#59: Thanks Stunsworth, for that link, I had seen it before. That set of articles is OK as far as they go but they dwell on the software and have not considered the analogue part of what they are doing as well as they might. They seem to think that RIAA is applicable to all sorts of EQ shapes too which it isn't, its a specific, defined standard.
Yes, that was the case with older discs. RIAA is a standardised form of pre-emphasis for discs of all kinds, 78 included, and was agreed and adopted in about 1954 and has been tinkered with since but is essentially the same now. Before 1954 RIAA did not exist and all companies had bespoke pre emphasis curves which some old pre amps deal with and now these software solutions appear to as well as part of their pull down menus. As well as the obligatory RIAA the cutting engineers certainly applied "equalisation" or EQ which was “adjust to taste” [house sound style] or to the known foibles of the cutting lathe in use and the known behaviour of vinyl pressing. Here’s a picture of the Abbey Road EQ box:
Yes, it is inherent in the amplitude shaping. Any network made of Rs, Ls and C s having a defined amplitude response has a corresponding defined phase response which is called its “minimum phase”. It is not to be confused with “linear phase” which requires that the network exhibits constant DELAY at all frequencies regardless of the amplitude response. To do this extra network elements are needed that will not affect the amplitude response but will adjust the phase.
In the “old days” when everything was analogue the networks that did the shaping at the cutter and also those complementary ones in pre –amps used RC networks to do the job. At the pre amp the shape is made by 2 poles and one zero. Here is a sketch of the shaping, the full black line is the nominal playback and the dotted one the record. The blue dotted line is the overall effective playback shape:
The first pole is at about 50 Hz [actually defined by a time constant not a frequency, in this case 3180 microseconds] and starts a linear declining amplitude response with frequency at a slope of 6 dB every octave [20dB per decade]. Below 50 Hz the response is flat unless some rumble control is also built in by the pre amp designer – it is not in the early RIAA but is in the later IEC version. Record companies often added a rumble filter themselves in their cutters [eg to deal with Kingsway Hall tube trains]. At 500 Hz [318 microseconds] there is a zero which breaks the response back to flat again for a short band of frequencies and then another pole resumes the downward linear slope again at 2122 Hz [75 microseconds]. Cutters often have built in HF filters too for protection, well beyond the audible range - I've shown one here at 50KHz. Each pole and zero has a 45 degree phase shift [at the turnover frequency, 90 degree well away from that] with the sense opposite for poles and zeros.
Remember that any phase shifts in the cutter end due to RIAA [the cutter will almost certainly have a complex phase response of its own as well if only from its mechanical and electrical resonances which should be up in the high frequencies well above audio – the same is true of the playback cartridge] will be automatically reversed at the pre amp [provided both are in spec and are implemented similarly] so if these software people use linear phase for the replay part it is a waste of time if the cutter didn’t do so as well. And what about that cartridge??
It seems to me that the case for doing RIAA in software is not proven and depends on other factors outside of the control of the software. That is not to say it cannot be done at all, just that I have doubts that it can be done well enough without control of a lot of additional parameters and that may be very difficult to achieve. It is true that the tolerances of the components in the analogue case need to be considered carefully if the record and playback responses are to match sufficiently well whereas in the digital case more precision is possible.
Originally posted by Dave2002
View Post
Was the phase response well defined in the RIAA curves? Digital signal processing can be very accurate with respect both to phase and to amplitude, though a theoretical accuracy can be blown out of the water if the issues raised about dynamic range dominate, which seems likely.
In the “old days” when everything was analogue the networks that did the shaping at the cutter and also those complementary ones in pre –amps used RC networks to do the job. At the pre amp the shape is made by 2 poles and one zero. Here is a sketch of the shaping, the full black line is the nominal playback and the dotted one the record. The blue dotted line is the overall effective playback shape:
The first pole is at about 50 Hz [actually defined by a time constant not a frequency, in this case 3180 microseconds] and starts a linear declining amplitude response with frequency at a slope of 6 dB every octave [20dB per decade]. Below 50 Hz the response is flat unless some rumble control is also built in by the pre amp designer – it is not in the early RIAA but is in the later IEC version. Record companies often added a rumble filter themselves in their cutters [eg to deal with Kingsway Hall tube trains]. At 500 Hz [318 microseconds] there is a zero which breaks the response back to flat again for a short band of frequencies and then another pole resumes the downward linear slope again at 2122 Hz [75 microseconds]. Cutters often have built in HF filters too for protection, well beyond the audible range - I've shown one here at 50KHz. Each pole and zero has a 45 degree phase shift [at the turnover frequency, 90 degree well away from that] with the sense opposite for poles and zeros.
Remember that any phase shifts in the cutter end due to RIAA [the cutter will almost certainly have a complex phase response of its own as well if only from its mechanical and electrical resonances which should be up in the high frequencies well above audio – the same is true of the playback cartridge] will be automatically reversed at the pre amp [provided both are in spec and are implemented similarly] so if these software people use linear phase for the replay part it is a waste of time if the cutter didn’t do so as well. And what about that cartridge??
It would seem best to do the RIAA in analogue, perhaps with a choice of different RIAA curves, and if wanted to do minor tweaking afterwards as at least the dynamic range limitations would be much more controlled and there'd be more room to play with in the digital domain.
Comment