[linux-audio-dev] Re: applying RIAA curves in software

[fons adriaensen]

On Tue, Oct 25, 2005 at 06:12:41PM -0500, Cornell III, Howard M wrote:

> I stand by my assertion that the RIAA record curve attenuates the low
> frequencies and amplifies the high frequencies.  The physical effects on
> the record are such that the attenuated low frequencies do not cause the
> cutting head trace to take up so much room on the master, and by the
> same reasoning, the normally low amplitudes at high frequencies are
> given more (physical) headroom.  Of course the reverse filter would
> flatten the frequency response and the use of the record filter
> optimizes the use of the master "real estate". 

It depends on how you look at it. 

Suppose you want the same amplitude on the disc (i.e. the same mechanical
displacement of the spiral track) for all frequencies. Then you need to
feed a typical disc cutter head with a signal that rises 6db per octave.
This is because the driving voltage controls its velocity (in the same
way really as with a DC motor), and not its amplitude.

If you play such a disc with a magnetic cartridge without any equaliser,
you willl get a response that again rises 6dB per octave, not because
of the EQ applied in the cutter, but because the cartride delivers a
voltage that is again proportional to velocity rather than amplitude.
To compensate for this, you apply -6dB/oct in the preamp.

So to have a flat frequency response _on disc_, we need to amplify the
HF when recording and attenuate it when playing back. That equalisation
is there purely because of the transducers used at either end.

Compared to such a flat system, the RIAA curve attenuates HF when 
recording and amplifies them on playback. The correction applied is
small compared to the filtering that was already required to have
a flat 'on disc' response, so the net result is still boosting HF
on record, and attenuating it at playback. But relative to the 
flat response, it's not preemphasis, but rather the opposite.

The confusion arises because the 'official' RIAA curves include the
transducer equalisation, or in other words, they are defined in terms
of voltages required to drive the cutter or obtained from the playback
cartridge, and not in terms of mechanical signal amplitude on the disc.

Hope this clears up the matter.

-- 
FA