Hi all, it is a lot of time I'm not writing to the list, but I read it ever=
y
day.
Concerning the bandwidth of digital recorders and sound boards we need to b=
e
very careful. Specification given by manufacturers don't tell the whole
truth and for extreme uses it is important to make some tests.
On the page http://www.unipv.it/cibra/res_techtest_uk.html you can see the
frequency response and the artifacts of antialiasing filters of some digita=
l
devices I'm using. Larger the "bell" in the graphs, smaller is the bandwidt=
h
you can safely record. The right side of the bell shows the response of the
antialiasing filter when fed with a sinusoid at constant level with
increasing frequency. The left side of the bell shows the artifacts
(aliases) produced by the sampling process when the input frequency is
higher than the Nyquist frequency, that is half of the sampling rate. Of
course the graphs shows the worst case, when a signal goes beyond Nyquist a=
t
a high constant level. Manufacturers design their filters considering that
in normal recordings of music programs taken with normal microphones it is
very unlikely to have so high level high frequencies. They don't consider
nature recordists willing to record dolphins and bats....
Even if you may not have strong tonals beyond Nyquist, it is often possible
to have harmonics going beyond Nyquist. Hope these plots will be useful Soo=
n
I would add more sound boards and recorders at 192k.
Gianni
2010/11/8 Raimund <>
>
>
> Hi Gus,
>
> Yes, to record a signal frequency of 100 kHz, one would need a sample rat=
e
> of lets say at least 220 kHz (205 kHz would be probably still too low). T=
he
> actually required minimum sample rate would depend on the properties of t=
he
> anti-aliasing filters of the specific recorder and the intensity of the
> potential signal components above 100 kHz.
>
> Some time ago I tested the aliasing behaviour of the SD 722 recorder (at =
a
> sample rate of 192 kHz) by recording a sine sweep signal ranging from 5 t=
o
> 130 kHz through its line input:
>
> http://www.avisoft.com/scartch/alias_722.gif
>
> It can be seen that its anti-aliasing filter is not very steep at all,
> which means that a small signal that appears in the recorded .wav file fo=
r
> instance at 70 kHz might actually be a 122 kHz signal (96-70+96=3D122).
>
> Bye the way, I enjoyed the stereo sample of your work very much.
>
>
> Raimund
>
> --- In <naturerecordists%40yahoogroups.co=
m>,
> "gus" <> wrote:
> >
> > Hi Raimund -
> > i think I know what you mean but Im not sure - what you are basically
> saying is that if say you wanted to record a frequency of ,say, 100khz yo=
u
> wouldnt just need a recorder that stated its max sample rate was 200khz b=
ut
> actually 205 khz or something like that (ie a little bit over) ?
> > Gus
>
>
>
--
Centro Interdisciplinare di Bioacustica e Ricerche Ambientali
Universit=E0 degli Studi di Pavia
Via Taramelli 24, 27100 Pavia
http://www.unipv.it/cibra
http://mammiferimarini.unipv.it
|
