[LAU] 3.5mm microphone into audio interface?

David Kastrup dak at gnu.org
Sun Apr 11 13:54:04 CEST 2021

Rodolfo Medina <rodolfo.medina at gmail.com> writes:

> David Kastrup <dak at gnu.org> writes:
>> Fons Adriaensen <fons at linuxaudio.org> writes:
>>> On Sat, Apr 10, 2021 at 12:25:09PM +0000, Rodolfo Medina wrote:
>>>> Thanks to Jeanette and Fons for their help.  Now please one more question:
>>>> I've got two such 3.5mm jack stereo microphones:
>>>>  https://www.amazon.it/...  etc
>>> The link describes this as a MONO omnidirectional mic.
> I said it was stereo because, when I record with it, the rsulting .wav file is
> 2-channel.  Then I concluded it was stereo.

Have you looked at the 2 channels to check whether there are actually
two significantly different significantly non-zero channels?  It is not
usual that the sound card detects how many sources are open or actually
connected, and it would be entirely unusual that the recording software
would act on such information rather than recording at least 2 channels
by default.

> What I want is, say, 4 different mono channels.  For example, if it's
> a string quartet, each instrument should have its own microphone and
> the resulting track should be mono.  If it is a piano record, I want
> two microphones, one for mid-trebles and one for mid-bass, 2 mono
> channels.  If it's voice and piano, I want 3 mono channels and so on.
> Do you think it'll be all right with the Behringer UMC404HD and that
> microphone of mine plus adapter?  That microphone is nice for its
> lightness too.

There is a high probability of muddy trebles, a significant noise floor
and an overemphasis on the 1k–4k frequency range: the main application
of cheap microphones like that is improving voice chat.  Since they have
less size limitations than a microphone built into a laptop and are not
in contact with the case that also holds hard disks (though those are
often SSDs these days) and cooling fans as well as various voltage
converters including for LCD backlight, they have a reasonable chance to
offer significant improvement there.  But it's a decidedly lower bar to
cross than picking up piano sound.

For piano, a good bet for close pickup are small-diaphragm but silent
condensor mics with omnidirectional characteristic.  You don't want to
use too many mics because of comb filtering problems.  For basically any
ensemble placement, you want comparatively directional microphones,
cardioid or hypercardioid (the latter tends to make it a bit easier to
mask other players in an ensemble arrangement at the cost of being more
finicky in placement).

I am going to go out on a limb here and state that I've seen surprising
value from Behringer B5 microphones (comes with omni and cardioid
capsules) for use with virtual rehearsals of an accordion ensemble.
Certainly to the degree where I could not justify getting the Oktava
small diaphragm condenser mics for everyone that I use myself.
Accordion is comparatively tricky in sound quality and easily sounds too
harsh and disharmonious in recordings (also tends to self-destruct under
lossy audio compression at surprisingly high bit rates).

Of course, close-captioning an accordion does not really require a low
noise floor, but the Behringer B5 should rarely be a dealbreaker in that
regard either, at least when considering the typical small diaphragm
condenser use case (you'd use large diaphragm for room/distance
recordings anyway).

>>> If you really want to go into multichannel audio better buy something
>>> of somewhat higher quality.
>> In particular, using omnidirectional microphones for 4-channel audio is
>> not likely to lead to a lot of differentiation unless you work with some
>> clever arrangement of baffles.
> What do you mean please by differentiation?

One channel sounds different from another.  Of course, if you are
placing your microphones per-instrument (rather than as a recording
group), that's not going to be an issue.

David Kastrup

More information about the Linux-audio-user mailing list