[LAU] USB audio interface support?

fons at kokkinizita.net fons at kokkinizita.net
Thu Nov 26 05:50:01 EST 2009

On Thu, Nov 26, 2009 at 05:07:16PM +0800, Ng Oon-Ee wrote:

> Sorry to sound like a broken record, but could anyone with knowledge in
> these matters advise me on:-
> 1) How much gain is needed by the Shure SM57 for comfortable recording
> with a low noise floor?
> 2) Whether the M-Audio Fast Track Pro, Tascam US-144, Alesis IO|2 or
> Edirol UA-4FX are able to provide pre-amping with that much gain?

The question to ask is not how much gain would be required.
All these interfaces will provide enough gain to the point
that increasing it will not improve S/N ratio.

The important question is then: how much noise do we have
in those conditions ?

Two values are important for this : the self noise of the mic,
and the equivalent input noise of the preamp.

For the SM57 the self noise is easy to find: 

Sensitivity = 1.9 mV/Pa or -54 dBV/Pa
The reference sound level is -94 dBPa, so the
mic will produce -148 dBV at that level.

The SM57 is a passive microphone, so its self noise
is determined by its impedance which is 310 ohm.
This will produce a noise voltage of around -130 dBV
in the audio range. 

Comparing these two numbers, the result is that the mic
will produce the equivalent of around 18 dB of acoustic
noise. For example if you record a sound of 78 dB SPL,
the S/N ratio will be 60 dB.

This is a typical figure for a mic of this type, which
is not really designed for low-noise use such as recording
classical music or soft instruments. It's used mainly for
solist voice, amplified instruments and drums. The best
low-noise mics would be some 10 dB lower.

The second parameter, equivalent input noise, is the
noise of a preamp, measured in some operational condition,
and assuming all that noise is generated by adding it at
the input. For a good preamp this will be very close to
reality. It will have the best EIN at max gain, and this
level will remain almost constant over a large gain range.

None of the four manufacturers provide this information,
nor any other data from which it could be computed. This
is no coincidence: for most 'consumer' type preamps the
EIN is not very good, and manufacturers want to avoid
that you compare their performance to the competition.

For my Edirol UA-5 I just measured it. For a source (mic)
impedance of 150 ohm it is around -120 dBV, at 310 ohm
it would probably be around -117 dBV. Now if you compare
this to the self noise of the mic, -130 dBV, the conclusion
is that the preamp will produce 13 dB more noise than the
mic. Not very good, but quite typical for any product in
this price range. 

Combining all this, using the SM57 with the Edirol UA-5
would produce a noise level that is equivalent to 31 dB
of acoustic noise. This is, in the end, the only figure
that matters, as it determines what you can record with
a good S/N ratio and what not.



Io lo dico sempre: l'Italia è troppo stretta e lunga.

More information about the Linux-audio-user mailing list