Phono EQ
Greg Berchin
Would this be the proper forum for very specific technical
questions about phonograph equalization? I don't mean just RIAA
EQ; I am looking for every bit of information that I can find
about EQ curves used before RIAA EQ was adopted.
I have found the following Websites that got me started ...
http://sound.westhost.com/project91.htm
http://www.vadlyd.dk/English/RIAA_and_78_RPM_preamp.html
http://www.hagtech.com/equalization.html
http://www.shellac.org/wams/wequal.html
http://www.smartdev.com/LT/compensation.htm
http://www.klaus-boening.de/html/phonostage.html
http://www.rfwilmut.clara.net/repro78/repro.html
http://www.hagtech.com/pdf/riaa.pdf
... but I have some specific questions about implementation that
are not answered in these places.
If anyone would care to engage in a discussion of this topic, or
to guide me to a place where such a discussion would be more
appropriate, I would appreciate it.
Thank you,
Greg Berchin
Kim_J...@volcanomail.com
and a Harman-Kardon solid-state preamp, all with selectable contours.
Is this info in the Radiotron Designer's book? I'm sure I've got a book
from the fifties, "Hi-Fi Design," something like that, that talks about
this.
Good topic. I miss the days of preamps with user-selectable contours.
Mark Oppat
all this info you mention is just a starting point in my opinion... why?
Because the curve applied only reverses the recording EQ, and does nothing
for the bad mics or uneven recording "mix" as it were, as well as the
imperfect recording heads, etc, etc.
I use the curves as a GUIDE only, then apply my trained ears. I want it to
sound NATURAL as possible, as if the instruments are there before me...which
is what recording is all about. Can all the instruments be heard properly?
Does the finished processing sound natural or forced?
When processing a recording, you have to make those many decisions...
fidelity vs noise is a big one. Adding reverb to a dry recording sometimes
brings the sound to life. Using a dynamic range expander can help.
Nothing beats a good computer program attached to the best analog gear you
can get. Get several size styluses, 2.7 , 3.0 and 3.3 are best, and a
decent turntable (the Numark TT1 and TT2 are good value today on eBay) .
We have a Packburn Preamp for playback... its designed for 78's.
Anyways, my buddy Dan Gutowski is the real "in the trench" guy on this. He
charges customers $1/ minute for finished work and its worth every penny
from what I have heard. He is at dg16ms26 at MSN dot com.
Mark Oppat
Antique Audio
"Greg Berchin" <76145...@compuswerve.com> wrote in message
news:gtpc71tm3mft925cg...@4ax.com...
Greg Berchin
<mop...@comcast.NOSPAMnet> wrote:
>>RE: 78 RPM processing...
>>all this info you mention is just a starting point in my opinion... why?
>>
>>Because the curve applied only reverses the recording EQ, and does nothing
>>for the bad mics or uneven recording "mix" as it were, as well as the
>>imperfect recording heads, etc, etc.
You are absolutely correct. But for the time being, at least, I
do not wish to address anything but recording/playback EQ. The
info in the links that I included in my original post is good, but
it is far from complete. I am trying first to fill in the gaps in
the technical info. After that, perhaps, I can address the issues
that you mention.
(Or perhaps not. The methods that you discuss are more artistic
than mathematical. My specialty is the mathematics. Whether I
can also serve as "artist" is debatable. But perhaps I can give
others, who really are "artists", the tools that they need.)
Thank you for your input,
Greg Berchin
Mark Oppat
you can not do this mathmatically, if so, all these recordings would be
perfect, and that just doesnt exist in musical recording. FORGET
mathmatics, it means nothing in this game, honestly. Use your mind and
your ear. Its challenging... but its the only way... or, hire an
experienced restorer to work with you.
What records are you trying to process that the information given is not
"complete"? What specifically are the "gaps" you seek young grasshopper?
You sound like the uber-nerd type, and you have to cast off those
shackles.
Remember that music was often recorded in odd spaces back then, too.
Orchestras jammed into "studios", jazz bands in what amounted to closets.
