K1
French L
South American N
Some sub-variants (systems B1, D1 and I1) were added right at the end of the analogue era, but these did not materially alter the overall picture.
In hindsight one could say that the Russians got it nearly right, missing only the proportional adjustment of the vestigial sideband, something that could have been remedied later on without too much pain. Had the Russian system been adopted in Western Europe, then I speculate that we might not have seen the subsequent multiplicity of 625/50 variants, and by the time that UHF transmissions arrived, it may have had too much inertia for anyone to seriously consider alternatives, which in any event would have offered relatively minor improvements. Even the French may have baulked at positive/AM had the rest of Western Europe long been using an 8 MHz channel, 6 MHz vision bandwidth negative/FM combination. However, the Western European compromise left wide open the door for doing better at UHF, and there were different views as to what “better” looked like, with no weight of established practice – in Western Europe anyway - to favour any one. Perhaps all that can be said is that for those countries new to 625/50, not to have done better than system B/G would have been a worse choice than doing better by any of the means proposed.
Be that as it may, system N may have been used anyway if the imperative to use a 6 MHz channel in Argentina and elsewhere in South America superseded all other considerations.
(*) The Donald Fink paper referred to is: “Perspectives on Television: The Role Played by the Two NTSC’s in Preparing Television Service for the American Public”. Apparently it was first published in the IEEE proceedings in 1976 September, then reprinted in 1999 as a Classic Paper. It is recommended reading from one of the outstanding TV systems and standards pioneers who was also an excellent technical writer.
Cheers,
Steve
