Re: [ARSCLIST] Non-RIAA preamp

[George Brock-Nannestad <pattac@xxxxxxxx>]

From: Patent Tactics, George Brock-Nannestad

Hi Jacob and others who reproduce mechanical recordings,

Jacob wrote:
> 
> The amplitude (loudness) of a frequency is not a function of the 
> amplitude of the groove, but rather the lateral speed of the 
> stylus. 

----- please use 'velocity' in this case, in which it is directed at right 
angles to the tangent to the groove. This makes it more efficient when we 
also want to talk about the speed of the carrier below the needle (the 
'linear' speed, although it varies 2:1 from outside-in). The relationship you 
refer to is only correct above the corner frequency of the pre/de-emphasis 
curve - this is where constant-velocity recording is aimed at. Constant-
velocity refers to the fact that the groove has the same angle to the 
baseline (non-modulated groove) at a zero crossing (sinusoidal signal) when 
the signal has a particular value, irrespective of the frequency. However, 
this means that the amplitude of the signal will fall proportionally to the 
frequency while maintaining the value. 

 The cartridge with its coil and magnet are effectively
> a "generator" - the faster the lateral movement, the more output
> the cartridge generates at any given frequency.  

----- what you can say is that it has a sensitivity that is proportional to 
frequency, or you could say that it differentiates the curveshape it traces. 
It is hence a velocity detector suitable for a constant-velocity recording.

When you reduce
> the lateral speed, you reduce the total dynamic range - the 
> difference between the softest and loudest sounds.

----- now, when the linear speed of the carrier is halved, then all 
velocities are halved. So if we remain in the constant-velocity domain, there 
is absolutely no change in dynamics.

> 
> When you half-speed transfer (not master), you reduce the output 
> in a way that compresses the relative amplitudes, reducing the 
> dynamic range.  Increasing the gain after adjusting pitch does 
> nothing to the dynamic range and simply offsets the compressed 
> signal.

----- the problem is that recording by the constant-velocity method, you 
cannot go down too far in frequency, because the amplitude increases when you 
go down in frequency, and it cannot increase to the degree that it begins to 
cut into the neighbouring turns of the groove. So, from a given corner 
frequency, the amplitude at recording is reduced inversely proportional to 
the frequency, an we obtain in effect a constant-amplitude recording below 
this corner frequency. The network used in the cutting amplifier is a simple 
first order network - a CR circuit (it is also of the type called "minimum 
phase", so it can be perfectly counteracted by an RC circuit). The corner 
frequency has moved around a bit, but 300 Hz or 250 Hz would be typical. 

----- the compression you perceive is probably due to the fact that when 
reproducing at half speed, you still use a corner frequency at the 
frequencies mentioned rather than 150 Hz or 125 Hz. Or that is what you will 
eventually have to do, if you have used the electromagnetic pickup with its 
constant-velocity characteristic without any de-emphasis.

> 
> The compressed signal is why it sounds "flat".  So some sort of
> expansion is required.
> 
> Half-speed mastering on the other hand takes this compression 
> into account, and they adjust the groove excursion accordingly
> so that the dynamic range is correct when played back at full
> speed.

----- precisely, they use CR circuit adjusted for 150 Hz or 125 Hz at 
recording, which then becomes correct with the electromagnetic pickup and the 
RIAA EQ. Actually it is only the lower part of RIAA, there is further 
manipulation at the high frequency end, but that need not bother us here, 
only we should remember that any half speed work needs filtering at half the 
frequency.

> 
> Perhaps the analogy is that half-speed transfers increase your
> sample rate density, but reduce your bit depth.

----- No, I use half-speed transfer for its better ability to follow 
transients, not the least the grit particles in shellac records, which behave 
like recorded velocities at for instance 100 cm/s. This means that the stylus 
does not have to be accelerated so much (acceleration increases proportional 
to the square of the frequency, and is really the cause of wear), and you do 
not excite spurious resonances in your cartridge so much.

----- finally, the groove shape is there, physically present in the surface 
of the disc. You can extract the information any way you want, and then 
manipulate the information by filtering, but obviously the filtering must 
have relevance for the frequency range you are treating

Kind regards,


George

-------------------------------------------------

> 
> Hope that explanation (and distinction between half-speed 
> masters and half-speed transfers) makes some sense.
> 
> Eric Jacobs
> Principal
> 
> The Audio Archive
> tel: 408.221.2128
> fax: 408.549.9867
> mailto:EricJ@xxxxxxxxxxxxxxxxxxx
> 
> 
> 
> -----Original Message-----
> From: phillip holmes [mailto:insuranceman@xxxxxxxxxx]
> Sent: Tuesday, September 12, 2006 4:24 PM
> To: ericj@xxxxxxxxxxxxxxxxxxx
> Subject: Re: [ARSCLIST] Non-RIAA preamp
> 
> 
> That seems counterintuitive.  It seems like it would be the opposite: 
> playing back at half speed would allow the playback equipment to better 
> track the transients.  I can hear that half speed mastered LPs sound 
> much better to me, and if half speed playback of vinyl sounds "flatter", would
> half speed playback of tape sound the same?  Or is it negated by doubling the
> speed of playback?  Wouldn't the same hold true if you played back a digital
> file at double speed?  Is it possible that it seems "flatter" when in
> actuality it's actually lower distortion, hence it sounds less aggressive (you
> know, lower distortion because the cartridge had an easier time of things)? 
> Have you analyzed things and found that it measures like you hear it?  I know
> I've been tricked into thinking things sound improved when they were actually
> worse (objectively measured to be worse).  Single ended tube amps "sound
> good", but measure awful and with the right music, sounds even worse.  I'd
> like to see a 5 watt single ended amp driving dynamic (cone) loudspeakers
> playing a Virgil Fox spectacular without crapping out (not an issue with horns
> due to efficiency).  Electrostatic speakers "sound" good, and in some ways
> they are lower in distortion (no box resonance, crossover issues, etc..).  But
> the comb filtering is very noticeable to a spectrum analyzer, though not to
> human ears.  I've usually preferred a well mastered and pressed record to a
> CD.  Some of that has to be 2nd order distortions that are added making it
> sound fuller or sweeter.  Some of my prejudice is that the high frequency
> response of a red book CD is not all that great (and MP3 is awful). Hey, I'm
> not contradicting you.  I am interested in why it sounds flatter. Phillip
> 
> Eric Jacobs wrote:
> > On 10/12/06, don cox wrote:
> >   
> >> On 10/09/06, phillip holmes wrote:
> >>     
> >>> In the phase encode/decode discussion of RIAA, did anyone mention that
> >>> speed is critical? If the disk was recorded slow or fast, you'll have a
> >>> hard time getting proper frequency response. 
> >>>       
> >> Playing at the wrong speed affects pitch, but not really frequency
> >> response.
> >>     
> >
> > Often overlooked is the fact that the dynamics change as well when
> > a recording is played back at the incorrect speed.  It is very
> > noticable in more extreme cases of speed change.  If you playback at 
> > half speed and then repitch the recording, it will sound "flatter" than if
> > it were played back at the correct speed.
> >
> > Has anyone tried to expand the dyanamics on a disc played back at 
> > half speed?
> >
> > Eric Jacobs
> > Principal
> >
> > The Audio Archive
> > tel: 408.221.2128
> > fax: 408.549.9867
> > mailto:EricJ@xxxxxxxxxxxxxxxxxxx
> >
> >
> >
> >