And, they recorded them DRY to provide more clarity on primitive machines.
You have to apply SOME "artistic" input to make these old recordings sound
natural and alive as possible. Looking at charts is but a small baby step
in the right direction, but mostly I ignore them and apply my years of live
listening and sound mixing experience. Its entirely subjective, of course.
BTW, a note to all: Get out and see more LIVE music. Its still the best,
usually. Especially the smaller venues that can really use your support. I
always make it a point to hit some when at "away" radio swaps. My fav is
the Neighborhood Theatre (usually folky and bluegrass) and Double Door
blues club in Charlotte NC when I am at the March CHarlotte AWA. In
Rochester NY the Dinosauer BBQ downtown has free live bands on Tues and
Weds, and great food. Lots of clubs in Baltimore (Fells Point, etc) near
the MAARC swap.
Mark Oppat
"Greg Berchin" <76145...@compuswerve.com> wrote in message
news:lb9d71960o80ds3so...@4ax.com...
Greg Berchin
<mop...@comcast.NOSPAMnet> wrote:
>>you can not do this mathmatically,
Mathematics described the recording EQ. Mathematics describe the
playback EQ. I am looking for the info about the EQ curves that
were supposed to have been used, to at least have a starting
point.
>>What specifically are the "gaps" you seek young grasshopper?
Okay.
Example: Many of the curves specify a turnover frequency below
which bass is boosted, but they do not specify at what frequency
this boost stops increasing. It is not possible for it to
increase at 6 dB per octave all the way down to DC; that would be
a pure integrator which has infinite gain at DC.
Example: Some of the specifications give impossible combinations,
like a bass turnover frequency of 150 Hz, 6 dB per octave slope,
with 15 dB boost at 50 Hz.
These are the kinds of gaps that I am trying to fill.
>> You sound like the uber-nerd type,
Give me a break. I can't tell you how many artistic types I've
worked with who didn't have a clue how much math went into their
art. We mathematician/engineer types give you tools that shield
you from all the math. We must do an awfully good job of it, if
you really think that mathematics don't matter.
>>You have to apply SOME "artistic" input to make these old recordings sound
>>natural and alive as possible.
I already acknowledged that. I am trying to give people like YOU
the tools that you need, so that you don't have to shoot in the
dark.
Greg Berchin
Mark Robinson
What are your specific questions?
Mark
"Greg Berchin" <76145...@compuswerve.com> wrote in message
news:gtpc71tm3mft925cg...@4ax.com...
Randy or Sherry Guttery
> I already acknowledged that. I am trying to give people like YOU
> the tools that you need, so that you don't have to shoot in the
> dark.
OK I'll bite - just exactly what kind of "tools" are you trying to "give
us"? What will it/they do that the "tools" already available won't do?
For example - take just one program I have - Adobe's Audition:
tremendous processing power; all kinds of "tools" for analyzing and
working on/with sound. Pretty nice program all on it's own.
Then add a few VST plug-ins - maybe a dozen different kinds of EQ;
(Parametric; pass; notch; shelf, etc.) reverbs; gates; companders; and
so on (I have about a hundred VST plug-ins currently).
And if you want to get REAL serious about restoration - then get Raygun;
or if your budget will handle it - something like Bias' SoundSoap Pro
plug-in ($600 -- $500 at Sweetwater Sound)-- which is specifically
designed for restoration work on records and other older technologies -
which can "learn" defects in a particular source and address those
without distorting / damaging the program material.
So - what "tools" are you intending to add to these already available?
best regards...
--
randy guttery
A Tender Tale - a page dedicated to those Ships and Crews
so vital to the United States Silent Service:
http://tendertale.com
Greg Berchin
<comc...@bellsouth.net> wrote:
>>OK I'll bite - just exactly what kind of "tools" are you trying to "give
>>us"? What will it/they do that the "tools" already available won't do?
If you are completely happy with the existing tools that are
available to you, then by all means, continue to use them. I
guess there is no room for improvement.
