>Rob Danielson, you wrote,
>
>>Some numbers that have proven to be very useful estimates for us are:
>>
>>Consumer MD mic pres with the gain set at maximum have an equivalent
>>self noise of 20-25dB(A).The Walkman DAT's are in the same
>>neighborhood.
><snip>
>
>Can you explain how this specification is derived?
>
>-Dan Dugan
>
Hi Dan--
Remember, max gain settings are the reference-- I looked at worst
case situations when noise from the mics and/or the pre has become
the primary obstacle to a better recording. Results at lower mic pre
record level settings would not apply.
These numbers have been borne out over the last 2 years with other
mics with other self noise ratings but the most specific test I did
was this: I took 2-Rode NT-3's with 17dB(A) self noise and a Sony
MS-957 with 25dB(A) self noise and made 1-20 minute recording of each
mic(s) on a sharp mt-90 at mic record level of 30 in at 2am in the
city (~35dB background level as I recall).
I later made a third recording running the rodes through an mp2 at
full gain and going line in on the MD recorder ("30" or unattenuated
line input level).
I transferred the first two files and located the quiet sections with
consistent saturations (about 10% for the Rodes as I recall). I
exported files of each and normalized each of these to -12 dB peaks.
I played the two files at a comfortable listening level through an
RME D-A and through a hafler power amp and yamaha reference monitors.
I played them alternating 2 second clips from each and also split
them up with one the Left channel and one on the Right. The files
made from the MS-957 had more noise and the spectrum of that noise
leaned towards the top end, 5Hz and up as I recall. (this was first
by ear and I did use audiograph but I don't remember it telling me
al ot). The noise from the file made with the Rode NT-3's seemed more
widely dispersed..
I took the third recording made with the Rodes going through a
cranked mp-2 [self noise ~ 5dB(A)] and normalized a quiet section to
-12dB peaks (the orig was close I recall because the mp2 has more
gain). I compared this file (made to better identify the noise
produced by the Rode NT-3) to that of the NT-3 recording made through
the MD mic pre. The noise produced by each was significantly
different in sonic character. The noise from the Rode NT3/MD mic pre
seemed crustier with more ~5-10K hiss low, more low end diffusion.
The noise of the Rodes/MP-2 seemed to have a thinner, higher hiss
component.
I deduced that the louder, crustier noise was that of the MD mic pre
and the rode mics mixed-- with a greater part of noise from the MD
mic pre because it of the difference in sonic character. (It
reminded me of the noise I get from the Rodes into my D7 mic pre,
cranked up). I felt comfortable assuming that mic pre noise was
somewhere in between the self noise figures for the rodes and the
sony mic and began using 20-22dB(A) as a guide number. If anything, I
think it could be higher,.. remember, when the pre gain is maximum.
A mic pre with a self noise of 22dB(A) is not too bad when you
consider that the mic2496 comutes to 22, the Tascam US-122 is 32; and
the $500 Really Nice Preamp is 12. (Bob Cain did these, not me)
Though we don't know if the manuf's numbers are made with the gain
cranked to max,.. There are still many inconsistencies in my working
numbers and manuf data so if anyone has thoughts or know of other
self noise measurements for common internal or portable mic preamps,
I'd appreciate them.
Rob D.
=3D =3D =3D =3D
--
Rob Danielson
Film Department
University of Wisconsin-Milwaukee
[Non-text portions of this message have been removed]
________________________________________________________________________
________________________________________________________________________
|