[ARSCLIST] MC vs MM results (was 2GB limit for audio file formats)

[Eric Jacobs <ericj@xxxxxxxxxxxxxxxxxxx>]

Goran Finberg:

> But this can be done even using a Shure M44-7 using 44.1 kHz sampling
> frequency.
>
> I may not like the scanning loss of the conical stylus nor the tracking
> distortion from it or the unrefined character of the M44-7 but I can
> easily remove the Clicks/Crackle/Noise from the final soundfile.


The ultimate proof is in the results, so let's go there!  Rather than
talk theory, I thought I'd make the discussion more interesting and concrete
by sharing real results from a real test that I performed.  I have performed
many hundreds of hours of tests fine tuning the performance of my transfer
system and processes, and the test I am offering is only one of these tests.

To anyone who can accept 2 MB of email attachments, I offer a collection
of 4 MP3 files that compare:

  - Shure M44-7 raw transfer
  - Shure M44-7 declicked
  - Ortofon SPU-GT raw transfer
  - Ortofon SPU-GT declicked

The work was done at 24/88.2, but obviously the final MP3 is much
lower resolution than that - and the difference in the results is
still very audible in the MP3.  You can even easily hear the
difference on cheap little computer speakers.  Although these tests
are voice transcriptions, I have performed the same tests with
musical work with the same results.

I have many more test files, where different independent variables
were tested to see their effects on the results.  The test results that
I am offering show the difference between MM and MC.  I don't have
the tests handy, but I also have test results showing the difference
between sample rates.

Even if my theory is wrong in the details, the results are real.
And because the experiments were controlled with only one independent
variable changing at a time, the effects of cartridge type on noise
reduction can be clearly heard.

The tests were run with absolutely everything constant in the transfer
system except:

  - cartridge loading
  - phono preamp gain
  - VTF

Even the stylus was identical - a 2.8 mil Truncated Elliptical sourced
from the same vendor.  I kept the noise reduction algorithm settings
the same for both tests, but I can offer anecdotally that no amount
of parameter adjustment on the MM transfer could remove the noise to
the same level as the MC without harming the signal.  And even then,
the noise artifacts from the MM transfer were still audible where the
same artifacts were not audible on the MC transfer.

I must also add that the tests where such dramatic results are
consistently achieved are with:

  - instantaneous mechanical recordings (ie. "acetates", etc.)
  - commercial 78s

On modern clean vinyl in good condition, the differences in noise
removal are far less dramatic and more subtle.

Finally, there are far more variables at work here than just cartridge
type.  The entire system configuration, including the selection of
phono preamp, ADC, and phono cables have an audible impact on the
level of noise reduction that can be achieved.

DISCLAIMER: You will not achieve the same results simply by switching
from an MM to an MC cartridge.  The results of these tests are the
product of a carefully configured system and its settings.  Literally
hundreds of combinations of equipment, settings, and processes have
been tested in my system in order to determine what can be actually
heard versus theoretically heard.  However, within the context of my
system, the difference between the MM and MC results are valid, and
those are the specific test results that are being shared.


Eric Jacobs
The Audio Archive