Greg Berchin
Greg Berchin
<mark...@verizon.net> wrote:
>>Hi Greg,
>>
>>What are your specific questions?
Thank you. I'll list a few as a start. The first two I already
mentioned in another post:
- Many of the curves specify a turnover frequency below which bass
is boosted, but they do not specify at what frequency this boost
stops increasing. It is not possible for it to increase at 6 dB
per octave all the way down to DC; that would be a pure integrator
which has infinite gain at DC. I am looking for any info that I
can find that would allow me to determine where each manufacturer
placed this "lower bass tip" frequency. Otherwise I'll just have
to set it to something not-too-unreasonable, like 10 Hz.
- Some of the specifications give impossible combinations, like a
bass turnover frequency of 150 Hz, 6 dB per octave slope, with 15
dB boost at 50 Hz. I am pretty sure that all of the EQ curves
used first-order (6 dB per octave) slope everywhere, so either the
turnover frequency has to be incorrect or the boost number has to
be incorrect. Any specific information that could help me to
determine which is correct and which is not would be appreciated.
- When a specific boost or cut is listed at a specific frequency,
is that the actual boost/cut produced by the real network, or the
approximate boost/cut predicted by the Bode plot?
- Two of the sources indicate that Neumann cutters added a fourth
high frequency corner with a time constant of 3.18 猶. This leads
to a number of questions: Did other cutter manufacturers do
something similar, and if so, what were the time constants that
they used? And, if the electronic rolloff was not introduced, at
what frequency did the mechanical rolloff become significant? And
was the mechanical rolloff just a rolloff (1st order), or was it a
resonance (2nd order)? If it was a resonance, what was its Q?
- Speaking of mechanical resonances, cutting machines must have a
low-frequency resonance below 20 Hz, just like a tonearm/cartridge
system does. What are the characteristics of that resonance? (Of
course, every system will be different.)
- Etc., etc., etc. But these are the questions that come off the
top of my head.
Many thanks,
Greg Berchin
Mark Robinson
I'll try to answer as many as I can to best to of my ability. Hopefully
others will jump in with better (or corrected) answers. See my comments
below. Are you trying to design analog or digital versions of these EQ's?
Also, have you looked at the Radiotron Designers Handbook? This source has
a good deal of technical data on disc record/playback EQ of the older era
and a detailed bibliography of sources.
> - Many of the curves specify a turnover frequency below which bass
> is boosted, but they do not specify at what frequency this boost
> stops increasing. It is not possible for it to increase at 6 dB
> per octave all the way down to DC; that would be a pure integrator
> which has infinite gain at DC. I am looking for any info that I
> can find that would allow me to determine where each manufacturer
> placed this "lower bass tip" frequency. Otherwise I'll just have
> to set it to something not-too-unreasonable, like 10 Hz.
You are correct. In the old days, these curves were not as well specified
as the modern RIAA curve. Clearly, you need to stop the boost at some
point. Most of these older recordings will not have much going on down that
low and the surfaces will have a good deal of rumble as well. So, you while
you can set a lower breakpoint down to 10hz, that might be too low. I'd
experiment to see what works or maybe allow this to be adjustable as needed.
>
> - Some of the specifications give impossible combinations, like a
> bass turnover frequency of 150 Hz, 6 dB per octave slope, with 15
> dB boost at 50 Hz. I am pretty sure that all of the EQ curves
> used first-order (6 dB per octave) slope everywhere, so either the
> turnover frequency has to be incorrect or the boost number has to
> be incorrect. Any specific information that could help me to
> determine which is correct and which is not would be appreciated.
AFIK, you are correct that all of the slopes used are 1st order, so either
you calculating wrong or there were errors in the specs.
> - When a specific boost or cut is listed at a specific frequency,
> is that the actual boost/cut produced by the real network, or the
> approximate boost/cut predicted by the Bode plot?
I believe it is the actual boost/cut at the noted frequency. The issue with
specifying things this way is the 0db reference point used. The time
constant approach avoids all of the questions since it specifies the
pole/zero locations directly. If you do a google search on the archives of
the group, you will find a discussion (maybe last year ??) about the use of
using the simplistic Bode plot lines as a desired curve to match.
> - Two of the sources indicate that Neumann cutters added a fourth
> high frequency corner with a time constant of 3.18 猶. This leads
> to a number of questions: Did other cutter manufacturers do
> something similar, and if so, what were the time constants that
> they used? And, if the electronic rolloff was not introduced, at
> what frequency did the mechanical rolloff become significant? And
> was the mechanical rolloff just a rolloff (1st order), or was it a
> resonance (2nd order)? If it was a resonance, what was its Q?
I think this is a similar issue with the low frequency stop point. There
may have been also correction used at the cutter to compensate for
losses/errors, but the end result needs to be flat when played back with the
specified curves. I don't think you njeed to be concerned about this.
> - Speaking of mechanical resonances, cutting machines must have a
> low-frequency resonance below 20 Hz, just like a tonearm/cartridge
> system does. What are the characteristics of that resonance? (Of
> course, every system will be different.)
I'm not sure if this is so since the cutter is rigidly mounted to the lead
screw assy. I would think a well designed cutter would try hard to
eliminate any bad system resonances. If not, you should be able to see them
in an averaged spectral plot of a disc.
> - Etc., etc., etc. But these are the questions that come off the
> top of my head.
Hope some of this is of use to you.
Mark
John Byrns
<mop...@comcast.NOSPAMnet> wrote:
> RE: 78 RPM processing...
> all this info you mention is just a starting point in my opinion... why?
>
> Because the curve applied only reverses the recording EQ, and does nothing
> for the bad mics or uneven recording "mix" as it were, as well as the
> imperfect recording heads, etc, etc.
This may be true of the "bad mics" and "uneven recording mix", but I don't
believe it was always the case with the "imperfect recording heads". My
understanding is that the characteristics of a specific cutter head were
taken into account in the actual electrical equalization applied in the
recording amplifier, so that the net result of the cutting system would
follow the recording curve being used. Different cutter heads got
different equalizer networks. Also I am given to understand that in some
of the early cutter systems the equalization was all done in the cutter
head mechanically, that would mainly be the low frequency roll off, in the
days before high frequency pre emphasis came into vogue.
The AES has several Anthologies that describe the history of record
cutting, and contain much interesting information.
> I use the curves as a GUIDE only, then apply my trained ears. I want it to
> sound NATURAL as possible, as if the instruments are there before me...which
> is what recording is all about. Can all the instruments be heard properly?
> Does the finished processing sound natural or forced?
>
> When processing a recording, you have to make those many decisions...
> fidelity vs noise is a big one. Adding reverb to a dry recording sometimes
> brings the sound to life. Using a dynamic range expander can help.
Adding reverb and other "artistic" inputs seems like a bad idea to me.
What happens if a few years down the road your "artistic" sensibilities
change and you no longer want the reverb? I would think a better plan
would be to first transcribe the disc to modern media, getting the
frequency balance as close as possible to the original, and removing
clicks, pops, and etc. Then save the result unscathed for future
"artistic" processing onto other media as your tastes of the moment may
dictate.
On the subject of equalization curves, I seem to remember seeing some that
specify a 3 dB per octave roll off. Does anyone know how the 3 dB roll
off was implemented, I suppose you could use a filter similar to those
used to generate pink noise, but that all seems rather complex, and I have
never seen such a circuit used in a phono preamp. Of course I assume home
phonos were simply voiced to sound "good" irrespective of the actual
recording curve, even if several were provided. Vintage professional
phonograph equipment seems to have used passive equalizers which consisted
of a box full of Inductors and Capacitors connected between the phono
pickup cartridge and a flat microphone input on a console. Resonant L and
C combinations could easily provide slopes greater than 6 dB per octave.
Regards,
John Byrns
Surf my web pages at, http://users.rcn.com/jbyrns/
Bruce Mercer
> > When processing a recording, you have to make those many decisions...
> > fidelity vs noise is a big one. Adding reverb to a dry recording
sometimes
> > brings the sound to life. Using a dynamic range expander can help.
>
> Adding reverb and other "artistic" inputs seems like a bad idea to me.
> What happens if a few years down the road your "artistic" sensibilities
> change and you no longer want the reverb? I would think a better plan
> would be to first transcribe the disc to modern media, getting the
> frequency balance as close as possible to the original, and removing
> clicks, pops, and etc. Then save the result unscathed for future
> "artistic" processing onto other media as your tastes of the moment may
> dictate.
ABSOLUTELY! Reverb? Oh my god, NO.
Bruce
Ken Doyle
reverb.
Lots of music sounds really neat recorded in a dead room.
Often, people weren't trying to make realistic records, they were making
cool sounding records, and certain "restorers", destroy the sound that sold
the record in the first place!
Ken D.
Mark Oppat
Greg wrote>>>>"> I already acknowledged that. I am trying to give people
like YOU
> the tools that you need, so that you don't have to shoot in the
> dark.
>
Fair enough, Greg. I dont mean to say all numbers mean nothing here. Just
that I hate it when I see audiophools totally absorbed in the details of the
hardware and ignore the whole reason for all of it:
THE MUSIC! As if knowing the "correct" perfect numbers will produce the
"best" playback...NOT!!!
I just want to bonk so many of them on the head, sort of like "Capt Kirk"
in the Saturday Night Live spoof of the Trekkies convention... "and, you,
have you ever kissed a girl"...that sort of thing.
Sounded like you were (are) headed in that direction... just want to keep
that extremism from foiling the ultimate goal...reproduction of a
performance. Or, in some cases, creation of a new version of a
performance... ala a DJ "remixing" a pop hit.
You can look at these numbers all you want, but in the past recording was
done in a million ways and there are dozens of other variables besides the
"EQ" curves. No matter what, you have to apply your mind and your ears. I
dont care what the numbers say.
And, someone else said., regarding adding reverb..... "what if later you
want to do it different" . OK.. so do it different again. As long as you
keep a dry copy, on CD or digitally, or the original disc...or whatever way
you want to archive... again... there are many many ways to do all this.
And, I am assuming we are not talking about commercial releases of what you
are doing anyways. But even it so, some engineer has to make a decision
as to what sounds "good" when those are done. Some do well, others, not so
well.
Mark Oppat
Greg Berchin
<mark...@verizon.net> wrote:
>>Are you trying to design analog or digital versions of these EQ's?
Digital. It offers the ability perform a single pass in which the
signal is sampled "dry" and archived. Then various EQ curves can
be tried without risking the original, and without generation
loss. I also have some tricks up my sleeve when it comes to
emulating analog filters in the digital domain -- beyond the
traditional bilinear transform or impulse invariance methods.
>>Also, have you looked at the Radiotron Designers Handbook?
No, but you are the second person to recommend it, so I am looking
forward to finding a copy.
>>You are correct. In the old days, these curves were not as well
>>specified as the modern RIAA curve. Clearly, you need to stop the
>>boost at some point.
As a practical matter; yes. Still, if the emphasis curve had a
true highpass characteristic (zero response at DC), then the
"correct" de-emphasis curve needs an integrator to compensate
properly. If the boost is stopped at, say, 10 or 20 Hz, then the
filter will be stable, but the phase response will be messed-up.
So there has to be a compromise, and I wonder if anybody really
cares about a few degrees of phase error at such low frequencies.
>>I'd experiment to see what works or maybe allow this to be adjustable
>>as needed.
Ultimately this may be the right answer to all questions.
>>AFIK, you are correct that all of the slopes used are 1st order, so
>>either you calculating wrong or there were errors in the specs.
I believe that there must be errors in the specs. Fifty Hz is
1.58 octaves below 150 Hz. With a slope of 6 dB per octave, the
boost will be 9.5 dB, not the specified 15 dB. So either the
turnover frequency is specified incorrectly, or the boost is
specified incorrectly. And I don't know which.
>>I believe it is the actual boost/cut at the noted frequency. The
>>issue with specifying things this way is the 0db reference point used.
>>The time constant approach avoids all of the questions since it
>>specifies the pole/zero locations directly.
Yes! I agree! Unfortunately I don't have time constants, only
corner frequencies and boost/cut quantities.
Haven't had a chance to do the google search that you mentioned,
but I will.
>>I think this is a similar issue with the low frequency stop point.
>>There may have been also correction used at the cutter to compensate
>>for losses/errors, but the end result needs to be flat when played
>>back with the specified curves. I don't think you njeed to be
>>concerned about this.
Good point. I may be a little overenthusiastic in my pursuit of
"perfection"!
>>Hope some of this is of use to you.
It certainly is. Thank you.
Greg
Mark Robinson
I did the same thing you are attempting. I wrote a VST plugin with
adjustable turnovers so that I could correct for any curve. I use
Audiomulch as the VST host for processing the audio. Any VST host would
work, but I like AudioMulch.
I built up a small PCB that I mounted inside my turntable. Its a flat
pre-amp to boost the mag cart (and properly terminate it) to line level so
that it can feed my audio card (RME). I used the matched z transform for my
filter rather than the bilinear transform due to the frequency warping the
later method produces. I can provide a Mathcad sheet with a comparison of
the analog vs. digital filter for the standard RIAA curve as 44.1khz sample
rate. If you are interested, e-mail me directly and I'll send it to you.
If you don't have Mathcad, I can also print it as a PDF. In this case you
won't have the "live" features of a Mathcad sheet, but it sill may be
useful. I can also provide the c source for my plugin.
> I believe that there must be errors in the specs. Fifty Hz is
> 1.58 octaves below 150 Hz. With a slope of 6 dB per octave, the
> boost will be 9.5 dB, not the specified 15 dB. So either the
> turnover frequency is specified incorrectly, or the boost is
> specified incorrectly. And I don't know which.
> >>I believe it is the actual boost/cut at the noted frequency. The
> >>issue with specifying things this way is the 0db reference point used.
> >>The time constant approach avoids all of the questions since it
> >>specifies the pole/zero locations directly.
>
> Yes! I agree! Unfortunately I don't have time constants, only
> corner frequencies and boost/cut quantities.
I'd go with the corner freq's and calc the time constants by tc = 1/(2 * PI
* turnover). I suspect the boost is not referenced to the 150hz point on
the graph, but instead to some midrange frequency reference of 1Khz or so.
I guessing that this point is 5db lower than the 150hz point and that
accounts for the error you point out.
Mark
Greg Berchin
<mark...@verizon.net> wrote:
>>I built up a small PCB that I mounted inside my turntable. Its a flat
>>pre-amp to boost the mag cart (and properly terminate it) to line level so
>>that it can feed my audio card (RME).
I'm thinking about something more-or-less similar, but using one
of the new generation of super-high-quality analog-to-digital
converters.
>>I used the matched z transform for my
>>filter rather than the bilinear transform due to the frequency warping the
>>later method produces. I can provide a Mathcad sheet with a comparison of
>>the analog vs. digital filter for the standard RIAA curve as 44.1khz sample
>>rate.
Traditional thought is that matched-Z-transform can do pretty well
with low-order systems (which these EQ curves are) and high
sampling rates (which 44.1 kHz is not). You probably match the
amplitude curve pretty well. I wonder how well you match phase?
>>I suspect the boost is not referenced to the 150hz point on
>>the graph, but instead to some midrange frequency reference of 1Khz or so.
Could be. Or it could just be that some of the information has
been lost, after more than fifty years. That's what makes this
frustrating.
Thanks,
Greg
Mark Robinson
You are right about the sample rate being a limitation. The magnitude
tracks very well until you get up towards the Nyquist point. IIRC, the
error approaches 3db at 20 khz. At 96Khz SR, the error is very small. I
did not check phase response ( I probably should). I'm no expert in filter
design. This was something I threw together quickly as a learning
experience.
Do you have any other suggestions for better transformation methods from s
domain to z? Is your plan to make a stand alone device rather than relying
on a desktop PC?
Since we know the basic filter topology required and the digital EQ is
easily adjusted, I don't think there is any problem in proceeding. The nice
thing about the digital implementation is that as better info becomes
available, you can easily apply it to any transfers. It also allows you to
adjust for any funny business by the different record companies by listening
and analyzing the results.
Mark
Tim Mullen
>Fair enough, Greg. I dont mean to say all numbers mean nothing here. Just
>that I hate it when I see audiophools totally absorbed in the details of the
>hardware and ignore the whole reason for all of it:
>THE MUSIC! As if knowing the "correct" perfect numbers will produce the
>"best" playback...NOT!!!
"Audiophoolery" got started, to a large extent, as a reaction
against products that measured well but sounded =terrible=. There
was a race for lowest THD, etc., etc., all while running test tones
through load resistors. I give the "phools" credit for the idea of
using live, acoustic music as a baseline.
>I just want to bonk so many of them on the head, sort of like "Capt Kirk"
>in the Saturday Night Live spoof of the Trekkies convention... "and, you,
>have you ever kissed a girl"...that sort of thing.
"Those two girls have the exact same measurements! Any difference
between them is all in your imagination!"
--
Tim Mullen
------------------------------------------------------------------
Am I in your basement? Looking for antique televisions, fans, etc.
------ finger this account or call anytime: (212)-463-0552 -------
Greg Berchin
<76145...@compuswerve.com> wrote:
>>I am looking for every bit of information that I can find
>>about EQ curves used before RIAA EQ was adopted.
A followup to my previous message:
I have been able to identify the following EQ curves. If anyone knows
of additional curves, or finds any errors, please post. (See
http://www.rfwilmut.clara.net/repro78/repro.html for definitions of
"Turnover" and "Tip" frequencies.)
I have a digital filter implementation for each of these, at 192 kHz
sampling rate, that matches the ideal analog filter ±0.025 dB amplitude
and ±0.25° phase from DC to over 75 kHz.
- Greg Berchin
Bass Tip/Bass Turnover/Treble Turnover combinations
---------------------------------------------------
/ 150/ 3.4k Decca
/ 150/ 5.8k early Decca
/ 200/ flat Westrex
/ 200/ 5.8k Columbia 1925
40/ 200/ 6.36k American 1025, Victor 1925 (some)
50/ 250/ flat Blumlein, HMV
/ 250/ flat Columbia (Eng.), EMI 1931
40/ 250/ 6.36k London FFRR 1949, FFRR 78
/ 300/ 1.6k Columbia 1938
/ 300/ 2k FFRR 1951
50/ 353/ 3.18k BSI
/ 375/ 2.5k Decca 1934
/ 375/ 5.8k Decca FFRR 1949, EMI
/ 375/ 6.36k Decca FFRR 1949, EMI, Victor 1925 (some)
70/ 400/ flat early 78 (mid-'30s), US mid 30
/ 400/ 2.5k old AES, Decca 1934, Mercury
100/ 450/ 3k FFRR 1953
/ 500/ flat early 78, Brunswick, Parlophone
100/ 500/ 1590 Columbia LP
/ 500/ 1590 early LP
/ 500/ 1.6k early LP, NAB, NARTB
70/ 500/ 2.5k EMI
/ 500/ 2.5k Capitol 1942, MGM, Victor 1947-1952
100/ 500/ 3k FFRR
50/ 500/ 3.18k CCIR
/ 500/ 3.18k London FFRR
/ 500/ 3.4k Concert Hall until 1952, Oiseau-Lyre until 1954
/ 500/ 5.8k Victor 1938-1947
50.05/500.5/2.122k RIAA
/ 629/ flat "629"
/ 800/ 2.5k early RCA